Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3151 commits
Select commit Hold shift + click to select a range
47cfe8e
Use the stable libtpu release for TF wheel runtime, instead of a nigh…
belitskiy May 9, 2025
1d07214
Update some tests.
tensorflower-gardener May 9, 2025
8c3ca34
Manually migrate deprecated references to `strings::StrCat` and `stri…
allanrenucci May 9, 2025
c6d1c09
Use low latency thread pool for async PjRT
tensorflower-gardener May 9, 2025
0786172
Add an API to compilation environments to initialize all known compil…
krishnaharidasan May 9, 2025
dcdf9b2
Fix race condition in tensorflow/python/data/flat_map_utils.cc.
yangustc07 May 9, 2025
fdde76d
Reverts 745718a9b6d5c65f8cf387283e4d853dbe095a87
junwhanahn May 9, 2025
67c3c16
[XLA] Simplify ArrayTypeSwitch to infer the return type from the func…
majnemer May 9, 2025
b41a5b9
Short circuit `StreamExecutorGpuCompiler::Compile` to `xla::PjRtClien…
junwhanahn May 9, 2025
67df22b
Add CommonPjRtClient::CreateUninitializedBuffer and remove
pschuh May 9, 2025
b545f0b
[XLA:benchmarks] Add a workflow file to generate GHA input matrices
juliagmt-google May 9, 2025
c058347
Enable building `upb` with Clang.
tensorflower-gardener May 9, 2025
5e66c26
Fix a crash when printing an invalid `Layout`.
tensorflower-gardener May 10, 2025
dd50c44
[PJRT:GPU] Remove unnecessary synchronization in donating a buffer.
pineapplejuice233 May 10, 2025
54be9ea
Provide flexibility to derived classes for specific backends w.r.t …
tensorflower-gardener May 10, 2025
73360c9
Explicitly return an error when attempting to serialize >2GiB protobu…
junwhanahn May 10, 2025
500425c
Automated Code Change
tensorflower-gardener May 10, 2025
43624eb
Automated Code Change
tensorflower-gardener May 10, 2025
aad4df2
Automated Code Change
tensorflower-gardener May 10, 2025
02df8ab
Automated Code Change
tensorflower-gardener May 10, 2025
bb6f9e4
XLA:CPU: If configured, use more memory types and change the
tensorflower-gardener May 10, 2025
c994700
Update GraphDef version to 2223.
tensorflower-gardener May 10, 2025
c2fd2dd
compat: Update forward compatibility horizon to 2025-05-10
tensorflower-gardener May 10, 2025
7f03283
Update CustomCallApiVersion for NopReturnToken in hlo_control_flow_fl…
tensorflower-gardener May 10, 2025
e0a1ef5
Reverts changelist 714988843
cota May 10, 2025
79c97f6
Fall back to the CPU tensor's shape when the device buffer has a dyna…
junwhanahn May 10, 2025
83b97e9
Automated Code Change
tensorflower-gardener May 10, 2025
e9c3902
Automated Code Change
tensorflower-gardener May 10, 2025
26a5c2c
Emit a metric for unified model id of loaded models
tensorflower-gardener May 10, 2025
d41fd6c
[async-pjrt] Make H2D callback nonblocking
tensorflower-gardener May 10, 2025
781baea
Remove the semaphore in ExecuteHelper
hhb May 10, 2025
8e1916e
Fix a bug where `ToLiteral` fails with a token buffer
junwhanahn May 10, 2025
e2ccfb8
[async-pjrt] Add semaphore to limit max inflight computations
tensorflower-gardener May 11, 2025
cfa0d92
Fix tfrt gpu client shutdown process
hhb May 11, 2025
d2cc3d0
compat: Update forward compatibility horizon to 2025-05-11
tensorflower-gardener May 11, 2025
ad21235
Update GraphDef version to 2224.
tensorflower-gardener May 11, 2025
48bff0c
Remove Tfrt from PJRT GPU buffer type names
hhb May 11, 2025
50bc4bb
Rename MaybeOwningGpuMemory to GpuDeviceMemory
hhb May 11, 2025
79b0452
Automated Code Change
tensorflower-gardener May 12, 2025
5d23567
Make the recently-added RegisterPluggableDevicePlugin type-safe.
tensorflower-gardener May 12, 2025
9279aed
Automated Code Change
tensorflower-gardener May 12, 2025
04e138c
Automated Code Change
tensorflower-gardener May 12, 2025
d928174
compat: Update forward compatibility horizon to 2025-05-12
tensorflower-gardener May 12, 2025
77d9977
Update GraphDef version to 2225.
tensorflower-gardener May 12, 2025
0634875
Handle non-normalized layouts for TMA.
Moerafaat May 12, 2025
9bb6a35
[XLA:CPU] Add op type id to ExecutionGraph::Operation
basioli-k May 12, 2025
df3427d
update sharding settings to remove empty shards in execution
metaflow May 12, 2025
d5dbfb6
Use `empty()` instead of `getValue().empty()` for manualAxesAttr.
ZixuanJiang May 12, 2025
2f3da4e
Automated Code Change
tensorflower-gardener May 12, 2025
420a899
[XLA:CPU] Graph visualization called only for executor owned by CpuEx…
basioli-k May 12, 2025
8d21875
#sdy Cleanup `removeAutoAxesToAvoidPadding` as this is now done in `-…
tomnatan30 May 12, 2025
cd6f5ff
Automated Code Change
tensorflower-gardener May 12, 2025
17abf83
Automated Code Change
tensorflower-gardener May 12, 2025
776c0c4
Automated Code Change
tensorflower-gardener May 12, 2025
ded9ab0
Automated Code Change
tensorflower-gardener May 12, 2025
f3d3f9f
Automated Code Change
tensorflower-gardener May 12, 2025
8035b2b
[xla:gpu] Remove packed_dim attribute from int4 tests
chsigg May 12, 2025
3b4f6bc
[XLA][graphviz] Graph visualization from a span of ops
basioli-k May 12, 2025
2b18a9a
[XLA:GPU] Add --force_auto_layout option to the multihost_hlo_runner
mooskagh May 12, 2025
f11011a
[XLA:GPU] Add custom kernel autotuning backend.
derdrdirk May 12, 2025
3c016ff
[XLA:GPU] Add simple triton support test for topK op
thcmbs May 12, 2025
575fd66
[XLA] Fix bug in collective permute HLO verifier.
olegshyshkov May 12, 2025
c540ec6
[XLA:GPU] update documentation and some comments around runtime varia…
metaflow May 12, 2025
2dfa91a
[XLA] Move execution graph renderer to execution_graph
basioli-k May 12, 2025
a177aa5
Fix a bug where TFRT GPU does not handle token literals correctly
junwhanahn May 12, 2025
c1814c0
[XLA:CPU] Expose an interface to get nested thunks
basioli-k May 12, 2025
b90c7e5
Introduce the `BUFFER` primitive type and update the `Shape` class to…
bixia1 May 12, 2025
f346ebc
[XLA:GPU] Only support cuBLAS custom calls in cuBLAS autotuning backend.
derdrdirk May 12, 2025
dd869af
fork stablehlo-quant-opt to tf-stablehlo-quant-opt without lite deps
ecalubaquib May 12, 2025
00053a0
Don't print a warning when a dependency repo is already defined.
rickeylev May 12, 2025
2ba2a9e
Use `VLOG()` directly instead of `if (VLOG_IS_ON(..))`
hhb May 12, 2025
2641089
[XLA:GPU] Add missing `device_memory_size` to XLA AOT GPU target test…
allanrenucci May 12, 2025
5bdbfeb
Increase random UUID space in the fingerprint proto.
tensorflower-gardener May 12, 2025
9678a62
Consistently apply preprocessing in HloRunnerAgnosticTestBase + friends.
nvgrw May 12, 2025
47d4996
Allow simplify_ici_dummy_variables to be triggered with TPUExecuteAnd…
tensorflower-gardener May 12, 2025
6b47a09
Add error checking for MakeRefType in FunctionInstantiationHelper
SeeForTwo May 12, 2025
3787c2e
Convert remaining `native.py_*` usages to load from rules_python
rickeylev May 12, 2025
0d63041
Remove TfrtCpuClient::BufferFromHostLiteral in favor of CommonPjRtCli…
pschuh May 12, 2025
7f6e2ee
[EagerTensor] Release GIL when calling potentially blocking methods.
mrry May 12, 2025
7b88aec
[XLA:Python] Use the released version of nanobind 2.7.0 rather than p…
hawkinsp May 12, 2025
8afb2a0
[XLA] Add F8E8M0FNU to IsF8Type
majnemer May 12, 2025
043ad27
Capture XLA flags from the environment in the ifrt proxy client
krishnaharidasan May 12, 2025
e9eb360
D2D copy with source device
hhb May 12, 2025
37b8693
Fix typo in `platform.h` doc string
ghpvnist May 12, 2025
ef9e12b
Disable eager:context_cross_platform_gpu_test GPU test
vkarampudi May 12, 2025
b29c868
Pin every NCCL communicator to its own thread.
mwhittaker May 12, 2025
d81564f
Implement Hadamard rotation reference as a custom op
paulinesho May 12, 2025
464eb2f
Add more VLOGs for D2D transfers
hhb May 13, 2025
c6dbbef
Move CreateViewOfDeviceBuffer and BufferFromHostBuffer into CommonPjR…
pschuh May 13, 2025
48c3aaf
[async-pjrt] Improve logging
tensorflower-gardener May 13, 2025
ba38f0a
Automated Code Change
tensorflower-gardener May 13, 2025
7a74d5d
Automated Code Change
tensorflower-gardener May 13, 2025
299bed7
Automated Code Change
tensorflower-gardener May 13, 2025
14d1f64
Avoid device lookup during PjRt compile
hhb May 13, 2025
026a74f
Automated Code Change
tensorflower-gardener May 13, 2025
12e50aa
Automated Code Change
tensorflower-gardener May 13, 2025
2016fbe
[XLA:GPU] Add triton support test for triangular-solve op
thcmbs May 13, 2025
2867876
Automated Code Change
tensorflower-gardener May 13, 2025
18cf322
Automated Code Change
tensorflower-gardener May 13, 2025
9d9d851
[XLA:GPU] Add more Traceme for the compilation steps
loislo May 13, 2025
0323632
Automated Code Change
tensorflower-gardener May 13, 2025
8242a30
Add proto serialization for ShapeIndex
beckerhe May 13, 2025
845bfdd
Automated Code Change
tensorflower-gardener May 13, 2025
88a1218
PR #26358: Support optimization barrier in HLO evaluator
sergey-kozub May 13, 2025
ae6d646
[XLA] Remove ShapeUtil::GetSubshapeOneIndex().
olegshyshkov May 13, 2025
5fcae0f
Update GraphDef version to 2226.
tensorflower-gardener May 13, 2025
8652c75
compat: Update forward compatibility horizon to 2025-05-13
tensorflower-gardener May 13, 2025
628265a
PR #26229: Enable dynamic-(update-)slice memcpy optimization.
jreiffers May 13, 2025
6653bbb
Automated Code Change
tensorflower-gardener May 13, 2025
ef41ee1
Fix build on python < 3.10
sohaibiftikhar May 13, 2025
00adee3
PR #26255: [Nvidia GPU] Nvshmem communicator API
Tixxx May 13, 2025
aa2914b
Automated Code Change
tensorflower-gardener May 13, 2025
ff5a598
[XLA:GPU] Improve int4 packing dimension detection
chsigg May 13, 2025
8a8a2b8
PR #26293: [GPU] Fix cuDNN CUDA graph updates.
sergachev May 13, 2025
3c947b5
[XLA:GPU] Enable HS optimisations at O1 **and above**.
allanrenucci May 13, 2025
27698a0
[XLA:CPU][roll-forward] Internal fixes.
basioli-k May 13, 2025
bfa4931
Allow non-square tiles in dynamic search space if one of the sides is…
gflegar May 13, 2025
cd4dd38
Automated Code Change
tensorflower-gardener May 13, 2025
1d38067
Add [de]serialization for `GemmConfig`.
khasanovaa May 13, 2025
f2500dc
Add proto serialization for GPU ConditionalThunk
apivovarov May 13, 2025
c259109
Set always_link for GpuCompilationEnvironment
krishnaharidasan May 13, 2025
065b986
[XLA:GPU] convert xla/backends/autotuner/backends/gpu:triton_test to …
metaflow May 13, 2025
d92ed60
[xla:gpu] Allow root to be a 'parameter' in generic emitter.
chsigg May 13, 2025
75e1305
[XLA:GPU] Fix one-shot all-reduce flag in CollectiveOpsE2ETest.
olegshyshkov May 13, 2025
e3555ea
[xla:gpu] add extract_tma_info pass
vwbaker May 13, 2025
919992b
[xla:ifrt] Rename references to ::Compile(module, options) that retur…
danielsuo May 13, 2025
58ffc49
Composite,CustomCall,DotGeneral,Dot,DynamicConv,DynamicGather,Dynamic…
abhigunj May 13, 2025
559e295
Add proto serialization for GPU WhileThunk
apivovarov May 13, 2025
2b1a776
[tosa] Fix `obj.isa` -> `mlir::isa` in an assertion (#92987)
mtrofin May 13, 2025
ee95b4a
Add support for flattening a custom-call result with nested tuples in…
tomnatan30 May 13, 2025
f9f2f3a
[XLA:CPU][XLA:GPU] Run FlattenCallGraph before CallInliner.
mkuperst May 13, 2025
db9a091
[XLA:GPU][Emitters] Use new heuristic for transpose emitter
tensorflower-gardener May 13, 2025
2732805
A reduce might produce a tuple. Need to guard it against a ShapeUtil:…
tensorflower-gardener May 13, 2025
096d88a
Fix race condition in test.
pschuh May 13, 2025
a430c65
[xla:ifrt] Rename references to ::Compile(module, options) that retur…
danielsuo May 13, 2025
c4a63f4
Fix a bug in `RandomPathGreedy` when tracking the objective value ove…
tensorflower-gardener May 13, 2025
82ceeb2
Add a function for directly constructing a DmaCopyChunk
pschuh May 13, 2025
82394a7
#sdy save shardings of the arguments on `CaseOp` during MHLO->HLO tra…
bartchr808 May 13, 2025
0733996
Add a trace type for Jax serving
SiqiaoWu1993 May 13, 2025
7fb6058
Updates ConvertToProblem() to gracefully handle solver requests conta…
tensorflower-gardener May 13, 2025
9f0cc98
Redefine vlog levels in tfrt_gpu_client
tensorflower-gardener May 13, 2025
ca0570f
Add optional 'miscompare callback' argument to literal equal test,…
tensorflower-gardener May 13, 2025
e9e8a09
[xla:ifrt] Rename references to ::Compile(module, options) that retur…
danielsuo May 13, 2025
db363df
PR #24871: [NVIDIA GPU] Fix CollectiveColorer memory space propagation
terryysun May 13, 2025
4abfd9b
[tensorflow] don't use deprecated XLA API
majnemer May 13, 2025
c403d57
Replace calls to <ctype.h> functions with their absl:: equivalents.
majnemer May 13, 2025
48dbdb8
[XLA:benchmarks] Add a workflow file for presubmit benchmarks and uti…
juliagmt-google May 13, 2025
dbd5a0d
Gather,Iota, Pad, PartitionId, RealDynamicSlice, ReducePrecision,Redu…
abhigunj May 13, 2025
016d8d9
fork tensorflow/python:unfreeze_constants to tensorflow/python:tf_unf…
ecalubaquib May 13, 2025
bb2fcc0
Inside Google, make XLA tests by default link statically on CI (to ca…
tensorflower-gardener May 13, 2025
18ffcb1
[XLA] Do not edit sort computation in-place when it has multiple uses.
mkuperst May 13, 2025
203d5a0
change visibility of xla::Semaphore
deqiangc May 13, 2025
16513b6
[XLA] Add some FlattenCallGraph tests.
mkuperst May 13, 2025
48370f7
Move ScheduleConfigProto to XLA third party
tensorflower-gardener May 13, 2025
d4b2f3c
Move templated implementation of RepeatedFieldSplitter into header
kevinbchen May 14, 2025
e5602c9
Enable Codegen for UnaryEinsumOp, add -ve test because UnaryEinsumOp …
abhigunj May 14, 2025
093d86c
Implement `layout` in IfrtProxyArray
tensorflower-gardener May 14, 2025
7b7df68
Do not insert unnecessary reshape if it does not change shape in `Con…
ZixuanJiang May 14, 2025
9eeef2b
Automated Code Change
tensorflower-gardener May 14, 2025
8760fd3
Automated Code Change
tensorflower-gardener May 14, 2025
a853d6d
Automated Code Change
tensorflower-gardener May 14, 2025
66a6e85
Automated Code Change
tensorflower-gardener May 14, 2025
88c273d
PR #25327: [XLA][CPU] Ensure correct scale is used when there are mul…
gaurides May 14, 2025
2e997c6
PR #26422: [GPU] Remove uses of absl::Status from CUB sort kernel.
sergachev May 14, 2025
9aab5b4
Automated Code Change
tensorflower-gardener May 14, 2025
5fe19c1
Disable cuDNN command buffer test on pre-Ampere GPUs
beckerhe May 14, 2025
0ef800a
PR #24733: [XLA:CPU][BugFix] Verify sufficient scratch allocation for…
akhilgoe May 14, 2025
869baf9
Update GraphDef version to 2227.
tensorflower-gardener May 14, 2025
da23473
compat: Update forward compatibility horizon to 2025-05-14
tensorflower-gardener May 14, 2025
6706f94
Automated Code Change
tensorflower-gardener May 14, 2025
2aa3e26
[XLA:GPU] Enable simple Triton Multi-output fusion by default
akuegel May 14, 2025
6891ede
[XLA:GPU] Add support for bf16 in one-shot all-reduce CUDA kernel.
olegshyshkov May 14, 2025
acb74d5
Automated Code Change
tensorflower-gardener May 14, 2025
122d3be
[XLA:GPU] Use Redzone buffers in cublas autotuning backend.
derdrdirk May 14, 2025
9c8af98
internal BUILD rule visibility
tensorflower-gardener May 14, 2025
060bd76
Add X64 support for `xla.sdy. GlobalToLocalShape` and `xla.sdy. Local…
tomnatan30 May 14, 2025
8ca5603
Reverts 2aa3e2626ebf6b6bd6d7a25d6b3de0fc714bc13c
akuegel May 14, 2025
e236a5a
Add `xla.sdy.PropagationBarrier` and `xla.sdy.ShardingGroup` to x64-r…
tomnatan30 May 14, 2025
8c9cfc2
[XLA:GPU][Emitters] Add transpose-reverse-reshape-bitcast to the list…
pifon2a May 14, 2025
c4fd0e8
[xla:gpu] Fork existing fusion_emitter_int4_device_test.cc and make i…
chsigg May 14, 2025
8a32a04
[XLA] Simplify call flattening.
mkuperst May 14, 2025
8c8d20d
[xla:gpu] Fix broadcast dimensions when hoisting bitcasts through it …
chsigg May 14, 2025
b5046ab
Use packed layouts for sub-byte types by default
junwhanahn May 14, 2025
4dd2604
Enable dynamic search space by default
gflegar May 14, 2025
5632374
[XLA] Clean-up called computation leftovers for TopK.
mkuperst May 14, 2025
91405b7
fork quantize_passes to use tf_passes
sgarciagoogle May 14, 2025
d72f781
Rollback of PR #24871
frgossen May 14, 2025
dc21446
Implement GELU reference kernel for INT16 activation
v-dziuba May 14, 2025
96a33fe
Remove unused `portable_gif_internal` dependency.
allanrenucci May 14, 2025
23bfcf7
Bump Shardy XLA commit.
tomnatan30 May 14, 2025
e713dc4
Rename queueing_delay_msecs metric to batch_queueing_delay_msecs.
tensorflower-gardener May 14, 2025
f3c92c7
[XLA:GPU] Add vectorization to one-shot all-reduce CUDA kernel.
olegshyshkov May 14, 2025
787d085
Remove support for dimension level type.
tensorflower-gardener May 14, 2025
c0c9c29
[ReplicaGroupV2] Use a tile assignment structure to store device grou…
tensorflower-gardener May 14, 2025
105d8c5
[ifrt] Refactor away from deprecated constructors
majnemer May 14, 2025
6337451
[XLA:GPU] cleanup FoldRTVarsAndConstructIndexingMap
metaflow May 14, 2025
41d98c7
Fix macos failure due to bound capture
tomnatan30 May 14, 2025
97e0ef4
Branch saved_model_import to use TFQuantDialect. Also split the heade…
fcoUnda May 14, 2025
18d1998
Remove unused and invalid legalize_tf_patterns_inc_gen.
vamsimanchala May 14, 2025
e1a70c5
[XLA:benchmarks] Trigger presubmit workflow by GitHub label `blocking…
juliagmt-google May 14, 2025
4b46c2c
Moving ScheduleConfig converters to third party.
tensorflower-gardener May 14, 2025
d8b65e8
Branch quantization/stablehlo/cc:saved_model_export to use TFQuantDia…
fcoUnda May 14, 2025
c373d90
[XLA:GPU] Pass typed pointers into all-reduce kernel.
olegshyshkov May 14, 2025
12f76eb
PR #23861: [ROCm] Use MIOPEN_BETA_API for miopenSetTensorDescriptorV2
hsharsha May 14, 2025
333f826
[GPU] Use proto_matchers in copy_thunk_test
apivovarov May 14, 2025
8a587ff
Disable eager:context_cross_platform_gpu_test test on TAP until it is…
vkarampudi May 14, 2025
63a7966
Add use_raw_buffers which allows switching the implementation to
pschuh May 14, 2025
824849a
No public description
tensorflower-gardener May 14, 2025
0051a59
Branch quantization/stablehlo/cc/calibration:component to use TFQuant…
fcoUnda May 14, 2025
b787c07
Integrate LLVM at llvm/llvm-project@741fef3a4453
tensorflower-gardener May 14, 2025
b292a0f
Branch quantization/stablehlo/cc:weight_only_ptq to use TFQuantDialect
fcoUnda May 14, 2025
9a06f50
Refactors three new functions -- GetFollowers() + GetAliases() + GetD…
tensorflower-gardener May 14, 2025
4dd52e1
Adds a schedule config to XLA's HloModuleConfig
tensorflower-gardener May 14, 2025
6f63b99
Disable MatrixUnaryFunctorGradientTest on linalg_grad_test
vkarampudi May 15, 2025
90b6b8d
Pass HashableReplicaGroupSpan into device_sets_per_replica_map as key…
tensorflower-gardener May 15, 2025
028fff0
Re-enable d2d test
sizhit2 May 15, 2025
8688728
No public description
tensorflower-gardener May 15, 2025
81d8fcd
Use the configured allocator during compilation
hhb May 15, 2025
dbd8b00
Make HloInstruction::name_ only take 8 bytes inline in the object
ddunl May 15, 2025
a6cf5de
Automated Code Change
tensorflower-gardener May 15, 2025
2b21a7d
Automated Code Change
tensorflower-gardener May 15, 2025
ae9a006
[XLA:GPU] Let TritonFusionNumericsVerifier handle tuples correctly.
akuegel May 15, 2025
24ba711
Automated Code Change
tensorflower-gardener May 15, 2025
37483e1
Reverts 88a1218b2a8c46669789055b6277d96a38506378
tensorflower-gardener May 15, 2025
628eaad
Reverts dbd8b0016c2b94c47603b12c6ed97d9b6be36c72
tensorflower-gardener May 15, 2025
fdcab34
Update GraphDef version to 2228.
tensorflower-gardener May 15, 2025
1dee7f3
compat: Update forward compatibility horizon to 2025-05-15
tensorflower-gardener May 15, 2025
808bc63
Automated Code Change
tensorflower-gardener May 15, 2025
39242d3
Fix determinism test to work with dynamic search space
gflegar May 15, 2025
3555392
[XLA:GPU] Fix Triton broadcast tiling propagation
loislo May 15, 2025
2896df5
Automated Code Change
tensorflower-gardener May 15, 2025
ce308a4
Automated Code Change
tensorflower-gardener May 15, 2025
80ace93
#sdy Properly handle token types in JAX and `ManualComputationOp`.
bartchr808 May 15, 2025
0295d0c
[XLA:GPU] Simplify TritonFusionNumericsVerifier.
akuegel May 15, 2025
2697201
Performance nit: Call `reserve` in `absl::Span` constructor.
allanrenucci May 15, 2025
68d52a2
Automated Code Change
tensorflower-gardener May 15, 2025
50a111a
[XLA:GPU] Remove flag-flip that is not necessary anymore.
loislo May 15, 2025
c7df113
Add test for [de]serialization of `BufferAllocationSlice`.
khasanovaa May 15, 2025
5e518a1
[XLA:GPU] Remove p2p rewriter in favor of xla_gpu_experimental_pipeli…
frgossen May 15, 2025
59e82e1
Log the number of threads that joined the rendezvous on time.
allanrenucci May 15, 2025
bd72f79
[XLA] Fix absl::string_view -> llvm::StringRef conversions to not rel…
d0k May 15, 2025
e21acad
[XLA:GPU] Reduce test time by reducing the tensor dimensions.
loislo May 15, 2025
5ba8da2
[Cleanup] Use CHECK_NOTNULL and CHECK_OK
frgossen May 15, 2025
4c83114
[XLA:GPU] Move collective_select_folder.cc
frgossen May 15, 2025
acc83b4
[xla:python] Restrict visibility of public targets.
danielsuo May 15, 2025
1a13c97
[XLA:GPU] Reduce the batch size of test tensor in fusion_emitter_int4…
loislo May 15, 2025
6ee0ce4
[xla:gpu] Use enum for swizzle mode in tma_descriptor attributes.
vwbaker May 15, 2025
97f83e5
Replace uses of zero-sized parameters with constants
WillFroom May 15, 2025
81cc34b
Automated Code Change
tensorflower-gardener May 15, 2025
5a809c2
Add debug to TFCI
MichaelHudgins May 15, 2025
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
  •  
  •  
  •  
The diff you're trying to view is too large. We only load the first 3000 changed files.
89 changes: 56 additions & 33 deletions .bazelrc
Original file line number Diff line number Diff line change
Expand Up @@ -125,6 +125,10 @@ build --config=short_logs
# TODO(mihaimaruseac): Document this option or remove if no longer needed
build --config=v2

# Precompiling results in some action conflicts. Disable it for now until
# the problematic targets are fixed.
build --@rules_python//python/config_settings:precompile=force_disabled

# TF now has `cc_shared_library` targets, so it needs the experimental flag
# TODO(rostam): Remove when `cc_shared_library` is enabled by default
common --experimental_cc_shared_library
Expand Down Expand Up @@ -159,15 +163,19 @@ build:android --host_crosstool_top=@bazel_tools//tools/cpp:toolchain
build:android_arm --config=android
build:android_arm --cpu=armeabi-v7a
build:android_arm --fat_apk_cpu=armeabi-v7a
build:android_arm --platforms=@org_tensorflow//tensorflow/tools/toolchains/android:armeabi-v7a
build:android_arm64 --config=android
build:android_arm64 --cpu=arm64-v8a
build:android_arm64 --fat_apk_cpu=arm64-v8a
build:android_arm64 --platforms=@org_tensorflow//tensorflow/tools/toolchains/android:arm64-v8a
build:android_x86 --config=android
build:android_x86 --cpu=x86
build:android_x86 --fat_apk_cpu=x86
build:android_x86 --platforms=@org_tensorflow//tensorflow/tools/toolchains/android:x86
build:android_x86_64 --config=android
build:android_x86_64 --cpu=x86_64
build:android_x86_64 --fat_apk_cpu=x86_64
build:android_x86_64 --platforms=@org_tensorflow//tensorflow/tools/toolchains/android:x86_64

# Build everything statically for Android since all static libs are later
# bundled together into a single .so for deployment.
Expand Down Expand Up @@ -200,6 +208,7 @@ build:apple-toolchain --host_crosstool_top=@local_config_apple_cc//:toolchain
# Settings for MacOS on ARM CPUs.
build:macos_arm64 --cpu=darwin_arm64
build:macos_arm64 --macos_minimum_os=11.0
build:macos_arm64 --platforms=@build_bazel_apple_support//configs/platforms:darwin_arm64

# iOS configs for each architecture and the fat binary builds.
build:ios --apple_platform_type=ios
Expand All @@ -208,14 +217,19 @@ build:ios --copt=-Wno-c++11-narrowing
build:ios --config=apple-toolchain
build:ios_armv7 --config=ios
build:ios_armv7 --cpu=ios_armv7
build:ios_armv7 --platforms=@org_tensorflow//tensorflow/tools/toolchains/ios:ios_armv7
build:ios_arm64 --config=ios
build:ios_arm64 --cpu=ios_arm64
build:ios_arm64 --platforms=@build_bazel_apple_support//configs/platforms:ios_arm64
build:ios_arm64e --config=ios
build:ios_arm64e --cpu=ios_arm64e
build:ios_arm64e --platforms=@build_bazel_apple_support//configs/platforms:ios_arm64e
build:ios_sim_arm64 --config=ios
build:ios_sim_arm64 --cpu=ios_sim_arm64
build:ios_sim_arm64 --platforms=@build_bazel_apple_support//configs/platforms:ios_sim_arm64
build:ios_x86_64 --config=ios
build:ios_x86_64 --cpu=ios_x86_64
build:ios_x86_64 --platforms=@build_bazel_apple_support//configs/platforms:ios_x86_64
build:ios_fat --config=ios
build:ios_fat --ios_multi_cpus=armv7,arm64,i386,x86_64

Expand All @@ -241,24 +255,24 @@ build:mkl_threadpool --define=tensorflow_mkldnn_contraction_kernel=0
build:mkl_threadpool --define=build_with_mkl_opensource=true
build:mkl_threadpool -c opt

# Config setting to build oneDNN with Compute Library for the Arm Architecture (ACL).
build:mkl_aarch64 --define=build_with_mkl_aarch64=true
build:mkl_aarch64 --define=build_with_openmp=true
build:mkl_aarch64 --define=build_with_acl=true
build:mkl_aarch64 -c opt

# Config setting to build oneDNN with Compute Library for the Arm Architecture (ACL).
# with Eigen threadpool support
build:mkl_aarch64_threadpool --define=build_with_mkl_aarch64=true
build:mkl_aarch64_threadpool --define=build_with_acl=true
build:mkl_aarch64_threadpool -c opt

# This is an alias for the mkl_aarch64_threadpool build.
build:mkl_aarch64 --config=mkl_aarch64_threadpool

# Default CUDA and CUDNN versions.
build:cuda_version --repo_env=HERMETIC_CUDA_VERSION="12.5.1"
build:cuda_version --repo_env=HERMETIC_CUDNN_VERSION="9.3.0"

# CUDA: This config refers to building CUDA op kernels with nvcc.
build:cuda --repo_env TF_NEED_CUDA=1
build:cuda --crosstool_top=@local_config_cuda//crosstool:toolchain
build:cuda --@local_config_cuda//:enable_cuda
# Default CUDA and CUDNN versions.
build:cuda --repo_env=HERMETIC_CUDA_VERSION="12.5.1"
build:cuda --repo_env=HERMETIC_CUDNN_VERSION="9.3.0"
build:cuda --config=cuda_version
# This flag is needed to include CUDA libraries.
build:cuda --@local_config_cuda//cuda:include_cuda_libs=true

Expand Down Expand Up @@ -288,8 +302,7 @@ build:cuda_clang --linkopt="-lm"

# Set up compilation CUDA version and paths and use the CUDA Clang toolchain.
build:cuda_clang_official --config=cuda_clang
build:cuda_clang_official --repo_env=HERMETIC_CUDA_VERSION="12.5.1"
build:cuda_clang_official --repo_env=HERMETIC_CUDNN_VERSION="9.3.0"
build:cuda_clang_official --config=cuda_version
build:cuda_clang_official --action_env=CLANG_CUDA_COMPILER_PATH="/usr/lib/llvm-18/bin/clang"
build:cuda_clang_official --crosstool_top="@local_config_cuda//crosstool:toolchain"

Expand Down Expand Up @@ -426,12 +439,8 @@ build:windows --dynamic_mode=off

# Default paths for TF_SYSTEM_LIBS
build:linux --define=PREFIX=/usr
build:linux --define=LIBDIR=$(PREFIX)/lib
build:linux --define=INCLUDEDIR=$(PREFIX)/include
build:linux --define=PROTOBUF_INCLUDE_PATH=$(PREFIX)/include
build:macos --define=PREFIX=/usr
build:macos --define=LIBDIR=$(PREFIX)/lib
build:macos --define=INCLUDEDIR=$(PREFIX)/include
build:macos --define=PROTOBUF_INCLUDE_PATH=$(PREFIX)/include
# TF_SYSTEM_LIBS do not work on windows.

Expand Down Expand Up @@ -494,20 +503,31 @@ build:avx_linux --copt=-mavx
build:avx_linux --host_copt=-mavx
build:avx_win --copt=/arch:AVX

build:win_clang_base --@com_google_protobuf//:use_dlls=True
build:win_clang_base --@com_google_absl//absl:use_dlls
build:win_clang_base --linkopt=/demangle:no --host_linkopt=/demangle:no
build:win_clang_base --linkopt=/errorlimit:0 --host_linkopt=/errorlimit:0
build:win_clang_base --copt=/clang:-Weverything
build:win_clang_base --host_copt=/clang:-Weverything
build:win_clang_base --compiler=clang-cl
build:win_clang_base --linkopt=/FORCE:MULTIPLE
build:win_clang_base --host_linkopt=/FORCE:MULTIPLE
build:win_clang_base --action_env=PATHEXT=.COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC;.PY;.PYW
test:win_clang_base --linkopt=/FORCE:MULTIPLE
test:win_clang_base --host_linkopt=/FORCE:MULTIPLE
test:win_clang_base --build_tests_only --keep_going --test_output=errors --verbose_failures=true --test_summary=short

build:win_clang --config=win_clang_base
build:win_clang --extra_toolchains=@local_config_cc//:cc-toolchain-x64_windows-clang-cl
build:win_clang --extra_execution_platforms=//tensorflow/tools/toolchains/win:x64_windows-clang-cl
build:win_clang --host_platform=//tensorflow/tools/toolchains/win:x64_windows-clang-cl

build:windows_x86_cpu_2022 --config=win_clang_base
build:windows_x86_cpu_2022 --crosstool_top="//tensorflow/tools/toolchains/win2022/20241118:toolchain"
build:windows_x86_cpu_2022 --extra_toolchains="//tensorflow/tools/toolchains/win2022/20241118:cc-toolchain-x64_windows-clang-cl"
build:windows_x86_cpu_2022 --extra_execution_platforms="//tensorflow/tools/toolchains/win2022:windows_ltsc2022_clang"
build:windows_x86_cpu_2022 --host_platform="//tensorflow/tools/toolchains/win2022:windows_ltsc2022_clang"
build:windows_x86_cpu_2022 --platforms="//tensorflow/tools/toolchains/win2022:windows_ltsc2022_clang"
build:windows_x86_cpu_2022 --copt=/clang:-Weverything
build:windows_x86_cpu_2022 --host_copt=/clang:-Weverything
build:windows_x86_cpu_2022 --compiler=clang-cl
build:windows_x86_cpu_2022 --linkopt=/FORCE:MULTIPLE
build:windows_x86_cpu_2022 --host_linkopt=/FORCE:MULTIPLE
test:windows_x86_cpu_2022 --linkopt=/FORCE:MULTIPLE
test:windows_x86_cpu_2022 --host_linkopt=/FORCE:MULTIPLE
test:windows_x86_cpu_2022 --action_env=PATHEXT=.COM;.EXE;.BAT;.CMD;.VBS;.VBE;.JS;.JSE;.WSF;.WSH;.MSC;.PY;.PYW
test:windows_x86_cpu_2022 --build_tests_only --keep_going --test_output=errors --verbose_failures=true --test_summary=short

# Options to build TensorFlow 1.x or 2.x.
# TODO(kanglan): Change v2's define to default behavior
Expand Down Expand Up @@ -581,6 +601,12 @@ build:rbe_linux_cpu --python_path="/usr/bin/python3"
# These you may need to change for your own GCP project.
common:rbe_linux_cpu --remote_instance_name=projects/tensorflow-testing/instances/default_instance

# Download CUDA/CUDNN redistributions to preserve the repositories cache between
# CPU and GPU builds.
# TODO(ybaturina): Uncomment when RBE is ready to support this.
# build:rbe_linux_cpu --repo_env USE_CUDA_REDISTRIBUTIONS=1
# build:rbe_linux_cpu --config=cuda_version

# TODO(kanglan): Remove it after toolchain update is complete.
build:rbe_linux_cpu_old --config=rbe_linux
build:rbe_linux_cpu_old --host_crosstool_top="@ubuntu20.04-gcc9_manylinux2014-cuda11.2-cudnn8.1-tensorrt7.2_config_cuda//crosstool:toolchain"
Expand All @@ -594,6 +620,7 @@ common:rbe_linux_cpu_old --remote_instance_name=projects/tensorflow-testing/inst

build:rbe_linux_cuda --config=cuda_clang_official
build:rbe_linux_cuda --config=rbe_linux_cpu
build:rbe_linux_cuda --repo_env=USE_CUDA_TAR_ARCHIVE_FILES=1
# For Remote build execution -- GPU configuration
build:rbe_linux_cuda --repo_env=REMOTE_GPU_TESTING=1

Expand Down Expand Up @@ -621,8 +648,10 @@ build:elinux --crosstool_top=@local_config_embedded_arm//:toolchain
build:elinux --host_crosstool_top=@bazel_tools//tools/cpp:toolchain
build:elinux_aarch64 --config=elinux
build:elinux_aarch64 --cpu=aarch64
build:elinux_aarch64 --platforms=@org_tensorflow//tensorflow/tools/toolchains/linux:linux_aarch64
build:elinux_armhf --config=elinux
build:elinux_armhf --cpu=armhf
build:elinux_armhf --platforms=@org_tensorflow//tensorflow/tools/toolchains/linux:linux_armhf
build:elinux_armhf --copt -mfp16-format=ieee

# Config-specific options should come above this line.
Expand Down Expand Up @@ -766,11 +795,6 @@ build:tf_public_macos_cache_push --config=tf_public_macos_cache --remote_upload_
# These are convenience config options that effectively declare TF's CI test suites. Look
# at the scripts of ci/official/ to see how TF's CI uses them.

# LIBTENSORFLOW TESTS are for building Libtensorflow archives. These are CUDA/CPU-agnostic.
test:linux_libtensorflow_test --config=cuda_wheel -- //tensorflow/tools/lib_package:libtensorflow_test //tensorflow/tools/lib_package:libtensorflow_java_test
build:linux_libtensorflow_build --config=cuda_wheel -- //tensorflow/tools/lib_package:libtensorflow.tar.gz //tensorflow/tools/lib_package:libtensorflow_jni.tar.gz //tensorflow/java:libtensorflow.jar //tensorflow/java:libtensorflow-src.jar //tensorflow/tools/lib_package:libtensorflow_proto.zip
build:windows_libtensorflow_build --config=cuda_wheel --config=windows_x86_cpu_2022 -- //:LICENSE //tensorflow:tensorflow.dll //tensorflow:tensorflow_dll_import_lib //tensorflow/tools/lib_package:clicenses_generate //tensorflow/java:tensorflow_jni.dll //tensorflow/tools/lib_package:jnilicenses_generate

# PYTHON TESTS run a suite of Python tests intended for verifying that the Python wheel
# will work properly. These are usually run Nightly or upon Release.
# CPU WHEEL
Expand Down Expand Up @@ -802,7 +826,7 @@ test:macos_x86_wheel_test --@local_xla//third_party/py:wheel_dependency=true --c
test:windows_x86_cpu_2022_wheel_test_filters --test_tag_filters=-no_windows,-windows_excluded,-no_oss,-oss_excluded,-gpu,-tpu,-benchmark-test,-v1only
test:windows_x86_cpu_2022_wheel_test_filters --build_tag_filters=-no_windows,-windows_excluded,-no_oss,-oss_excluded,-benchmark-test,-v1only
test:windows_x86_cpu_2022_wheel_test_filters --test_lang_filters=cc,py --test_size_filters=small,medium --test_timeout="300,450,1200,3600"
test:windows_x86_cpu_2022_wheel_test --build_tests_only --config=windows_x86_cpu_pycpp_test_filters -- //tensorflow/... -//tensorflow/java/... -//tensorflow/lite/... -//tensorflow/compiler/...
test:windows_x86_cpu_2022_wheel_test --build_tests_only --config=windows_x86_cpu_pycpp_test_filters -- //tensorflow/... //tensorflow/tools/pip_package:prebuilt_wheel_import_api_packages_test_cpu -//tensorflow/java/... -//tensorflow/lite/... -//tensorflow/compiler/...

# PYCPP TESTS run a suite of Python and C++ tests to verify general correctness over
# the whole TF code base. These are usually run continuously or upon presubmit.
Expand Down Expand Up @@ -853,12 +877,11 @@ build:cross_compile_macos_x86_pycpp_test --config=macos_x86_pycpp_test
build:cross_compile_macos_x86_pycpp_test -//tensorflow/core/kernels:quantized_conv_ops_test -//tensorflow/core/kernels:quantized_matmul_op_test -//tensorflow/python/ops:quantized_conv_ops_test -//tensorflow/tools/graph_transforms:transforms_test -//tensorflow/python/tools:aot_compiled_test
# WINDOWS X86-64 CPU PYCPP
build:windows_x86_cpu_2022_pycpp_test_build_opts --copt=/d2ReducedOptimizeHugeFunctions --host_copt=/d2ReducedOptimizeHugeFunctions --dynamic_mode=off
build:windows_x86_cpu_2022_pycpp_test_build_opts_debug --config=windows_x86_cpu_2022_pycpp_test_build_opts --linkopt=/demangle:no --host_linkopt=/demangle:no --linkopt=/errorlimit:0 --host_linkopt=/errorlimit:0
test:windows_x86_cpu_2022_pycpp_test_filters --test_tag_filters=-no_windows,-windows_excluded,-no_oss,-tf_tosa,-oss_excluded,-gpu,-tpu,-benchmark-test,-v1only
test:windows_x86_cpu_2022_pycpp_test_filters --build_tag_filters=-no_windows,-windows_excluded,-no_oss,-tf_tosa,-oss_excluded,-benchmark-test,-v1only
build:windows_x86_cpu_2022_pycpp_test_filters --build_tag_filters=-no_windows,-windows_excluded,-no_oss,-tf_tosa,-oss_excluded,-benchmark-test,-v1only
test:windows_x86_cpu_2022_pycpp_test_filters --test_lang_filters=cc,py --test_size_filters=small,medium --test_timeout="300,450,1200,3600"
test:windows_x86_cpu_2022_pycpp_test_opts --config=windows_x86_cpu_2022_pycpp_test_build_opts --build_tests_only
test:windows_x86_cpu_2022_pycpp_test --config=windows_x86_cpu_2022_pycpp_test_opts --config=windows_x86_cpu_2022_pycpp_test_filters -- //tensorflow/... -//tensorflow/java/... -//tensorflow/lite/... -//tensorflow/compiler/...
test:windows_x86_cpu_2022_pycpp_test --config=windows_x86_cpu_2022_pycpp_test_opts --config=windows_x86_cpu_2022_pycpp_test_filters -- //tensorflow/... //tensorflow/tools/pip_package:import_api_packages_test_cpu -//tensorflow/java/... -//tensorflow/lite/... -//tensorflow/compiler/...

# END TF TEST SUITE OPTIONS

Expand Down
2 changes: 1 addition & 1 deletion .github/workflows/osv-scanner-scheduled.yml
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,7 @@ permissions:
jobs:
scan-scheduled:
if: github.repository == 'tensorflow/tensorflow'
uses: "google/osv-scanner-action/.github/workflows/osv-scanner-reusable.yml@v1.9.2"
uses: "google/osv-scanner-action/.github/workflows/osv-scanner-reusable.yml@v2.0.1"
with:
scan-args: |-
--lockfile=requirements.txt:./requirements_lock_3_9.txt
Expand Down
2 changes: 1 addition & 1 deletion .github/workflows/pylint-presubmit.yml
Original file line number Diff line number Diff line change
Expand Up @@ -38,7 +38,7 @@ jobs:
run: |
echo Changed files: ${{ steps.get_file_changes.outputs.files }}
- name: Set up Python 3.9
uses: actions/setup-python@42375524e23c412d93fb67b49958b491fce71c38 # v5.4.0
uses: actions/setup-python@a26af69be951a213d495a4c3e4e4022e16d87065 # v5.6.0
with:
python-version: "3.9"
- name: Install Python dependencies
Expand Down
2 changes: 1 addition & 1 deletion .github/workflows/release-branch-cherrypick.yml
Original file line number Diff line number Diff line change
Expand Up @@ -58,7 +58,7 @@ jobs:
echo "SHORTSHA=$(git log -1 ${{ github.event.inputs.git_commit }} --format="%h")" >> "$GITHUB_OUTPUT"
echo "TITLE=$(git log -1 ${{ github.event.inputs.git_commit }} --format="%s")" >> "$GITHUB_OUTPUT"
- name: Create Pull Request with changes
uses: peter-evans/create-pull-request@dd2324fc52d5d43c699a5636bcf19fceaa70c284 # v7.0.7
uses: peter-evans/create-pull-request@271a8d0340265f705b14b6d32b9829c1cb33d45e # v7.0.8
with:
title: '${{ github.event.inputs.release_branch }} cherry-pick: ${{ steps.cherrypick.outputs.SHORTSHA }} "${{ steps.cherrypick.outputs.TITLE }}"'
committer: TensorFlow Release Automation <jenkins@tensorflow.org>
Expand Down
4 changes: 2 additions & 2 deletions .github/workflows/scorecards-analysis.yml
Original file line number Diff line number Diff line change
Expand Up @@ -55,7 +55,7 @@ jobs:
# Upload the results as artifacts (optional). Commenting out will disable uploads of run results in SARIF
# format to the repository Actions tab.
- name: "Upload artifact"
uses: actions/upload-artifact@4cec3d8aa04e39d1a68397de0c4cd6fb9dce8ec1 # v4.6.1
uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
with:
name: SARIF file
path: results.sarif
Expand All @@ -64,6 +64,6 @@ jobs:
# Upload the results to GitHub's code scanning dashboard (optional).
# Commenting out will disable upload of results to your repo's Code Scanning dashboard
- name: "Upload to code-scanning"
uses: github/codeql-action/upload-sarif@b56ba49b26e50535fa1e7f7db0f4f7b4bf65d80d # v3.28.10
uses: github/codeql-action/upload-sarif@28deaeda66b76a05916b6923827895f2b14ab387 # v3.28.16
with:
sarif_file: results.sarif
86 changes: 0 additions & 86 deletions .github/workflows/sigbuild-docker-branch.yml

This file was deleted.

Loading
Loading