experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
STT-tensorflow/tensorflow/compiler/xla/service
History
Deven Desai 868395d7d0 Updating XLA code to account for the device lib files location change in ROCm 3.9 
The location of the ROCm device lib files is changing in ROCm 3.9

Current (ROCm 3.8 and before) location is $ROCM_PATH/lib

```
root@ixt-rack-04:/opt/rocm-3.8.0# find . -name *.bc
./lib/oclc_isa_version_701.amdgcn.bc
./lib/ocml.amdgcn.bc
./lib/oclc_daz_opt_on.amdgcn.bc
./lib/oclc_isa_version_700.amdgcn.bc
./lib/oclc_isa_version_810.amdgcn.bc
./lib/oclc_unsafe_math_off.amdgcn.bc
./lib/oclc_wavefrontsize64_off.amdgcn.bc
./lib/oclc_isa_version_803.amdgcn.bc
./lib/oclc_isa_version_1011.amdgcn.bc
./lib/oclc_isa_version_1012.amdgcn.bc
./lib/opencl.amdgcn.bc
./lib/oclc_unsafe_math_on.amdgcn.bc
./lib/oclc_isa_version_1010.amdgcn.bc
./lib/oclc_finite_only_off.amdgcn.bc
./lib/oclc_correctly_rounded_sqrt_on.amdgcn.bc
./lib/oclc_daz_opt_off.amdgcn.bc
./lib/oclc_isa_version_802.amdgcn.bc
./lib/ockl.amdgcn.bc
./lib/oclc_isa_version_906.amdgcn.bc
./lib/oclc_isa_version_1030.amdgcn.bc
./lib/oclc_correctly_rounded_sqrt_off.amdgcn.bc
./lib/hip.amdgcn.bc
./lib/oclc_isa_version_908.amdgcn.bc
./lib/oclc_isa_version_900.amdgcn.bc
./lib/oclc_isa_version_702.amdgcn.bc
./lib/oclc_wavefrontsize64_on.amdgcn.bc
./lib/hc.amdgcn.bc
./lib/oclc_isa_version_902.amdgcn.bc
./lib/oclc_isa_version_801.amdgcn.bc
./lib/oclc_finite_only_on.amdgcn.bc
./lib/oclc_isa_version_904.amdgcn.bc
```

New (ROCm 3.9 and above) location is $ROCM_PATH/amdgcn/bitcode
```
root@ixt-hq-99:/opt/rocm-3.9.0-3703# find -name *.bc
./amdgcn/bitcode/oclc_isa_version_700.bc
./amdgcn/bitcode/ocml.bc
./amdgcn/bitcode/oclc_isa_version_1030.bc
./amdgcn/bitcode/oclc_isa_version_1010.bc
./amdgcn/bitcode/oclc_isa_version_904.bc
./amdgcn/bitcode/hip.bc
./amdgcn/bitcode/hc.bc
./amdgcn/bitcode/oclc_daz_opt_off.bc
./amdgcn/bitcode/oclc_wavefrontsize64_off.bc
./amdgcn/bitcode/oclc_wavefrontsize64_on.bc
./amdgcn/bitcode/oclc_isa_version_900.bc
./amdgcn/bitcode/oclc_isa_version_1012.bc
./amdgcn/bitcode/oclc_isa_version_702.bc
./amdgcn/bitcode/oclc_daz_opt_on.bc
./amdgcn/bitcode/oclc_unsafe_math_off.bc
./amdgcn/bitcode/ockl.bc
./amdgcn/bitcode/oclc_isa_version_803.bc
./amdgcn/bitcode/oclc_isa_version_908.bc
./amdgcn/bitcode/oclc_isa_version_802.bc
./amdgcn/bitcode/oclc_correctly_rounded_sqrt_off.bc
./amdgcn/bitcode/oclc_finite_only_on.bc
./amdgcn/bitcode/oclc_isa_version_701.bc
./amdgcn/bitcode/oclc_unsafe_math_on.bc
./amdgcn/bitcode/oclc_isa_version_902.bc
./amdgcn/bitcode/oclc_finite_only_off.bc
./amdgcn/bitcode/opencl.bc
./amdgcn/bitcode/oclc_isa_version_906.bc
./amdgcn/bitcode/oclc_isa_version_810.bc
./amdgcn/bitcode/oclc_isa_version_801.bc
./amdgcn/bitcode/oclc_correctly_rounded_sqrt_on.bc
./amdgcn/bitcode/oclc_isa_version_1011.bc
```

Also not the change in the filename(s)

This commit updates the XLA code, that has the device lib path + filename(s) hardcoded, to account for the change in location / filename
2020-12-04 02:26:58 +00:00
..
cpu
[XLA:CPU] Fix OrcV2 on Windows
2020-10-21 11:03:05 -07:00
g3doc
[XLA] Remove unsupported sparse layout
2020-01-02 18:04:52 -08:00
gpu
Updating XLA code to account for the device lib files location change in ROCm 3.9
2020-12-04 02:26:58 +00:00
interpreter
Remove dependencies on aliases in tensorflow/core/BUILD
2020-10-11 21:03:09 -07:00
llvm_ir
Use a pair of (building block, index) as key for the cache.
2020-10-21 04:20:37 -07:00
mlir_gpu
Fix the MHLO to LMHLO lowering of 'gather'
2020-10-19 15:28:32 -07:00
spmd
[XLA:SPMD] Add basic support for SPMD FFT.
2020-10-21 11:19:54 -07:00
algebraic_simplifier_proof_distributive_property.py
…
algebraic_simplifier_test.cc
[XLA] Convert Abs(a)*Abs(a) to a*a and add an option to allow for numerically unsafe algebraic simplifications
2020-09-06 12:51:40 -07:00
algebraic_simplifier.cc
[XLA] Preserve the layout of the slice shape during algebraic_simplification.
2020-10-08 13:28:46 -07:00
algebraic_simplifier.h
[XLA] Convert Abs(a)*Abs(a) to a*a and add an option to allow for numerically unsafe algebraic simplifications
2020-09-06 12:51:40 -07:00
all_gather_decomposer_test.cc
[XLA] Fix some all-gather issues.
2020-05-19 17:55:08 -07:00
all_gather_decomposer.cc
[XLA] Fix some all-gather issues.
2020-05-19 17:55:08 -07:00
all_gather_decomposer.h
[XLA] Fix some all-gather issues.
2020-05-19 17:55:08 -07:00
all_reduce_combiner_test.cc
Verify replica count from AllReduce replica group config
2020-04-27 08:53:44 -07:00
all_reduce_combiner.cc
[XLA] Skip AllReduceCombiner when threshold is 0
2020-08-14 19:34:44 -07:00
all_reduce_combiner.h
[XLA:GPU] Add an AllReduceCombiner pass, that merges AllReduce operations.
2020-02-19 22:07:07 -08:00
all_reduce_simplifier_test.cc
[XLA] replace all-reduces with a singleton replica_group_size with the operand.
2020-09-11 08:58:29 -07:00
all_reduce_simplifier.cc
[XLA] replace all-reduces with a singleton replica_group_size with the operand.
2020-09-11 08:58:29 -07:00
all_reduce_simplifier.h
…
allocation_tracker.cc
[XLA] Don't pass on_host_shape to ShapedBuffer/ScopedShapedBuffer inside XLA.
2020-10-08 12:00:12 -07:00
allocation_tracker.h
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
ar_crs_combiner_test.cc
Verify replica count from AllReduce replica group config
2020-04-27 08:53:44 -07:00
ar_crs_combiner.cc
This CL optimizes C++11 range-based for loops where the variable is copied in each iteration but it would suffice to obtain it by const reference. This is only applied to loop variables of types that are expensive to copy which means they are not trivially copyable or have a non-trivial copy constructor or destructor.
2020-04-06 17:55:23 -07:00
ar_crs_combiner.h
Combine cross-replica / cross-partition AllReduce after SPMD partition
2019-12-03 13:32:47 -08:00
backend.cc
…
backend.h
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
batch_dot_simplification_test.cc
…
batch_dot_simplification.cc
…
batch_dot_simplification.h
…
batchnorm_expander_test.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
batchnorm_expander.cc
…
batchnorm_expander.h
…
bfloat16_conversion_folding_test.cc
Enable using global ids for replica groups in AllReduce
2020-02-21 19:01:52 -08:00
bfloat16_conversion_folding.cc
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
bfloat16_conversion_folding.h
…
bfloat16_normalization_test.cc
Add optional layout constraint to AllToAll
2020-04-27 18:37:21 -07:00
bfloat16_normalization.cc
…
bfloat16_normalization.h
…
bfloat16_propagation_test.cc
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
bfloat16_propagation.cc
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
bfloat16_propagation.h
[XLA] Support conditional in bf16 propagation
2020-04-05 16:46:27 -07:00
bfloat16_support.cc
[XLA] Introducing AllGather HLO and a decomposition pass
2020-05-09 12:22:17 -07:00
bfloat16_support.h
…
buffer_assignment_test.cc
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
buffer_assignment.cc
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
buffer_assignment.h
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
buffer_value_containers.h
…
buffer_value.cc
[XLA] BufferValue::Color now type aliases int64.
2020-05-27 09:35:14 -07:00
buffer_value.h
[XLA] Deprecate BufferValue::color.
2020-06-29 14:37:43 -07:00
BUILD
Remove dependencies on aliases in tensorflow/core/BUILD
2020-10-11 21:03:09 -07:00
call_graph_test.cc
…
call_graph.cc
This CL optimizes C++11 range-based for loops where the variable is copied in each iteration but it would suffice to obtain it by const reference. This is only applied to loop variables of types that are expensive to copy which means they are not trivially copyable or have a non-trivial copy constructor or destructor.
2020-04-06 17:55:23 -07:00
call_graph.h
…
call_inliner_test.cc
Convert absl string_view constants to const, rather than constexpr.
2020-02-05 11:15:53 -08:00
call_inliner.cc
[XLA:CPU] Restrict CallInliner to functions with a single call site
2020-01-23 09:19:45 -08:00
call_inliner.h
[XLA:CPU] Restrict CallInliner to functions with a single call site
2020-01-23 09:19:45 -08:00
channel_tracker.cc
…
channel_tracker.h
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
cholesky_expander.cc
[XLA] Add complex number support to HLO cholesky decomposition implementation.
2020-09-18 07:08:30 -07:00
cholesky_expander.h
[XLA] Make the inner block kernel of CholeskyExpander override-able.
2020-08-31 09:33:25 -07:00
collective_ops_utils.cc
[XLA:CPU/GPU] Implement all-reduce and/or of pred
2020-03-30 11:10:37 -07:00
collective_ops_utils.h
[NFC] Eliminate reference to HLO Inst from NcclAllReduceThunk
2020-10-06 09:50:08 -07:00
comparison_expander.cc
Adding total-order comparison support in proto and HloInstruction.
2020-08-10 09:40:29 -07:00
comparison_expander.h
Adding total-order comparison support in proto and HloInstruction.
2020-08-10 09:40:29 -07:00
compilation_cache.cc
…
compilation_cache.h
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
compilation_stats.cc
…
compilation_stats.h
…
compile_only_service.cc
Add a module config option to enable hlo deduplication.
2020-08-03 12:42:04 -07:00
compile_only_service.h
…
compiler.cc
[mlir] Add FusionOp to XLA HLO -> LMHLO
2020-10-13 14:20:52 -07:00
compiler.h
[mlir] Add FusionOp to XLA HLO -> LMHLO
2020-10-13 14:20:52 -07:00
computation_layout.cc
This CL adds the following changes:
2020-07-01 16:45:34 -07:00
computation_layout.h
This CL adds the following changes:
2020-07-01 16:45:34 -07:00
computation_placer.cc
…
computation_placer.h
minor spelling tweaks
2019-12-09 18:21:12 +09:00
conditional_canonicalizer_test.cc
Canonicalize dense array conditional into tuple conditional with one element.
2020-07-17 11:55:43 -07:00
conditional_canonicalizer.cc
Fix conditional canonicalizer to use clone instead of mutating existing conditional's shape.
2020-07-27 13:44:30 -07:00
conditional_canonicalizer.h
Canonicalize dense array conditional into tuple conditional with one element.
2020-07-17 11:55:43 -07:00
conditional_code_motion_test.cc
Disabling the hoisting of converts in the presence of conditionals, where the branch roots are not of the same type.
2020-09-30 16:45:07 -07:00
conditional_code_motion.cc
Disabling the hoisting of converts in the presence of conditionals, where the branch roots are not of the same type.
2020-09-30 16:45:07 -07:00
conditional_code_motion.h
[XLA] Fix bugs and Improve conditional code motion to support moving multiple streams of instructions inside conditionals.
2020-09-18 16:42:52 -07:00
conditional_simplifier_test.cc
[XLA] Add token support to conditional_simplifier
2020-06-02 11:00:01 -07:00
conditional_simplifier.cc
Don't rewrite conditionals with empty branches into select.
2020-07-28 20:09:37 -07:00
conditional_simplifier.h
…
conditional_to_select_test.cc
…
conditional_to_select.cc
…
conditional_to_select.h
…
convolution_4d_expander_test.cc
Convert 4D convolutions with trivial dimensions to lower-dimensional convolutions.
2020-01-28 02:40:48 -08:00
convolution_4d_expander.cc
Convert 4D convolutions with trivial dimensions to lower-dimensional convolutions.
2020-01-28 02:40:48 -08:00
convolution_4d_expander.h
Convert 4D convolutions with trivial dimensions to lower-dimensional convolutions.
2020-01-28 02:40:48 -08:00
convolution_group_converter_test.cc
[XLA] Fix the condition for rewriting batch group convolutions to include when the input batch is not equal to the batch group count.
2020-04-08 16:50:13 -07:00
convolution_group_converter.cc
HLO creation utils API change to allow passing batch_group_count.
2020-09-03 11:31:12 -07:00
convolution_group_converter.h
Fix confusing way of specifying filter_expansion parameter.
2020-01-08 00:43:03 -08:00
copy_insertion_test.cc
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
copy_insertion.cc
[XLA] [NFC] Refactor LiveRangeBefore
2020-10-07 12:35:14 -07:00
copy_insertion.h
…
custom_call_target_registry.cc
…
custom_call_target_registry.h
…
defuser_test.cc
…
defuser.cc
…
defuser.h
…
despecializer.cc
…
despecializer.h
…
dfs_hlo_visitor_with_default_test.cc
…
dfs_hlo_visitor_with_default.h
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
dfs_hlo_visitor.cc
…
dfs_hlo_visitor.h
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
dot_as_convolution_util.cc
[XLA:SPMD] Fix SPMD issue with replicated depthwise convolution.
2020-09-16 10:08:20 -07:00
dot_as_convolution_util.h
[XLA:SPMD] 1st step to refactor convolution_handler.
2020-09-11 19:33:47 -07:00
dot_decomposer_test.cc
Improve DotDecomposer to not add unnecessary non-contracting dimensions.
2020-05-26 05:49:19 -07:00
dot_decomposer.cc
Polish some comments in dot decomposer.
2020-06-15 14:41:41 -07:00
dot_decomposer.h
…
dump.cc
lsc. no new code.
2020-04-16 04:23:49 -07:00
dump.h
[XLA] Add a configurable limit to the number of HLO dumps
2020-02-03 22:32:17 -08:00
dynamic_dimension_inference_test.cc
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
dynamic_dimension_inference.cc
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
dynamic_dimension_inference.h
Simplify reshape handling in dynamic dimension inference.
2020-07-23 14:49:24 -07:00
dynamic_index_splitter_test.cc
…
dynamic_index_splitter.cc
…
dynamic_index_splitter.h
…
dynamic_padder_test.cc
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
dynamic_padder.cc
Add metric for auto outside compilation & dynamic padder.
2020-10-15 16:11:28 -07:00
dynamic_padder.h
[XLA]Add HLO infrastructure to support dynamic op lowering.
2020-04-30 17:00:49 -07:00
dynamic_parameter_binding_test.cc
…
dynamic_parameter_binding.cc
…
dynamic_parameter_binding.h
…
dynamic_update_slice_test.cc
…
elemental_ir_emitter_test.cc
[XLA] Fix type conversion from and to bf16 and improve test coverage.
2019-11-26 10:31:34 -08:00
elemental_ir_emitter.cc
Remove dead code; NFC
2020-07-27 00:30:23 -07:00
elemental_ir_emitter.h
Enable XLA:CPU fast math for min/max by default to be similar to TF's behavior.
2020-06-17 01:27:08 -07:00
executable.cc
[XLA] Don't pass on_host_shape to ShapedBuffer/ScopedShapedBuffer inside XLA.
2020-10-08 12:00:12 -07:00
executable.h
[XLA] Don't pass on_host_shape to ShapedBuffer/ScopedShapedBuffer inside XLA.
2020-10-08 12:00:12 -07:00
execution_tracker.cc
…
execution_tracker.h
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
flatten_call_graph_test.cc
…
flatten_call_graph.cc
…
flatten_call_graph.h
…
fusion_node_indexing_evaluation_test.cc
Evaluate the maximum code duplication in fusion nodes instead of average.
2020-09-02 07:36:23 -07:00
fusion_node_indexing_evaluation.cc
Refactor the code to avoid duplication (NFC).
2020-10-01 05:16:36 -07:00
fusion_node_indexing_evaluation.h
Refactor the code to avoid duplication (NFC).
2020-10-01 05:16:36 -07:00
fusion_queue.h
…
gather_expander_test.cc
Expand simple gathers into dynamic-slice.
2020-08-07 16:06:09 -07:00
gather_expander.cc
Expand simple gathers into dynamic-slice.
2020-08-07 16:06:09 -07:00
gather_expander.h
Expand simple gathers into dynamic-slice.
2020-08-07 16:06:09 -07:00
generic_transfer_manager.cc
[XLA] Don't pass on_host_shape to ShapedBuffer/ScopedShapedBuffer inside XLA.
2020-10-08 12:00:12 -07:00
generic_transfer_manager.h
[XLA] Implement dynamic input and output in DynamicPadder.
2019-12-12 10:31:20 -08:00
heap_simulator_test.cc
[XLA] Use templates in heap simulator to allow opaque type to be different than HloValue (NFC)
2020-08-14 15:58:46 -07:00
heap_simulator.cc
[XLA] Use templates in heap simulator to allow opaque type to be different than HloValue (NFC)
2020-08-14 15:58:46 -07:00
heap_simulator.h
[XLA] Use templates in heap simulator to allow opaque type to be different than HloValue (NFC)
2020-08-14 15:58:46 -07:00
hlo_alias_analysis_test.cc
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
hlo_alias_analysis.cc
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
hlo_alias_analysis.h
Internal change
2020-09-12 16:03:35 -07:00
hlo_buffer.cc
…
hlo_buffer.h
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_casting_utils_test.cc
…
hlo_casting_utils.h
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_clone_context.h
…
hlo_computation_test.cc
…
hlo_computation.cc
Add add new option in HloComputation::AddInstruction to add instruction with different name.
2020-10-08 13:08:57 -07:00
hlo_computation.h
Add add new option in HloComputation::AddInstruction to add instruction with different name.
2020-10-08 13:08:57 -07:00
hlo_constant_folding_test.cc
Replace absl::Span where we use initializer lists in tests with std::vector.
2020-02-28 17:37:42 -08:00
hlo_constant_folding.cc
…
hlo_constant_folding.h
…
hlo_cost_analysis_test.cc
[XLA] Separate bytes accessed in cost analysis by operands and outputs.
2019-10-23 16:02:21 -07:00
hlo_cost_analysis.cc
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
hlo_cost_analysis.h
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
hlo_creation_utils_test.cc
…
hlo_creation_utils.cc
HLO creation utils API change to allow passing batch_group_count.
2020-09-03 11:31:12 -07:00
hlo_creation_utils.h
HLO creation utils API change to allow passing batch_group_count.
2020-09-03 11:31:12 -07:00
hlo_cse_test.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_cse.cc
[XLA] Avoid hash collisions in CseHash.
2020-02-20 00:01:30 -08:00
hlo_cse.h
…
hlo_dataflow_analysis_test.cc
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
hlo_dataflow_analysis.cc
Internal change
2020-09-24 10:22:25 -07:00
hlo_dataflow_analysis.h
Internal change
2020-09-24 10:22:25 -07:00
hlo_dce_test.cc
…
hlo_dce.cc
[XLA] Introducing AllGather HLO and a decomposition pass
2020-05-09 12:22:17 -07:00
hlo_dce.h
[XLA] Resubmit DCE change
2020-03-03 17:43:00 -08:00
hlo_domain_isolator.cc
…
hlo_domain_isolator.h
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_domain_map.cc
…
hlo_domain_map.h
…
hlo_domain_metadata.h
…
hlo_domain_remover.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_domain_remover.h
…
hlo_domain_test.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_domain_verifier.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_domain_verifier.h
…
hlo_element_type_converter_test.cc
…
hlo_element_type_converter.cc
…
hlo_element_type_converter.h
…
hlo_evaluator_test.cc
[XLA] Add support for bf16 map() to HloEvaluator
2020-04-22 05:35:36 -07:00
hlo_evaluator_typed_visitor_bfloat16.cc
…
hlo_evaluator_typed_visitor_bool.cc
…
hlo_evaluator_typed_visitor_complex64.cc
…
hlo_evaluator_typed_visitor_complex128.cc
…
hlo_evaluator_typed_visitor_double.cc
…
hlo_evaluator_typed_visitor_float.cc
…
hlo_evaluator_typed_visitor_half.cc
…
hlo_evaluator_typed_visitor_int8.cc
…
hlo_evaluator_typed_visitor_int16.cc
…
hlo_evaluator_typed_visitor_int32.cc
…
hlo_evaluator_typed_visitor_int64.cc
…
hlo_evaluator_typed_visitor_uint8.cc
…
hlo_evaluator_typed_visitor_uint16.cc
…
hlo_evaluator_typed_visitor_uint32.cc
…
hlo_evaluator_typed_visitor_uint64.cc
…
hlo_evaluator_typed_visitor.h
[XLA] Remove unused indices vector in HandleScatter in hlo_evaluator.
2020-10-09 13:06:43 -07:00
hlo_evaluator.cc
[XLA] Implement S8,S16,U16 support for Literal::GetIntegralAsS64
2020-08-06 14:48:38 -07:00
hlo_evaluator.h
[XLA:CPU] Wire up C64/C128 matmul to Eigen
2020-05-27 09:56:35 -07:00
hlo_execution_profile_data.proto
…
hlo_execution_profile_test.cc
Rewrite the pattern we want to match in hlo_execution_profile_test.
2020-01-27 05:22:03 -08:00
hlo_execution_profile.cc
[XLA/GPU] Remove uses of Thunk::hlo_instruction() for profiling.
2020-07-10 15:31:01 -07:00
hlo_execution_profile.h
[XLA/GPU] Remove uses of Thunk::hlo_instruction() for profiling.
2020-07-10 15:31:01 -07:00
hlo_get_dimension_size_rewriter.h
…
hlo_graph_dumper_test.cc
…
hlo_graph_dumper.cc
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
hlo_graph_dumper.h
Roll forward "Add a show_fusion_subcomputations command to interactive_graphviz" with fix
2020-05-27 11:27:39 -07:00
hlo_input_output_alias_config_test.cc
[XLA] Unify aliasing types
2020-07-07 16:07:00 -07:00
hlo_input_output_alias_config.cc
[Resubmit] If an input-output pair is configured to be must-alias(off by default), they must be aliased at runtime.
2020-08-07 14:06:08 -07:00
hlo_input_output_alias_config.h
Fix priting of input/output must/may aliasing.
2020-08-11 13:48:12 -07:00
hlo_instruction_test.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_instruction.cc
Internal change
2020-09-24 10:22:25 -07:00
hlo_instruction.h
Merge pull request #43746 from trentlo:clear_opnds_for_removed_hlos
2020-10-07 11:58:45 -07:00
hlo_instructions.cc
Internal change
2020-09-24 10:22:25 -07:00
hlo_instructions.h
Internal change
2020-09-24 10:22:25 -07:00
hlo_lexer.cc
[XLA] Add support for parsing negative nans as constants in HLO parser.
2020-10-06 13:12:47 -07:00
hlo_lexer.h
[XLA] Add support for parsing negative nans as constants in HLO parser.
2020-10-06 13:12:47 -07:00
hlo_live_range_test.cc
[XLA] Unify aliasing types
2020-07-07 16:07:00 -07:00
hlo_live_range.cc
[XLA] Fix live range flattening for while loops.
2019-11-01 15:50:08 -07:00
hlo_live_range.h
[XLA] Implement memory space allocation across sequential calls (e.g. while).
2019-11-20 14:24:43 -08:00
hlo_liveness_analysis_test.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_liveness_analysis.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_liveness_analysis.h
…
hlo_matchers_test.cc
[XLA] Add an optional bool is_cross_program_prefetch field to kCopyStart HLOs.
2020-08-21 15:10:38 -07:00
hlo_matchers.cc
…
hlo_matchers.h
[XLA] Add AllGather to the HLO matchers.
2020-05-21 22:29:00 -07:00
hlo_memory_scheduler_test.cc
[XLA] Use templates in heap simulator to allow opaque type to be different than HloValue (NFC)
2020-08-14 15:58:46 -07:00
hlo_memory_scheduler.cc
If an outfeed instruction has a config, schedule it early. If an infeed instruction has a config, schedule it late.
2020-06-03 22:15:37 -07:00
hlo_memory_scheduler.h
…
hlo_module_config.cc
Rename GenerateCubinForTfCode to GenerateGpuBinaryForTfCode and add ROCm support for it
2020-08-26 13:50:01 +00:00
hlo_module_config.h
Add a module config option to enable hlo deduplication.
2020-08-03 12:42:04 -07:00
hlo_module_dce_test.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_module_dce.cc
…
hlo_module_dce.h
…
hlo_module_group_metadata.cc
…
hlo_module_group_metadata.h
…
hlo_module_group_test.cc
…
hlo_module_group_util.cc
…
hlo_module_group_util.h
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_module_group.cc
[XLA] Add option to print out hlo pass hash.
2020-02-27 14:54:06 -08:00
hlo_module_group.h
[XLA] Add option to print out hlo pass hash.
2020-02-27 14:54:06 -08:00
hlo_module_test.cc
…
hlo_module.cc
Add a module config option to enable hlo deduplication.
2020-08-03 12:42:04 -07:00
hlo_module.h
This CL adds the following changes:
2020-07-01 16:45:34 -07:00
hlo_opcode_test.cc
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
hlo_opcode.cc
…
hlo_opcode.h
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
hlo_ordering_test.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_ordering.cc
…
hlo_ordering.h
…
hlo_parser_test.cc
[XLA] Add support for parsing negative nans as constants in HLO parser.
2020-10-06 13:12:47 -07:00
hlo_parser.cc
[XLA] Add support for parsing negative nans as constants in HLO parser.
2020-10-06 13:12:47 -07:00
hlo_parser.h
…
hlo_pass_fix.h
Better warning. With that we can just to replay_computation directly for further analysis.
2020-10-06 11:01:41 -07:00
hlo_pass_interface.h
…
hlo_pass_pipeline_test.cc
…
hlo_pass_pipeline.cc
[XLA] [NFC] Log time taken by each pass inside an HloPassPipeline
2020-10-09 12:08:26 -07:00
hlo_pass_pipeline.h
[XLA] Stop verifying the HLO inside the pass pipeline on optimized builds.
2020-04-02 08:19:57 -07:00
hlo_phi_graph_test.cc
Properly support nest phi reduction in reverse order.
2020-06-17 18:47:22 -07:00
hlo_phi_graph.cc
Properly support nest phi reduction in reverse order.
2020-06-17 18:47:22 -07:00
hlo_phi_graph.h
Properly support nest phi reduction in reverse order.
2020-06-17 18:47:22 -07:00
hlo_profile_printer_data.proto
…
hlo_profile_printer.cc
…
hlo_profile_printer.h
…
hlo_proto_util_test.cc
…
hlo_proto_util.cc
…
hlo_proto_util.h
…
hlo_query.cc
Add some XLA frontend attribute names.
2020-01-16 19:01:38 -08:00
hlo_query.h
Add some XLA frontend attribute names.
2020-01-16 19:01:38 -08:00
hlo_reachability_test.cc
…
hlo_reachability.cc
Allow split-phase operations for HloReachabilityMap where we can
2020-06-14 14:46:08 -07:00
hlo_reachability.h
Allow split-phase operations for HloReachabilityMap where we can
2020-06-14 14:46:08 -07:00
hlo_rematerialization_test_utils_test.cc
…
hlo_rematerialization_test_utils.h
…
hlo_rematerialization_test.cc
Express remat: Only consider remat with nodes that create big buffers.
2020-09-23 16:29:09 -07:00
hlo_rematerialization.cc
Add add new option in HloComputation::AddInstruction to add instruction with different name.
2020-10-08 13:08:57 -07:00
hlo_rematerialization.h
Express remat: Only consider remat with nodes that create big buffers.
2020-09-23 16:29:09 -07:00
hlo_replication_analysis_test.cc
Support x64<->x32 instructions in hlo replication analysis.
2020-10-01 13:25:16 -07:00
hlo_replication_analysis.cc
Support x64<->x32 instructions in hlo replication analysis.
2020-10-01 13:25:16 -07:00
hlo_replication_analysis.h
Combine cross-replica / cross-partition AllReduce after SPMD partition
2019-12-03 13:32:47 -08:00
hlo_runner.cc
[XLA] Don't pass on_host_shape to ShapedBuffer/ScopedShapedBuffer inside XLA.
2020-10-08 12:00:12 -07:00
hlo_runner.h
Extend hlo_runner and hlo_test_base to support running different executables
2020-08-14 17:20:31 -07:00
hlo_schedule_test.cc
…
hlo_schedule.cc
…
hlo_schedule.h
…
hlo_sharding_metadata.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
hlo_sharding_metadata.h
…
hlo_sharding_test.cc
…
hlo_sharding_util_test.cc
[XLA:SPMD] Improve resharding
2020-07-22 23:18:17 -07:00
hlo_sharding_util.cc
[XLA] Avoid quadratic behavior in DevicesForSharding
2020-09-24 20:08:08 -07:00
hlo_sharding_util.h
[XLA:SPMD] Fix vector indices sharding in GatherIndexSharding.
2020-08-14 21:58:47 -07:00
hlo_sharding.cc
[XLA] Fix trivial partial sharding
2020-09-29 18:33:54 -07:00
hlo_sharding.h
[XLA] Partial sharding and propagation for broadcast/reduce
2020-08-09 23:53:14 -07:00
hlo_subcomputation_unification_test.cc
Fix some HLO bugs in tests detected when using VerifiedHloModule.
2019-10-31 08:01:08 -07:00
hlo_subcomputation_unification.cc
…
hlo_subcomputation_unification.h
…
hlo_value.cc
[XLA] BufferValue::Color now type aliases int64.
2020-05-27 09:35:14 -07:00
hlo_value.h
[XLA] Add hash function to HloUse/HloPosition.
2020-07-29 23:27:09 -07:00
hlo_verifier_test.cc
[XLA] Make mixed-precision DUS a verifier failure even in mixed precision mode.
2020-09-21 13:08:15 -07:00
hlo_verifier.cc
Internal change
2020-09-24 10:22:25 -07:00
hlo_verifier.h
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
hlo.proto
Internal change
2020-09-24 10:22:25 -07:00
human_readable_profile_builder.cc
…
human_readable_profile_builder.h
…
indexed_array_analysis_test.cc
Remove 1 + "string" cases and undo -Wno-string-plus-int.
2020-02-05 16:58:37 -08:00
indexed_array_analysis.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
indexed_array_analysis.h
…
instruction_fusion_test.cc
rng loop fusion
2020-06-05 10:34:16 -07:00
instruction_fusion.cc
Clear reachability map after fusion.
2020-09-15 12:45:47 -07:00
instruction_fusion.h
Add caching for reused parameters.
2020-09-01 06:49:41 -07:00
layout_assignment_test.cc
Make linear layout more explicit.
2020-06-18 06:55:02 -07:00
layout_assignment.cc
Internal change
2020-08-21 13:30:52 -07:00
layout_assignment.h
Internal change
2020-08-21 13:30:52 -07:00
llvm_compiler.cc
…
llvm_compiler.h
Update paths to LLVM.
2019-11-04 13:53:39 -08:00
local_service.cc
Add a module config option to enable hlo deduplication.
2020-08-03 12:42:04 -07:00
local_service.h
Change LocalClient::Compile to support returning multiple executables (one per partition).
2020-01-29 01:03:57 -08:00
logical_buffer_analysis.cc
[XLA] Change CopyStart return type to (target, source, context)
2020-01-28 16:11:12 -08:00
logical_buffer_analysis.h
[XLA] Change CopyStart return type to (target, source, context)
2020-01-28 16:11:12 -08:00
logical_buffer.cc
[XLA] BufferValue::Color now type aliases int64.
2020-05-27 09:35:14 -07:00
logical_buffer.h
…
logistic_expander_test.cc
[XLA] Add expander pass for kLogistic.
2020-06-24 22:32:37 -07:00
logistic_expander.cc
[XLA] Add expander pass for kLogistic.
2020-06-24 22:32:37 -07:00
logistic_expander.h
[XLA] Add expander pass for kLogistic.
2020-06-24 22:32:37 -07:00
map_inliner_test.cc
…
map_inliner.cc
…
map_inliner.h
…
maybe_owning_device_memory.cc
Make XRT CPU/GPU use MaybeOwning buffer interface, so the new copy protection CL won't break aliasing.
2020-06-22 12:32:43 -07:00
maybe_owning_device_memory.h
Make XRT CPU/GPU use MaybeOwning buffer interface, so the new copy protection CL won't break aliasing.
2020-06-22 12:32:43 -07:00
memory_space_assignment_best_fit_repacker_test.cc
[XLA] Implement a repacker that wraps heap simulator's best fit algorithm.
2020-08-18 09:21:37 -07:00
memory_space_assignment_best_fit_repacker.cc
[XLA] Implement a repacker that wraps heap simulator's best fit algorithm.
2020-08-18 09:21:37 -07:00
memory_space_assignment_best_fit_repacker.h
[XLA] Implement a repacker that wraps heap simulator's best fit algorithm.
2020-08-18 09:21:37 -07:00
memory_space_assignment_repacking.h
[XLA] Use templates in heap simulator to allow opaque type to be different than HloValue (NFC)
2020-08-14 15:58:46 -07:00
memory_space_assignment_test.cc
[XLA] Add back required assignment for conditional outputs after every repack.
2020-10-16 17:50:07 -07:00
memory_space_assignment_utils.cc
Internal change
2020-09-24 10:22:25 -07:00
memory_space_assignment_utils.h
[XLA] Exposes the IsValueAllowedInAlternateMemory() for better reuse.
2020-08-18 14:26:53 -07:00
memory_space_assignment.cc
[XLA] Add back required assignment for conditional outputs after every repack.
2020-10-16 17:50:07 -07:00
memory_space_assignment.h
[XLA] Add back required assignment for conditional outputs after every repack.
2020-10-16 17:50:07 -07:00
memory_space_propagation_test.cc
[XLA] Propagate memory spaces recursively inside nested fusions.
2020-06-18 13:49:23 -07:00
memory_space_propagation.cc
[XLA] Propagate memory spaces recursively inside nested fusions.
2020-06-18 13:49:23 -07:00
memory_space_propagation.h
[XLA] Propagate memory spaces recursively inside nested fusions.
2020-06-18 13:49:23 -07:00
multi_output_fusion.cc
[XLA] Try committing DUS buffer sharing again with fixes.
2020-09-17 09:20:11 -07:00
multi_output_fusion.h
[Resubmit] Fix several issues of multi output fusion.
2020-02-18 20:50:32 -08:00
name_uniquer_test.cc
…
name_uniquer.cc
[-Wsign-compare] warning fixes batch 4
2020-06-24 21:58:00 +00:00
name_uniquer.h
…
op_expander_pass.cc
…
op_expander_pass.h
minor spelling tweaks
2019-12-09 18:21:12 +09:00
optimize_input_output_buffer_alias_test.cc
Automated rollback of commit afd980b3ce068c30bd3a43d54929e2590be8bd29
2020-02-12 14:31:05 -08:00
optimize_input_output_buffer_alias.cc
[XLA] Unify aliasing types
2020-07-07 16:07:00 -07:00
optimize_input_output_buffer_alias.h
Automated rollback of commit afd980b3ce068c30bd3a43d54929e2590be8bd29
2020-02-12 14:31:05 -08:00
pattern_matcher_gmock_test.cc
[XLA] Remove unsupported sparse layout
2020-01-02 18:04:52 -08:00
pattern_matcher_gmock.h
…
pattern_matcher_test.cc
[XLA] Remove unsupported sparse layout
2020-01-02 18:04:52 -08:00
pattern_matcher.h
[XLA] Add select and scatter to pattern matchers.
2020-10-20 10:35:38 -07:00
platform_util.cc
Cleanup PlatformUtil and MultiPlatformManager. NFC.
2019-11-08 20:01:51 -08:00
platform_util.h
Cleanup PlatformUtil and MultiPlatformManager. NFC.
2019-11-08 20:01:51 -08:00
qr_expander.cc
[XLA] Add support for complex numbers to Qr decomposition expander.
2020-09-22 19:36:07 -07:00
qr_expander.h
[XLA] Change the QR decomposition implementation to lower to a CustomCall operator that is later expanded by a new QrExpander pass. This allows backends to lower the Qr decomposition differently if they choose.
2020-09-21 17:35:34 -07:00
reshape_mover_test.cc
…
reshape_mover.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
reshape_mover.h
…
rng_bit_generator_expander.cc
Rewrite output state to match the input.
2020-04-29 15:51:55 -07:00
rng_bit_generator_expander.h
Introduce new RngBitGenerator HLO
2020-02-06 05:06:44 -08:00
rng_expander.cc
Introduce new RngBitGenerator HLO
2020-02-06 05:06:44 -08:00
rng_expander.h
[XLA] Cache expanded RNG computations
2020-01-23 10:39:48 -08:00
root_instruction_sinker_test.cc
[XLA] Add a RootInstructionSinker pass.
2020-04-22 14:46:15 -07:00
root_instruction_sinker.cc
[XLA] Add a RootInstructionSinker pass.
2020-04-22 14:46:15 -07:00
root_instruction_sinker.h
[XLA] Add a RootInstructionSinker pass.
2020-04-22 14:46:15 -07:00
scatter_expander_test.cc
[XLA] Expand simple scatter operations into dynamic-update-slice.
2020-08-18 13:21:28 -07:00
scatter_expander.cc
[XLA] Expand simple scatter operations into dynamic-update-slice.
2020-08-18 13:21:28 -07:00
scatter_expander.h
[XLA] Expand simple scatter operations into dynamic-update-slice.
2020-08-18 13:21:28 -07:00
service_executable_run_options.h
…
service.cc
Merge pull request #41923 from tg-at-google:wsign-compare-semi-final-xla-service1-split5
2020-08-03 21:47:29 -07:00
service.h
[XLA] Wire through static device assignments from the Python client to the compiler
2020-02-21 17:42:43 -08:00
shape_inference_test.cc
Don't allow reductions to have repeated reduction dimensions
2020-07-01 16:57:30 -07:00
shape_inference.cc
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
shape_inference.h
Introduce dynamic reshape op.
2020-08-14 15:48:19 -07:00
shaped_buffer_test.cc
[XLA] Don't pass on_host_shape to ShapedBuffer/ScopedShapedBuffer inside XLA.
2020-10-08 12:00:12 -07:00
shaped_buffer.cc
[XLA] Don't pass on_host_shape to ShapedBuffer/ScopedShapedBuffer inside XLA.
2020-10-08 12:00:12 -07:00
shaped_buffer.h
[XLA] Don't pass on_host_shape to ShapedBuffer/ScopedShapedBuffer inside XLA.
2020-10-08 12:00:12 -07:00
sharding_propagation_test.cc
[XLA:SPMD] Support normal convolution partition weights at both input feature
2020-09-14 09:48:16 -07:00
sharding_propagation.cc
[XLA:SPMD] Support normal convolution partition weights at both input feature
2020-09-14 09:48:16 -07:00
sharding_propagation.h
[XLA] Move sharding propagation to third party
2020-05-22 18:18:30 -07:00
slice_sinker_test.cc
…
slice_sinker.cc
…
slice_sinker.h
…
slow_operation_alarm.cc
Change std::call_once to absl::call_once
2020-01-29 08:40:19 -08:00
slow_operation_alarm.h
Add absl/base/attributes.h to slow_operation_alarm.h as ABSL_MUST_USE_RESULT is used
2020-02-12 18:22:02 -08:00
sort_simplifier_test.cc
…
sort_simplifier.cc
…
sort_simplifier.h
…
source_map_util.h
…
space_to_batch_converter_test.cc
Extend space to batch to apply to larger batch sizes
2020-10-06 09:56:50 -07:00
space_to_batch_converter.cc
Extend space to batch to apply to larger batch sizes
2020-10-06 09:56:50 -07:00
space_to_batch_converter.h
Extend space to batch to apply to larger batch sizes
2020-10-06 09:56:50 -07:00
stable_sort_expander_test.cc
…
stable_sort_expander.cc
…
stable_sort_expander.h
…
stream_pool_test.cc
…
stream_pool.cc
…
stream_pool.h
Prefixing TensorFlow thread annotation macros with TF_.
2020-03-05 08:42:01 -08:00
topk_rewriter_test.cc
[XLA] Rewrite 1d sort to TopK
2020-08-05 02:40:46 -07:00
topk_rewriter.cc
[XLA] Rewrite 1d sort to TopK
2020-08-05 02:40:46 -07:00
topk_rewriter.h
[XLA] Add TopK rewriter pass
2020-08-05 01:44:08 -07:00
transfer_manager.cc
[XLA] Don't pass on_host_shape to ShapedBuffer/ScopedShapedBuffer inside XLA.
2020-10-08 12:00:12 -07:00
transfer_manager.h
[XLA] Don't pass on_host_shape to ShapedBuffer/ScopedShapedBuffer inside XLA.
2020-10-08 12:00:12 -07:00
transpose_folding_test.cc
[XLA] Teach TransposeFolding to fold dots with batch dimensions
2020-10-08 09:13:33 -07:00
transpose_folding.cc
[XLA] Teach TransposeFolding to fold dots with batch dimensions
2020-10-08 09:13:33 -07:00
transpose_folding.h
[XLA] Add the ability to disable the Reduce(Reshape(X)) to Reduce(X) rewrite and have TransposeFolding default to folding all operands of Convolution and Dot
2020-03-07 00:36:17 -08:00
tree_reduction_rewriter.cc
…
tree_reduction_rewriter.h
minor spelling tweaks
2019-12-09 18:21:12 +09:00
triangular_solve_expander_test.cc
Merge pull request #39690 from whoozle:triangular-solve-expander-block-size
2020-07-15 06:57:41 -07:00
triangular_solve_expander.cc
[XLA] Refactor triangular solve expander to make InvertDiagBlocks overridable.
2020-08-27 19:46:58 -07:00
triangular_solve_expander.h
[XLA] Refactor triangular solve expander to make InvertDiagBlocks overridable.
2020-08-27 19:46:58 -07:00
tuple_points_to_analysis_test.cc
[XLA] Add an optional bool is_cross_program_prefetch field to kCopyStart HLOs.
2020-08-21 15:10:38 -07:00
tuple_points_to_analysis.cc
[XLA] Change CopyStart return type to (target, source, context)
2020-01-28 16:11:12 -08:00
tuple_points_to_analysis.h
[XLA] Change CopyStart return type to (target, source, context)
2020-01-28 16:11:12 -08:00
tuple_simplifier_test.cc
…
tuple_simplifier.cc
[XLA:HLO] Small refactoring and more comments in tuple_simplifier.
2020-06-08 11:52:07 -07:00
tuple_simplifier.h
[XLA:HLO] Small refactoring and more comments in tuple_simplifier.
2020-06-08 11:52:07 -07:00
tuple_util_test.cc
…
tuple_util.cc
…
tuple_util.h
[XLA] Add a RootInstructionSinker pass.
2020-04-22 14:46:15 -07:00
while_loop_analysis_test.cc
…
while_loop_analysis.cc
…
while_loop_analysis.h
…
while_loop_constant_sinking_test.cc
…
while_loop_constant_sinking.cc
minor spelling tweaks
2019-12-09 18:21:12 +09:00
while_loop_constant_sinking.h
…
while_loop_invariant_code_motion_test.cc
…
while_loop_invariant_code_motion.cc
[XLA] Correct WhileLoopInvariantCodeMotion log messages.
2020-05-07 13:35:38 -07:00
while_loop_invariant_code_motion.h
…
while_loop_simplifier_test.cc
While loop simplification to remove repeated parameters.
2020-08-28 09:18:08 -07:00
while_loop_simplifier.cc
While loop simplification to remove repeated parameters.
2020-08-28 09:18:08 -07:00
while_loop_simplifier.h
…
while_loop_trip_count_annotator_test.cc
…
while_loop_trip_count_annotator.cc
…
while_loop_trip_count_annotator.h
…
while_util_test.cc
Fix some HLO bugs in tests detected when using VerifiedHloModule.
2019-10-31 08:01:08 -07:00
while_util.cc
[XLA][TF2XLA] Support tensor list with dynamic dimension.
2020-04-22 14:37:18 -07:00
while_util.h
[XLA][TF2XLA] Support tensor list with dynamic dimension.
2020-04-22 14:37:18 -07:00
zero_sized_hlo_elimination_test.cc
Fix some HLO bugs in tests detected when using VerifiedHloModule.
2019-10-31 08:01:08 -07:00
zero_sized_hlo_elimination.cc
…
zero_sized_hlo_elimination.h
…