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

Vulkan Fixes (#5223)

Browse source 
* Fix Vulkan F16 models

* Fix Vulkan context shift crash

* Add Vulkan to common.cpp dump_non_result_info_yaml function

* Fix bug in Vulkan CPY op

* Fix small matrix multiplication errors in AMD GPUs on Windows or with amdvlk

Co-authored-by: Engininja2 <139037756+Engininja2@users.noreply.github.com>

---------

Co-authored-by: Engininja2 <139037756+Engininja2@users.noreply.github.com>
This commit is contained in:
0cc4m 2024-01-31 11:44:19 +01:00 • committed by GitHub
parent d62520eb2c
commit f8e9140cb4
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 704 additions and 1267 deletions
Download patch file Download diff file Expand all files Collapse all files

4
ggml_vk_generate_shaders.py
View file

@ -19,8 +19,8 @@ shader_int8_ext = """
# Type-specific defines
shader_f16_defines = """
#define QUANT_K 32
#define QUANT_R 2
#define QUANT_K 1
#define QUANT_R 1
#define A_TYPE float16_t
"""