Skip to content
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

[GPU] gemm data type align #2874

Open
wants to merge 6 commits into
base: main
Choose a base branch
from
Open

[GPU] gemm data type align #2874

wants to merge 6 commits into from

Conversation

dyoussif
Copy link
Contributor

addresses failures in MFDNN-13096.

Fixup some strategies with incorrect alignment by running reinfo.
Also, in the case where strategy with larger data types is selected (for example SSS kernel for f32:s8:f16 case), keep alignment requirements according to the smaller data type.

@dyoussif dyoussif requested a review from a team as a code owner March 12, 2025 22:49
@github-actions github-actions bot added the platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel label Mar 12, 2025
@dyoussif
Copy link
Contributor Author

make test
disable device_cpu
disable benchdnn_all
enable benchdnn_matmul

@srazumov
Copy link
Contributor

make test Lightweight_scans

@srazumov
Copy link
Contributor

make test
disable device_cpu
disable benchdnn_all
enable benchdnn_matmul

@echeresh echeresh force-pushed the dyoussif/gemm_data_align branch from 2d152a6 to 893b1df Compare March 14, 2025 17:02
@dyoussif
Copy link
Contributor Author

make test
disable device_cpu
disable benchdnn_all
enable benchdnn_matmul

@dyoussif dyoussif force-pushed the dyoussif/gemm_data_align branch from 893b1df to 4769aa1 Compare March 17, 2025 17:51
@dyoussif
Copy link
Contributor Author

make test
disable device_cpu
disable run_scans
disable benchdnn_all
enable benchdnn_nightly
enable benchdnn_matmul
enable benchdnn_ip
enable arch_xe-lp
enable arch_xe2-lpg
enable arch_xe3-lpg

1 similar comment
@dyoussif
Copy link
Contributor Author

make test
disable device_cpu
disable run_scans
disable benchdnn_all
enable benchdnn_nightly
enable benchdnn_matmul
enable benchdnn_ip
enable arch_xe-lp
enable arch_xe2-lpg
enable arch_xe3-lpg

@dyoussif dyoussif force-pushed the dyoussif/gemm_data_align branch from 254f265 to 6f61324 Compare March 18, 2025 20:00
@echeresh echeresh force-pushed the dyoussif/gemm_data_align branch 2 times, most recently from f0aed5e to 28883c4 Compare March 19, 2025 01:21
@dyoussif dyoussif force-pushed the dyoussif/gemm_data_align branch from 28883c4 to f17a098 Compare March 22, 2025 00:29
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
platform:gpu-intel Codeowner: @oneapi-src/onednn-gpu-intel
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants