frozenleaves/oneDNN
1
0
Fork 0
mirror of https://github.com/uxlfoundation/oneDNN.git synced 2025-10-20 10:03:50 +08:00
Code Issues Packages Projects Releases Wiki Activity
20,988 Commits 484 Branches 211 Tags

Default Branch

main

846ba7c1d4 · x64: matmul: Enable AVX2 matmul weight decompression · Updated 2025-10-18 22:47:57 +08:00
 ZIP  TAR.GZ

Branches

dzarukin/rls-v3.8/restore_brgemm_perf

d1a336fe78 · fixup: cpu: x64: brgemm: fix handling big offsets · Updated 2025-10-19 13:16:23 +08:00
1748
129
 ZIP  TAR.GZ
rls-v3.10

024dcf2913 · x64: matmul: fix scales types check · Updated 2025-10-18 22:47:37 +08:00
128
44
 ZIP  TAR.GZ
gh-pages

9437870b96 · gh-pages: update documentation · Updated 2025-10-18 05:51:23 +08:00
20988
1
 ZIP  TAR.GZ
georgena/hybrid_cache_topology

4d0b907709 · x64: cpu: conv: hybrid cache topology win bug fix for non-hybrid cpu · Updated 2025-10-18 05:18:46 +08:00
219
4
 ZIP  TAR.GZ
dev-v3.8-thunk-preview

ae73a8bdad · Backport: cpu: x64: fix perf issue for f32 conv (#4108) · Updated 2025-10-18 03:12:55 +08:00
1748
171
 ZIP  TAR.GZ
petercad/disable_fos

3c29a00c50 · xe: jit: gemm: disable FOI/FOS usage for perf evaluation · Updated 2025-10-18 01:17:28 +08:00
19
1
 ZIP  TAR.GZ
vgvozdeva/release_notes_update_v3.10

c56a140dc1 · doc: fix grammatical error · Updated 2025-10-18 00:15:46 +08:00
128
26
 ZIP  TAR.GZ
mgouicem/main/mxfp_example

1b96fcbe4b · examples: add mxfp8 example · Updated 2025-10-17 21:38:11 +08:00
20
1
 ZIP  TAR.GZ
ryo-not-rio/ci

8d145d6648 · ci: aarch64: build oneDNN in a composite action and add c8g runners · Updated 2025-10-17 21:30:18 +08:00
20
2
 ZIP  TAR.GZ
asimonov/rnn_matmul_f32_avx512

85bfd3abb0 · cpu: rnn: implement merged iter calculation using matmul · Updated 2025-10-17 20:32:45 +08:00
20
11
 ZIP  TAR.GZ
asimonov/f32_gemv_for_brgemm_matmul_use_avx2_for_avx512

e3c08d0784 · cpu: x64: matmul: enable GEMV code path to brgemm matmul for avx512 · Updated 2025-10-17 20:30:32 +08:00
20
1
 ZIP  TAR.GZ
asimonov/f32_gemv_for_brgemm_matmul_avx512

d5aca6bfda · cpu: x64: matmul: enable GEMV code path to brgemm matmul for avx512 · Updated 2025-10-17 20:29:19 +08:00
20
1
 ZIP  TAR.GZ
ryo-not-rio/jit_sve_dw

af847ff248 · cpu: aarch64: conv: Enable SVE 128 jit uni depthwise convolutions · Updated 2025-10-17 17:54:38 +08:00
29
1
 ZIP  TAR.GZ
syurkevi/concat_block_write1_bugfix

540067c082 · xe: concat: add ext_idx calculation for blocked write=1 cases · Updated 2025-10-17 09:55:19 +08:00
28
1
 ZIP  TAR.GZ
kealanba/dyn_scale_main

0e6e3fe091 · xe: gemm: enable jit dynamic scaling · Updated 2025-10-17 08:03:08 +08:00
28
6
 ZIP  TAR.GZ
skazakov/host-scalar-gemm-EXP

7574d4ce9f · printing · Updated 2025-10-17 07:01:19 +08:00
21
1
 ZIP  TAR.GZ
dzarukin/try_pimpl_for_async_reference_2

4cb0e039cd · graph: backend: dnnl: use dnnl_primitive_execute_without_tp_hook · Updated 2025-10-17 06:12:49 +08:00
128
76
 ZIP  TAR.GZ
spalicki/l0_backend

633a03d736 · gpu: intel: add Level Zero backend · Updated 2025-10-17 01:28:55 +08:00
28
1
 ZIP  TAR.GZ
asztukow/pool-diff-dt-src-dst

5077856778 · cpu: pool: enable different dt on src and dst · Updated 2025-10-16 16:15:42 +08:00
30
1
 ZIP  TAR.GZ
dyoussif/gemm_gs_fix

3749ecc4a6 · xe: gemm: jit: use stable sort compare function · Updated 2025-10-16 08:59:18 +08:00
32
1
 ZIP  TAR.GZ