From e29acf74fea996014380d59d31aa504ae8964258 Mon Sep 17 00:00:00 2001
From: Acly <aclysia@gmail.com>
Date: Sat, 4 Oct 2025 11:42:56 +0200
Subject: [PATCH 1/4] vulkan : incremental shader builds (#16341)

* vulkan (DRAFT): split shader generation by GLSL source file, to improve incremental build times

* support dep-files so shaders are recompiled if their included files change

* rename shader files which are used as "headers" to use .glsl extension
* move glslc extension detection shaders to separate folders
* the above is to prevent them from getting glob'd with the actual compute shaders that need to be compiled

* vulkan : only write embedded shader .hpp/.cpp when they change

* avoid recompiling ggml-vulkan.cpp when editing shaders
* pass single --source argument instead of --input-dir & --filter to shader gen
* check for source file match earlier

* fix hang in vulkan-shaders-gen when there are compilation errors

* early out did not decrement compile_count

* clean up

* fix glslc integer dot product test

* unconditionally write the embedded shader cpp output

* replace output filepath in generated dep-files to match output in CMakeLists

---------

Co-authored-by: Jeff Bolz <jbolz@nvidia.com>
---
 ggml/src/ggml-vulkan/CMakeLists.txt           |  45 ++-
 ggml/src/ggml-vulkan/vulkan-shaders/acc.comp  |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/add.comp  |   4 +-
 .../ggml-vulkan/vulkan-shaders/add_id.comp    |   2 +-
 .../ggml-vulkan/vulkan-shaders/argmax.comp    |   4 +-
 .../ggml-vulkan/vulkan-shaders/argsort.comp   |   2 +-
 .../src/ggml-vulkan/vulkan-shaders/clamp.comp |   4 +-
 .../ggml-vulkan/vulkan-shaders/concat.comp    |   4 +-
 .../vulkan-shaders/contig_copy.comp           |   4 +-
 .../ggml-vulkan/vulkan-shaders/conv2d_dw.comp |   2 +-
 .../ggml-vulkan/vulkan-shaders/conv2d_mm.comp |   2 +-
 .../vulkan-shaders/conv_transpose_1d.comp     |   2 +-
 ggml/src/ggml-vulkan/vulkan-shaders/copy.comp |   4 +-
 .../vulkan-shaders/copy_from_quant.comp       |   6 +-
 .../vulkan-shaders/copy_to_quant.comp         |   8 +-
 ggml/src/ggml-vulkan/vulkan-shaders/cos.comp  |   4 +-
 .../vulkan-shaders/count_equal.comp           |   4 +-
 .../vulkan-shaders/dequant_f32.comp           |   2 +-
 ...{dequant_funcs.comp => dequant_funcs.glsl} |   2 +-
 ..._funcs_cm2.comp => dequant_funcs_cm2.glsl} |   2 +-
 .../{dequant_head.comp => dequant_head.glsl}  |   2 +-
 .../vulkan-shaders/dequant_iq1_m.comp         |   2 +-
 .../vulkan-shaders/dequant_iq1_s.comp         |   2 +-
 .../vulkan-shaders/dequant_iq2_s.comp         |   2 +-
 .../vulkan-shaders/dequant_iq2_xs.comp        |   2 +-
 .../vulkan-shaders/dequant_iq2_xxs.comp       |   2 +-
 .../vulkan-shaders/dequant_iq3_s.comp         |   2 +-
 .../vulkan-shaders/dequant_iq3_xxs.comp       |   2 +-
 .../vulkan-shaders/dequant_iq4_nl.comp        |   2 +-
 .../vulkan-shaders/dequant_iq4_xs.comp        |   2 +-
 .../vulkan-shaders/dequant_mxfp4.comp         |   2 +-
 .../vulkan-shaders/dequant_q2_k.comp          |   2 +-
 .../vulkan-shaders/dequant_q3_k.comp          |   2 +-
 .../vulkan-shaders/dequant_q4_0.comp          |   2 +-
 .../vulkan-shaders/dequant_q4_1.comp          |   2 +-
 .../vulkan-shaders/dequant_q4_k.comp          |   2 +-
 .../vulkan-shaders/dequant_q5_0.comp          |   2 +-
 .../vulkan-shaders/dequant_q5_1.comp          |   2 +-
 .../vulkan-shaders/dequant_q5_k.comp          |   2 +-
 .../vulkan-shaders/dequant_q6_k.comp          |   2 +-
 .../vulkan-shaders/dequant_q8_0.comp          |   2 +-
 .../vulkan-shaders/diag_mask_inf.comp         |   2 +-
 ggml/src/ggml-vulkan/vulkan-shaders/div.comp  |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/exp.comp  |   6 +-
 .../bfloat16.comp}                            |   0
 .../coopmat.comp}                             |   0
 .../coopmat2.comp}                            |   0
 .../integer_dot.comp}                         |   0
 .../vulkan-shaders/flash_attn.comp            |   4 +-
 ...sh_attn_base.comp => flash_attn_base.glsl} |   0
 .../vulkan-shaders/flash_attn_cm1.comp        |   4 +-
 .../vulkan-shaders/flash_attn_cm2.comp        |   6 +-
 .../src/ggml-vulkan/vulkan-shaders/geglu.comp |   4 +-
 .../ggml-vulkan/vulkan-shaders/geglu_erf.comp |   4 +-
 .../vulkan-shaders/geglu_quick.comp           |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/gelu.comp |   4 +-
 .../ggml-vulkan/vulkan-shaders/gelu_erf.comp  |   4 +-
 .../vulkan-shaders/gelu_quick.comp            |   4 +-
 ...ary_head.comp => generic_binary_head.glsl} |   4 +-
 .../{generic_head.comp => generic_head.glsl}  |   0
 ...nary_head.comp => generic_unary_head.glsl} |   0
 .../ggml-vulkan/vulkan-shaders/get_rows.comp  |   4 +-
 .../vulkan-shaders/get_rows_quant.comp        |   6 +-
 .../{glu_head.comp => glu_head.glsl}          |   2 +-
 .../{glu_main.comp => glu_main.glsl}          |   0
 .../vulkan-shaders/group_norm.comp            |   4 +-
 .../vulkan-shaders/hardsigmoid.comp           |   4 +-
 .../ggml-vulkan/vulkan-shaders/hardswish.comp |   4 +-
 .../ggml-vulkan/vulkan-shaders/im2col.comp    |   5 +-
 .../ggml-vulkan/vulkan-shaders/im2col_3d.comp |   5 +-
 .../ggml-vulkan/vulkan-shaders/l2_norm.comp   |   4 +-
 .../vulkan-shaders/leaky_relu.comp            |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/mul.comp  |   4 +-
 .../vulkan-shaders/mul_mat_vec.comp           |   2 +-
 ...at_vec_base.comp => mul_mat_vec_base.glsl} |   4 +-
 .../vulkan-shaders/mul_mat_vec_iq1_m.comp     |   2 +-
 .../vulkan-shaders/mul_mat_vec_iq1_s.comp     |   2 +-
 .../vulkan-shaders/mul_mat_vec_iq2_s.comp     |   2 +-
 .../vulkan-shaders/mul_mat_vec_iq2_xs.comp    |   2 +-
 .../vulkan-shaders/mul_mat_vec_iq2_xxs.comp   |   2 +-
 .../vulkan-shaders/mul_mat_vec_iq3_s.comp     |   2 +-
 .../vulkan-shaders/mul_mat_vec_iq3_xxs.comp   |   2 +-
 .../vulkan-shaders/mul_mat_vec_q2_k.comp      |   2 +-
 .../vulkan-shaders/mul_mat_vec_q3_k.comp      |   2 +-
 .../vulkan-shaders/mul_mat_vec_q4_k.comp      |   2 +-
 .../vulkan-shaders/mul_mat_vec_q5_k.comp      |   2 +-
 .../vulkan-shaders/mul_mat_vec_q6_k.comp      |   2 +-
 .../vulkan-shaders/mul_mat_vecq.comp          |   4 +-
 .../ggml-vulkan/vulkan-shaders/mul_mm.comp    |   4 +-
 .../vulkan-shaders/mul_mm_cm2.comp            |   6 +-
 .../{mul_mm_funcs.comp => mul_mm_funcs.glsl}  |   0
 .../ggml-vulkan/vulkan-shaders/mul_mmq.comp   |   4 +-
 ...{mul_mmq_funcs.comp => mul_mmq_funcs.glsl} |   2 +-
 .../ggml-vulkan/vulkan-shaders/multi_add.comp |   6 +-
 ggml/src/ggml-vulkan/vulkan-shaders/norm.comp |   4 +-
 .../vulkan-shaders/opt_step_adamw.comp        |   4 +-
 .../vulkan-shaders/opt_step_sgd.comp          |   2 +-
 ggml/src/ggml-vulkan/vulkan-shaders/pad.comp  |   2 +-
 .../ggml-vulkan/vulkan-shaders/pool2d.comp    |   2 +-
 .../vulkan-shaders/quantize_q8_1.comp         |   2 +-
 .../src/ggml-vulkan/vulkan-shaders/reglu.comp |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/relu.comp |   4 +-
 .../ggml-vulkan/vulkan-shaders/repeat.comp    |   4 +-
 .../vulkan-shaders/repeat_back.comp           |   4 +-
 .../ggml-vulkan/vulkan-shaders/rms_norm.comp  |   4 +-
 .../vulkan-shaders/rms_norm_back.comp         |   4 +-
 .../vulkan-shaders/rms_norm_partials.comp     |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/roll.comp |   4 +-
 .../{rope_head.comp => rope_head.glsl}        |   4 +-
 .../vulkan-shaders/rope_multi.comp            |   2 +-
 .../ggml-vulkan/vulkan-shaders/rope_neox.comp |   2 +-
 .../ggml-vulkan/vulkan-shaders/rope_norm.comp |   2 +-
 .../vulkan-shaders/rope_vision.comp           |   2 +-
 .../vulkan-shaders/{rte.comp => rte.glsl}     |   0
 .../src/ggml-vulkan/vulkan-shaders/scale.comp |   4 +-
 .../ggml-vulkan/vulkan-shaders/sigmoid.comp   |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/silu.comp |   4 +-
 .../ggml-vulkan/vulkan-shaders/silu_back.comp |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/sin.comp  |   4 +-
 .../ggml-vulkan/vulkan-shaders/soft_max.comp  |   2 +-
 .../vulkan-shaders/soft_max_back.comp         |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/sqrt.comp |   4 +-
 .../ggml-vulkan/vulkan-shaders/square.comp    |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/sub.comp  |   4 +-
 .../ggml-vulkan/vulkan-shaders/sum_rows.comp  |   2 +-
 .../ggml-vulkan/vulkan-shaders/swiglu.comp    |   4 +-
 .../vulkan-shaders/swiglu_oai.comp            |   4 +-
 ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp |   4 +-
 .../vulkan-shaders/timestep_embedding.comp    |   2 +-
 .../vulkan-shaders/{types.comp => types.glsl} |   0
 .../ggml-vulkan/vulkan-shaders/upscale.comp   |   2 +-
 .../vulkan-shaders/{utils.comp => utils.glsl} |   0
 .../vulkan-shaders/vulkan-shaders-gen.cpp     | 294 +++++++++++-------
 133 files changed, 404 insertions(+), 315 deletions(-)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{dequant_funcs.comp => dequant_funcs.glsl} (99%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{dequant_funcs_cm2.comp => dequant_funcs_cm2.glsl} (99%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{dequant_head.comp => dequant_head.glsl} (91%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{test_bfloat16_support.comp => feature-tests/bfloat16.comp} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{test_coopmat_support.comp => feature-tests/coopmat.comp} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{test_coopmat2_support.comp => feature-tests/coopmat2.comp} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{test_integer_dot_support.comp => feature-tests/integer_dot.comp} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{flash_attn_base.comp => flash_attn_base.glsl} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{generic_binary_head.comp => generic_binary_head.glsl} (97%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{generic_head.comp => generic_head.glsl} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{generic_unary_head.comp => generic_unary_head.glsl} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{glu_head.comp => glu_head.glsl} (95%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{glu_main.comp => glu_main.glsl} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{mul_mat_vec_base.comp => mul_mat_vec_base.glsl} (99%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{mul_mm_funcs.comp => mul_mm_funcs.glsl} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{mul_mmq_funcs.comp => mul_mmq_funcs.glsl} (99%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{rope_head.comp => rope_head.glsl} (97%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{rte.comp => rte.glsl} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{types.comp => types.glsl} (100%)
 rename ggml/src/ggml-vulkan/vulkan-shaders/{utils.comp => utils.glsl} (100%)

diff --git a/ggml/src/ggml-vulkan/CMakeLists.txt b/ggml/src/ggml-vulkan/CMakeLists.txt
index b97e7bf99..83a83887b 100644
--- a/ggml/src/ggml-vulkan/CMakeLists.txt
+++ b/ggml/src/ggml-vulkan/CMakeLists.txt
@@ -1,5 +1,6 @@
 cmake_minimum_required(VERSION 3.19)
 cmake_policy(SET CMP0114 NEW)
+cmake_policy(SET CMP0116 NEW)
 
 find_package(Vulkan COMPONENTS glslc REQUIRED)
 
@@ -54,25 +55,25 @@ if (Vulkan_FOUND)
     # Test all shader extensions
     test_shader_extension_support(
         "GL_KHR_cooperative_matrix"
-        "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders/test_coopmat_support.comp"
+        "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders/feature-tests/coopmat.comp"
         "GGML_VULKAN_COOPMAT_GLSLC_SUPPORT"
     )
 
     test_shader_extension_support(
         "GL_NV_cooperative_matrix2"
-        "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders/test_coopmat2_support.comp"
+        "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders/feature-tests/coopmat2.comp"
         "GGML_VULKAN_COOPMAT2_GLSLC_SUPPORT"
     )
 
     test_shader_extension_support(
         "GL_EXT_integer_dot_product"
-        "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders/test_integer_dot_support.comp"
+        "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders/feature-tests/integer_dot.comp"
         "GGML_VULKAN_INTEGER_DOT_GLSLC_SUPPORT"
     )
 
     test_shader_extension_support(
         "GL_EXT_bfloat16"
-        "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders/test_bfloat16_support.comp"
+        "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders/feature-tests/bfloat16.comp"
         "GGML_VULKAN_BFLOAT16_GLSLC_SUPPORT"
     )
 
@@ -160,7 +161,6 @@ if (Vulkan_FOUND)
     set (_ggml_vk_genshaders_dir "${CMAKE_BINARY_DIR}/$<CONFIG>")
     set (_ggml_vk_genshaders_cmd "${_ggml_vk_genshaders_dir}/vulkan-shaders-gen${_ggml_vk_host_suffix}")
     set (_ggml_vk_header     "${CMAKE_CURRENT_BINARY_DIR}/ggml-vulkan-shaders.hpp")
-    set (_ggml_vk_source     "${CMAKE_CURRENT_BINARY_DIR}/ggml-vulkan-shaders.cpp")
     set (_ggml_vk_input_dir  "${CMAKE_CURRENT_SOURCE_DIR}/vulkan-shaders")
     set (_ggml_vk_output_dir "${CMAKE_CURRENT_BINARY_DIR}/vulkan-shaders.spv")
 
@@ -176,24 +176,35 @@ if (Vulkan_FOUND)
 
     add_custom_command(
         OUTPUT ${_ggml_vk_header}
-               ${_ggml_vk_source}
-
         COMMAND ${_ggml_vk_genshaders_cmd}
-            --glslc      ${Vulkan_GLSLC_EXECUTABLE}
-            --input-dir  ${_ggml_vk_input_dir}
             --output-dir ${_ggml_vk_output_dir}
             --target-hpp ${_ggml_vk_header}
-            --target-cpp ${_ggml_vk_source}
-            --no-clean
-
-        DEPENDS ${_ggml_vk_shader_files}
-                ${_ggml_vk_shaders_gen_sources}
+        DEPENDS ${_ggml_vk_shaders_gen_sources}
                 vulkan-shaders-gen
-
-        COMMENT "Generate vulkan shaders"
+        COMMENT "Generate vulkan shaders header"
     )
+    target_sources(ggml-vulkan PRIVATE ${_ggml_vk_header})
 
-    target_sources(ggml-vulkan PRIVATE ${_ggml_vk_source} ${_ggml_vk_header})
+    foreach (file_full ${_ggml_vk_shader_files})
+        get_filename_component(file ${file_full} NAME)
+        set (_ggml_vk_target_cpp "${CMAKE_CURRENT_BINARY_DIR}/${file}.cpp")
+
+        add_custom_command(
+            OUTPUT  ${_ggml_vk_target_cpp}
+            DEPFILE ${_ggml_vk_target_cpp}.d
+            COMMAND ${_ggml_vk_genshaders_cmd}
+                --glslc      ${Vulkan_GLSLC_EXECUTABLE}
+                --source     ${file_full}
+                --output-dir ${_ggml_vk_output_dir}
+                --target-hpp ${_ggml_vk_header}
+                --target-cpp ${_ggml_vk_target_cpp}
+            DEPENDS ${file_full}
+                    ${_ggml_vk_shaders_gen_sources}
+                    vulkan-shaders-gen
+            COMMENT "Generate vulkan shaders for ${file}"
+        )
+        target_sources(ggml-vulkan PRIVATE ${_ggml_vk_target_cpp})
+    endforeach()
 
 else()
     message(WARNING "Vulkan not found")
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp b/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp
index d896f1ef0..5084a70ed 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/acc.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_binary_head.comp"
+#include "types.glsl"
+#include "generic_binary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/add.comp b/ggml/src/ggml-vulkan/vulkan-shaders/add.comp
index 00cf2dd62..3bcfe6908 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/add.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/add.comp
@@ -6,8 +6,8 @@
 #extension GL_KHR_shader_subgroup_basic : enable
 #endif
 
-#include "types.comp"
-#include "generic_binary_head.comp"
+#include "types.glsl"
+#include "generic_binary_head.glsl"
 
 const uint num_threads = 256;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/add_id.comp b/ggml/src/ggml-vulkan/vulkan-shaders/add_id.comp
index 3ae8f0116..495249d5f 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/add_id.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/add_id.comp
@@ -2,7 +2,7 @@
 
 #extension GL_EXT_control_flow_attributes : require
 
-#include "types.comp"
+#include "types.glsl"
 
 layout (push_constant) uniform parameter
 {
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/argmax.comp b/ggml/src/ggml-vulkan/vulkan-shaders/argmax.comp
index a1d4c240d..7c1287767 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/argmax.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/argmax.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/argsort.comp b/ggml/src/ggml-vulkan/vulkan-shaders/argsort.comp
index dc53a401e..c81b84452 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/argsort.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/argsort.comp
@@ -1,7 +1,7 @@
 #version 450
 #extension GL_EXT_control_flow_attributes : enable
 
-#include "types.comp"
+#include "types.glsl"
 
 layout(constant_id = 0) const int BLOCK_SIZE = 1024;
 layout(constant_id = 1) const int BLOCK_SIZE_LOG2 = 10;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/clamp.comp b/ggml/src/ggml-vulkan/vulkan-shaders/clamp.comp
index 1e5cb8dae..653431895 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/clamp.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/clamp.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/concat.comp b/ggml/src/ggml-vulkan/vulkan-shaders/concat.comp
index 9ee2f1fae..e40469838 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/concat.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/concat.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_binary_head.comp"
+#include "types.glsl"
+#include "generic_binary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp b/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp
index 6567a8c54..ca1a3ac25 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/contig_copy.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 #extension GL_EXT_control_flow_attributes : require
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_dw.comp b/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_dw.comp
index 938c74da5..70a301488 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_dw.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_dw.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "types.comp"
+#include "types.glsl"
 
 layout (push_constant) uniform parameter
 {
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_mm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_mm.comp
index 44a64ddc8..0367e80bb 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_mm.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/conv2d_mm.comp
@@ -11,7 +11,7 @@
 #    extension GL_KHR_shader_subgroup_shuffle : enable
 #endif
 
-#include "types.comp"
+#include "types.glsl"
 
 // shape notation: [dim(N), ..., dim(0)] -- stride(dim(j)) >= stride(dim(i)) if i > j
 layout(binding = 0) readonly buffer A {
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/conv_transpose_1d.comp b/ggml/src/ggml-vulkan/vulkan-shaders/conv_transpose_1d.comp
index b17b4e83e..5217e18bd 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/conv_transpose_1d.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/conv_transpose_1d.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "types.comp"
+#include "types.glsl"
 
 layout (binding = 0) readonly buffer A {A_TYPE data_a[];};   // src0 - kernel:    [K, Cout, Cin]
 layout (binding = 1) readonly buffer B {B_TYPE data_b[];};   // src1 - input:     [L, Cin]
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp b/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp
index f476a2e3d..9f8bfd3c1 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/copy.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/copy_from_quant.comp b/ggml/src/ggml-vulkan/vulkan-shaders/copy_from_quant.comp
index 978d43003..06df50952 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/copy_from_quant.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/copy_from_quant.comp
@@ -1,8 +1,8 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
-#include "dequant_funcs.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
+#include "dequant_funcs.glsl"
 
 #if defined(DATA_A_IQ4_NL) || defined(DATA_A_MXFP4)
 // 16 invocations needed for init_iq_shmem
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/copy_to_quant.comp b/ggml/src/ggml-vulkan/vulkan-shaders/copy_to_quant.comp
index bc2e1f2df..b8c40eec1 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/copy_to_quant.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/copy_to_quant.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "rte.comp"
-#include "types.comp"
+#include "rte.glsl"
+#include "types.glsl"
 
 #if defined(SET_ROWS) && QUANT_K == 1
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
@@ -14,7 +14,7 @@ const uint BLOCK_SIZE = 32;
 layout (binding = 0) readonly buffer S {float data_s[];};
 
 #if defined(SET_ROWS)
-#include "generic_binary_head.comp"
+#include "generic_binary_head.glsl"
 layout (binding = 1) readonly buffer C {B_TYPE data_i[];};
 layout (binding = 2) writeonly buffer Q {A_TYPE data_q[];};
 
@@ -25,7 +25,7 @@ layout (binding = 2) writeonly buffer Q {A_TYPE data_q[];};
 #endif
 
 #else
-#include "generic_unary_head.comp"
+#include "generic_unary_head.glsl"
 layout (binding = 1) writeonly buffer Q {A_TYPE data_q[];};
 #endif
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp b/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp
index 0b8d02f58..db6865db9 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/cos.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/count_equal.comp b/ggml/src/ggml-vulkan/vulkan-shaders/count_equal.comp
index d9345497c..e75df6675 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/count_equal.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/count_equal.comp
@@ -2,8 +2,8 @@
 
 #extension GL_EXT_control_flow_attributes : enable
 
-#include "types.comp"
-#include "generic_head.comp"
+#include "types.glsl"
+#include "generic_head.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_f32.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_f32.comp
index a4d3fca55..765afffa8 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_f32.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_f32.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.glsl
similarity index 99%
rename from ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.glsl
index 73fef4fa6..0d98f5a9d 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs.glsl
@@ -2,7 +2,7 @@
 #extension GL_EXT_shader_explicit_arithmetic_types_int8 : require
 #endif
 
-#include "types.comp"
+#include "types.glsl"
 
 #if defined(A_TYPE_PACKED16)
 layout (binding = 0) readonly buffer A_PACKED16 {A_TYPE_PACKED16 data_a_packed16[];};
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.glsl
similarity index 99%
rename from ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.glsl
index 706540fd8..6a5bb4574 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_funcs_cm2.glsl
@@ -1,5 +1,5 @@
 
-#include "types.comp"
+#include "types.glsl"
 
 layout(buffer_reference, std430, buffer_reference_align = 2) buffer decodeBufQ4_0 {
    block_q4_0_packed16 block;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.glsl
similarity index 91%
rename from ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.glsl
index 8d806435b..addceafad 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_head.glsl
@@ -10,4 +10,4 @@ layout (push_constant) uniform parameter
     uint nel;
 } p;
 
-#include "types.comp"
+#include "types.glsl"
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_m.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_m.comp
index b604c1881..637c95fa3 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_m.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_m.comp
@@ -2,7 +2,7 @@
 
 #extension GL_EXT_shader_explicit_arithmetic_types_int16 : require
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_s.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_s.comp
index fd1e4e30d..d1cbc5e9d 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_s.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq1_s.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_s.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_s.comp
index 127c7b642..78490162c 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_s.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_s.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xs.comp
index a08331c40..9b8ce0a7f 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xs.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xs.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xxs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xxs.comp
index 0ae9acd02..aacf07d0f 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xxs.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq2_xxs.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_s.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_s.comp
index e4f42be94..f2c20b1d2 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_s.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_s.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_xxs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_xxs.comp
index 19c7fdeef..671c1f4a0 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_xxs.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq3_xxs.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp
index 46d9ad15e..8f7833eab 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_nl.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_xs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_xs.comp
index f930852a4..a31369977 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_xs.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_iq4_xs.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_mxfp4.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_mxfp4.comp
index ee496e9d5..ffba5a77d 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_mxfp4.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_mxfp4.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q2_k.comp
index d4e4e6bae..58dc2e5df 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q2_k.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 64, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q3_k.comp
index 3661f771c..0c90be8b4 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q3_k.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 64, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_0.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_0.comp
index 408185327..b92b29213 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_0.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_0.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_1.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_1.comp
index 2f27eee68..6b63cbe58 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_1.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_1.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_k.comp
index 1370db365..8b7be557e 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q4_k.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 32, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_0.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_0.comp
index b20b80529..f1b0bac87 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_0.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_0.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_1.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_1.comp
index dc59fe3b7..c495b31f1 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_1.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_1.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_k.comp
index 3f3b839e1..6bc04670f 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q5_k.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 64, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q6_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q6_k.comp
index 9cf34256e..c8d6fcb49 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q6_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q6_k.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 64, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q8_0.comp b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q8_0.comp
index bd1344a88..10844ddf7 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q8_0.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/dequant_q8_0.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "dequant_head.comp"
+#include "dequant_head.glsl"
 
 layout(local_size_x = 256, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp b/ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp
index 26d8bc22a..9cef8a8ec 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/diag_mask_inf.comp
@@ -10,7 +10,7 @@ layout (push_constant) uniform parameter
     uint n_past;
 } p;
 
-#include "types.comp"
+#include "types.glsl"
 
 layout(local_size_x = 1, local_size_y = 512, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/div.comp b/ggml/src/ggml-vulkan/vulkan-shaders/div.comp
index 9fb69c6c1..572472f8a 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/div.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/div.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_binary_head.comp"
+#include "types.glsl"
+#include "generic_binary_head.glsl"
 
 const uint num_threads = 256;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/exp.comp b/ggml/src/ggml-vulkan/vulkan-shaders/exp.comp
index a3941372a..b69d4ddb0 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/exp.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/exp.comp
@@ -1,8 +1,8 @@
 #version 450
 
-#include "rte.comp"
-#include "generic_head.comp"
-#include "types.comp"
+#include "rte.glsl"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/test_bfloat16_support.comp b/ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/bfloat16.comp
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/test_bfloat16_support.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/bfloat16.comp
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat_support.comp b/ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/coopmat.comp
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat_support.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/coopmat.comp
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat2_support.comp b/ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/coopmat2.comp
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/test_coopmat2_support.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/coopmat2.comp
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/test_integer_dot_support.comp b/ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/integer_dot.comp
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/test_integer_dot_support.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/feature-tests/integer_dot.comp
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp b/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp
index e42475026..62acbf107 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp
@@ -8,8 +8,8 @@
 
 #extension GL_KHR_shader_subgroup_shuffle : enable
 
-#include "types.comp"
-#include "flash_attn_base.comp"
+#include "types.glsl"
+#include "flash_attn_base.glsl"
 
 const uint32_t HSK_per_thread = HSK / D_split;
 const uint32_t HSV_per_thread = HSV / D_split;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_base.comp b/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_base.glsl
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_base.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_base.glsl
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp b/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp
index 0507df2d8..2066a05b3 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp
@@ -10,8 +10,8 @@
 #extension GL_KHR_memory_scope_semantics : enable
 #extension GL_KHR_cooperative_matrix : enable
 
-#include "types.comp"
-#include "flash_attn_base.comp"
+#include "types.glsl"
+#include "flash_attn_base.glsl"
 
 const uint32_t HSK_per_thread = HSK / D_split;
 const uint32_t HSV_per_thread = HSV / D_split;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp b/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp
index a65553a48..910da1ab0 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp
@@ -16,9 +16,9 @@
 #extension GL_KHR_shader_subgroup_vote : enable
 #extension GL_EXT_null_initializer : enable
 
-#include "types.comp"
-#include "dequant_funcs_cm2.comp"
-#include "flash_attn_base.comp"
+#include "types.glsl"
+#include "dequant_funcs_cm2.glsl"
+#include "flash_attn_base.glsl"
 
 layout (binding = 0) readonly buffer Q {uint8_t data_q[];};
 layout (binding = 1) readonly buffer K {uint8_t data_k[];};
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/geglu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/geglu.comp
index f4268ed24..e017b5036 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/geglu.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/geglu.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "glu_head.comp"
+#include "glu_head.glsl"
 
 const float GELU_COEF_A    = 0.044715f;
 const float SQRT_2_OVER_PI = 0.79788456080286535587989211986876f;
@@ -10,4 +10,4 @@ float op(float a, float b) {
     return 0.5f*a*(2.0f - 2.0f / (exp(2 * val) + 1)) * b;
 }
 
-#include "glu_main.comp"
+#include "glu_main.glsl"
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/geglu_erf.comp b/ggml/src/ggml-vulkan/vulkan-shaders/geglu_erf.comp
index cbd4cb36b..759a1848f 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/geglu_erf.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/geglu_erf.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "glu_head.comp"
+#include "glu_head.glsl"
 
 // based on Abramowitz and Stegun formula 7.1.26 or similar Hastings' approximation
 // ref: https://www.johndcook.com/blog/python_erf/
@@ -24,4 +24,4 @@ float op(float a, float b) {
     return 0.5f * a * (1.0f + erf_approx) * b;
 }
 
-#include "glu_main.comp"
+#include "glu_main.glsl"
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/geglu_quick.comp b/ggml/src/ggml-vulkan/vulkan-shaders/geglu_quick.comp
index 3a2a6897b..c4032ab21 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/geglu_quick.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/geglu_quick.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "glu_head.comp"
+#include "glu_head.glsl"
 
 const float GELU_QUICK_COEF = -1.702f;
 
@@ -8,4 +8,4 @@ float op(float a, float b) {
     return a * (1.0f / (1.0f + exp(GELU_QUICK_COEF * a))) * b;
 }
 
-#include "glu_main.comp"
+#include "glu_main.glsl"
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/gelu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/gelu.comp
index 4cc7a68ca..a95c2525c 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/gelu.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/gelu.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/gelu_erf.comp b/ggml/src/ggml-vulkan/vulkan-shaders/gelu_erf.comp
index 5fd5a5e70..58375aba0 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/gelu_erf.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/gelu_erf.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/gelu_quick.comp b/ggml/src/ggml-vulkan/vulkan-shaders/gelu_quick.comp
index e6e6fcfd2..bfdfe2182 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/gelu_quick.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/gelu_quick.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.glsl
similarity index 97%
rename from ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.glsl
index 750e78575..99595fc68 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/generic_binary_head.glsl
@@ -1,8 +1,8 @@
 #extension GL_EXT_shader_16bit_storage : require
 #extension GL_EXT_control_flow_attributes : require
 
-#include "rte.comp"
-#include "utils.comp"
+#include "rte.glsl"
+#include "utils.glsl"
 
 layout (push_constant) uniform parameter
 {
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/generic_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/generic_head.glsl
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/generic_head.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/generic_head.glsl
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/generic_unary_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/generic_unary_head.glsl
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/generic_unary_head.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/generic_unary_head.glsl
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/get_rows.comp b/ggml/src/ggml-vulkan/vulkan-shaders/get_rows.comp
index 7ef75cd7a..76d83041c 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/get_rows.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/get_rows.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_binary_head.comp"
+#include "types.glsl"
+#include "generic_binary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/get_rows_quant.comp b/ggml/src/ggml-vulkan/vulkan-shaders/get_rows_quant.comp
index 339f905fc..9dba437ed 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/get_rows_quant.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/get_rows_quant.comp
@@ -2,9 +2,9 @@
 
 #extension GL_EXT_control_flow_attributes : enable
 
-#include "types.comp"
-#include "generic_binary_head.comp"
-#include "dequant_funcs.comp"
+#include "types.glsl"
+#include "generic_binary_head.glsl"
+#include "dequant_funcs.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/glu_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/glu_head.glsl
similarity index 95%
rename from ggml/src/ggml-vulkan/vulkan-shaders/glu_head.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/glu_head.glsl
index 51d70869d..216898934 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/glu_head.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/glu_head.glsl
@@ -1,6 +1,6 @@
 #extension GL_EXT_shader_16bit_storage : require
 
-#include "rte.comp"
+#include "rte.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/glu_main.comp b/ggml/src/ggml-vulkan/vulkan-shaders/glu_main.glsl
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/glu_main.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/glu_main.glsl
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/group_norm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/group_norm.comp
index b6a0d5645..bdf97dbb5 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/group_norm.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/group_norm.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 #define BLOCK_SIZE 512
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/hardsigmoid.comp b/ggml/src/ggml-vulkan/vulkan-shaders/hardsigmoid.comp
index 1da252cc6..b4dbdf314 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/hardsigmoid.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/hardsigmoid.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/hardswish.comp b/ggml/src/ggml-vulkan/vulkan-shaders/hardswish.comp
index 3afc58827..1ec315915 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/hardswish.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/hardswish.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp b/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp
index f0f19a019..1827d647a 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/im2col.comp
@@ -3,9 +3,8 @@
 #extension GL_EXT_shader_16bit_storage : require
 #extension GL_EXT_control_flow_attributes : require
 
-#include "rte.comp"
-
-#include "types.comp"
+#include "rte.glsl"
+#include "types.glsl"
 
 layout (push_constant) uniform parameter
 {
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/im2col_3d.comp b/ggml/src/ggml-vulkan/vulkan-shaders/im2col_3d.comp
index 9faa636ac..4bf8b4ca0 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/im2col_3d.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/im2col_3d.comp
@@ -4,9 +4,8 @@
 #extension GL_EXT_control_flow_attributes : require
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "rte.comp"
-
-#include "types.comp"
+#include "rte.glsl"
+#include "types.glsl"
 
 layout (push_constant) uniform parameter
 {
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/l2_norm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/l2_norm.comp
index deba8c398..83ef2f879 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/l2_norm.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/l2_norm.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 #define BLOCK_SIZE 512
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/leaky_relu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/leaky_relu.comp
index d90a99aea..b281e855c 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/leaky_relu.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/leaky_relu.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul.comp
index 43de19df8..02ef1eace 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_binary_head.comp"
+#include "types.glsl"
+#include "generic_binary_head.glsl"
 
 const uint num_threads = 256;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp
index bb429dd59..9a03925cf 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec.comp
@@ -2,7 +2,7 @@
 
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.glsl
similarity index 99%
rename from ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.glsl
index f761391ea..450dee040 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_base.glsl
@@ -11,7 +11,7 @@
 #define EXPERT_COUNT 8
 #endif
 
-#include "types.comp"
+#include "types.glsl"
 
 #ifndef MMQ
 layout (binding = 0) readonly buffer A {A_TYPE data_a[];};
@@ -32,7 +32,7 @@ layout (binding = 2) writeonly buffer D {D_TYPE data_d[];};
 layout (binding = 3) readonly buffer IDS {int data_ids[];};
 #endif
 
-#include "dequant_funcs.comp"
+#include "dequant_funcs.glsl"
 
 layout (push_constant) uniform parameter
 {
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_m.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_m.comp
index e4acbd4f9..4cb292380 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_m.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_m.comp
@@ -1,7 +1,7 @@
 #version 450
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_s.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_s.comp
index 309da0991..0b74b3321 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_s.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq1_s.comp
@@ -1,7 +1,7 @@
 #version 450
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_s.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_s.comp
index 8d01536fa..e424af12c 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_s.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_s.comp
@@ -1,7 +1,7 @@
 #version 450
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xs.comp
index c49604324..0cd906dbb 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xs.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xs.comp
@@ -1,7 +1,7 @@
 #version 450
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xxs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xxs.comp
index 94d4b92e1..71bd72d17 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xxs.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq2_xxs.comp
@@ -1,7 +1,7 @@
 #version 450
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_s.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_s.comp
index f021e4047..a4b9ab1f9 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_s.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_s.comp
@@ -1,7 +1,7 @@
 #version 450
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_xxs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_xxs.comp
index 3fe9dc3a4..40849c691 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_xxs.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_iq3_xxs.comp
@@ -1,7 +1,7 @@
 #version 450
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
index 423ceb8a3..03ed25d3b 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q2_k.comp
@@ -1,7 +1,7 @@
 #version 450
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
index e91724a28..528f224d8 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q3_k.comp
@@ -1,7 +1,7 @@
 #version 450
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
index f9cde0648..21d07d2e5 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q4_k.comp
@@ -2,7 +2,7 @@
 
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
index 6c84ef3cd..9e46c89a1 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q5_k.comp
@@ -2,7 +2,7 @@
 
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
index d53d9ee0a..d7a7f6426 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vec_q6_k.comp
@@ -2,7 +2,7 @@
 
 #extension GL_EXT_shader_explicit_arithmetic_types_int32 : require
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vecq.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vecq.comp
index 8fb314fa0..64293f6ec 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vecq.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mat_vecq.comp
@@ -6,13 +6,13 @@
 #define MMQ
 #define B_TYPE block_q8_1_x4
 
-#include "mul_mat_vec_base.comp"
+#include "mul_mat_vec_base.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
 #define K_PER_ITER 8
 
-#include "mul_mmq_funcs.comp"
+#include "mul_mmq_funcs.glsl"
 
 uint a_offset, b_offset, d_offset;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp
index 3cb24412d..85400ac5f 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm.comp
@@ -28,7 +28,7 @@
 #extension GL_EXT_shader_explicit_arithmetic_types_int16 : require
 #endif
 
-#include "types.comp"
+#include "types.glsl"
 
 #ifndef LOAD_VEC_A
 #define LOAD_VEC_A 1
@@ -195,7 +195,7 @@ void load_row_ids(uint expert_idx, bool nei0_is_pow2, uint ic) {
 shared ACC_TYPE coopmat_stage[TM * TN * NUM_WARPS];
 #endif
 
-#include "mul_mm_funcs.comp"
+#include "mul_mm_funcs.glsl"
 
 void main() {
 #ifdef NEEDS_INIT_IQ_SHMEM
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp
index 0e3065e01..2e04baa44 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_cm2.comp
@@ -18,8 +18,8 @@
 #extension GL_EXT_bfloat16 : enable
 #endif
 
-#include "types.comp"
-#include "utils.comp"
+#include "types.glsl"
+#include "utils.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
@@ -71,7 +71,7 @@ layout (binding = 2) writeonly buffer D {D_TYPE data_d[];};
 #if QUANT_K > 1
 #define DECODEFUNCA , dequantFuncA
 
-#include "dequant_funcs_cm2.comp"
+#include "dequant_funcs_cm2.glsl"
 
 #else
 #define DECODEFUNCA
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_funcs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_funcs.glsl
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_funcs.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/mul_mm_funcs.glsl
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq.comp
index f36add62a..b5d761c0b 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq.comp
@@ -20,7 +20,7 @@
 #extension GL_EXT_shader_explicit_arithmetic_types_int16 : require
 #endif
 
-#include "types.comp"
+#include "types.glsl"
 
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
 
@@ -110,7 +110,7 @@ shared u16vec2 row_ids[4096];
 shared ACC_TYPE coopmat_stage[TM * TN * NUM_WARPS];
 #endif
 
-#include "mul_mmq_funcs.comp"
+#include "mul_mmq_funcs.glsl"
 
 void main() {
 #ifdef NEEDS_INIT_IQ_SHMEM
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.comp b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.glsl
similarity index 99%
rename from ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.glsl
index cdfb230f4..fe71eb131 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/mul_mmq_funcs.glsl
@@ -2,7 +2,7 @@
 #extension GL_EXT_shader_explicit_arithmetic_types_int16 : require
 #extension GL_EXT_shader_explicit_arithmetic_types_int8 : require
 
-#include "types.comp"
+#include "types.glsl"
 
 // Each iqs value maps to a 32-bit integer
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/multi_add.comp b/ggml/src/ggml-vulkan/vulkan-shaders/multi_add.comp
index 854a2ad81..1e8f694a7 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/multi_add.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/multi_add.comp
@@ -8,9 +8,9 @@
 #extension GL_KHR_shader_subgroup_basic : enable
 #endif
 
-#include "rte.comp"
-#include "types.comp"
-#include "utils.comp"
+#include "rte.glsl"
+#include "types.glsl"
+#include "utils.glsl"
 
 layout (push_constant) uniform parameter2
 {
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/norm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/norm.comp
index 6627a50bd..cc3ea0b76 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/norm.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/norm.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 #define BLOCK_SIZE 512
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_adamw.comp b/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_adamw.comp
index e0214fe76..1f05f922c 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_adamw.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_adamw.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_sgd.comp b/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_sgd.comp
index 6426dedee..1251f9cc6 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_sgd.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/opt_step_sgd.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "generic_head.comp"
+#include "generic_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/pad.comp b/ggml/src/ggml-vulkan/vulkan-shaders/pad.comp
index 0d81220c7..f3c817687 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/pad.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/pad.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "types.comp"
+#include "types.glsl"
 
 layout (push_constant) uniform parameter
 {
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/pool2d.comp b/ggml/src/ggml-vulkan/vulkan-shaders/pool2d.comp
index b6124411a..d9d7166e3 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/pool2d.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/pool2d.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "types.comp"
+#include "types.glsl"
 
 #extension GL_EXT_shader_16bit_storage : require
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/quantize_q8_1.comp b/ggml/src/ggml-vulkan/vulkan-shaders/quantize_q8_1.comp
index 145c9fbdc..0f3c6ca87 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/quantize_q8_1.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/quantize_q8_1.comp
@@ -17,7 +17,7 @@ layout (push_constant) uniform parameter
     uint ne;
 } p;
 
-#include "types.comp"
+#include "types.glsl"
 
 layout(constant_id = 0) const uint GROUP_SIZE = 32;
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/reglu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/reglu.comp
index 0073d8f76..86be2669a 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/reglu.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/reglu.comp
@@ -1,9 +1,9 @@
 #version 450
 
-#include "glu_head.comp"
+#include "glu_head.glsl"
 
 float op(float a, float b) {
     return max(a, 0.0f) * b;
 }
 
-#include "glu_main.comp"
+#include "glu_main.glsl"
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/relu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/relu.comp
index 4f806270c..5725cef23 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/relu.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/relu.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/repeat.comp b/ggml/src/ggml-vulkan/vulkan-shaders/repeat.comp
index 1568b141d..8f4b9a868 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/repeat.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/repeat.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/repeat_back.comp b/ggml/src/ggml-vulkan/vulkan-shaders/repeat_back.comp
index d86279934..87df78294 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/repeat_back.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/repeat_back.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp
index 41197e930..d5b211ffa 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_binary_head.comp"
-#include "types.comp"
+#include "generic_binary_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 #define BLOCK_SIZE 512
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_back.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_back.comp
index 76009f3df..87707fc14 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_back.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_back.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 #define BLOCK_SIZE 512
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_partials.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_partials.comp
index ba4677c29..4618b2c7e 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_partials.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/rms_norm_partials.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_binary_head.comp"
-#include "types.comp"
+#include "generic_binary_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 #extension GL_KHR_shader_subgroup_arithmetic : enable
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/roll.comp b/ggml/src/ggml-vulkan/vulkan-shaders/roll.comp
index b9abe8ded..68fbd0c7b 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/roll.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/roll.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.glsl
similarity index 97%
rename from ggml/src/ggml-vulkan/vulkan-shaders/rope_head.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/rope_head.glsl
index 00e203e73..50fc1f1e2 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/rope_head.glsl
@@ -1,8 +1,8 @@
-#include "types.comp"
+#include "types.glsl"
 
 #extension GL_EXT_shader_16bit_storage : require
 
-#include "rte.comp"
+#include "rte.glsl"
 
 layout(local_size_x = 1, local_size_y = 256, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/rope_multi.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rope_multi.comp
index 5808710cc..111286b49 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/rope_multi.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/rope_multi.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "rope_head.comp"
+#include "rope_head.glsl"
 
 void main() {
     const uint i0 = 2*gl_GlobalInvocationID.y;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp
index 366a7b1c4..06e095bef 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/rope_neox.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "rope_head.comp"
+#include "rope_head.glsl"
 
 void main() {
     const uint i0 = 2*gl_GlobalInvocationID.y;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp
index 9643bca96..6ba957540 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/rope_norm.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "rope_head.comp"
+#include "rope_head.glsl"
 
 void main() {
     const uint i0 = 2*gl_GlobalInvocationID.y;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/rope_vision.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rope_vision.comp
index cedacc4d1..d37d1c104 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/rope_vision.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/rope_vision.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "rope_head.comp"
+#include "rope_head.glsl"
 
 void main() {
     const uint i0 = 2*gl_GlobalInvocationID.y;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/rte.comp b/ggml/src/ggml-vulkan/vulkan-shaders/rte.glsl
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/rte.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/rte.glsl
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp b/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp
index f10b0a02b..35ec726a0 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/scale.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 const uint num_threads = 128;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/sigmoid.comp b/ggml/src/ggml-vulkan/vulkan-shaders/sigmoid.comp
index 5c9e5c350..32298d43c 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/sigmoid.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/sigmoid.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/silu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/silu.comp
index 4d36f88e0..7d1cc6f45 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/silu.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/silu.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/silu_back.comp b/ggml/src/ggml-vulkan/vulkan-shaders/silu_back.comp
index f9afa9b13..e5d949ff1 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/silu_back.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/silu_back.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp b/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp
index d7c15a169..61f17b2f0 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/sin.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp b/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp
index 5f20a1ee7..dca0d896b 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/soft_max.comp
@@ -23,7 +23,7 @@ layout (push_constant) uniform parameter
     uint has_sinks;
 } p;
 
-#include "types.comp"
+#include "types.glsl"
 
 layout(constant_id = 0) const uint BLOCK_SIZE = 32;
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/soft_max_back.comp b/ggml/src/ggml-vulkan/vulkan-shaders/soft_max_back.comp
index 144ea58e6..d873332ee 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/soft_max_back.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/soft_max_back.comp
@@ -2,8 +2,8 @@
 
 #extension GL_EXT_control_flow_attributes : enable
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 layout(constant_id = 0) const uint BLOCK_SIZE = 32;
 layout(local_size_x_id = 0, local_size_y = 1, local_size_z = 1) in;
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/sqrt.comp b/ggml/src/ggml-vulkan/vulkan-shaders/sqrt.comp
index 4bc697b9b..70daad6c5 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/sqrt.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/sqrt.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/square.comp b/ggml/src/ggml-vulkan/vulkan-shaders/square.comp
index ef43598ba..4eb56afcb 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/square.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/square.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "types.comp"
-#include "generic_unary_head.comp"
+#include "types.glsl"
+#include "generic_unary_head.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/sub.comp b/ggml/src/ggml-vulkan/vulkan-shaders/sub.comp
index 72353cc32..bc924b520 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/sub.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/sub.comp
@@ -2,8 +2,8 @@
 
 #extension GL_EXT_shader_16bit_storage : require
 
-#include "types.comp"
-#include "generic_binary_head.comp"
+#include "types.glsl"
+#include "generic_binary_head.glsl"
 
 const uint num_threads = 256;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/sum_rows.comp b/ggml/src/ggml-vulkan/vulkan-shaders/sum_rows.comp
index 759204afa..bc22aa7bd 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/sum_rows.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/sum_rows.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "types.comp"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/swiglu.comp b/ggml/src/ggml-vulkan/vulkan-shaders/swiglu.comp
index a28e7c6cc..4fee433a1 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/swiglu.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/swiglu.comp
@@ -1,9 +1,9 @@
 #version 450
 
-#include "glu_head.comp"
+#include "glu_head.glsl"
 
 float op(float a, float b) {
     return a / (1.0f + exp(-a)) * b;
 }
 
-#include "glu_main.comp"
+#include "glu_main.glsl"
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/swiglu_oai.comp b/ggml/src/ggml-vulkan/vulkan-shaders/swiglu_oai.comp
index 970750eec..bda9dea21 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/swiglu_oai.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/swiglu_oai.comp
@@ -1,6 +1,6 @@
 #version 450
 
-#include "glu_head.comp"
+#include "glu_head.glsl"
 
 float op(float a, float b) {
     float xi = min(a, p.limit);
@@ -11,4 +11,4 @@ float op(float a, float b) {
     return out_glu;
 }
 
-#include "glu_main.comp"
+#include "glu_main.glsl"
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp b/ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp
index 8a6f868f5..7b5eb413b 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/tanh.comp
@@ -1,7 +1,7 @@
 #version 450
 
-#include "generic_head.comp"
-#include "types.comp"
+#include "generic_head.glsl"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/timestep_embedding.comp b/ggml/src/ggml-vulkan/vulkan-shaders/timestep_embedding.comp
index ce8e09442..160556545 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/timestep_embedding.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/timestep_embedding.comp
@@ -9,7 +9,7 @@ layout (push_constant) uniform parameter
     uint max_period;
 } p;
 
-#include "types.comp"
+#include "types.glsl"
 
 #extension GL_EXT_control_flow_attributes : enable
 #define BLOCK_SIZE 256
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/types.comp b/ggml/src/ggml-vulkan/vulkan-shaders/types.glsl
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/types.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/types.glsl
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp b/ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp
index 74771def0..154a2172d 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/upscale.comp
@@ -9,7 +9,7 @@ layout (push_constant) uniform parameter
     float sf0; float sf1; float sf2; float sf3;
 } p;
 
-#include "types.comp"
+#include "types.glsl"
 
 layout(local_size_x = 512, local_size_y = 1, local_size_z = 1) in;
 
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/utils.comp b/ggml/src/ggml-vulkan/vulkan-shaders/utils.glsl
similarity index 100%
rename from ggml/src/ggml-vulkan/vulkan-shaders/utils.comp
rename to ggml/src/ggml-vulkan/vulkan-shaders/utils.glsl
diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp b/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp
index 84bb9df9a..e2726f1fa 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp
@@ -34,13 +34,13 @@
 
 std::mutex lock;
 std::vector<std::pair<std::string, std::string>> shader_fnames;
+std::locale c_locale("C");
 
 std::string GLSLC = "glslc";
-std::string input_dir = "vulkan-shaders";
+std::string input_filepath = "";
 std::string output_dir = "/tmp";
-std::string target_hpp = "ggml-vulkan-shaders.hpp";
-std::string target_cpp = "ggml-vulkan-shaders.cpp";
-bool no_clean = false;
+std::string target_hpp = "";
+std::string target_cpp = "";
 
 const std::vector<std::string> type_names = {
     "f32",
@@ -75,6 +75,7 @@ enum MatMulIdType {
 };
 
 namespace {
+
 void execute_command(const std::string& command, std::string& stdout_str, std::string& stderr_str) {
 #ifdef _WIN32
     HANDLE stdout_read, stdout_write;
@@ -232,16 +233,87 @@ std::string basename(const std::string &path) {
     return path.substr(path.find_last_of("/\\") + 1);
 }
 
+std::stringstream make_generic_stringstream() {
+    std::stringstream ss;
+    ss.imbue(c_locale);
+    return ss;
+}
+
+std::string read_binary_file(const std::string& path, bool may_not_exist = false) {
+    FILE* f = fopen(path.c_str(), "rb");
+    if (!f) {
+        if (!may_not_exist) {
+            std::cerr << "Error opening file: " << path << " (" << strerror(errno) << ")\n";
+        }
+        return {};
+    }
+
+    fseek(f, 0, SEEK_END);
+    size_t size = ftell(f);
+    fseek(f, 0, SEEK_SET);
+
+    std::string data(size, '\0');
+    size_t read_size = fread(data.data(), 1, size, f);
+    fclose(f);
+    if (read_size != size) {
+        std::cerr << "Error reading file: " << path << " (" << strerror(errno) << ")\n";
+        return {};
+    }
+
+    return data;
+}
+
+void write_binary_file(const std::string& path, const std::string& content) {
+    FILE* f = fopen(path.c_str(), "wb");
+    if (!f) {
+        std::cerr << "Error opening file for writing: " << path << " (" << strerror(errno) << ")\n";
+        return;
+    }
+
+    size_t write_size = fwrite(content.data(), 1, content.size(), f);
+    fclose(f);
+    if (write_size != content.size()) {
+        std::cerr << "Error writing file: " << path << " (" << strerror(errno) << ")\n";
+        return;
+    }
+}
+
+void write_file_if_changed(const std::string& path, const std::string& content) {
+    std::string existing = read_binary_file(path, true);
+    if (existing != content) {
+        write_binary_file(path, content);
+    }
+}
+
+
 // variables to track number of compiles in progress
 static uint32_t compile_count = 0;
 static std::mutex compile_count_mutex;
 static std::condition_variable compile_count_cond;
+static bool generate_dep_file = true;
 
-void string_to_spv_func(const std::string& _name, const std::string& in_fname, const std::map<std::string, std::string>& defines, bool fp16 = true, bool coopmat = false, bool coopmat2 = false, bool f16acc = false) {
-    std::string name = _name + (f16acc ? "_f16acc" : "") + (coopmat ? "_cm1" : "") + (coopmat2 ? "_cm2" : (fp16 ? "" : "_fp32"));
-    std::string out_fname = join_paths(output_dir, name + ".spv");
-    std::string in_path = join_paths(input_dir, in_fname);
+void decrement_compile_count(uint32_t * count) {
+    if (count) {
+        std::lock_guard<std::mutex> guard(compile_count_mutex);
+        assert(compile_count > 0);
+        compile_count--;
+        compile_count_cond.notify_all();
+    }
+}
 
+using compile_count_guard = std::unique_ptr<uint32_t, decltype(&decrement_compile_count)>;
+
+compile_count_guard acquire_compile_slot() {
+    // wait until fewer than N compiles are in progress.
+    // 16 is an arbitrary limit, the goal is to avoid "failed to create pipe" errors.
+    uint32_t N = 16;
+    std::unique_lock<std::mutex> guard(compile_count_mutex);
+    compile_count_cond.wait(guard, [N] { return compile_count < N; });
+    compile_count++;
+    return compile_count_guard(&compile_count, &decrement_compile_count);
+}
+
+void string_to_spv_func(std::string name, std::string in_path, std::string out_path, std::map<std::string, std::string> defines, bool coopmat, bool dep_file, compile_count_guard slot) {
     std::string target_env = (name.find("_cm2") != std::string::npos) ? "--target-env=vulkan1.3" : "--target-env=vulkan1.2";
 
     // disable spirv-opt for coopmat shaders for https://github.com/ggerganov/llama.cpp/issues/10734
@@ -249,11 +321,17 @@ void string_to_spv_func(const std::string& _name, const std::string& in_fname, c
     std::string opt_level = (coopmat || name.find("bf16") != std::string::npos) ? "" : "-O";
 
     #ifdef _WIN32
-        std::vector<std::string> cmd = {GLSLC, "-fshader-stage=compute", target_env, opt_level, "\"" + in_path + "\"", "-o", "\"" + out_fname + "\""};
+        std::vector<std::string> cmd = {GLSLC, "-fshader-stage=compute", target_env, opt_level, "\"" + in_path + "\"", "-o", "\"" + out_path + "\""};
     #else
-        std::vector<std::string> cmd = {GLSLC, "-fshader-stage=compute", target_env, opt_level, in_path, "-o",  out_fname};
+        std::vector<std::string> cmd = {GLSLC, "-fshader-stage=compute", target_env, opt_level, in_path, "-o",  out_path};
     #endif
 
+    if (dep_file) {
+        cmd.push_back("-MD");
+        cmd.push_back("-MF");
+        cmd.push_back("\"" + target_cpp + ".d\"");
+    }
+
     #ifdef GGML_VULKAN_SHADER_DEBUG_INFO
         cmd.push_back("-g");
     #endif
@@ -281,17 +359,23 @@ void string_to_spv_func(const std::string& _name, const std::string& in_fname, c
             return;
         }
 
+        if (dep_file) {
+            // replace .spv output path with the embed .cpp path which is used as output in CMakeLists.txt
+            std::string dep = read_binary_file(target_cpp + ".d", true);
+            if (!dep.empty()) {
+                size_t pos = dep.find(out_path);
+                if (pos != std::string::npos) {
+                    dep.replace(pos, out_path.length(), target_cpp);
+                }
+                write_binary_file(target_cpp + ".d", dep);
+            }
+        }
+
         std::lock_guard<std::mutex> guard(lock);
-        shader_fnames.push_back(std::make_pair(name, out_fname));
+        shader_fnames.push_back(std::make_pair(name, out_path));
     } catch (const std::exception& e) {
         std::cerr << "Error executing command for " << name << ": " << e.what() << std::endl;
     }
-    {
-        std::lock_guard<std::mutex> guard(compile_count_mutex);
-        assert(compile_count > 0);
-        compile_count--;
-    }
-    compile_count_cond.notify_all();
 }
 
 std::map<std::string, std::string> merge_maps(const std::map<std::string, std::string>& a, const std::map<std::string, std::string>& b) {
@@ -301,18 +385,24 @@ std::map<std::string, std::string> merge_maps(const std::map<std::string, std::s
 }
 
 static std::vector<std::future<void>> compiles;
-void string_to_spv(const std::string& _name, const std::string& in_fname, const std::map<std::string, std::string>& defines, bool fp16 = true, bool coopmat = false, bool coopmat2 = false, bool f16acc = false) {
-    {
-        // wait until fewer than N compiles are in progress.
-        // 16 is an arbitrary limit, the goal is to avoid "failed to create pipe" errors.
-        uint32_t N = 16;
-        std::unique_lock<std::mutex> guard(compile_count_mutex);
-        while (compile_count >= N) {
-            compile_count_cond.wait(guard);
-        }
-        compile_count++;
+void string_to_spv(std::string name, const std::string& source, const std::map<std::string, std::string>& defines, bool fp16 = true, bool coopmat = false, bool coopmat2 = false, bool f16acc = false) {
+    name = name + (f16acc ? "_f16acc" : "") + (coopmat ? "_cm1" : "") + (coopmat2 ? "_cm2" : (fp16 ? "" : "_fp32"));
+    std::string out_path = join_paths(output_dir, name + ".spv");
+
+    if (input_filepath == "") {
+        // No input source to compile, only generate header for all shaders
+        shader_fnames.push_back(std::pair(name, out_path));
+        return;
+    } else if (basename(input_filepath) != source) {
+        // Only compile shader variants matching the input filename
+        return;
     }
-    compiles.push_back(std::async(string_to_spv_func, _name, in_fname, defines, fp16, coopmat, coopmat2, f16acc));
+
+    compile_count_guard slot = acquire_compile_slot();
+    compiles.push_back(std::async(
+        string_to_spv_func, name, input_filepath, out_path, defines, coopmat, generate_dep_file, std::move(slot)));
+    // Don't write the same dep file from multiple processes
+    generate_dep_file = false;
 }
 
 void matmul_shaders(bool fp16, MatMulIdType matmul_id_type, bool coopmat, bool coopmat2, bool f16acc) {
@@ -485,7 +575,6 @@ void matmul_shaders(bool fp16, MatMulIdType matmul_id_type, bool coopmat, bool c
 }
 
 void process_shaders() {
-    std::cout << "ggml_vulkan: Generating and compiling shaders to SPIR-V" << std::endl;
     std::map<std::string, std::string> base_dict = {{"FLOAT_TYPE", "float"}};
 
     // matmul
@@ -837,11 +926,11 @@ void process_shaders() {
 }
 
 void write_output_files() {
-    FILE* hdr = fopen(target_hpp.c_str(), "w");
-    FILE* src = fopen(target_cpp.c_str(), "w");
+    std::stringstream hdr = make_generic_stringstream();
+    std::stringstream src = make_generic_stringstream();
 
-    fprintf(hdr, "#include <cstdint>\n\n");
-    fprintf(src, "#include \"%s\"\n\n", basename(target_hpp).c_str());
+    hdr << "#include <cstdint>\n\n";
+    src << "#include \"" << basename(target_hpp) << "\"\n\n";
 
     std::sort(shader_fnames.begin(), shader_fnames.end());
     for (const auto& pair : shader_fnames) {
@@ -853,91 +942,85 @@ void write_output_files() {
             const std::string& path = pair.second;
         #endif
 
-        FILE* spv = fopen(path.c_str(), "rb");
-        if (!spv) {
-            std::cerr << "Error opening SPIR-V file: " << path << " (" << strerror(errno) << ")\n";
-            continue;
-        }
+        hdr << "extern const uint64_t " << name << "_len;\n";
+        hdr << "extern const unsigned char " << name << "_data[];\n\n";
 
-        fseek(spv, 0, SEEK_END);
-        size_t size = ftell(spv);
-        fseek(spv, 0, SEEK_SET);
+        if (input_filepath != "") {
+            std::string data = read_binary_file(path);
+            if (data.empty()) {
+                continue;
+            }
 
-        std::vector<unsigned char> data(size);
-        size_t read_size = fread(data.data(), 1, size, spv);
-        fclose(spv);
-        if (read_size != size) {
-            std::cerr << "Error reading SPIR-V file: " << path << " (" << strerror(errno) << ")\n";
-            continue;
-        }
-
-        fprintf(hdr, "extern unsigned char %s_data[%zu];\n", name.c_str(), size);
-        fprintf(hdr, "const uint64_t %s_len = %zu;\n\n", name.c_str(), size);
-
-        fprintf(src, "unsigned char %s_data[%zu] = {\n", name.c_str(), size);
-        for (size_t i = 0; i < size; ++i) {
-            fprintf(src, "0x%02x,", data[i]);
-            if ((i + 1) % 12 == 0) fprintf(src, "\n");
-        }
-        fprintf(src, "\n};\n\n");
-
-        if (!no_clean) {
-            std::remove(path.c_str());
+            src << "const uint64_t " << name << "_len = " << data.size() << ";\n";
+            src << "const unsigned char " << name << "_data[" << data.size() << "] = {\n" << std::hex;
+            auto bytes = reinterpret_cast<const uint8_t*>(data.data());
+            for (size_t i = 0; i < data.size(); ++i) {
+                src << "0x" << static_cast<int>(bytes[i]) << ",";
+                if ((i + 1) % 12 == 0) src << "\n";
+            }
+            src << std::dec << "\n};\n\n";
         }
     }
 
     std::string suffixes[2] = {"_f32", "_f16"};
-    for (const char *op : {"add", "sub", "mul", "div", "add_rms"}) {
-        fprintf(hdr, "extern unsigned char *%s_data[2][2][2][2];\n", op);
-        fprintf(hdr, "extern uint64_t %s_len[2][2][2][2];\n", op);
-        std::string data = "unsigned char *" + std::string(op) + "_data[2][2][2][2] = ";
-        std::string len = "uint64_t " + std::string(op) + "_len[2][2][2][2] = ";
+    for (auto op : {"add", "sub", "mul", "div", "add_rms"}) {
+        hdr << "extern const void * " << op << "_data[2][2][2][2];\n";
+        hdr << "extern const uint64_t " << op << "_len[2][2][2][2];\n";
+
+        std::string op_file = op == "add_rms" ? "add.comp" : std::string(op) + ".comp";
+        if (basename(input_filepath) != op_file) {
+            continue;
+        }
+        std::stringstream data = make_generic_stringstream();
+        std::stringstream len  = make_generic_stringstream();
+        data << "const void * " << op << "_data[2][2][2][2] = ";
+        len  << "const uint64_t " << op << "_len[2][2][2][2] = ";
         for (uint32_t t0 = 0; t0 < 2; ++t0) {
             if (t0 == 0) {
-                data += "{";
-                len += "{";
+                data << "{";
+                len  << "{";
             }
             for (uint32_t t1 = 0; t1 < 2; ++t1) {
                 if (t1 == 0) {
-                    data += "{";
-                    len += "{";
+                    data << "{";
+                    len  << "{";
                 }
                 for (uint32_t t2 = 0; t2 < 2; ++t2) {
                     if (t2 == 0) {
-                        data += "{";
-                        len += "{";
+                        data << "{";
+                        len  << "{";
                     }
                     for (uint32_t rte = 0; rte < 2; ++rte) {
                         if (rte == 0) {
-                            data += "{";
-                            len += "{";
+                            data << "{";
+                            len  << "{";
                         }
-                        data += op + suffixes[t0] + suffixes[t1] + suffixes[t2] + ((rte != 0) ? "_rte" : "");
-                        len  += op + suffixes[t0] + suffixes[t1] + suffixes[t2] + ((rte != 0) ? "_rte" : "");
-                        data += "_data,";
-                        len  += "_len,";
+                        data << op << suffixes[t0] << suffixes[t1] << suffixes[t2] << ((rte != 0) ? "_rte" : "");
+                        len  << op << suffixes[t0] << suffixes[t1] << suffixes[t2] << ((rte != 0) ? "_rte" : "");
+                        data << "_data,";
+                        len  << "_len,";
                         if (rte == 1) {
-                            data += "}, ";
-                            len += "}, ";
+                            data << "}, ";
+                            len  << "}, ";
                         }
                     }
                     if (t2 == 1) {
-                        data += "}, ";
-                        len += "}, ";
+                        data << "}, ";
+                        len  << "}, ";
                     }
                 }
                 if (t1 == 1) {
-                    data += "}, ";
-                    len += "}, ";
+                    data << "}, ";
+                    len  << "}, ";
                 }
             }
             if (t0 == 1) {
-                data += "};\n";
-                len += "};\n";
+                data << "};\n";
+                len  << "};\n";
             }
         }
-        fputs(data.c_str(), src);
-        fputs(len.c_str(), src);
+        src << data.str();
+        src << len.str();
     }
 
     std::vector<std::string> btypes = {"f16", "f32"};
@@ -951,20 +1034,25 @@ void write_output_files() {
         if (btype == "q8_1" && !is_legacy_quant(tname)) {
             continue;
         }
-        fprintf(hdr, "extern unsigned char *arr_dmmv_%s_%s_f32_data[3];\n", tname.c_str(), btype.c_str());
-        fprintf(hdr, "extern uint64_t arr_dmmv_%s_%s_f32_len[3];\n", tname.c_str(), btype.c_str());
-        std::string data = "unsigned char *arr_dmmv_" + tname + "_" + btype + "_f32_data[3] = {mul_mat_vec_" + tname + "_" + btype + "_f32_data, mul_mat_vec_" + tname + "_" + btype + "_f32_subgroup_data, mul_mat_vec_" + tname + "_" + btype + "_f32_subgroup_no_shmem_data};\n";
-        std::string len =  "uint64_t arr_dmmv_"       + tname + "_" + btype + "_f32_len[3] =  {mul_mat_vec_" + tname + "_" + btype + "_f32_len,  mul_mat_vec_" + tname + "_" + btype + "_f32_subgroup_len, mul_mat_vec_" + tname + "_" + btype + "_f32_subgroup_no_shmem_len};\n";
-        fputs(data.c_str(), src);
-        fputs(len.c_str(), src);
+        hdr << "extern const void * arr_dmmv_"   << tname << "_" << btype << "_f32_data[3];\n";
+        hdr << "extern const uint64_t arr_dmmv_" << tname << "_" << btype << "_f32_len[3];\n";
+        if (basename(input_filepath) == "mul_mat_vec.comp") {
+            src << "const void * arr_dmmv_"   << tname << "_" << btype << "_f32_data[3] = {mul_mat_vec_" << tname << "_" << btype << "_f32_data, mul_mat_vec_" << tname << "_" << btype << "_f32_subgroup_data, mul_mat_vec_" << tname << "_" << btype << "_f32_subgroup_no_shmem_data};\n";
+            src << "const uint64_t arr_dmmv_" << tname << "_" << btype << "_f32_len[3] =  {mul_mat_vec_" << tname << "_" << btype << "_f32_len,  mul_mat_vec_" << tname << "_" << btype << "_f32_subgroup_len, mul_mat_vec_"  << tname << "_" << btype << "_f32_subgroup_no_shmem_len};\n";
+        }
     }
     }
 
-    fclose(hdr);
-    fclose(src);
-}
+    if (input_filepath == "") {
+        write_file_if_changed(target_hpp, hdr.str());
+    }
+    if (target_cpp != "") {
+        write_binary_file(target_cpp, src.str());
+    }
 }
 
+} // namespace
+
 int main(int argc, char** argv) {
     std::map<std::string, std::string> args;
     for (int i = 1; i < argc; ++i) {
@@ -982,8 +1070,8 @@ int main(int argc, char** argv) {
     if (args.find("--glslc") != args.end()) {
         GLSLC = args["--glslc"]; // Path to glslc
     }
-    if (args.find("--input-dir") != args.end()) {
-        input_dir = args["--input-dir"]; // Directory containing shader sources
+    if (args.find("--source") != args.end()) {
+        input_filepath = args["--source"]; // The shader source file to compile
     }
     if (args.find("--output-dir") != args.end()) {
         output_dir = args["--output-dir"]; // Directory for containing SPIR-V output
@@ -994,14 +1082,6 @@ int main(int argc, char** argv) {
     if (args.find("--target-cpp") != args.end()) {
         target_cpp = args["--target-cpp"]; // Path to generated cpp file
     }
-    if (args.find("--no-clean") != args.end()) {
-        no_clean = true; // Keep temporary SPIR-V files in output-dir after build
-    }
-
-    if (!directory_exists(input_dir)) {
-        std::cerr << "\"" << input_dir << "\" must be a valid directory containing shader sources" << std::endl;
-        return EXIT_FAILURE;
-    }
 
     if (!directory_exists(output_dir)) {
         if (!create_directory(output_dir)) {

From 898acba6816ad23b6a9491347d30e7570bffadfd Mon Sep 17 00:00:00 2001
From: Radoslav Gerganov <rgerganov@gmail.com>
Date: Sat, 4 Oct 2025 12:49:16 +0300
Subject: [PATCH 2/4] rpc : add support for multiple devices (#16276)

* rpc : add support for multiple devices

Allow rpc-server to expose multiple devices from a single endpoint.
Change RPC protocol to include device identifier where needed.

closes: #15210

* fixes

* use ggml_backend_reg_t

* address review comments

* fix llama-bench backend report

* address review comments, change device naming

* fix cmd order
---
 common/arg.cpp                    |  16 +-
 ggml/include/ggml-backend.h       |   2 +
 ggml/include/ggml-rpc.h           |  17 +-
 ggml/src/ggml-backend-impl.h      |   3 -
 ggml/src/ggml-rpc/ggml-rpc.cpp    | 401 +++++++++++++++++++++---------
 tools/llama-bench/llama-bench.cpp |  51 ++--
 tools/rpc/rpc-server.cpp          | 158 ++++++------
 7 files changed, 403 insertions(+), 245 deletions(-)

diff --git a/common/arg.cpp b/common/arg.cpp
index 577048c20..a020ac441 100644
--- a/common/arg.cpp
+++ b/common/arg.cpp
@@ -1615,18 +1615,14 @@ static void add_rpc_devices(const std::string & servers) {
     if (!rpc_reg) {
         throw std::invalid_argument("failed to find RPC backend");
     }
-    typedef ggml_backend_dev_t (*ggml_backend_rpc_add_device_t)(const char * endpoint);
-    ggml_backend_rpc_add_device_t ggml_backend_rpc_add_device_fn = (ggml_backend_rpc_add_device_t) ggml_backend_reg_get_proc_address(rpc_reg, "ggml_backend_rpc_add_device");
-    if (!ggml_backend_rpc_add_device_fn) {
-        throw std::invalid_argument("failed to find RPC device add function");
+    typedef ggml_backend_reg_t (*ggml_backend_rpc_add_server_t)(const char * endpoint);
+    ggml_backend_rpc_add_server_t ggml_backend_rpc_add_server_fn = (ggml_backend_rpc_add_server_t) ggml_backend_reg_get_proc_address(rpc_reg, "ggml_backend_rpc_add_server");
+    if (!ggml_backend_rpc_add_server_fn) {
+        throw std::invalid_argument("failed to find RPC add server function");
     }
     for (const auto & server : rpc_servers) {
-        ggml_backend_dev_t dev = ggml_backend_rpc_add_device_fn(server.c_str());
-        if (dev) {
-            ggml_backend_device_register(dev);
-        } else {
-            throw std::invalid_argument("failed to register RPC device");
-        }
+        auto reg = ggml_backend_rpc_add_server_fn(server.c_str());
+        ggml_backend_register(reg);
     }
 }
 
diff --git a/ggml/include/ggml-backend.h b/ggml/include/ggml-backend.h
index 62b6d65e5..f1b740785 100644
--- a/ggml/include/ggml-backend.h
+++ b/ggml/include/ggml-backend.h
@@ -215,6 +215,8 @@ extern "C" {
     // Backend registry
     //
 
+    GGML_API void ggml_backend_register(ggml_backend_reg_t reg);
+
     GGML_API void ggml_backend_device_register(ggml_backend_dev_t device);
 
     // Backend (reg) enumeration
diff --git a/ggml/include/ggml-rpc.h b/ggml/include/ggml-rpc.h
index 1e6741127..72eff0027 100644
--- a/ggml/include/ggml-rpc.h
+++ b/ggml/include/ggml-rpc.h
@@ -7,26 +7,25 @@
 extern "C" {
 #endif
 
-#define RPC_PROTO_MAJOR_VERSION    2
+#define RPC_PROTO_MAJOR_VERSION    3
 #define RPC_PROTO_MINOR_VERSION    0
 #define RPC_PROTO_PATCH_VERSION    0
 #define GGML_RPC_MAX_SERVERS       16
 
 // backend API
-GGML_BACKEND_API ggml_backend_t ggml_backend_rpc_init(const char * endpoint);
+GGML_BACKEND_API ggml_backend_t ggml_backend_rpc_init(const char * endpoint, uint32_t device);
 GGML_BACKEND_API bool ggml_backend_is_rpc(ggml_backend_t backend);
 
-GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint);
+GGML_BACKEND_API ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint, uint32_t device);
 
-GGML_BACKEND_API void ggml_backend_rpc_get_device_memory(const char * endpoint, size_t * free, size_t * total);
+GGML_BACKEND_API void ggml_backend_rpc_get_device_memory(const char * endpoint, uint32_t device, size_t * free, size_t * total);
 
-GGML_BACKEND_API void ggml_backend_rpc_start_server(ggml_backend_t backend, const char * endpoint,
-                                                    const char * cache_dir,
-                                                    size_t free_mem, size_t total_mem);
+GGML_BACKEND_API void ggml_backend_rpc_start_server(const char * endpoint, const char * cache_dir,
+                                                    size_t n_threads, size_t n_devices,
+                                                    ggml_backend_dev_t * devices, size_t * free_mem, size_t * total_mem);
 
 GGML_BACKEND_API ggml_backend_reg_t ggml_backend_rpc_reg(void);
-
-GGML_BACKEND_API ggml_backend_dev_t ggml_backend_rpc_add_device(const char * endpoint);
+GGML_BACKEND_API ggml_backend_reg_t ggml_backend_rpc_add_server(const char * endpoint);
 
 #ifdef  __cplusplus
 }
diff --git a/ggml/src/ggml-backend-impl.h b/ggml/src/ggml-backend-impl.h
index 07784d6f6..6792ba986 100644
--- a/ggml/src/ggml-backend-impl.h
+++ b/ggml/src/ggml-backend-impl.h
@@ -209,9 +209,6 @@ extern "C" {
         void * context;
     };
 
-    // Internal backend registry API
-    GGML_API void ggml_backend_register(ggml_backend_reg_t reg);
-
     // Add backend dynamic loading support to the backend
 
     // Initialize the backend
diff --git a/ggml/src/ggml-rpc/ggml-rpc.cpp b/ggml/src/ggml-rpc/ggml-rpc.cpp
index f99681c84..1a8739e78 100644
--- a/ggml/src/ggml-rpc/ggml-rpc.cpp
+++ b/ggml/src/ggml-rpc/ggml-rpc.cpp
@@ -105,9 +105,12 @@ enum rpc_cmd {
     RPC_CMD_INIT_TENSOR,
     RPC_CMD_GET_ALLOC_SIZE,
     RPC_CMD_HELLO,
+    RPC_CMD_DEVICE_COUNT,
     RPC_CMD_COUNT,
 };
 
+static_assert(RPC_CMD_HELLO == 14, "RPC_CMD_HELLO must be always 14");
+
 // Try RPC_CMD_SET_TENSOR_HASH first when data size is larger than this threshold
 const size_t HASH_THRESHOLD = 10 * 1024 * 1024;
 
@@ -117,7 +120,12 @@ struct rpc_msg_hello_rsp {
     uint8_t patch;
 };
 
+struct rpc_msg_device_count_rsp {
+    uint32_t device_count;
+};
+
 struct rpc_msg_get_alloc_size_req {
+    uint32_t   device;
     rpc_tensor tensor;
 };
 
@@ -130,6 +138,7 @@ struct rpc_msg_init_tensor_req {
 };
 
 struct rpc_msg_alloc_buffer_req {
+    uint32_t device;
     uint64_t size;
 };
 
@@ -138,10 +147,18 @@ struct rpc_msg_alloc_buffer_rsp {
     uint64_t remote_size;
 };
 
+struct rpc_msg_get_alignment_req {
+    uint32_t device;
+};
+
 struct rpc_msg_get_alignment_rsp {
     uint64_t alignment;
 };
 
+struct rpc_msg_get_max_size_req {
+    uint32_t device;
+};
+
 struct rpc_msg_get_max_size_rsp {
     uint64_t max_size;
 };
@@ -192,6 +209,10 @@ struct rpc_msg_graph_compute_rsp {
     uint8_t result;
 };
 
+struct rpc_msg_get_device_memory_req {
+    uint32_t device;
+};
+
 struct rpc_msg_get_device_memory_rsp {
     uint64_t free_mem;
     uint64_t total_mem;
@@ -207,13 +228,15 @@ static ggml_guid_t ggml_backend_rpc_guid() {
 
 struct ggml_backend_rpc_buffer_type_context {
     std::string endpoint;
+    uint32_t    device;
     std::string name;
-    size_t alignment;
-    size_t max_size;
+    size_t      alignment;
+    size_t      max_size;
 };
 
 struct ggml_backend_rpc_context {
     std::string endpoint;
+    uint32_t    device;
     std::string name;
 };
 
@@ -653,7 +676,7 @@ static const char * ggml_backend_rpc_buffer_type_name(ggml_backend_buffer_type_t
 
 static ggml_backend_buffer_t ggml_backend_rpc_buffer_type_alloc_buffer(ggml_backend_buffer_type_t buft, size_t size) {
     ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)buft->context;
-    rpc_msg_alloc_buffer_req request = {size};
+    rpc_msg_alloc_buffer_req request = {buft_ctx->device, size};
     rpc_msg_alloc_buffer_rsp response;
     auto sock = get_socket(buft_ctx->endpoint);
     bool status = send_rpc_cmd(sock, RPC_CMD_ALLOC_BUFFER, &request, sizeof(request), &response, sizeof(response));
@@ -669,9 +692,10 @@ static ggml_backend_buffer_t ggml_backend_rpc_buffer_type_alloc_buffer(ggml_back
     }
 }
 
-static size_t get_alignment(const std::shared_ptr<socket_t> & sock) {
+static size_t get_alignment(const std::shared_ptr<socket_t> & sock, uint32_t device) {
+    rpc_msg_get_alignment_req request = {device};
     rpc_msg_get_alignment_rsp response;
-    bool status = send_rpc_cmd(sock, RPC_CMD_GET_ALIGNMENT, nullptr, 0, &response, sizeof(response));
+    bool status = send_rpc_cmd(sock, RPC_CMD_GET_ALIGNMENT, &request, sizeof(request), &response, sizeof(response));
     RPC_STATUS_ASSERT(status);
     return response.alignment;
 }
@@ -681,9 +705,10 @@ static size_t ggml_backend_rpc_buffer_type_get_alignment(ggml_backend_buffer_typ
     return buft_ctx->alignment;
 }
 
-static size_t get_max_size(const std::shared_ptr<socket_t> & sock) {
+static size_t get_max_size(const std::shared_ptr<socket_t> & sock, uint32_t device) {
+    rpc_msg_get_max_size_req request = {device};
     rpc_msg_get_max_size_rsp response;
-    bool status = send_rpc_cmd(sock, RPC_CMD_GET_MAX_SIZE, nullptr, 0, &response, sizeof(response));
+    bool status = send_rpc_cmd(sock, RPC_CMD_GET_MAX_SIZE, &request, sizeof(request), &response, sizeof(response));
     RPC_STATUS_ASSERT(status);
     return response.max_size;
 }
@@ -700,7 +725,7 @@ static size_t ggml_backend_rpc_buffer_type_get_alloc_size(ggml_backend_buffer_ty
         auto sock = get_socket(buft_ctx->endpoint);
 
         rpc_msg_get_alloc_size_req request;
-
+        request.device = buft_ctx->device;
         request.tensor = serialize_tensor(tensor);
 
         rpc_msg_get_alloc_size_rsp response;
@@ -754,7 +779,7 @@ static void add_tensor(ggml_tensor * tensor, std::vector<rpc_tensor> & tensors,
     tensors.push_back(serialize_tensor(tensor));
 }
 
-static void serialize_graph(const ggml_cgraph * cgraph, std::vector<uint8_t> & output) {
+static void serialize_graph(uint32_t device, const ggml_cgraph * cgraph, std::vector<uint8_t> & output) {
     uint32_t n_nodes = cgraph->n_nodes;
     std::vector<rpc_tensor> tensors;
     std::unordered_set<ggml_tensor*> visited;
@@ -762,24 +787,29 @@ static void serialize_graph(const ggml_cgraph * cgraph, std::vector<uint8_t> & o
         add_tensor(cgraph->nodes[i], tensors, visited);
     }
     // serialization format:
-    // | n_nodes (4 bytes) | nodes (n_nodes * sizeof(uint64_t) | n_tensors (4 bytes) | tensors (n_tensors * sizeof(rpc_tensor)) |
+    // | device (4 bytes) | n_nodes (4 bytes) | nodes (n_nodes * sizeof(uint64_t) | n_tensors (4 bytes) | tensors (n_tensors * sizeof(rpc_tensor)) |
     uint32_t n_tensors = tensors.size();
-    int output_size = sizeof(uint32_t) + n_nodes * sizeof(uint64_t) + sizeof(uint32_t) + n_tensors * sizeof(rpc_tensor);
+    int output_size = 2*sizeof(uint32_t) + n_nodes * sizeof(uint64_t) + sizeof(uint32_t) + n_tensors * sizeof(rpc_tensor);
     output.resize(output_size, 0);
-    memcpy(output.data(), &n_nodes, sizeof(n_nodes));
+    uint8_t * dest = output.data();
+    memcpy(dest, &device, sizeof(device));
+    dest += sizeof(device);
+    memcpy(dest, &n_nodes, sizeof(n_nodes));
+    dest += sizeof(n_nodes);
     for (uint32_t i = 0; i < n_nodes; i++) {
-        memcpy(output.data() + sizeof(n_nodes) + i * sizeof(uint64_t), &cgraph->nodes[i], sizeof(uint64_t));
+        memcpy(dest + i * sizeof(uint64_t), &cgraph->nodes[i], sizeof(uint64_t));
     }
-    uint32_t * out_ntensors = (uint32_t *)(output.data() + sizeof(n_nodes) + n_nodes * sizeof(uint64_t));
-    *out_ntensors = n_tensors;
-    rpc_tensor * out_tensors = (rpc_tensor *)(output.data() + sizeof(n_nodes) + n_nodes * sizeof(uint64_t) + sizeof(uint32_t));
+    dest += n_nodes * sizeof(uint64_t);
+    memcpy(dest, &n_tensors, sizeof(n_tensors));
+    dest += sizeof(n_tensors);
+    rpc_tensor * out_tensors = (rpc_tensor *)dest;
     memcpy(out_tensors, tensors.data(), n_tensors * sizeof(rpc_tensor));
 }
 
 static enum ggml_status ggml_backend_rpc_graph_compute(ggml_backend_t backend, ggml_cgraph * cgraph) {
     ggml_backend_rpc_context * rpc_ctx = (ggml_backend_rpc_context *)backend->context;
     std::vector<uint8_t> input;
-    serialize_graph(cgraph, input);
+    serialize_graph(rpc_ctx->device, cgraph, input);
     rpc_msg_graph_compute_rsp response;
     auto sock = get_socket(rpc_ctx->endpoint);
     bool status = send_rpc_cmd(sock, RPC_CMD_GRAPH_COMPUTE, input.data(), input.size(), &response, sizeof(response));
@@ -804,12 +834,13 @@ static ggml_backend_i ggml_backend_rpc_interface = {
     /* .graph_optimize          = */ NULL,
 };
 
-ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint) {
+ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint, uint32_t device) {
     static std::mutex mutex;
     std::lock_guard<std::mutex> lock(mutex);
+    std::string buft_name = "RPC" + std::to_string(device) + "[" + std::string(endpoint) + "]";
     // NOTE: buffer types are allocated and never freed; this is by design
     static std::unordered_map<std::string, ggml_backend_buffer_type_t> buft_map;
-    auto it = buft_map.find(endpoint);
+    auto it = buft_map.find(buft_name);
     if (it != buft_map.end()) {
         return it->second;
     }
@@ -818,34 +849,37 @@ ggml_backend_buffer_type_t ggml_backend_rpc_buffer_type(const char * endpoint) {
         GGML_LOG_ERROR("Failed to connect to %s\n", endpoint);
         return nullptr;
     }
-    size_t alignment = get_alignment(sock);
-    size_t max_size = get_max_size(sock);
+    size_t alignment = get_alignment(sock, device);
+    size_t max_size = get_max_size(sock, device);
     ggml_backend_rpc_buffer_type_context * buft_ctx = new ggml_backend_rpc_buffer_type_context {
         /* .endpoint  = */ endpoint,
-        /* .name      = */ "RPC[" + std::string(endpoint) + "]",
+        /* .device    = */ device,
+        /* .name      = */ buft_name,
         /* .alignment = */ alignment,
         /* .max_size  = */ max_size
     };
-
+    auto reg = ggml_backend_rpc_add_server(endpoint);
     ggml_backend_buffer_type_t buft = new ggml_backend_buffer_type {
         /* .iface   = */ ggml_backend_rpc_buffer_type_interface,
-        /* .device  = */ ggml_backend_rpc_add_device(endpoint),
+        /* .device  = */ ggml_backend_reg_dev_get(reg, device),
         /* .context = */ buft_ctx
     };
-    buft_map[endpoint] = buft;
+    buft_map[buft_name] = buft;
     return buft;
 }
 
-ggml_backend_t ggml_backend_rpc_init(const char * endpoint) {
+ggml_backend_t ggml_backend_rpc_init(const char * endpoint, uint32_t device) {
+    std::string dev_name = "RPC" + std::to_string(device) + "[" + std::string(endpoint) + "]";
     ggml_backend_rpc_context * ctx = new ggml_backend_rpc_context {
-        /* .endpoint  = */ endpoint,
-        /* .name      = */ "RPC[" + std::string(endpoint) + "]",
+        /* .endpoint = */ endpoint,
+        /* .device   = */ device,
+        /* .name     = */ dev_name
     };
-
+    auto reg = ggml_backend_rpc_add_server(endpoint);
     ggml_backend_t backend = new ggml_backend {
         /* .guid    = */ ggml_backend_rpc_guid(),
         /* .iface   = */ ggml_backend_rpc_interface,
-        /* .device  = */ ggml_backend_rpc_add_device(endpoint),
+        /* .device  = */ ggml_backend_reg_dev_get(reg, device),
         /* .context = */ ctx
     };
     return backend;
@@ -855,37 +889,39 @@ bool ggml_backend_is_rpc(ggml_backend_t backend) {
     return backend != NULL && ggml_guid_matches(backend->guid, ggml_backend_rpc_guid());
 }
 
-static void get_device_memory(const std::shared_ptr<socket_t> & sock, size_t * free, size_t * total) {
+static void get_device_memory(const std::shared_ptr<socket_t> & sock, uint32_t device, size_t * free, size_t * total) {
+    rpc_msg_get_device_memory_req request;
+    request.device = device;
     rpc_msg_get_device_memory_rsp response;
-    bool status = send_rpc_cmd(sock, RPC_CMD_GET_DEVICE_MEMORY, nullptr, 0, &response, sizeof(response));
+    bool status = send_rpc_cmd(sock, RPC_CMD_GET_DEVICE_MEMORY, &request, sizeof(request), &response, sizeof(response));
     RPC_STATUS_ASSERT(status);
     *free = response.free_mem;
     *total = response.total_mem;
 }
 
-void ggml_backend_rpc_get_device_memory(const char * endpoint, size_t * free, size_t * total) {
+void ggml_backend_rpc_get_device_memory(const char * endpoint, uint32_t device, size_t * free, size_t * total) {
     auto sock = get_socket(endpoint);
     if (sock == nullptr) {
         *free = 0;
         *total = 0;
         return;
     }
-    get_device_memory(sock, free, total);
+    get_device_memory(sock, device, free, total);
 }
 
 // RPC server-side implementation
 
 class rpc_server {
 public:
-    rpc_server(ggml_backend_t backend, const char * cache_dir)
-        : backend(backend), cache_dir(cache_dir) {
+    rpc_server(std::vector<ggml_backend_t> backends, const char * cache_dir)
+        : backends(std::move(backends)), cache_dir(cache_dir) {
     }
     ~rpc_server();
 
     void hello(rpc_msg_hello_rsp & response);
-    void alloc_buffer(const rpc_msg_alloc_buffer_req & request, rpc_msg_alloc_buffer_rsp & response);
-    void get_alignment(rpc_msg_get_alignment_rsp & response);
-    void get_max_size(rpc_msg_get_max_size_rsp & response);
+    bool alloc_buffer(const rpc_msg_alloc_buffer_req & request, rpc_msg_alloc_buffer_rsp & response);
+    bool get_alignment(const rpc_msg_get_alignment_req & request, rpc_msg_get_alignment_rsp & response);
+    bool get_max_size(const rpc_msg_get_max_size_req & request, rpc_msg_get_max_size_rsp & response);
     bool buffer_get_base(const rpc_msg_buffer_get_base_req & request, rpc_msg_buffer_get_base_rsp & response);
     bool free_buffer(const rpc_msg_free_buffer_req & request);
     bool buffer_clear(const rpc_msg_buffer_clear_req & request);
@@ -906,7 +942,7 @@ private:
                               std::unordered_map<uint64_t, struct ggml_tensor*> & tensor_map);
 
 
-    ggml_backend_t backend;
+    std::vector<ggml_backend_t> backends;
     const char * cache_dir;
     std::unordered_set<ggml_backend_buffer_t> buffers;
 };
@@ -919,6 +955,10 @@ void rpc_server::hello(rpc_msg_hello_rsp & response) {
 }
 
 bool rpc_server::get_alloc_size(const rpc_msg_get_alloc_size_req & request, rpc_msg_get_alloc_size_rsp & response) {
+    uint32_t dev_id = request.device;
+    if (dev_id >= backends.size()) {
+        return false;
+    }
     ggml_backend_buffer_type_t buft;
     struct ggml_init_params params {
         /*.mem_size   =*/ ggml_tensor_overhead(),
@@ -935,10 +975,10 @@ bool rpc_server::get_alloc_size(const rpc_msg_get_alloc_size_req & request, rpc_
         GGML_LOG_ERROR("Null tensor pointer passed to server get_alloc_size function.\n");
         return false;
     }
-    LOG_DBG("[%s] buffer: %p, data: %p\n", __func__, (void*)tensor->buffer, tensor->data);
+    LOG_DBG("[%s] device: %d, buffer: %p, data: %p\n", __func__, dev_id, (void*)tensor->buffer, tensor->data);
     if (tensor->buffer == nullptr) {
         //No buffer allocated.
-        buft = ggml_backend_get_default_buffer_type(backend);
+        buft = ggml_backend_get_default_buffer_type(backends[dev_id]);
     } else {
         buft = tensor->buffer->buft;
     }
@@ -948,33 +988,49 @@ bool rpc_server::get_alloc_size(const rpc_msg_get_alloc_size_req & request, rpc_
     return true;
 }
 
-void rpc_server::alloc_buffer(const rpc_msg_alloc_buffer_req & request, rpc_msg_alloc_buffer_rsp & response) {
-    ggml_backend_buffer_type_t buft = ggml_backend_get_default_buffer_type(backend);
+bool rpc_server::alloc_buffer(const rpc_msg_alloc_buffer_req & request, rpc_msg_alloc_buffer_rsp & response) {
+    uint32_t dev_id = request.device;
+    if (dev_id >= backends.size()) {
+        return false;
+    }
+    ggml_backend_buffer_type_t buft = ggml_backend_get_default_buffer_type(backends[dev_id]);
     ggml_backend_buffer_t buffer = ggml_backend_buft_alloc_buffer(buft, request.size);
     response.remote_ptr = 0;
     response.remote_size = 0;
     if (buffer != nullptr) {
         response.remote_ptr = reinterpret_cast<uint64_t>(buffer);
         response.remote_size = buffer->size;
-        LOG_DBG("[%s] size: %" PRIu64 " -> remote_ptr: %" PRIx64 ", remote_size: %" PRIu64 "\n", __func__, request.size, response.remote_ptr, response.remote_size);
+        LOG_DBG("[%s] device: %d, size: %" PRIu64 " -> remote_ptr: %" PRIx64 ", remote_size: %" PRIu64 "\n",
+            __func__, dev_id, request.size, response.remote_ptr, response.remote_size);
         buffers.insert(buffer);
     } else {
-        LOG_DBG("[%s] size: %" PRIu64 " -> failed\n", __func__, request.size);
+        LOG_DBG("[%s] device: %d, size: %" PRIu64 " -> failed\n", __func__, dev_id, request.size);
     }
+    return true;
 }
 
-void rpc_server::get_alignment(rpc_msg_get_alignment_rsp & response) {
-    ggml_backend_buffer_type_t buft = ggml_backend_get_default_buffer_type(backend);
+bool rpc_server::get_alignment(const rpc_msg_get_alignment_req & request, rpc_msg_get_alignment_rsp & response) {
+    uint32_t dev_id = request.device;
+    if (dev_id >= backends.size()) {
+        return false;
+    }
+    ggml_backend_buffer_type_t buft = ggml_backend_get_default_buffer_type(backends[dev_id]);
     size_t alignment = ggml_backend_buft_get_alignment(buft);
-    LOG_DBG("[%s] alignment: %lu\n", __func__, alignment);
+    LOG_DBG("[%s] device: %d, alignment: %lu\n", __func__, dev_id, alignment);
     response.alignment = alignment;
+    return true;
 }
 
-void rpc_server::get_max_size(rpc_msg_get_max_size_rsp & response) {
-    ggml_backend_buffer_type_t buft = ggml_backend_get_default_buffer_type(backend);
+bool rpc_server::get_max_size(const rpc_msg_get_max_size_req & request, rpc_msg_get_max_size_rsp & response) {
+    uint32_t dev_id = request.device;
+    if (dev_id >= backends.size()) {
+        return false;
+    }
+    ggml_backend_buffer_type_t buft = ggml_backend_get_default_buffer_type(backends[dev_id]);
     size_t max_size = ggml_backend_buft_get_max_size(buft);
-    LOG_DBG("[%s] max_size: %lu\n", __func__, max_size);
+    LOG_DBG("[%s] device: %d, max_size: %lu\n", __func__, dev_id, max_size);
     response.max_size = max_size;
+    return true;
 }
 
 bool rpc_server::buffer_get_base(const rpc_msg_buffer_get_base_req & request, rpc_msg_buffer_get_base_rsp & response) {
@@ -1332,23 +1388,33 @@ ggml_tensor * rpc_server::create_node(uint64_t id,
 
 bool rpc_server::graph_compute(const std::vector<uint8_t> & input, rpc_msg_graph_compute_rsp & response) {
     // serialization format:
-    // | n_nodes (4 bytes) | nodes (n_nodes * sizeof(uint64_t) | n_tensors (4 bytes) | tensors (n_tensors * sizeof(rpc_tensor)) |
-    if (input.size() < sizeof(uint32_t)) {
+    // | device (4 bytes) | n_nodes (4 bytes) | nodes (n_nodes * sizeof(uint64_t) | n_tensors (4 bytes) | tensors (n_tensors * sizeof(rpc_tensor)) |
+    if (input.size() < 2*sizeof(uint32_t)) {
+        return false;
+    }
+    const uint8_t * src = input.data();
+    uint32_t device;
+    memcpy(&device, src, sizeof(device));
+    src += sizeof(device);
+    if (device >= backends.size()) {
         return false;
     }
     uint32_t n_nodes;
-    memcpy(&n_nodes, input.data(), sizeof(n_nodes));
-    if (input.size() < sizeof(uint32_t) + n_nodes*sizeof(uint64_t) + sizeof(uint32_t)) {
+    memcpy(&n_nodes, src, sizeof(n_nodes));
+    src += sizeof(n_nodes);
+    if (input.size() < 2*sizeof(uint32_t) + n_nodes*sizeof(uint64_t) + sizeof(uint32_t)) {
         return false;
     }
-    const uint64_t * nodes = (const uint64_t *)(input.data() + sizeof(n_nodes));
+    const uint64_t * nodes = (const uint64_t *)src;
+    src += n_nodes*sizeof(uint64_t);
     uint32_t n_tensors;
-    memcpy(&n_tensors, input.data() + sizeof(n_nodes) + n_nodes*sizeof(uint64_t), sizeof(n_tensors));
-    if (input.size() < sizeof(uint32_t) + n_nodes*sizeof(uint64_t) + sizeof(uint32_t) + n_tensors*sizeof(rpc_tensor)) {
+    memcpy(&n_tensors, src, sizeof(n_tensors));
+    src += sizeof(n_tensors);
+    if (input.size() < 2*sizeof(uint32_t) + n_nodes*sizeof(uint64_t) + sizeof(uint32_t) + n_tensors*sizeof(rpc_tensor)) {
         return false;
     }
-    const rpc_tensor * tensors = (const rpc_tensor *)(input.data() + sizeof(n_nodes) + n_nodes*sizeof(uint64_t) + sizeof(n_tensors));
-    LOG_DBG("[%s] n_nodes: %u, n_tensors: %u\n", __func__, n_nodes, n_tensors);
+    const rpc_tensor * tensors = (const rpc_tensor *)src;
+    LOG_DBG("[%s] device: %u, n_nodes: %u, n_tensors: %u\n", __func__, device, n_nodes, n_tensors);
 
     size_t buf_size = ggml_tensor_overhead()*(n_nodes + n_tensors) + ggml_graph_overhead_custom(n_nodes, false);
 
@@ -1380,7 +1446,7 @@ bool rpc_server::graph_compute(const std::vector<uint8_t> & input, rpc_msg_graph
             return false;
         }
     }
-    ggml_status status = ggml_backend_graph_compute(backend, graph);
+    ggml_status status = ggml_backend_graph_compute(backends[device], graph);
     response.result = status;
     return true;
 }
@@ -1391,9 +1457,9 @@ rpc_server::~rpc_server() {
     }
 }
 
-static void rpc_serve_client(ggml_backend_t backend, const char * cache_dir,
-                             sockfd_t sockfd, size_t free_mem, size_t total_mem) {
-    rpc_server server(backend, cache_dir);
+static void rpc_serve_client(const std::vector<ggml_backend_t> & backends, const char * cache_dir,
+                             sockfd_t sockfd, const std::vector<size_t> & free_mem, const std::vector<size_t> & total_mem) {
+    rpc_server server(backends, cache_dir);
     uint8_t cmd;
     if (!recv_data(sockfd, &cmd, 1)) {
         return;
@@ -1425,13 +1491,26 @@ static void rpc_serve_client(ggml_backend_t backend, const char * cache_dir,
                 // HELLO command is handled above
                 return;
             }
+            case RPC_CMD_DEVICE_COUNT: {
+                if (!recv_msg(sockfd, nullptr, 0)) {
+                    return;
+                }
+                rpc_msg_device_count_rsp response;
+                response.device_count = backends.size();
+                if (!send_msg(sockfd, &response, sizeof(response))) {
+                    return;
+                }
+                break;
+            }
             case RPC_CMD_ALLOC_BUFFER: {
                 rpc_msg_alloc_buffer_req request;
                 if (!recv_msg(sockfd, &request, sizeof(request))) {
                     return;
                 }
                 rpc_msg_alloc_buffer_rsp response;
-                server.alloc_buffer(request, response);
+                if (!server.alloc_buffer(request, response)) {
+                    return;
+                }
                 if (!send_msg(sockfd, &response, sizeof(response))) {
                     return;
                 }
@@ -1452,22 +1531,28 @@ static void rpc_serve_client(ggml_backend_t backend, const char * cache_dir,
                 break;
             }
             case RPC_CMD_GET_ALIGNMENT: {
-                if (!recv_msg(sockfd, nullptr, 0)) {
+                rpc_msg_get_alignment_req request;
+                if (!recv_msg(sockfd, &request, sizeof(request))) {
                     return;
                 }
                 rpc_msg_get_alignment_rsp response;
-                server.get_alignment(response);
+                if (!server.get_alignment(request, response)) {
+                    return;
+                }
                 if (!send_msg(sockfd, &response, sizeof(response))) {
                     return;
                 }
                 break;
             }
             case RPC_CMD_GET_MAX_SIZE: {
-                if (!recv_msg(sockfd, nullptr, 0)) {
+                rpc_msg_get_max_size_req request;
+                if (!recv_msg(sockfd, &request, sizeof(request))) {
                     return;
                 }
                 rpc_msg_get_max_size_rsp response;
-                server.get_max_size(response);
+                if (!server.get_max_size(request, response)) {
+                    return;
+                }
                 if (!send_msg(sockfd, &response, sizeof(response))) {
                     return;
                 }
@@ -1593,12 +1678,19 @@ static void rpc_serve_client(ggml_backend_t backend, const char * cache_dir,
                 break;
             }
             case RPC_CMD_GET_DEVICE_MEMORY: {
-                if (!recv_msg(sockfd, nullptr, 0)) {
+                rpc_msg_get_device_memory_req request;
+                if (!recv_msg(sockfd, &request, sizeof(request))) {
+                    return;
+                }
+                auto dev_id = request.device;
+                if (dev_id >= backends.size()) {
                     return;
                 }
                 rpc_msg_get_device_memory_rsp response;
-                response.free_mem = free_mem;
-                response.total_mem = total_mem;
+                response.free_mem = free_mem[dev_id];
+                response.total_mem = total_mem[dev_id];
+                LOG_DBG("[get_device_mem] device: %u, free_mem: %" PRIu64 ", total_mem: %" PRIu64 "\n", dev_id,
+                    response.free_mem, response.total_mem);
                 if (!send_msg(sockfd, &response, sizeof(response))) {
                     return;
                 }
@@ -1612,16 +1704,41 @@ static void rpc_serve_client(ggml_backend_t backend, const char * cache_dir,
     }
 }
 
-void ggml_backend_rpc_start_server(ggml_backend_t backend, const char * endpoint,
-                                   const char * cache_dir,
-                                   size_t free_mem, size_t total_mem) {
+void ggml_backend_rpc_start_server(const char * endpoint, const char * cache_dir,
+                                   size_t n_threads, size_t n_devices,
+                                   ggml_backend_dev_t * devices, size_t * free_mem, size_t * total_mem) {
+    if (n_devices == 0 || devices == nullptr || free_mem == nullptr || total_mem == nullptr) {
+        fprintf(stderr, "Invalid arguments to ggml_backend_rpc_start_server\n");
+        return;
+    }
+    std::vector<ggml_backend_t> backends;
+    std::vector<size_t> free_mem_vec(free_mem, free_mem + n_devices);
+    std::vector<size_t> total_mem_vec(total_mem, total_mem + n_devices);
     printf("Starting RPC server v%d.%d.%d\n",
         RPC_PROTO_MAJOR_VERSION,
         RPC_PROTO_MINOR_VERSION,
         RPC_PROTO_PATCH_VERSION);
     printf("  endpoint       : %s\n", endpoint);
     printf("  local cache    : %s\n", cache_dir ? cache_dir : "n/a");
-    printf("  backend memory : %zu MB\n", free_mem / (1024 * 1024));
+    printf("Devices:\n");
+    for (size_t i = 0; i < n_devices; i++) {
+        auto dev = devices[i];
+        printf("  %s: %s (%zu MiB, %zu MiB free)\n", ggml_backend_dev_name(dev), ggml_backend_dev_description(dev),
+               total_mem[i] / 1024 / 1024, free_mem[i] / 1024 / 1024);
+        auto backend = ggml_backend_dev_init(dev, nullptr);
+        if (!backend) {
+            fprintf(stderr, "Failed to create backend for device %s\n", dev->iface.get_name(dev));
+            return;
+        }
+        backends.push_back(backend);
+        ggml_backend_reg_t reg = dev ? ggml_backend_dev_backend_reg(dev) : nullptr;
+        if (reg) {
+            auto ggml_backend_set_n_threads_fn = (ggml_backend_set_n_threads_t) ggml_backend_reg_get_proc_address(reg, "ggml_backend_set_n_threads");
+            if (ggml_backend_set_n_threads_fn) {
+                ggml_backend_set_n_threads_fn(backend, n_threads);
+            }
+        }
+    }
 
     std::string host;
     int port;
@@ -1649,22 +1766,27 @@ void ggml_backend_rpc_start_server(ggml_backend_t backend, const char * endpoint
             fprintf(stderr, "Failed to accept client connection\n");
             return;
         }
-        printf("Accepted client connection, free_mem=%zu, total_mem=%zu\n", free_mem, total_mem);
+        printf("Accepted client connection\n");
         fflush(stdout);
-        rpc_serve_client(backend, cache_dir, client_socket->fd, free_mem, total_mem);
+        rpc_serve_client(backends, cache_dir, client_socket->fd, free_mem_vec, total_mem_vec);
         printf("Client connection closed\n");
         fflush(stdout);
     }
 #ifdef _WIN32
     WSACleanup();
 #endif
+    for (auto backend : backends) {
+        ggml_backend_free(backend);
+    }
 }
 
 // device interface
 
 struct ggml_backend_rpc_device_context {
     std::string endpoint;
+    uint32_t    device;
     std::string name;
+    std::string description;
 };
 
 static const char * ggml_backend_rpc_device_get_name(ggml_backend_dev_t dev) {
@@ -1676,15 +1798,13 @@ static const char * ggml_backend_rpc_device_get_name(ggml_backend_dev_t dev) {
 static const char * ggml_backend_rpc_device_get_description(ggml_backend_dev_t dev) {
     ggml_backend_rpc_device_context * ctx = (ggml_backend_rpc_device_context *)dev->context;
 
-    return ctx->name.c_str();
+    return ctx->description.c_str();
 }
 
 static void ggml_backend_rpc_device_get_memory(ggml_backend_dev_t dev, size_t * free, size_t * total) {
     ggml_backend_rpc_device_context * ctx = (ggml_backend_rpc_device_context *)dev->context;
 
-    ggml_backend_rpc_get_device_memory(ctx->endpoint.c_str(), free, total);
-
-    GGML_UNUSED(dev);
+    ggml_backend_rpc_get_device_memory(ctx->endpoint.c_str(), ctx->device, free, total);
 }
 
 static enum ggml_backend_dev_type ggml_backend_rpc_device_get_type(ggml_backend_dev_t dev) {
@@ -1710,7 +1830,7 @@ static void ggml_backend_rpc_device_get_props(ggml_backend_dev_t dev, struct ggm
 static ggml_backend_t ggml_backend_rpc_device_init(ggml_backend_dev_t dev, const char * params) {
     ggml_backend_rpc_device_context * ctx = (ggml_backend_rpc_device_context *)dev->context;
 
-    return ggml_backend_rpc_init(ctx->endpoint.c_str());
+    return ggml_backend_rpc_init(ctx->endpoint.c_str(), ctx->device);
 
     GGML_UNUSED(params);
 }
@@ -1718,7 +1838,7 @@ static ggml_backend_t ggml_backend_rpc_device_init(ggml_backend_dev_t dev, const
 static ggml_backend_buffer_type_t ggml_backend_rpc_device_get_buffer_type(ggml_backend_dev_t dev) {
     ggml_backend_rpc_device_context * ctx = (ggml_backend_rpc_device_context *)dev->context;
 
-    return ggml_backend_rpc_buffer_type(ctx->endpoint.c_str());
+    return ggml_backend_rpc_buffer_type(ctx->endpoint.c_str(), ctx->device);
 
     GGML_UNUSED(dev);
 }
@@ -1736,7 +1856,7 @@ static bool ggml_backend_rpc_device_supports_buft(ggml_backend_dev_t dev, ggml_b
     }
     ggml_backend_rpc_buffer_type_context * buft_ctx = (ggml_backend_rpc_buffer_type_context *)buft->context;
     ggml_backend_rpc_device_context * dev_ctx = (ggml_backend_rpc_device_context *)dev->context;
-    return buft_ctx->endpoint == dev_ctx->endpoint;
+    return buft_ctx->endpoint == dev_ctx->endpoint && buft_ctx->device == dev_ctx->device;
 }
 
 static const struct ggml_backend_device_i ggml_backend_rpc_device_i = {
@@ -1759,28 +1879,34 @@ static const struct ggml_backend_device_i ggml_backend_rpc_device_i = {
 
 // backend reg interface
 
-static const char * ggml_backend_rpc_reg_get_name(ggml_backend_reg_t reg) {
-    return "RPC";
+struct ggml_backend_rpc_reg_context {
+    std::string                     name;
+    std::vector<ggml_backend_dev_t> devices;
+};
 
-    GGML_UNUSED(reg);
+static const char * ggml_backend_rpc_reg_get_name(ggml_backend_reg_t reg) {
+    ggml_backend_rpc_reg_context * ctx = (ggml_backend_rpc_reg_context *)reg->context;
+    return ctx ? ctx->name.c_str() : "RPC";
 }
 
 static size_t ggml_backend_rpc_reg_get_device_count(ggml_backend_reg_t reg) {
-    return 0;
-
-    GGML_UNUSED(reg);
+    ggml_backend_rpc_reg_context * ctx = (ggml_backend_rpc_reg_context *)reg->context;
+    return ctx ? ctx->devices.size() : 0;
 }
 
 static ggml_backend_dev_t ggml_backend_rpc_reg_get_device(ggml_backend_reg_t reg, size_t index) {
-    GGML_ABORT("The RPC backend does not have enumerated devices - use ggml_backend_add_device instead");
-
-    GGML_UNUSED(reg);
-    GGML_UNUSED(index);
+    ggml_backend_rpc_reg_context * ctx = (ggml_backend_rpc_reg_context *)reg->context;
+    if (ctx == nullptr) {
+        GGML_ABORT("The RPC backend does not have enumerated devices - use ggml_backend_rpc_add_server instead");
+    } else {
+        GGML_ASSERT(index < ctx->devices.size());
+        return ctx->devices[index];
+    }
 }
 
 static void * ggml_backend_rpc_get_proc_address(ggml_backend_reg_t reg, const char * name) {
-    if (std::strcmp(name, "ggml_backend_rpc_add_device") == 0) {
-        return (void *)ggml_backend_rpc_add_device;
+    if (std::strcmp(name, "ggml_backend_rpc_add_server") == 0) {
+        return (void *)ggml_backend_rpc_add_server;
     }
     if (std::strcmp(name, "ggml_backend_rpc_start_server") == 0) {
         return (void *)ggml_backend_rpc_start_server;
@@ -1807,30 +1933,61 @@ ggml_backend_reg_t ggml_backend_rpc_reg(void) {
     return &ggml_backend_rpc_reg;
 }
 
-ggml_backend_dev_t ggml_backend_rpc_add_device(const char * endpoint) {
-    static std::unordered_map<std::string, ggml_backend_dev_t> dev_map;
-
-    static std::mutex mutex;
-    std::lock_guard<std::mutex> lock(mutex);
-
-    if (dev_map.find(endpoint) != dev_map.end()) {
-        return dev_map[endpoint];
-    }
-
-    ggml_backend_rpc_device_context * ctx = new ggml_backend_rpc_device_context {
-        /* .endpoint = */ endpoint,
-        /* .name     = */ "RPC[" + std::string(endpoint) + "]",
-    };
-
-    ggml_backend_dev_t dev = new ggml_backend_device {
-        /* .iface   = */ ggml_backend_rpc_device_i,
-        /* .reg     = */ ggml_backend_rpc_reg(),
-        /* .context = */ ctx,
-    };
-
-    dev_map[endpoint] = dev;
-
-    return dev;
+static uint32_t ggml_backend_rpc_get_device_count(const char * endpoint) {
+    auto sock = get_socket(endpoint);
+    rpc_msg_device_count_rsp response;
+    bool status = send_rpc_cmd(sock, RPC_CMD_DEVICE_COUNT, nullptr, 0, &response, sizeof(response));
+    RPC_STATUS_ASSERT(status);
+    return response.device_count;
 }
 
+static const ggml_backend_reg_i ggml_backend_rpc_reg_interface = {
+    /* .get_name          = */ ggml_backend_rpc_reg_get_name,
+    /* .get_device_count  = */ ggml_backend_rpc_reg_get_device_count,
+    /* .get_device        = */ ggml_backend_rpc_reg_get_device,
+    /* .get_proc_address  = */ ggml_backend_rpc_get_proc_address,
+};
+
+ggml_backend_reg_t ggml_backend_rpc_add_server(const char * endpoint) {
+    static std::unordered_map<std::string, ggml_backend_reg_t> reg_map;
+    static std::mutex mutex;
+    static uint32_t dev_id = 0;
+    std::lock_guard<std::mutex> lock(mutex);
+    if (reg_map.find(endpoint) != reg_map.end()) {
+        return reg_map[endpoint];
+    }
+    uint32_t dev_count = ggml_backend_rpc_get_device_count(endpoint);
+    if (dev_count == 0) {
+        return nullptr;
+    }
+    ggml_backend_rpc_reg_context * ctx = new ggml_backend_rpc_reg_context;
+    ctx->name = "RPC[" + std::string(endpoint) + "]";
+    for (uint32_t ind = 0; ind < dev_count; ind++) {
+        std::string dev_name = "RPC" + std::to_string(dev_id);
+        std::string dev_desc = std::string(endpoint);
+        ggml_backend_rpc_device_context * dev_ctx = new ggml_backend_rpc_device_context {
+            /* .endpoint    = */ endpoint,
+            /* .device      = */ ind,
+            /* .name        = */ dev_name,
+            /* .description = */ dev_desc
+        };
+
+        ggml_backend_dev_t dev = new ggml_backend_device {
+            /* .iface   = */ ggml_backend_rpc_device_i,
+            /* .reg     = */ ggml_backend_rpc_reg(),
+            /* .context = */ dev_ctx,
+        };
+        ctx->devices.push_back(dev);
+        dev_id++;
+    }
+    ggml_backend_reg_t reg = new ggml_backend_reg {
+        /* .api_version = */ GGML_BACKEND_API_VERSION,
+        /* .iface       = */ ggml_backend_rpc_reg_interface,
+        /* .context     = */ ctx
+    };
+    reg_map[endpoint] = reg;
+    return reg;
+}
+
+
 GGML_BACKEND_DL_IMPL(ggml_backend_rpc_reg)
diff --git a/tools/llama-bench/llama-bench.cpp b/tools/llama-bench/llama-bench.cpp
index 275ba367c..89bc01b48 100644
--- a/tools/llama-bench/llama-bench.cpp
+++ b/tools/llama-bench/llama-bench.cpp
@@ -168,7 +168,7 @@ static std::vector<ggml_backend_dev_t> parse_devices_arg(const std::string & val
     return devices;
 }
 
-static std::vector<ggml_backend_dev_t> register_rpc_device_list(const std::string & servers) {
+static void register_rpc_server_list(const std::string & servers) {
     auto rpc_servers = string_split<std::string>(servers, ',');
     if (rpc_servers.empty()) {
         throw std::invalid_argument("no RPC servers specified");
@@ -179,36 +179,15 @@ static std::vector<ggml_backend_dev_t> register_rpc_device_list(const std::strin
         throw std::invalid_argument("failed to find RPC backend");
     }
 
-    using add_rpc_device_fn = ggml_backend_dev_t (*)(const char * endpoint);
-    auto * ggml_backend_rpc_add_device_fn = (add_rpc_device_fn) ggml_backend_reg_get_proc_address(rpc_reg, "ggml_backend_rpc_add_device");
-    if (!ggml_backend_rpc_add_device_fn) {
-        throw std::invalid_argument("failed to find RPC device add function");
+    using add_rpc_server_fn = ggml_backend_reg_t (*)(const char * endpoint);
+    auto * ggml_backend_rpc_add_server_fn = (add_rpc_server_fn) ggml_backend_reg_get_proc_address(rpc_reg, "ggml_backend_rpc_add_server");
+    if (!ggml_backend_rpc_add_server_fn) {
+        throw std::invalid_argument("failed to find RPC add server function");
     }
-
-    static std::unordered_set<std::string> registered;
-    std::vector<ggml_backend_dev_t> devices;
     for (const auto & server : rpc_servers) {
-        ggml_backend_dev_t dev = nullptr;
-
-        std::string name = string_format("RPC[%s]", server.c_str());
-
-        if (registered.find(server) != registered.end()) {
-            dev = ggml_backend_dev_by_name(name.c_str());
-        }
-
-        if (!dev) {
-            dev = ggml_backend_rpc_add_device_fn(server.c_str());
-            if (!dev) {
-                throw std::invalid_argument(string_format("failed to add RPC device for server '%s'", server.c_str()));
-            }
-            ggml_backend_device_register(dev);
-            registered.insert(server);
-        }
-
-        devices.push_back(dev);
+        auto reg = ggml_backend_rpc_add_server_fn(server.c_str());
+        ggml_backend_register(reg);
     }
-
-    return devices;
 }
 
 static std::string devices_to_string(const std::vector<ggml_backend_dev_t> & devices) {
@@ -714,7 +693,7 @@ static cmd_params parse_cmd_params(int argc, char ** argv) {
                     break;
                 }
                 try {
-                    register_rpc_device_list(argv[i]);
+                    register_rpc_server_list(argv[i]);
                 } catch (const std::exception & e) {
                     fprintf(stderr, "error: %s\n", e.what());
                     invalid_param = true;
@@ -1368,13 +1347,23 @@ struct test {
 
     static std::string get_backend() {
         std::vector<std::string> backends;
+        bool                     rpc_used = false;
         for (size_t i = 0; i < ggml_backend_reg_count(); i++) {
             auto *      reg  = ggml_backend_reg_get(i);
             std::string name = ggml_backend_reg_name(reg);
-            if (name != "CPU") {
-                backends.push_back(ggml_backend_reg_name(reg));
+            if (string_starts_with(name, "RPC")) {
+                if (ggml_backend_reg_dev_count(reg) > 0) {
+                    rpc_used = true;
+                }
+            } else {
+                if (name != "CPU") {
+                    backends.push_back(ggml_backend_reg_name(reg));
+                }
             }
         }
+        if (rpc_used) {
+            backends.push_back("RPC");
+        }
         return backends.empty() ? "CPU" : join(backends, ",");
     }
 
diff --git a/tools/rpc/rpc-server.cpp b/tools/rpc/rpc-server.cpp
index dc8e077f3..088515612 100644
--- a/tools/rpc/rpc-server.cpp
+++ b/tools/rpc/rpc-server.cpp
@@ -22,6 +22,7 @@
 #include <filesystem>
 #include <algorithm>
 #include <thread>
+#include <regex>
 
 namespace fs = std::filesystem;
 
@@ -131,24 +132,24 @@ static std::string fs_get_cache_directory() {
 }
 
 struct rpc_server_params {
-    std::string host        = "127.0.0.1";
-    int         port        = 50052;
-    size_t      backend_mem = 0;
-    bool        use_cache   = false;
-    int         n_threads   = std::max(1U, std::thread::hardware_concurrency()/2);
-    std::string device;
+    std::string              host        = "127.0.0.1";
+    int                      port        = 50052;
+    bool                     use_cache   = false;
+    int                      n_threads   = std::max(1U, std::thread::hardware_concurrency()/2);
+    std::vector<std::string> devices;
+    std::vector<size_t>      dev_mem;
 };
 
 static void print_usage(int /*argc*/, char ** argv, rpc_server_params params) {
     fprintf(stderr, "Usage: %s [options]\n\n", argv[0]);
     fprintf(stderr, "options:\n");
-    fprintf(stderr, "  -h, --help                show this help message and exit\n");
-    fprintf(stderr, "  -t,      --threads        number of threads for the CPU backend (default: %d)\n", params.n_threads);
-    fprintf(stderr, "  -d DEV,  --device         device to use\n");
-    fprintf(stderr, "  -H HOST, --host HOST      host to bind to (default: %s)\n", params.host.c_str());
-    fprintf(stderr, "  -p PORT, --port PORT      port to bind to (default: %d)\n", params.port);
-    fprintf(stderr, "  -m MEM,  --mem MEM        backend memory size (in MB)\n");
-    fprintf(stderr, "  -c,      --cache          enable local file cache\n");
+    fprintf(stderr, "  -h, --help                       show this help message and exit\n");
+    fprintf(stderr, "  -t, --threads N                  number of threads for the CPU device (default: %d)\n", params.n_threads);
+    fprintf(stderr, "  -d, --device <dev1,dev2,...>     comma-separated list of devices\n");
+    fprintf(stderr, "  -H, --host HOST                  host to bind to (default: %s)\n", params.host.c_str());
+    fprintf(stderr, "  -p, --port PORT                  port to bind to (default: %d)\n", params.port);
+    fprintf(stderr, "  -m, --mem <M1,M2,...>            memory size for each device (in MB)\n");
+    fprintf(stderr, "  -c, --cache                      enable local file cache\n");
     fprintf(stderr, "\n");
 }
 
@@ -174,17 +175,17 @@ static bool rpc_server_params_parse(int argc, char ** argv, rpc_server_params &
             if (++i >= argc) {
                 return false;
             }
-            params.device = argv[i];
-            if (ggml_backend_dev_by_name(params.device.c_str()) == nullptr) {
-                fprintf(stderr, "error: unknown device: %s\n", params.device.c_str());
-                fprintf(stderr, "available devices:\n");
-                for (size_t i = 0; i < ggml_backend_dev_count(); i++) {
-                    auto * dev = ggml_backend_dev_get(i);
-                    size_t free, total;
-                    ggml_backend_dev_memory(dev, &free, &total);
-                    printf("  %s: %s (%zu MiB, %zu MiB free)\n", ggml_backend_dev_name(dev), ggml_backend_dev_description(dev), total / 1024 / 1024, free / 1024 / 1024);
+            const std::regex regex{ R"([,/]+)" };
+            std::string dev_str = argv[i];
+            std::sregex_token_iterator iter(dev_str.begin(), dev_str.end(), regex, -1);
+            std::sregex_token_iterator end;
+            for ( ; iter != end; ++iter) {
+                try {
+                    params.devices.push_back(*iter);
+                } catch (const std::exception & ) {
+                    fprintf(stderr, "error: invalid device: %s\n", iter->str().c_str());
+                    return false;
                 }
-                return false;
             }
         } else if (arg == "-p" || arg == "--port") {
             if (++i >= argc) {
@@ -200,7 +201,19 @@ static bool rpc_server_params_parse(int argc, char ** argv, rpc_server_params &
             if (++i >= argc) {
                 return false;
             }
-            params.backend_mem = std::stoul(argv[i]) * 1024 * 1024;
+            const std::regex regex{ R"([,/]+)" };
+            std::string mem_str = argv[i];
+            std::sregex_token_iterator iter(mem_str.begin(), mem_str.end(), regex, -1);
+            std::sregex_token_iterator end;
+            for ( ; iter != end; ++iter) {
+                try {
+                    size_t mem = std::stoul(*iter) * 1024 * 1024;
+                    params.dev_mem.push_back(mem);
+                } catch (const std::exception & ) {
+                    fprintf(stderr, "error: invalid memory size: %s\n", iter->str().c_str());
+                    return false;
+                }
+            }
         } else if (arg == "-h" || arg == "--help") {
             print_usage(argc, argv, params);
             exit(0);
@@ -213,45 +226,46 @@ static bool rpc_server_params_parse(int argc, char ** argv, rpc_server_params &
     return true;
 }
 
-static ggml_backend_t create_backend(const rpc_server_params & params) {
-    ggml_backend_t backend = nullptr;
+static std::vector<ggml_backend_dev_t> get_devices(const rpc_server_params & params) {
+    std::vector<ggml_backend_dev_t> devices;
+    if (!params.devices.empty()) {
+        for (auto device : params.devices) {
+            ggml_backend_dev_t dev = ggml_backend_dev_by_name(device.c_str());
+            if (dev) {
+                devices.push_back(dev);
+            } else {
+                fprintf(stderr, "error: unknown device: %s\n", device.c_str());
+                fprintf(stderr, "available devices:\n");
+                for (size_t i = 0; i < ggml_backend_dev_count(); i++) {
+                    auto * dev = ggml_backend_dev_get(i);
+                    size_t free, total;
+                    ggml_backend_dev_memory(dev, &free, &total);
+                    printf("  %s: %s (%zu MiB, %zu MiB free)\n", ggml_backend_dev_name(dev), ggml_backend_dev_description(dev), total / 1024 / 1024, free / 1024 / 1024);
+                }
+                return {};
+            }
+        }
+    }
 
-    if (!params.device.empty()) {
-        ggml_backend_dev_t dev = ggml_backend_dev_by_name(params.device.c_str());
+    // Try non-CPU devices first
+    if (devices.empty()) {
+        for (size_t i = 0; i < ggml_backend_dev_count(); i++) {
+            ggml_backend_dev_t dev = ggml_backend_dev_get(i);
+            if (ggml_backend_dev_type(dev) != GGML_BACKEND_DEVICE_TYPE_CPU) {
+                devices.push_back(dev);
+            }
+        }
+    }
+
+    // If there are no accelerators, fallback to CPU device
+    if (devices.empty()) {
+        ggml_backend_dev_t dev = ggml_backend_dev_by_type(GGML_BACKEND_DEVICE_TYPE_CPU);
         if (dev) {
-            backend = ggml_backend_dev_init(dev, nullptr);
-            if (!backend) {
-                fprintf(stderr, "Failed to create backend for device %s\n", params.device.c_str());
-                return nullptr;
-            }
+            devices.push_back(dev);
         }
     }
 
-    if (!backend) {
-        backend = ggml_backend_init_best();
-    }
-
-    if (backend) {
-        fprintf(stderr, "%s: using %s backend\n", __func__, ggml_backend_name(backend));
-
-        // set the number of threads
-        ggml_backend_dev_t dev = ggml_backend_get_device(backend);
-        ggml_backend_reg_t reg = dev ? ggml_backend_dev_backend_reg(dev) : nullptr;
-        if (reg) {
-            auto ggml_backend_set_n_threads_fn = (ggml_backend_set_n_threads_t) ggml_backend_reg_get_proc_address(reg, "ggml_backend_set_n_threads");
-            if (ggml_backend_set_n_threads_fn) {
-                ggml_backend_set_n_threads_fn(backend, params.n_threads);
-            }
-        }
-    }
-
-    return backend;
-}
-
-static void get_backend_memory(ggml_backend_t backend, size_t * free_mem, size_t * total_mem) {
-    ggml_backend_dev_t dev = ggml_backend_get_device(backend);
-    GGML_ASSERT(dev != nullptr);
-    ggml_backend_dev_memory(dev, free_mem, total_mem);
+    return devices;
 }
 
 int main(int argc, char * argv[]) {
@@ -273,18 +287,23 @@ int main(int argc, char * argv[]) {
         fprintf(stderr, "\n");
     }
 
-    ggml_backend_t backend = create_backend(params);
-    if (!backend) {
-        fprintf(stderr, "Failed to create backend\n");
+    auto devices = get_devices(params);
+    if (devices.empty()) {
+        fprintf(stderr, "No devices found\n");
         return 1;
     }
     std::string endpoint = params.host + ":" + std::to_string(params.port);
-    size_t free_mem, total_mem;
-    if (params.backend_mem > 0) {
-        free_mem = params.backend_mem;
-        total_mem = params.backend_mem;
-    } else {
-        get_backend_memory(backend, &free_mem, &total_mem);
+    std::vector<size_t> free_mem, total_mem;
+    for (size_t i = 0; i < devices.size(); i++) {
+        if (i < params.dev_mem.size()) {
+            free_mem.push_back(params.dev_mem[i]);
+            total_mem.push_back(params.dev_mem[i]);
+        } else {
+            size_t free, total;
+            ggml_backend_dev_memory(devices[i], &free, &total);
+            free_mem.push_back(free);
+            total_mem.push_back(total);
+        }
     }
     const char * cache_dir = nullptr;
     std::string cache_dir_str;
@@ -309,8 +328,7 @@ int main(int argc, char * argv[]) {
         return 1;
     }
 
-    start_server_fn(backend, endpoint.c_str(), cache_dir, free_mem, total_mem);
-
-    ggml_backend_free(backend);
+    start_server_fn(endpoint.c_str(), cache_dir, params.n_threads, devices.size(),
+        devices.data(), free_mem.data(), total_mem.data());
     return 0;
 }

From f39283960b58a92ecc0c72567711318b20e22b55 Mon Sep 17 00:00:00 2001
From: Radoslav Gerganov <rgerganov@gmail.com>
Date: Sat, 4 Oct 2025 16:22:45 +0300
Subject: [PATCH 3/4] rpc : check src buffer when copying tensor (#16421)

Only dst buffer is guaranteed to be an RPC buffer. Add check for the src
one.
---
 ggml/src/ggml-rpc/ggml-rpc.cpp | 37 ++++++++++++++++++++--------------
 1 file changed, 22 insertions(+), 15 deletions(-)

diff --git a/ggml/src/ggml-rpc/ggml-rpc.cpp b/ggml/src/ggml-rpc/ggml-rpc.cpp
index 1a8739e78..aad48d62a 100644
--- a/ggml/src/ggml-rpc/ggml-rpc.cpp
+++ b/ggml/src/ggml-rpc/ggml-rpc.cpp
@@ -631,23 +631,30 @@ static void ggml_backend_rpc_buffer_get_tensor(ggml_backend_buffer_t buffer, con
     RPC_STATUS_ASSERT(status);
 }
 
+static bool ggml_backend_buffer_is_rpc(ggml_backend_buffer_t buffer) {
+    return buffer->iface.free_buffer == ggml_backend_rpc_buffer_free_buffer;
+}
+
 static bool ggml_backend_rpc_buffer_cpy_tensor(ggml_backend_buffer_t buffer, const ggml_tensor * src, ggml_tensor * dst) {
-    // check if src and dst are on the same server
-    ggml_backend_buffer_t src_buffer = src->buffer;
-    ggml_backend_rpc_buffer_context * src_ctx = (ggml_backend_rpc_buffer_context *)src_buffer->context;
-    ggml_backend_buffer_t dst_buffer = dst->buffer;
-    ggml_backend_rpc_buffer_context * dst_ctx = (ggml_backend_rpc_buffer_context *)dst_buffer->context;
-    if (src_ctx->sock != dst_ctx->sock) {
-        return false;
+    if (ggml_backend_buffer_is_rpc(src->buffer)) {
+        // check if src and dst are on the same server
+        ggml_backend_buffer_t src_buffer = src->buffer;
+        ggml_backend_rpc_buffer_context * src_ctx = (ggml_backend_rpc_buffer_context *)src_buffer->context;
+        ggml_backend_buffer_t dst_buffer = dst->buffer;
+        ggml_backend_rpc_buffer_context * dst_ctx = (ggml_backend_rpc_buffer_context *)dst_buffer->context;
+        if (src_ctx->sock != dst_ctx->sock) {
+            return false;
+        }
+        ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context;
+        rpc_msg_copy_tensor_req request;
+        request.src = serialize_tensor(src);
+        request.dst = serialize_tensor(dst);
+        rpc_msg_copy_tensor_rsp response;
+        bool status = send_rpc_cmd(ctx->sock, RPC_CMD_COPY_TENSOR, &request, sizeof(request), &response, sizeof(response));
+        RPC_STATUS_ASSERT(status);
+        return response.result;
     }
-    ggml_backend_rpc_buffer_context * ctx = (ggml_backend_rpc_buffer_context *)buffer->context;
-    rpc_msg_copy_tensor_req request;
-    request.src = serialize_tensor(src);
-    request.dst = serialize_tensor(dst);
-    rpc_msg_copy_tensor_rsp response;
-    bool status = send_rpc_cmd(ctx->sock, RPC_CMD_COPY_TENSOR, &request, sizeof(request), &response, sizeof(response));
-    RPC_STATUS_ASSERT(status);
-    return response.result;
+    return false;
 }
 
 static void ggml_backend_rpc_buffer_clear(ggml_backend_buffer_t buffer, uint8_t value) {

From 86df2c9ae4f2f1ee63d2558a9dc797b98524639b Mon Sep 17 00:00:00 2001
From: Eve <139727413+netrunnereve@users.noreply.github.com>
Date: Sat, 4 Oct 2025 20:04:27 +0000
Subject: [PATCH 4/4] vulkan: use a more appropriate amount of threads when
 generating shaders (#16418)

* use a more flexible amount of threads

* fix windows compile and 0 thread case

* nominmax
---
 ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp b/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp
index e2726f1fa..f0cc24ff3 100644
--- a/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp
+++ b/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp
@@ -1,5 +1,3 @@
-
-
 #include <iostream>
 #include <fstream>
 #include <sstream>
@@ -22,6 +20,7 @@
 #include <sys/types.h>
 
 #ifdef _WIN32
+    #define NOMINMAX
     #include <windows.h>
     #include <direct.h> // For _mkdir on Windows
 #else
@@ -306,7 +305,7 @@ using compile_count_guard = std::unique_ptr<uint32_t, decltype(&decrement_compil
 compile_count_guard acquire_compile_slot() {
     // wait until fewer than N compiles are in progress.
     // 16 is an arbitrary limit, the goal is to avoid "failed to create pipe" errors.
-    uint32_t N = 16;
+    uint32_t N = std::max(1u, std::min(16u, std::thread::hardware_concurrency()));
     std::unique_lock<std::mutex> guard(compile_count_mutex);
     compile_count_cond.wait(guard, [N] { return compile_count < N; });
     compile_count++;