-
Notifications
You must be signed in to change notification settings - Fork 11.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Compile bug: libggml-vulkan.so reproducible builds issue #11306
Labels
Comments
jeffbolznv
added a commit
to jeffbolznv/llama.cpp
that referenced
this issue
Jan 20, 2025
jeffbolznv
added a commit
to jeffbolznv/llama.cpp
that referenced
this issue
Jan 20, 2025
Hi, please try #11315 |
0cc4m
pushed a commit
that referenced
this issue
Jan 23, 2025
anagri
pushed a commit
to BodhiSearch/llama.cpp
that referenced
this issue
Jan 26, 2025
tinglou
pushed a commit
to tinglou/llama.cpp
that referenced
this issue
Feb 13, 2025
arthw
pushed a commit
to arthw/llama.cpp
that referenced
this issue
Feb 26, 2025
mglambda
pushed a commit
to mglambda/llama.cpp
that referenced
this issue
Mar 8, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Git commit
667d728
Operating systems
Linux
GGML backends
Vulkan
Problem description & steps to reproduce
I'm compiling the
llamacpp
openSUSE package and found that thelibggml-vulkan.so
differs in every build.My analysis points to
https://github.com/ggerganov/llama.cpp/blob/bd38dde/ggml/src/ggml-vulkan/vulkan-shaders/vulkan-shaders-gen.cpp#L502
as the place where some sorting is missing to provide deterministic output.
First Bad Commit
No response
Compile command
we build with cmake - see https://github.com/bmwiedemann/openSUSE/blob/0da890b/packages/l/llamacpp/llamacpp.spec#L123 for details.
Relevant log output
Generated source-code varied (randomly) thusly: /home/abuild/rpmbuild/BUILD.unsorted/llamacpp-4501/build/ggml/src/ggml-vulkan/ggml-vulkan-shaders.cpp @@ -1,8 +1,892 @@ #include "ggml-vulkan-shaders.hpp" -unsigned char matmul_f16_f32_aligned_fp32_data[10276] = { +unsigned char matmul_q4_1_f16_fp32_data[10564] = { ...
The text was updated successfully, but these errors were encountered: