Skip to content

[hipBLASLt] [TensileLite] Add tail loop support in subtile path for BF16 for all k sizes #7661

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.

Sign up for GitHub

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

Open
bnemanich wants to merge 63 commits into
base: users/bnemanich/subtile-tailloop-k32-rebased
Choose a base branch
from
Open

[hipBLASLt] [TensileLite] Add tail loop support in subtile path for BF16 for all k sizes #7661

Changes from all commits
Commits
Show all changes
63 commits
Select commit Hold shift + click to select a range
3b488dd
subtile: allow low ASEM for bf16 any-K tail loop
bnemanich May 20, 2026
357d9ae
subtile: add bf16 any-K unit + yaml regression tests
bnemanich May 20, 2026
fd3846b
subtile: add BufferLoadB16/BufferLoadU16 rocisa bindings
bnemanich May 22, 2026
48486b4
subtile: add largemt anyK regression yaml + unit pins per sebvince's …
bnemanich May 23, 2026
4fd44e9
subtile: replay bf16-anyk tail-mask evolution onto K32-on-develop
bnemanich May 25, 2026
9004947
docs(rocsolver): update install pages for 7.13 (#7364)
peterjunpark May 25, 2026
a889cf3
[stinkytofu] Prepare for TheRock subproject integration (#7556)
KKyang May 25, 2026
91ee252
docs(rocrand): update install pages for 7.13 (#7366)
peterjunpark May 25, 2026
9d7c386
[CI] [Documentation] Fix docs synchronization pipeline introduced by …
alexxu-amd May 25, 2026
2cec672
[dnn-providers] Run test_name_validator.py as a ctest for providers (…
adickin-amd May 25, 2026
5ff4d1a
[rocFFT] Batched distributed transform in MPI sample
af-ayala May 25, 2026
c158fde
rocfft: Replace rocm-smi with amd-smi in perf scripts (#7696)
Abuudiii May 25, 2026
268d30e
Add FP32-to-FP8 conversion with stochastic rounding
StaceyLai May 4, 2026
900b47b
Support FP32 to FP8 stochastic rounding conversion without v_prng_b32
StaceyLai Nov 11, 2025
6824cec
Fix CI: add StochasticRounding to test_storeD_roundtrip ProblemType
StaceyLai May 25, 2026
739b4a0
subtile: scrub reviewer-name references from bf16-anyk delta
bnemanich May 25, 2026
c3895f4
subtile: trim narrative comments in tail scaffold
bnemanich May 25, 2026
02827ba
subtile: replace long MIT header with short form on new test files
bnemanich May 25, 2026
857751b
subtile: move tail SRD tighteners to Components/Subtile/SubtileTailSr…
bnemanich May 25, 2026
1429817
subtile: move tail-mask helpers to Components/Subtile/SubtileTailMask
bnemanich May 25, 2026
bd205dc
subtile: move tail-loop scaffold to Components/Subtile/SubtileTailSca…
bnemanich May 25, 2026
7ae6c53
subtile: drop reviewer-name trailer from extracted scaffold
bnemanich May 25, 2026
2531c99
subtile: fold test_solution_subtile_anyk_largemt into test_solution_s…
bnemanich May 25, 2026
7da12b2
Adapt tolerances for spsm / sptrsm on HawkPoint for ill-conditioned m…
amontoison May 25, 2026
f4471a8
Fix yaml type mismatch in library logic for gfx1152/gfx1153/gfx1200 (…
Alex-Vasile May 25, 2026
60e5be1
Fix yaml type mismatch in library logic for aquavanjaram (a) (3/13 of…
Alex-Vasile May 26, 2026
2828780
[stinkytofu] Add comgr support for runtime toolchain capability probi…
KKyang May 26, 2026
402dbad
[CK_TILE] Use Persistent Scheduling for FMHA BWD Group Deterministic …
DDEle May 26, 2026
b9db673
Update instruction: running clang-tidy (#7701)
KKyang May 26, 2026
a0e9f50
[hipDNN] ALMIOPEN-1869 Add optional hipdnn-frontend Python bindings t…
tvy-amd May 26, 2026
32ccae3
Enable HalfPLR for MXF8 in gfx1250 (#7453)
boringmorning May 26, 2026
9296d81
[tensilelite][stinkytofu] Fix PGR1 token bug (#7730)
hcman2 May 26, 2026
18ee0d8
[hipsparselt] Refactor LRVWMetadata (#7487)
leowu2017 May 26, 2026
54aed1e
[CK] Add rocm_ck spec factories: GemmSpec, makeSpec() (#7180)
shumway May 26, 2026
0fab8d8
[CK TILE] Unification Work – Add MFMA specialisations for `fp64_t` (#…
yungshengtu May 26, 2026
c659ffd
subtile: reject single-wave WG=(1,1) + large WT + K-tail at Solution …
bnemanich May 26, 2026
5177400
[stinkytofu] Make // and ; comment-stripping block-comment aware
darrenhsieh-amd May 20, 2026
ef165a5
[stinkytofu] Add RaiseVgprMsbPass with Insert byte-encoding fix (#7727)
darrenhsieh-amd May 26, 2026
45583bd
[CK_TILE][FMHA] Improve precision of mxfp4 FMHA with fp6 for matrix P…
ex-rzr May 26, 2026
c939faf
[hipDNN] ALMIOPEN-1869 Enable clang-tidy for Python bindings in CI (#…
tvy-amd May 26, 2026
e4d5f04
[tensilelite] Add testpaths and norecursedirs to pytest.ini (#7571)
talumbau May 26, 2026
7e27be9
Add MIOpen integration test for batchnorm unhappy activation (#7404)
Aleksandar301 May 26, 2026
d23f097
[ci] bump TheRock hash to `974db70` (2026-05-18) (#7582)
kailash-khalasi May 26, 2026
45eb1a1
docs(hipfft): update install pages for 7.13 (#7375)
peterjunpark May 26, 2026
ddda8ac
[CK_TILE] Add save_matrix_txt() and extract HostTensor I/O to free fu…
AviralGoelAMD May 26, 2026
8bc1843
Fix ZeroDivisionError and silent failure when TransposeLDS=0 is incom…
aadeshamd May 26, 2026
ef8a4cf
[ALMIOPEN-1951] [miopen] Fix install RPATH for MIOpenDriver and CK ba…
SreecharanGundaboluAMD May 26, 2026
ffbde87
[MIOpen] Add JSON performance logs for MIOpen convolution driver comm…
jdcampbe May 26, 2026
867bece
[CK_TILE] Adding steps in Stream-K Tile Engine (#6511)
arai713 May 26, 2026
5b3f4b7
[CK_TILE] Stream-K XCD remapping (#4279)
assistant-librarian[bot] May 26, 2026
52f486a
[rocblas] Fix install.sh/rmake.py when CMAKE_GENERATOR=Ninja is set i…
evedovelli May 21, 2026
e916514
Add missing dependency package to Dockerfiles
evedovelli May 25, 2026
258c1fb
Bump urllib3 from 2.6.3 to 2.7.0 in /shared/tensile/docs/sphinx
dependabot[bot] May 11, 2026
7c0d7aa
[Hipblaslt] [Subtiling] Add non-uniform partition size to Logical Sch…
sebvince May 26, 2026
07c4e5e
consistently weaken new k==0 test so we don't verify that alpha is ig…
TorreZuk May 21, 2026
d3f057b
[MIOpen] Add initial MIOpen support for gfx1250 (#7587)
SreecharanGundaboluAMD May 26, 2026
99c8e5b
Bump gitpython from 3.1.49 to 3.1.50 in /shared/tensile/docs/sphinx
dependabot[bot] May 9, 2026
f96e909
[tensilelite] Fix test_PlaceholderMerge xfail for TheRock CI (#7716)
archana-ramalingam May 26, 2026
62d3a26
[hip-kernel-provider] Remove hip includes from RTC kernels (#7563)
EwanC May 26, 2026
e85adb5
Merge branch 'develop' into users/bnemanich/subtile-bf16-anyk
bnemanich May 26, 2026
f9dd411
subtile: add PGR=1 to largemt anyk yaml to clear post-develop-merge V…
bnemanich May 27, 2026
b88565f
subtile: make SubtileTailSrdTighten swizzle-size factor explicit
bnemanich May 27, 2026
784ea30
subtile: trim verbose comments in tail SRD tighten swizzle refactor
bnemanich May 27, 2026

Sorry, this diff is taking too long to generate.

It may be too large to display on GitHub.