onnxruntime
[x86] matmulnbit x64 kernel for 8bits
#24491
Merged

[x86] matmulnbit x64 kernel for 8bits #24491

fajin-corp merged 44 commits into main from fajin/matmul8bit_x64_kernel
fajin-corp
fajin-corp fajin-corp requested a review 319 days ago
github-actions
github-actions commented on 2025-04-21
github-advanced-security
github-advanced-security commented on 2025-04-21
liqunfu
liqunfu dismissed these changes on 2025-04-21
fajin-corp fajin-corp dismissed their stale review via 33af3895 319 days ago
fajin-corp added quant8 interface
75d4ab55
fajin-corp added q8 packb and blocksum
d5ef23ec
fajin-corp added interface for sq8 int8 matmul
71d1f1cd
fajin-corp fix prepack stride
caf4737c
fajin-corp finished q8 matmul m2 n4
ced7c66f
fajin-corp finished Q8Int8GemmR2xC1BlkLen16Avx2
fc3e9799
fajin-corp finished Q8Int8GemmR2xC1BlkLen16Avx2
517b9891
fajin-corp finished block16 avx2/vnni
e5da9d40
fajin-corp finished Q8Int8GemmR2xC4BlkLen32Avx2
d270008a
fajin-corp finished Q8Int8GemmR2xC1BlkLen32Avx2
4d45bd84
fajin-corp finished Q8Int8GemmR1xC4BlkLen32Avx2
13097f0b
fajin-corp finished Q8Int8GemmR1xC1BlkLen32Avx2
1287666a
fajin-corp finished block64 avx2/vnni
8e24a764
fajin-corp added avx512/vnni kernel interface
04071b30
fajin-corp finished Q8Int8GemmR2xC4BlkLen16Avx512
23cb557a
fajin-corp finished MlasQ8Int8GemmKernelBlkLen16Avx512
1548d95d
fajin-corp finished MlasQ8Int8GemmKernelBlkLen32Avx512
d0df455d
fajin-corp finished MlasQ8Int8GemmKernelBlkLen64Avx512
d7ae11bc
fajin-corp finished MlasQ8Int8GemmKernelBlkLen128Avx512
702962fe
fajin-corp fixed 512 vnni build
30118126
fajin-corp added prepack ut
02f74e44
fajin-corp added avx flags
89a38948
fajin-corp finished ut
6c64086f
fajin-corp passed prepack ut
d17c840c
fajin-corp fixing gemm ut
4cdf43bd
fajin-corp passed gemm vnni ut
b3ec1648
fajin-corp fixed ut
8d5abe11
fajin-corp added benchmark
f99ebffc
fajin-corp fixed avx2vnni ut
f11de12a
fajin-corp debugging ut
1cb47c8e
fajin-corp passed matmulnbits op ut
d6a15050
fajin-corp fix linting
8ab12ccf
fajin-corp try to fix ci
a9f186af
fajin-corp fajin-corp force pushed from 066e0731 to a9f186af 319 days ago
fajin-corp fix gcc pragma
cb4398d6
fajin-corp resolving ut build error
ac17924a
fajin-corp fix ut size
ede8f759
github-actions
github-actions commented on 2025-04-22
fajin-corp fix linting
a5332fd8
fajin-corp reduce test count
809e8dbc
fajin-corp configure cpu ep for ut
367c5685
fajin-corp fix dml and webgpu error
30cbf729
fajin-corp fix coreml ut
daf30649
github-actions
github-actions commented on 2025-04-23
fajin-corp fix linting
5e973df8
fajin-corp reduce test count
159f7f1c
fajin-corp fajin-corp closed this 317 days ago
fajin-corp fajin-corp reopened this 317 days ago
yihonglyu
yihonglyu commented on 2025-04-23
tianleiwu tianleiwu added release:1.22.0
fajin-corp update comments
d99c6f68
liqunfu
liqunfu
liqunfu approved these changes on 2025-04-24
fajin-corp fajin-corp merged 7801c513 into main 317 days ago
fajin-corp fajin-corp deleted the fajin/matmul8bit_x64_kernel branch 317 days ago
fajin-corp
snnn snnn removed release:1.22.0
snnn

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone