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

Rebase #1900 #2009

Merged
merged 916 commits into from
Sep 30, 2022
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
916 commits
Select commit Hold shift + click to select a range
8e074f4
hash update - bug fix for branches (#83865)
clee2000 Aug 23, 2022
4e90526
[FSDP] Remove unneeded checks (#83150)
rohan-varma Aug 22, 2022
b29a074
[BE] Revert distributed change in https://github.com/pytorch/pytorch/…
rohan-varma Aug 22, 2022
6cf7eb0
Transpose scheduler small dim sizes better support (#1910)
zasdfgbnm Aug 23, 2022
04d8da8
Optimize transpose copy on CPU using fbgemm transpose (#83327)
CaoE Aug 23, 2022
20cf109
Grouped grid welford (#1921)
naoyam Aug 23, 2022
3dfb8df
[ONNX] Use `errors.SymbolicValueError` for more context (#83332)
justinchuby Aug 23, 2022
a419e48
[quant][fx] Add support for quantized matmul (#83885)
jerryzh168 Aug 23, 2022
3c3c89e
Misc fixes/tuning for transpose scheduler (#1912)
zasdfgbnm Aug 23, 2022
dd67d52
[nn] split rnn_utils test from test_nn.py (#83675)
kshitij12345 Aug 23, 2022
7c20ad3
[optim] rprop: handle complex params as independent real params (#83858)
kshitij12345 Aug 23, 2022
ed949e2
[xla hash update] update the pinned xla hash (#83899)
pytorchmergebot Aug 23, 2022
247468b
[ROCm] More Sparse UTs enablement and more hipification mappings. (#7…
jpvillam-amd Aug 23, 2022
4dfa6d2
Normalize DLPack stride to 1 where shape < 2 (#83158)
mattip Aug 23, 2022
58170fb
Remove DBR quantization from the codebase (#83642)
vkuzo Aug 22, 2022
7ebdb4c
Refactored ops on size to be dispatcher ops (#83719)
Chillee Aug 23, 2022
0270a70
Fix stride issue with faketensors (#83822)
Chillee Aug 23, 2022
14a53e6
Nullary RNGOp (#1892)
zasdfgbnm Aug 23, 2022
a315a2c
[ROCm] restore MIOpen benchmark flag default to true (#82656)
jeffdaily Aug 23, 2022
7e38684
Update retry action to latest version (#83911)
janeyx99 Aug 23, 2022
e4af53c
[PyTorch] Remove unused sstream/string includes from c10/macros/Macro…
swolchok Aug 22, 2022
df04841
[functorch] add linalg cross batch rule (#83759)
Aug 22, 2022
01434c2
Improve DistanceKernel.cu (#83811)
chengscott Aug 23, 2022
75ec7b7
reinplace pass: bugfix for output node replacement (#83845)
bdhirsh Aug 22, 2022
8db04c1
reinplace pass: special handling for view_scatter ops (#83846)
bdhirsh Aug 22, 2022
8793cd2
Move ATenNVRTC.h include from `jit_utils.h` to `jit_utils.cpp` (#83886)
malfet Aug 23, 2022
cb488e6
Allow None arguments for elementwise type promotion wrapper and fix c…
IvanYashchuk Aug 23, 2022
9beddde
Enable NCCL_DESYNC_DEBUG when TORCH_DISTRIBUTED_DEBUG=DETAIL (#83881)
kwen2501 Aug 23, 2022
7f0198e
Strenghten preconditions of linalg.cross (#83798)
lezcano Aug 23, 2022
8184359
Fix view_func replay in no-grad mode (#83872)
soulitzer Aug 23, 2022
732255f
[vulkan] Add VMA as a third_party subrepo (#83906)
SS-JIA Aug 23, 2022
5204b8e
[torchgen] Add documentation for `autogen` keyword (#83610)
larryliu0820 Aug 22, 2022
8f9ae35
remove assertEqualIgnoreTypes from test/distributions/test_distributi…
liuk22 Aug 23, 2022
9095030
[fix] edge case in `MaxPool1d` and add ErrorInputs (#83553)
khushi-411 Aug 23, 2022
a802603
[complex] conv_transpose1d (#79694)
kshitij12345 Aug 23, 2022
bbe803c
Revert "Strenghten preconditions of linalg.cross (#83798)"
pytorchmergebot Aug 23, 2022
25dd2a0
Fix load_extra_only api for flatbuffers and enable flatbuffers in mob…
Aug 23, 2022
4808bda
Prefer signal from land checks over PR signals (#83715)
ZainRizvi Aug 23, 2022
8fae702
Don't introduce new overload for SymInt (#83628)
ezyang Aug 23, 2022
21bc77c
Remove CoreMLMemoryObserver (#83703)
jmdetloff Aug 23, 2022
7c8d265
ci: Remove dead code related to android uploads (#83930)
seemethere Aug 23, 2022
8907217
[fx][pass infra] Adding error catching (#83933)
angelayi Aug 23, 2022
fc470cf
Back out "Support regex-style matching for Any and Oneof (#82853)" (#…
Aug 24, 2022
591222f
Fix use-dict-literal lint (#83718)
kit1980 Aug 24, 2022
84e45e7
Revert "Optimize transpose copy on CPU using fbgemm transpose (#83327)"
pytorchmergebot Aug 24, 2022
58f61d5
Add hypothesis to requirements.txt (#83740)
ezyang Aug 23, 2022
c47e045
[fbia] Keep Track of full qualified name before and after remote shar…
nanoax Aug 24, 2022
84f0411
add merge blocking to ci: sev template (#83940)
clee2000 Aug 24, 2022
6edcf8e
Move nnapi code from ATen common code to specific library (#83748)
kirklandsign Aug 24, 2022
7fdc2f7
Task: T129772171 remove assertEqualIgnoreTypes from test/test_nn.py (…
Ishan-Rajgarhia Aug 24, 2022
2c76d05
[Nested Tensor] Make offset copy and move assignment more explicit. (…
drisspg Aug 24, 2022
754d7f0
Remove conj kernels for real dtypes (#80374)
peterbell10 Aug 23, 2022
df70714
[BE][CUDA] Use packed_accessor64 (#83949)
malfet Aug 24, 2022
0491e1a
Support returning symbolic strides from t.stride() in Python (#83842)
ezyang Aug 23, 2022
67aed39
Support the XPU backend untyped storage (#83952)
chengjunlu Aug 24, 2022
432c508
Support NCCL Premul Sum (#81272)
crcrpar Aug 24, 2022
0ae298f
Test type promotion assertignoretypes (#83867)
Aug 24, 2022
fa241fd
[Profiler] record nn.Module's parameters (#83209)
slgong-fb Aug 24, 2022
ce7a9f9
[xla hash update] update the pinned xla hash (#83967)
pytorchmergebot Aug 24, 2022
b34e3b9
Fix `ir_utils::hasBlockSync` + misc fixes in transpose scheduler (#1924)
zasdfgbnm Aug 24, 2022
4a18d0a
Fix LTC build warnings (#83955)
henrytwo Aug 24, 2022
0bdcfcb
Strenghten preconditions of linalg.cross (#83798)
lezcano Aug 24, 2022
4737b33
Make linalg.inv composite of linalg.solve (#80074)
lezcano Aug 24, 2022
3e6e0a1
Support a stable double backward on linalg.det for real inputs (#80217)
lezcano Aug 24, 2022
4eb02e8
[LTC] Add custom lazy tensor save function (#83294)
antoniojkim Aug 24, 2022
7a81525
move pooling test from test_nn to test/nn/test_pooling (#83915)
kshitij12345 Aug 24, 2022
3f88171
[ONNX] Remove static None graph output (#82623)
BowenBao Aug 24, 2022
30a5583
[TorchTidy Fix] Don't try to collect strides for non-strided tensors …
Gamrix Aug 24, 2022
5c49c7b
[WIP] Validating input_col for certain datapipes (#80267)
bushshrub Aug 24, 2022
0831813
support more symintnode operations (#83877)
Krovatkin Aug 23, 2022
2b805e3
add arithmetic ops (#83878)
Krovatkin Aug 23, 2022
b842670
logical ops (#83879)
Krovatkin Aug 23, 2022
5e01fb9
strip SymIntNodes off in the mobile builds (#83938)
Krovatkin Aug 23, 2022
d79ccb7
[pthreadpool] Cap max thread count to fix TSAN issues (#83950)
digantdesai Aug 24, 2022
3a9ae51
Skip NCCL slimming for cxx11 libtorch builds (#83959)
malfet Aug 24, 2022
f7e668b
add hud link to merge failure message (#83946)
clee2000 Aug 24, 2022
b745e5f
Check all CUDA API calls for errors in benchmarks/cpp/nvfuser (#74920…
r-barnes Aug 24, 2022
a8a36c4
[frontend] Fix tensor list alias annotation (#84005)
larryliu0820 Aug 24, 2022
a5564c4
Suppress Anomaly mode warning message (#83966)
SherlockNoMad Aug 24, 2022
ef782e7
Support BF16 for fast layernorm (#83971)
thomasw21 Aug 24, 2022
7b1a056
Map new CUDA error handling to HIP (#75032) (#83953)
r-barnes Aug 24, 2022
a741927
Improve Normalization.cuh (#83871)
chengscott Aug 24, 2022
67f0940
Check all CUDA API calls for errors in test/ (#74921) (#83954)
r-barnes Aug 24, 2022
f38a32c
remove duplicate WarpReduceSum (#83757)
chengscott Aug 24, 2022
d565772
Set python build-docs timeout to 30 minutes and cpp build-docs timeou…
huydhn Aug 24, 2022
f5bfa4d
[ROCm] Enable test_multiprocessing tests (#82356)
jaglinux Aug 24, 2022
1eff853
Pin conda to 4.13.0 (#83991)
janeyx99 Aug 24, 2022
37d3db7
Deletes CCACHE_DISABLE and SCCACHE_DISABLE from nccl.cmake (#84007)
janeyx99 Aug 24, 2022
3b11b80
Named pipe based watchdog timer (#83695)
bchen2020 Aug 24, 2022
e722566
Enhance add_out_dense_sparse_cpu for hybrid sparse tensor (#23057)
JianpingChen066 Aug 24, 2022
bf8d5e8
Pretty print stack trace with gm.print_readable() (#83706)
SherlockNoMad Aug 24, 2022
101709f
Add comments for block_reduce.cuh (#83825)
chengscott Aug 24, 2022
cef522a
Add docstring type guidelines for list & tuple to `CONTRIBUTING.md` (…
ProGamerGov Aug 24, 2022
4a6726a
use condensed disabled tests file (#84017)
clee2000 Aug 25, 2022
5321bf5
Revert "Make linalg.inv composite of linalg.solve (#80074)"
pytorchmergebot Aug 25, 2022
7a02ee5
Revert "[xla hash update] update the pinned xla hash (#83967)"
pytorchmergebot Aug 25, 2022
a7edf71
Revert "Don't introduce new overload for SymInt (#83628)"
pytorchmergebot Aug 25, 2022
6b59759
[Quant] Vectorize scalar remainder in quantized kernel for normalizat…
Xia-Weiwen Aug 25, 2022
e90db17
Increase timeout for linux binary builds (#84008)
atalman Aug 25, 2022
b21a6ff
[NVFuser] Upstream push 0811 (#83239)
jjsjann123 Aug 23, 2022
b6ba419
[TorchTidy] Adding support for unique tensor identifiers (#80266)
Gamrix Aug 24, 2022
a013597
fix oneDNN channels_last path issue (#83653)
XiaobingSuper Aug 25, 2022
460636a
[caffe2] Remove last clang-for-cuda sources (#84021)
andrewjcg Aug 25, 2022
1f61c39
Revert "Support NCCL Premul Sum (#81272)"
pytorchmergebot Aug 25, 2022
d5af2a7
Revert "[TorchTidy] Adding support for unique tensor identifiers (#80…
pytorchmergebot Aug 25, 2022
2000eba
NCCL: Re-enable parallel builds (#83696)
peterbell10 Aug 24, 2022
c03f8ab
[fx+scripting] Adding num_iter_1 and num_iter_2 params LearningRate o…
Aug 25, 2022
8b8942b
Fix dumb make_fx issue (#84011)
Chillee Aug 25, 2022
c00f0c8
[fx] add deferred weights (xl_weight) and tracing for xl_embedding_ba…
alexbeloi Aug 25, 2022
02c3781
Enable cache action for lint workflow (#84026)
huydhn Aug 25, 2022
63cbdc9
switching the exact check to isinstance check (#84023)
Krovatkin Aug 25, 2022
e3c89d0
Disable autocast cache during aotdispatch (#84035)
Chillee Aug 25, 2022
f5a3515
Make linalg.inv composite of linalg.solve (#80074)
lezcano Aug 25, 2022
31f1517
add qscheme check for quantization observer (#80126)
XiaobingSuper Aug 25, 2022
00cb184
[functorch] add batching rule for fill_.Tensor (#84015)
zou3519 Aug 24, 2022
56fef4e
fix `NoneType` object has no attribute `python_exit_status` (#83985)
ejguan Aug 25, 2022
e2f75d6
Decomposition - batch_norm, save_mean and save_variance always float3…
anijain2305 Aug 25, 2022
a83d7d8
enable qlinear dynamic parallelization with fbgemm (#84033)
jianyuh Aug 25, 2022
c92e5ac
[quant][ao_migration] `torch.nn.quantized.modules` → `torch.ao.nn.qua…
Aug 25, 2022
d32a762
[quant][ao_migration] `torch.nn.quantized.dynamic` → `torch.ao.nn.qua…
Aug 25, 2022
b1455f9
[quant][ao_migration] `torch.nn.quantized._reference` → `torch.ao.nn.…
Aug 25, 2022
29e83b6
[quant][ao_migration] `torch.nn.quantizable` → `torch.ao.nn.quantizab…
Aug 25, 2022
2f04ba2
[quant][ao_migration] `torch.nn.qat` → `torch.ao.nn.qat` (#78716)
Aug 25, 2022
86e134d
disable c10::SymIntNode tests on mobile (#84066)
Krovatkin Aug 25, 2022
caaa723
[GHF][BE] Move merge rules to yaml (#84065)
malfet Aug 25, 2022
0c6a616
run functorch decomps after functionalization when enabled (#83992)
bdhirsh Aug 25, 2022
b82c74d
functionalization: support inplace views on inputs (#83993)
bdhirsh Aug 25, 2022
4c19981
[DataPipe] Reset Shuffler's iterator when NotStarted (#83535)
ejguan Aug 24, 2022
3c2a078
[ONNX] Assign ONNXScopeName during function substituion (#82039)
BowenBao Aug 24, 2022
ced2ca8
Torch cond operator, python dispatch, pyoperator (#83154)
voznesenskym Aug 25, 2022
9f4626e
[vulkan] use VMA at third-party (#83934)
SS-JIA Aug 25, 2022
562a021
[GHF] Land validation should not change default branch (#84084)
malfet Aug 25, 2022
bf25a14
[ONNX] Add runtime type checking to `export` (#83673)
justinchuby Aug 25, 2022
85f82f7
example program for paper intro (#83945)
migeed-z Aug 25, 2022
a4a55f5
New TORCH_UCC_BLOCKING_WAIT env variable (#81791)
zasdfgbnm Aug 25, 2022
7e5c76d
Make graph_module.print_readable() discoverable (#83960)
SherlockNoMad Aug 25, 2022
b35e7c5
Fix FSDP not all outputs used in loss (#83195)
rohan-varma Aug 25, 2022
d2f3740
Silence namedtuple warning in dist (#84072)
rohan-varma Aug 25, 2022
9790d90
Don't introduce new overload for SymInt (#83628)
ezyang Aug 26, 2022
15b560a
Fix missing include for size_t (#84088)
zasdfgbnm Aug 26, 2022
089101f
Fix small typo in cuda.rst (#84012)
cjauvin Aug 26, 2022
993d8bb
Use size to check same tensor sizes in reduce_scatter and allgather (…
minsii Aug 26, 2022
b247dcf
Separate kernel compilation API from kernel execution API (#1914)
jjsjann123 Aug 26, 2022
f03ab28
Use an unused variable (#84073)
kit1980 Aug 26, 2022
d54fad5
Remove unreachable except block (#84070)
kit1980 Aug 26, 2022
88de85e
Upstream cherry pick fixes 0811 (#1934)
jjsjann123 Aug 26, 2022
ddedc29
[xla hash update] update the pinned xla hash (#84043)
pytorchmergebot Aug 26, 2022
9a236c7
Made some minor cleanups to decompositions (#83814)
Chillee Aug 26, 2022
cddf96c
Fix preconditions of adaptive_avg_pooling2d (#84061)
lezcano Aug 26, 2022
65ea3d0
[composite compliance] cov, corrcoef (#82954)
kshitij12345 Aug 26, 2022
b429a17
Enable -Wunused-local-typedefs (#83708)
peterbell10 Aug 25, 2022
a089114
Use C10_HAS_CPP_ATTRIBUTE to simplify nodiscard definition (#83976)
peterbell10 Aug 25, 2022
e9dff85
[functorch] add lstsq batch rule (#82325)
Aug 25, 2022
c0a187a
do not use deprecated functions (#1935)
naoyam Aug 26, 2022
b4ac2c8
Map IterationDomains through view operations. (#1919)
csarofeen Aug 26, 2022
582c083
mac circleci workflows (#82780)
clee2000 Aug 26, 2022
c6348a7
Add type hints to torch.save, torch.load (#83937)
adamjstewart Aug 26, 2022
527a160
Expose ProcessGroup::Work.wait() API to TorchScript (#83303)
mrshenli Aug 26, 2022
f93446a
Update proxy_tensor.py to support List input/output (#83302)
mrshenli Aug 26, 2022
ec5b83f
Make allreduce compatible with fx ProxyTensor (#84126)
mrshenli Aug 26, 2022
63132a0
Propagate permissive mapping information into indexing pass (#1929)
shmsong Aug 26, 2022
681c387
[ONNX] Clean up patch functions (#83136)
justinchuby Aug 25, 2022
014a333
[Profiler][Minor] Extend Python bindings (#83622)
Aug 26, 2022
61b9d8f
[Profiler][Trivial] Add null handling to `AppendOnlyList::copy` memcp…
Aug 26, 2022
6a58603
Update Dynamo pin (#83829)
anijain2305 Aug 26, 2022
b5c2b0b
make job pass even if monitoring script fails (#84068)
clee2000 Aug 26, 2022
6a36662
[ONNX] Export node and value with scope name (#82040)
BowenBao Aug 26, 2022
e0f0c8e
Add support to traverse all python collection objects (#84079)
ejguan Aug 26, 2022
3fae89d
Read via FileAdapter when loading files in torch if not flatbuffer (#…
Aug 26, 2022
82efb0e
Enable cache action for windows and other minor workflows (#84093)
huydhn Aug 26, 2022
aa4be48
[Nested Tensor] do not use at::cuda::getDefaultCUDAStream() (#84134)
jeffdaily Aug 26, 2022
ef98f36
Fix a bug (#1936)
naoyam Aug 26, 2022
eb2fa2e
[fx][pass] Fix type of exception (#84094)
angelayi Aug 25, 2022
f4dc7b3
[Profiler][Trivial] Cleanup ExperimentalConfig (#83890)
Aug 26, 2022
69e9f90
[Profiler] Add `disabled` and `global` methods to ProfilerConfig. (#8…
Aug 26, 2022
a423c96
[DataPipe] Convert MapDataPipe.shuffle to IterDataPipe (#83202)
ejguan Aug 26, 2022
bed85cc
[Prim] Implement group_norm_backward (#84037)
malfet Aug 27, 2022
38e5e4a
Revert "[xla hash update] update the pinned xla hash (#84043)"
pytorchmergebot Aug 27, 2022
c7edcd6
Revert "Don't introduce new overload for SymInt (#83628)"
pytorchmergebot Aug 27, 2022
43620b7
[AOT Autograd] Redirect named_parameters to original mod (#84157)
anijain2305 Aug 27, 2022
092fe71
[Nested Tensor] detach (#84078)
drisspg Aug 27, 2022
adc9a1e
Enforce explicit ProcessGroup passed into DefaultState (#84105)
rohan-varma Aug 26, 2022
e33897c
_to_copy decomp (#84108)
ngimel Aug 27, 2022
6446da1
[ONNX] Fix type annotations and enable type checking for all apis (#8…
justinchuby Aug 27, 2022
7e7694b
Add nvprims.var_mean (#83508)
IvanYashchuk Aug 27, 2022
c032b09
[xla hash update] update the pinned xla hash (#84164)
pytorchmergebot Aug 27, 2022
b8fe0ed
Revert "Make allreduce compatible with fx ProxyTensor (#84126)"
pytorchmergebot Aug 27, 2022
0dab160
Fix softmax bwd sizes. (#1890)
csarofeen Aug 27, 2022
1d0c267
Test `rand` in a fusion with zero tensor input (#1932)
zasdfgbnm Aug 27, 2022
371f282
Improve trivial reduction merge support (#1931)
zasdfgbnm Aug 27, 2022
56c00fd
Double support on all expression evaluators (#1937)
zasdfgbnm Aug 27, 2022
b2fd01e
arange support (#1933)
zasdfgbnm Aug 27, 2022
c9b144f
Replace assertEqualIgnoreTypes from common_methods_invocations.py (#8…
kuttire42 Aug 28, 2022
b078d24
Nvfuser to copy decomp to prim (#83782)
jjsjann123 Aug 28, 2022
89330aa
Tensor factories must set the output shape as its input (#1939)
zasdfgbnm Aug 28, 2022
71cd3fa
Revert "[xla hash update] update the pinned xla hash (#84164)"
pytorchmergebot Aug 28, 2022
b159a52
Revert "Add nvprims.var_mean (#83508)"
pytorchmergebot Aug 28, 2022
d8cc836
Revert "[ONNX] Fix type annotations and enable type checking for all …
pytorchmergebot Aug 28, 2022
4c254c0
Fix arange when step is negative (#1942)
naoyam Aug 29, 2022
aab10bc
Merge remote-tracking branch 'upstream/viable/strict' into HEAD
jjsjann123 Aug 29, 2022
6310948
Add full, full_like, zeros, zeros_like, ones, ones_like (#1943)
zasdfgbnm Aug 29, 2022
ac4de38
Merge pull request #1945 from csarofeen/master_merge_0828
csarofeen Aug 30, 2022
208262b
Move detection of self mapping IDs to IterDomainGraph from (#1941)
naoyam Aug 31, 2022
992e17c
test the groups the same order as they are merged (#1949)
shmsong Sep 1, 2022
ddc01e4
Exclude unsupported data types (#1951)
naoyam Sep 2, 2022
90a51f2
Some indexing cleanups, Add eye support (#1940)
zasdfgbnm Sep 2, 2022
8eafc54
Fix detection of unmappable root domains (#1952)
naoyam Sep 2, 2022
357ba22
Fill allocation with nan on tests (#1956)
zasdfgbnm Sep 2, 2022
7d1d7c8
TVDomainGuard factory (#1953)
zasdfgbnm Sep 2, 2022
df3393a
Some cleanup (#1957)
zasdfgbnm Sep 3, 2022
f7bc341
Remove unused variables (#1955)
zasdfgbnm Sep 3, 2022
a3ecb33
Improve the comments at the beginning of index_compute.h (#1946)
zasdfgbnm Sep 5, 2022
45e95fd
Allow splitting inner-most ID to create virtual innermost ID in trans…
zasdfgbnm Sep 6, 2022
93505bc
WAR on index mapping when exact and permissive maps differ (#1960)
shmsong Sep 8, 2022
c40202b
Fix dump effective bandwidth (#1962)
zasdfgbnm Sep 9, 2022
c668e13
Upstream push ci fixes (#1965)
jjsjann123 Sep 9, 2022
d9420e4
View scheduling (#1928)
csarofeen Sep 9, 2022
b7a206e
Move scheduler vectorize utilities into their own file (#1959)
csarofeen Sep 10, 2022
bd93578
Enable transpose scheduler (#1927)
zasdfgbnm Sep 11, 2022
b1bd32c
Minor fix (#1967)
naoyam Sep 12, 2022
306d4a6
Fix canScheduleCompileTime check of transpose scheduler (#1969)
zasdfgbnm Sep 13, 2022
425dce2
Add a null scheduler that helps segmenting away no-op schedules (#1835)
shmsong Sep 13, 2022
e96aacf
Enable Transpose operation (#1882)
rdspring1 Sep 13, 2022
eabe8d8
Segment self mapping fusions (#1954)
zasdfgbnm Sep 14, 2022
634820c
Add support for some empty fusion (#1981)
zasdfgbnm Sep 14, 2022
eb1dad1
Remove non-const functions, remove GpuLower instance on build, pass i…
csarofeen Sep 16, 2022
f262d9c
Add support for uniform RNG (#1986)
zasdfgbnm Sep 26, 2022
1d9858c
Minor cleanup (#1992)
naoyam Sep 26, 2022
7f42537
Fix missing thread predicates
naoyam Sep 27, 2022
8f1c7f5
Minor cleanup lower_unroll.cpp (#1994)
zasdfgbnm Sep 27, 2022
15f2f6d
Move ConcretizedBroadcastDomains to shared_ptr in GpuLower. (#1988)
csarofeen Sep 27, 2022
fcf8c09
Cleanup of lower_utils.cpp: Isolate out GpuLower usage (#1989)
csarofeen Sep 27, 2022
42ccc52
Minor build fix. (#1996)
csarofeen Sep 27, 2022
4cbe0db
Improve divisible split detection (#1970)
csarofeen Sep 27, 2022
482386c
cleanup (#1997)
naoyam Sep 27, 2022
fc235b0
Just fixes comments (#1998)
naoyam Sep 27, 2022
a7a7d57
Fix build problem (#1999)
naoyam Sep 27, 2022
3ca21eb
More strict validation (#2000)
naoyam Sep 27, 2022
dc45835
Test util cleanup (#2003)
naoyam Sep 28, 2022
93124a3
Merge branch 'devel' of github.com:csarofeen/pytorch into speculative…
zasdfgbnm Sep 28, 2022
ebeb201
fix merge
zasdfgbnm Sep 28, 2022
cde6e4d
fix merge
zasdfgbnm Sep 28, 2022
022c443
format
zasdfgbnm Sep 28, 2022
a43cb20
Make inlining even more modular (#2004)
zasdfgbnm Sep 28, 2022
967aa77
Contiguous indexing for View operations (#1990)
csarofeen Sep 28, 2022
45045cd
Enable tests previously disabled due to an aliasing bug (#2005)
naoyam Sep 28, 2022
0f9f0b4
Add matmul benchmark (#2007)
zasdfgbnm Sep 29, 2022
c90b90f
Merge branch 'devel' of github.com:csarofeen/pytorch into speculative…
zasdfgbnm Sep 29, 2022
3417e8e
Merge branch 'speculative_index' of github.com:csarofeen/pytorch into…
zasdfgbnm Sep 29, 2022
d942787
Merge branch 'speculative_index' into matmul_swizzle_gen-rebase
zasdfgbnm Sep 30, 2022
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
The table of contents is too big for display.
Diff view
Diff view
  •  
  •  
  •  
1 change: 1 addition & 0 deletions .circleci/cimodel/data/dimensions.py
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,7 @@
"102",
"113",
"116",
"117",
]

ROCM_VERSIONS = [
Expand Down
25 changes: 8 additions & 17 deletions .circleci/cimodel/data/simple/ios_definitions.py
Original file line number Diff line number Diff line change
Expand Up @@ -11,7 +11,7 @@ def __init__(self, name, custom_build_name=""):

def render(self):
extra_parts = [self.custom_build_name] if len(self.custom_build_name) > 0 else []
return "_".join([self.name] + extra_parts)
return "-".join([self.name] + extra_parts).replace("_", "-")


def get_platform(arch_variant_name):
Expand All @@ -25,30 +25,25 @@ def __init__(self, xcode_version, arch_variant, is_org_member_context=True, extr
self.is_org_member_context = is_org_member_context
self.extra_props = extra_props

def gen_name_parts(self, with_version_dots):

version_parts = self.xcode_version.render_dots_or_parts(with_version_dots)
build_variant_suffix = "_".join([self.arch_variant.render(), "build"])

def gen_name_parts(self):
version_parts = self.xcode_version.render_dots_or_parts("-")
build_variant_suffix = self.arch_variant.render()
return [
"pytorch",
"ios",
] + version_parts + [
build_variant_suffix,
]

def gen_job_name(self):
return "_".join(self.gen_name_parts(False))
return "-".join(self.gen_name_parts())

def gen_tree(self):

platform_name = get_platform(self.arch_variant.name)

props_dict = {
"build_environment": "-".join(self.gen_name_parts(True)),
"name": self.gen_job_name(),
"build_environment": self.gen_job_name(),
"ios_arch": self.arch_variant.name,
"ios_platform": platform_name,
"name": self.gen_job_name(),
}

if self.is_org_member_context:
Expand All @@ -63,16 +58,12 @@ def gen_tree(self):
WORKFLOW_DATA = [
IOSJob(XCODE_VERSION, ArchVariant("x86_64"), is_org_member_context=False, extra_props={
"lite_interpreter": miniutils.quote(str(int(True)))}),
IOSJob(XCODE_VERSION, ArchVariant("x86_64", "full_jit"), is_org_member_context=False, extra_props={
"lite_interpreter": miniutils.quote(str(int(False)))}),
IOSJob(XCODE_VERSION, ArchVariant("arm64"), extra_props={
"lite_interpreter": miniutils.quote(str(int(True)))}),
IOSJob(XCODE_VERSION, ArchVariant("arm64", "metal"), extra_props={
"use_metal": miniutils.quote(str(int(True))),
"lite_interpreter": miniutils.quote(str(int(True)))}),
IOSJob(XCODE_VERSION, ArchVariant("arm64", "full_jit"), extra_props={
"lite_interpreter": miniutils.quote(str(int(False)))}),
IOSJob(XCODE_VERSION, ArchVariant("arm64", "custom"), extra_props={
IOSJob(XCODE_VERSION, ArchVariant("arm64", "custom-ops"), extra_props={
"op_list": "mobilenetv2.yaml",
"lite_interpreter": miniutils.quote(str(int(True)))}),
IOSJob(XCODE_VERSION, ArchVariant("x86_64", "coreml"), is_org_member_context=False, extra_props={
Expand Down
105 changes: 97 additions & 8 deletions .circleci/cimodel/data/simple/macos_definitions.py
Original file line number Diff line number Diff line change
@@ -1,3 +1,7 @@
from collections import OrderedDict
from cimodel.lib.miniutils import quote


class MacOsJob:
def __init__(self, os_version, is_build=False, is_test=False, extra_props=tuple()):
# extra_props is tuple type, because mutable data structures for argument defaults
Expand All @@ -11,10 +15,14 @@ def gen_tree(self):
non_phase_parts = ["pytorch", "macos", self.os_version, "py3"]

extra_name_list = [name for name, exist in self.extra_props.items() if exist]
full_job_name_list = non_phase_parts + extra_name_list + [
'build' if self.is_build else None,
'test' if self.is_test else None,
]
full_job_name_list = (
non_phase_parts
+ extra_name_list
+ [
"build" if self.is_build else None,
"test" if self.is_test else None,
]
)

full_job_name = "_".join(list(filter(None, full_job_name_list)))

Expand All @@ -41,12 +49,93 @@ def gen_tree(self):
"10_13",
is_build=True,
is_test=True,
extra_props=tuple({
"lite_interpreter": True
}.items()),
)
extra_props=tuple({"lite_interpreter": True}.items()),
),
]


def get_new_workflow_jobs():
return [
OrderedDict(
{
"mac_build": OrderedDict(
{
"name": "macos-12-py3-x86-64-build",
"build-environment": "macos-12-py3-x86-64",
"xcode-version": quote("13.3.1"),
}
)
}
),
OrderedDict(
{
"mac_test": OrderedDict(
{
"name": "macos-12-py3-x86-64-test-1-2-default",
"build-environment": "macos-12-py3-x86-64",
"xcode-version": quote("13.3.1"),
"shard-number": quote("1"),
"num-test-shards": quote("2"),
"requires": ["macos-12-py3-x86-64-build"],
}
)
}
),
OrderedDict(
{
"mac_test": OrderedDict(
{
"name": "macos-12-py3-x86-64-test-2-2-default",
"build-environment": "macos-12-py3-x86-64",
"xcode-version": quote("13.3.1"),
"shard-number": quote("2"),
"num-test-shards": quote("2"),
"requires": ["macos-12-py3-x86-64-build"],
}
)
}
),
OrderedDict(
{
"mac_test": OrderedDict(
{
"name": "macos-12-py3-x86-64-test-1-1-functorch",
"build-environment": "macos-12-py3-x86-64",
"xcode-version": quote("13.3.1"),
"shard-number": quote("1"),
"num-test-shards": quote("1"),
"test-config": "functorch",
"requires": ["macos-12-py3-x86-64-build"],
}
)
}
),
OrderedDict(
{
"mac_build": OrderedDict(
{
"name": "macos-12-py3-x86-64-lite-interpreter-build-test",
"build-environment": "macos-12-py3-lite-interpreter-x86-64",
"xcode-version": quote("13.3.1"),
"build-generates-artifacts": "false",
}
)
}
),
OrderedDict(
{
"mac_build": OrderedDict(
{
"name": "macos-12-py3-arm64-build",
"build-environment": "macos-12-py3-arm64",
"xcode-version": quote("13.3.1"),
"python-version": quote("3.9.12"),
}
)
}
),
]


def get_workflow_jobs():
return [item.gen_tree() for item in WORKFLOW_DATA]
8 changes: 4 additions & 4 deletions .circleci/cimodel/data/simple/nightly_ios.py
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,7 @@ def __init__(self,
def get_phase_name(self):
return "upload" if self.is_upload else "build"

def get_common_name_pieces(self, with_version_dots):
def get_common_name_pieces(self, sep):

extra_name_suffix = [self.get_phase_name()] if self.is_upload else []

Expand All @@ -24,7 +24,7 @@ def get_common_name_pieces(self, with_version_dots):
common_name_pieces = [
"ios",
] + extra_name + [
] + ios_definitions.XCODE_VERSION.render_dots_or_parts(with_version_dots) + [
] + ios_definitions.XCODE_VERSION.render_dots_or_parts(sep) + [
"nightly",
self.variant,
"build",
Expand All @@ -33,14 +33,14 @@ def get_common_name_pieces(self, with_version_dots):
return common_name_pieces

def gen_job_name(self):
return "_".join(["pytorch"] + self.get_common_name_pieces(False))
return "_".join(["pytorch"] + self.get_common_name_pieces(None))

def gen_tree(self):
build_configs = BUILD_CONFIGS_FULL_JIT if self.is_full_jit else BUILD_CONFIGS
extra_requires = [x.gen_job_name() for x in build_configs] if self.is_upload else []

props_dict = {
"build_environment": "-".join(["libtorch"] + self.get_common_name_pieces(True)),
"build_environment": "-".join(["libtorch"] + self.get_common_name_pieces(".")),
"requires": extra_requires,
"context": "org-member",
"filters": {"branches": {"only": "nightly"}},
Expand Down
20 changes: 20 additions & 0 deletions .circleci/cimodel/data/simple/upload_test_stats_definition.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,20 @@
from typing import OrderedDict


def get_workflow_job():
return [
OrderedDict(
{
"upload_test_stats": OrderedDict(
{
"name": "upload test status",
"requires": [
"macos-12-py3-x86-64-test-1-2-default",
"macos-12-py3-x86-64-test-2-2-default",
"macos-12-py3-x86-64-test-1-1-functorch",
],
}
)
}
),
]
14 changes: 7 additions & 7 deletions .circleci/cimodel/data/simple/util/versions.py
Original file line number Diff line number Diff line change
@@ -1,3 +1,6 @@
from typing import Optional


class MultiPartVersion:
def __init__(self, parts, prefix=""):
self.parts = parts
Expand All @@ -13,14 +16,11 @@ def prefixed_parts(self):
else:
return [self.prefix]

def render_dots(self):
return ".".join(self.prefixed_parts())

def render_dots_or_parts(self, with_dots):
if with_dots:
return [self.render_dots()]
else:
def render_dots_or_parts(self, sep: Optional[str] = None):
if sep is None:
return self.prefixed_parts()
else:
return [sep.join(self.prefixed_parts())]


class CudaVersion(MultiPartVersion):
Expand Down
Loading