vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Vulkan k-quant mmq and ggml-backend offload functionality (#6155)

Browse source 
* Fix Vulkan no kv offload incoherence

* Add k-quant mul mat mat shaders

* Rework working buffer allocation, reduces vram use noticeably

Clean up cpu assist code, replaced with ggml-backend offload function

* Default to all dedicated GPUs

* Add fallback for integrated GPUs if no dedicated GPUs are found

* Add debug info which device is allocating memory

* Fix Intel dequant issue

Fix validation issue

* Fix Vulkan GGML_OP_GET_ROWS implementation

* Clean up merge artifacts

* Remove Vulkan warning
This commit is contained in:
0cc4m 2024-03-29 17:29:21 +01:00 • committed by GitHub
parent d48ccf3ad4
commit ba0c7c70ab
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
7 changed files with 25616 additions and 3192 deletions
Download patch file Download diff file Expand all files Collapse all files

19
llama.cpp
View file

@ -2121,10 +2121,6 @@ struct llama_context {
ggml_backend_free(backend);
}
#ifdef GGML_USE_VULKAN
ggml_vk_free_cpu_assist();
#endif
ggml_backend_buffer_free(buf_output);
}
@ -14131,7 +14127,20 @@ struct llama_context * llama_new_context_with_model(
}
}
#elif defined(GGML_USE_VULKAN)
if (model->n_gpu_layers > 0) {
if (model->split_mode == LLAMA_SPLIT_MODE_ROW) {
LLAMA_LOG_ERROR("%s: Row split not supported. Failed to initialize Vulkan backend\n", __func__);
llama_free(ctx);
return nullptr;
}
if (model->split_mode == LLAMA_SPLIT_MODE_NONE) {
ggml_backend_t backend = ggml_backend_vk_init(0);
if (backend == nullptr) {
LLAMA_LOG_ERROR("%s: failed to initialize Vulkan backend\n", __func__);
llama_free(ctx);
return nullptr;
}
ctx->backends.push_back(backend);
} else {
for (int device = 0; device < ggml_backend_vk_get_device_count(); ++device) {
ggml_backend_t backend = ggml_backend_vk_init(device);
if (backend == nullptr) {