vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-13 18:39:48 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions 1

llama : add high-throughput mode (#14363)

Browse source 
* kv-cache : prepare K/V buffers for separation

ggml-ci

* batched-bench : fix oob write

ggml-ci

* llama : add "virtual sequences"

ggml-ci

* llama : use "stream" vs "virtual sequence"

ggml-ci

* graph : fix stream splitting when KV cache is not used

ggml-ci

* kv-cache : add multi-stream save/load support

ggml-ci

* llama : add "--attn-streams" flag

ggml-ci

* kv-cache : fix handling when find_slot fails

ggml-ci

* kv-cache : restore find_slot impl

ggml-ci

* kv-cache : add comments

* kv-cache : add bounds checks for sequence id

ggml-ci

* cont : add n_seq_max to batch allocr

ggml-ci

* kv-cache : perform stream copies lazily after llama_synchronize

ggml-ci

* kv-cache : avoid throwing exceptions across the C boundary

ggml-ci

* CUDA: 4D FlashAttention support (#14628)

* CUDA: 4D FlashAttention support

* CUDA: fix WMMA FA kernel

* llama : rename attn_streams -> kv_unified

ggml-ci

* common : rename kv_split -> kv_unified

ggml-ci

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
This commit is contained in:
Georgi Gerganov 2025-07-16 16:35:42 +03:00 • committed by GitHub
parent ab14019821
commit 225e7a1438
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
30 changed files with 1080 additions and 460 deletions
Download patch file Download diff file Expand all files Collapse all files

19
src/llama-model.cpp
View file

@ -16647,7 +16647,18 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
} else {
const auto padding = llama_kv_cache_unified::get_padding(cparams);
cparams.n_ctx = GGML_PAD(cparams.n_ctx, padding);
uint32_t n_ctx_per_stream = cparams.n_ctx;
if (!cparams.kv_unified) {
n_ctx_per_stream = (cparams.n_ctx + cparams.n_seq_max - 1)/cparams.n_seq_max;
n_ctx_per_stream = GGML_PAD(n_ctx_per_stream, padding);
cparams.n_ctx = n_ctx_per_stream*cparams.n_seq_max;
} else {
n_ctx_per_stream = GGML_PAD(n_ctx_per_stream, padding);
cparams.n_ctx = n_ctx_per_stream;
}
LLAMA_LOG_DEBUG("%s: n_ctx = %u (padded)\n", __func__, cparams.n_ctx);
@ -16661,7 +16672,8 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
!cparams.flash_attn,
cparams.offload_kqv,
params.swa_full,
cparams.n_ctx,
cparams.kv_unified,
n_ctx_per_stream,
cparams.n_seq_max,
cparams.n_ubatch,
padding);
@ -16675,7 +16687,8 @@ llama_memory_i * llama_model::create_memory(const llama_memory_params & params,
params.type_v,
!cparams.flash_attn,
cparams.offload_kqv,
cparams.n_ctx,
cparams.kv_unified,
n_ctx_per_stream,
cparams.n_seq_max,
padding,
hparams.n_swa,