koboldcpp/.github/workflows
shivamkumard-ctrl 2e2d99cfd2
ci: Add support for CUDA 13.4 ARM64 builds for Windows (#26650)
* ci: Add support for CUDA 13.4 ARM64 builds for Windows

Added an architecture-specific CUDA 13.4 Windows build entry targeting ARM64.
Added a CMake configuration to enable ARM64 CUDA cross-compilation from an x64 Windows environment using the x64-hosted CUDA and MSVC toolchain while linking against the ARM64 CUDA import libraries to produce ggml-cuda.dll.
Validated the self-hosted Windows x64 workflow, including toolkit acquisition, CMake configuration, ARM64 CUDA cross-compilation, and packaging. Runtime validation was performed separately on a native ARM64 RTX Spark system using TinyLlama 1.1B Q4_K_M to verify the generated binaries.
The ARM64 CUDA job builds only the ggml-cuda.dll backend (LLAMA_BUILD_SERVER=OFF). The release consists of two packages: the main ARM64 release package, which combines the existing ARM64 CPU outputs with ggml-cuda.dll, and a separate runtime package containing the required CUDA runtime libraries (cudart64_13.dll, cublas64_13.dll, and cublasLt64_13.dll).
The CUDA 13.4 setup uses NVIDIA Developer Preview component archives instead of the GA component downloads used by the existing CUDA setups and will require updates once CUDA 13.4 reaches GA.

* ci: cleans up to align with x64 CUDA setup

- Moves CUDA-specific CMake options into matrix defines.
- Keeps the CUB 3DOT2 option only for CUDA 12.4.
- Removes runtime argument construction and the unnecessary server option.
- Aligns ARM64 CUDA runtime packaging with the existing robocopy approach.
- Generalizes the ARM64 release label from CUDA 13.4 to CUDA 13.

* ci: Set CUDA job name as version-architecture pair

* mark as preview

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>

---------

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@huggingface.co>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
2026-08-10 11:46:44 +03:00
..
ai-issues.yml ai : update gh permissions (#20895) 2026-03-23 13:21:41 +02:00
bench.yml.disabled llama : move end-user examples to tools directory (#13249) 2025-05-02 20:27:13 +02:00
build-3rd-party.yml common : fix env names to all have LLAMA_ARG_ prefix (#23778) 2026-05-27 14:52:47 +03:00
build-and-test-snapdragon.yml snapdragon: bump toolchain docker to v0.7 to fix ui build issues (#23680) 2026-05-25 10:57:43 -07:00
build-android.yml ci : refactor (#23789) 2026-05-28 09:44:25 +03:00
build-apple.yml build : remove GGML_METAL_USE_BF16 from all build scripts (#26604) 2026-08-05 10:44:34 +02:00
build-cache.yml openvino: Update to OV 2026.2.1, self-contained release packages, operator improvements (#24974) 2026-06-26 15:07:19 +03:00
build-cann.yml common : fix env names to all have LLAMA_ARG_ prefix (#23778) 2026-05-27 14:52:47 +03:00
build-cmake-pkg.yml ci : move most slim jobs to self-hosted runners (#23619) 2026-05-25 08:11:19 +03:00
build-cpu.yml vulkan: update vulkan sdk to 1.4.357.0 (#26303) 2026-07-31 07:27:03 -05:00
build-cross.yml ci : reduce (disable SYCL and CANN builds/releases) (#23705) 2026-05-26 15:21:21 +03:00
build-cuda-ubuntu.yml ci: rm GGML_HIP_ROCWMMA_FATTN (#26760) 2026-08-09 18:15:28 +08:00
build-cuda-windows.yml ci: rm GGML_HIP_ROCWMMA_FATTN (#26760) 2026-08-09 18:15:28 +08:00
build-ibm.yml common : fix env names to all have LLAMA_ARG_ prefix (#23778) 2026-05-27 14:52:47 +03:00
build-msys.yml ci : build-msys job slimming [no ci] (#24157) 2026-06-05 07:57:36 +02:00
build-opencl.yml ci : refactor (#23789) 2026-05-28 09:44:25 +03:00
build-openvino.yml openvino: Update to OV 2026.2.1, self-contained release packages, operator improvements (#24974) 2026-06-26 15:07:19 +03:00
build-riscv.yml ci : refactor (#23789) 2026-05-28 09:44:25 +03:00
build-rpc.yml ci: remove redundant or duplicate jobs (#23927) 2026-06-01 06:32:17 +03:00
build-sanitize.yml ci: fix the ctest sanitize runs (#26593) 2026-08-10 09:31:28 +03:00
build-self-hosted.yml ci : onboard AMD ROCm CI with gfx1151 fixes (#26544) 2026-08-06 10:43:26 +02:00
build-sycl.yml [SYCL] Fix CI build & release for SYCL backend (#24387) 2026-06-12 09:30:24 +03:00
build-virtgpu.yml ggml-virtgpu: Add a GHA build check (#22943) 2026-05-11 21:38:22 +08:00
build-vulkan.yml ci: fix vulkan llvmpipe runs (#26533) 2026-08-04 03:28:57 +00:00
build-wasm.yml ggml-webgpu: Fix WASM compilation with OpenMP (#25943) 2026-07-25 17:37:18 -07:00
build-webgpu.yml ggml-webgpu: Fix WASM compilation with OpenMP (#25943) 2026-07-25 17:37:18 -07:00
check-vendor.yml ci : move most slim jobs to self-hosted runners (#23619) 2026-05-25 08:11:19 +03:00
close-issue.yml ci: Also exempt 'security' tag from auto-close (#21844) 2026-04-14 01:18:44 +08:00
code-style.yml ci : move most slim jobs to self-hosted runners (#23619) 2026-05-25 08:11:19 +03:00
copilot-setup-steps.yml ci : limit requirements versions (#20980) 2026-03-25 10:55:37 +02:00
docker.yml docker : prebuild web UI for s390x build [no release] (#24829) 2026-06-20 05:54:42 -05:00
editorconfig.yml ci : move most slim jobs to self-hosted runners (#23619) 2026-05-25 08:11:19 +03:00
gguf-publish.yml gguf-py : bump version to 0.19.0 (#22664) 2026-05-06 14:46:14 +02:00
hip-quality-check.yml Make hip quality check run on all changes (#25403) 2026-07-08 14:38:51 +02:00
labeler.yml ci : use new 1vCPU runner for lightweight jobs (#19107) 2026-01-26 15:22:49 +01:00
pr-draft-label.yml ci: add pr-draft-label (#26801) 2026-08-09 16:51:21 +02:00
pre-tokenizer-hashes.yml ci : fix pre-tokenizer-hashes check (#23651) 2026-05-25 10:41:25 +02:00
python-check-requirements.yml ci : move most slim jobs to self-hosted runners (#23619) 2026-05-25 08:11:19 +03:00
python-lint.yml ci : move most slim jobs to self-hosted runners (#23619) 2026-05-25 08:11:19 +03:00
python-type-check.yml ci : move most slim jobs to self-hosted runners (#23619) 2026-05-25 08:11:19 +03:00
release.yml ci: Add support for CUDA 13.4 ARM64 builds for Windows (#26650) 2026-08-10 11:46:44 +03:00
server-sanitize.yml common : fix env names to all have LLAMA_ARG_ prefix (#23778) 2026-05-27 14:52:47 +03:00
server-self-hosted.yml CUDA: Support CUDA Virtual Devices (#25228) 2026-07-16 13:37:35 +03:00
server.yml speculative : fix n_outputs_max and remove draft-simple auto-enable (#23988) 2026-06-01 22:26:58 +03:00
ui-build-self-hosted.yml server: clean up static assets handling (#24550) 2026-06-13 11:51:20 +02:00
ui-build.yml server: clean up static assets handling (#24550) 2026-06-13 11:51:20 +02:00
ui-publish.yml server: remove loading.html (#25500) 2026-07-10 14:42:17 +02:00
ui-self-hosted.yml ui: PWA support (#23871) 2026-06-12 15:53:26 +02:00
ui.yml ui: PWA support (#23871) 2026-06-12 15:53:26 +02:00
update-ops-docs.yml ci : move most slim jobs to self-hosted runners (#23619) 2026-05-25 08:11:19 +03:00
winget.yml ci : bump komac version (#24396) 2026-06-10 09:45:20 +02:00