A machine learning compiler for GPUs, CPUs, and ML accelerators
This report summarizes scanner evidence for openxla/xla, a C++ project with 4,580 GitHub stars. SynthScan examined 441,521 lines of code across 2991 source files, recording 3841 pattern matches distributed across 19 syntactic categories. The legacy adjusted pattern signal is 9.8 (Moderate pattern signal); it is not an authorship estimate.
Use the findings to ask whether maintainers can explain intent, verify dependencies, simplify unnecessary scaffolding, and focus review effort. Risk Payload v1 adds ownership, supply-chain, over-engineering, and review-effort vectors when available. These are prompts for human verification, not probabilities, security findings, or measured maintenance costs.
Use each vector to guide manual verification of the underlying findings.
Can maintainers explain intent, assumptions, and trade-offs? 3644 finding(s).
Are imports declared and package/API names verified? 13 finding(s).
Are wrappers, guards, and repeated structures necessary? 438 finding(s).
Which changes need focused human review? 3841 finding(s).
This chart tracks the legacy adjusted pattern score across scan runs. Changes can reflect source changes, scan coverage, or detector/scoring versions; they do not measure AI adoption or prove that a team refactored code. Compare runs only when their scoring versions are compatible. The dashed line tracks the raw pattern-hit count.
Groups findings by configured severity weight. Severity is not a confidence estimate, security rating, or proof that a package is missing. Verify imports against the project's manifests and inspect each finding in its local design context.
This horizontal bar chart shows where configured pattern scores accumulate across directories. Treat the bars as locations to inspect, not as evidence of authorship or code quality. Use the file-level findings to decide whether a directory needs a focused ownership, dependency, complexity, or review check.
The scanner identified 3841 distinct pattern matches across 19 syntactic categories. Each entry below represents a discrete location in the source code where the engine recorded a detector match to inspect in context. Expand any category row to inspect the individual file paths, line numbers, code snippets, and the lexical context (CODE, COMMENT, or STRING) in which each match was detected.
Reading the findings table: The Severity column indicates the configured finding weight (CRITICAL / HIGH / MEDIUM / LOW). The Context column identifies whether the match occurred inside executable code, an inline comment, or a string literal — legacy comment-context matches receive a ×1.5 score weight. The ⚡ bolt icon marks clustered matches: three or more patterns within a 10-line window, each receiving an additional ×1.5 density multiplier under the legacy scoring model.
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | bazel_downloader.cfg | 1 | # Copyright 2026 The OpenXLA Authors. All Rights Reserved. | COMMENT |
| LOW | configure.py | 1 | #!/usr/bin/env python3 | COMMENT |
| LOW | .pre-commit-config.yaml | 1 | # Copyright 2026 The OpenXLA Authors. All Rights Reserved. | COMMENT |
| LOW | xla/literal_pool.h | 21 | #include <vector> | COMMENT |
| LOW | xla/shape_partition.h | 21 | #include <vector> | COMMENT |
| LOW | xla/shape_partition.h | 41 | // [0, 1), [1, 2), [2, 3), [3, 4), [4, 5) [5, 8) | COMMENT |
| LOW | xla/printer.h | 21 | #include <string> | COMMENT |
| LOW | xla/literal_comparison.h | 21 | COMMENT | |
| LOW | xla/literal_comparison.h | 61 | const LiteralSlice& expected, const LiteralSlice& actual, | COMMENT |
| LOW | xla/literal_comparison.h | 81 | // tuple which contains a complex/floating-point tensor, Near() is equivalent to | COMMENT |
| LOW | xla/executable_run_options.h | 141 | // values are: 0 to # of devices - 1. These are the logical device ordinals, | COMMENT |
| LOW | xla/text_literal_reader.h | 21 | #include "absl/status/statusor.h" | COMMENT |
| LOW | xla/array4d.h | 21 | #include <initializer_list> | COMMENT |
| LOW | xla/array4d.h | 41 | // | COMMENT |
| LOW | xla/shape_tree.h | 21 | #include <memory> | COMMENT |
| LOW | xla/shape_tree.h | 41 | // XLA Shape and holds a value of type T for each subshape. | COMMENT |
| LOW | xla/shape_tree.h | 61 | // efficiency, you can construct it with a const Shape* to avoid copies. In this | COMMENT |
| LOW | xla/fp_util.h | 21 | #include <algorithm> | COMMENT |
| LOW | xla/shape_layout.h | 21 | #include "absl/status/status.h" | COMMENT |
| LOW | xla/shape_layout.h | 41 | // Assigns the layouts in this ShapeLayout to the Layout fields of the given | COMMENT |
| LOW | xla/lit.cfg.py | 1 | # Copyright 2019 The OpenXLA Authors. | COMMENT |
| LOW | xla/sh_test_with_runfiles.py | 1 | # Copyright 2025 The OpenXLA Authors. | COMMENT |
| LOW | xla/permutation_util.h | 21 | #include <cstddef> | COMMENT |
| LOW | xla/debug_options_flags.h | 21 | #include "absl/log/log.h" | COMMENT |
| LOW | xla/debug_options_flags.h | 61 | }; | COMMENT |
| LOW | xla/debug_options_flags.h | 81 | // `debug_options_flags.cc` | COMMENT |
| LOW | xla/debug_options_flags.h | 101 | // Checks whether the pass fuel was explicitly set. | COMMENT |
| LOW | xla/debug_options_flags.h | 121 | // before the program exits, we'll print a warning. | COMMENT |
| LOW | xla/metric_table_report.h | 21 | COMMENT | |
| LOW | xla/metric_table_report.h | 41 | // * 10.51% %convolution.202 | COMMENT |
| LOW | xla/metric_table_report.h | 81 | void SetShowAllEntries(); | COMMENT |
| LOW | xla/tuple_tree.h | 21 | #include <initializer_list> | COMMENT |
| LOW | xla/tuple_tree.h | 161 | // is defined by the nesting of tuples. | COMMENT |
| LOW | xla/parse_flags_from_env.h | 21 | // non-whitespace in the variable value is not '-'), a file named by that | COMMENT |
| LOW | xla/parse_flags_from_env.h | 41 | // Examples: | COMMENT |
| LOW | xla/literal.h | 21 | #include <complex> | COMMENT |
| LOW | xla/literal.h | 41 | #include "absl/status/status_macros.h" | COMMENT |
| LOW | xla/literal.h | 121 | int64_t total_size_bytes() const; | COMMENT |
| LOW | xla/literal.h | 161 | COMMENT | |
| LOW | xla/literal.h | 281 | std::optional<double> GetAsDouble( | COMMENT |
| LOW | xla/literal.h | 321 | COMMENT | |
| LOW | xla/literal.h | 341 | // are considered equal to true/false; other values are not considered equal | COMMENT |
| LOW | xla/literal.h | 481 | std::unique_ptr<Literal> CloneToUnique() const; | COMMENT |
| LOW | xla/literal.h | 501 | COMMENT | |
| LOW | xla/literal.h | 521 | // This literal must be an array. | COMMENT |
| LOW | xla/literal.h | 541 | // start_indices and limit_indices must be the rank of the literal, and the | COMMENT |
| LOW | xla/literal.h | 561 | // See comments on ArrayValueState for detailed explanation. | COMMENT |
| LOW | xla/literal.h | 841 | private: | COMMENT |
| LOW | xla/literal.h | 1341 | COMMENT | |
| LOW | xla/literal.h | 1381 | COMMENT | |
| LOW | xla/literal.h | 1461 | absl::Status Populate(Generator&& generator); | COMMENT |
| LOW | xla/literal.h | 1601 | // allocated in this literal for the subshape at dest_shape_index is | COMMENT |
| LOW | xla/literal.h | 1681 | COMMENT | |
| LOW | xla/literal.h | 1781 | } | COMMENT |
| LOW | xla/shape.h | 21 | #include <cstdint> | COMMENT |
| LOW | xla/shape.h | 41 | #include "xla/util.h" | COMMENT |
| LOW | xla/shape.h | 81 | // Precondition: | COMMENT |
| LOW | xla/shape.h | 181 | // shapes are traversed recursively. | COMMENT |
| LOW | xla/shape.h | 241 | // Precondition: | COMMENT |
| LOW | xla/shape.h | 301 | } | COMMENT |
| 3203 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | bazel_downloader.cfg | 14 | # ============================================================================== | COMMENT |
| MEDIUM | configure.py | 15 | # ============================================================================== | COMMENT |
| MEDIUM | .pre-commit-config.yaml | 14 | # ============================================================================ | COMMENT |
| MEDIUM | xla/pjrt/distributed/testdata/mtls/generate.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | xla/python_api/types_.py | 14 | # ====================================== | COMMENT |
| MEDIUM | xla/python_api/xla_literal.py | 14 | # ====================================== | COMMENT |
| MEDIUM | xla/python_api/xla_literal_test.py | 14 | # ====================================== | COMMENT |
| MEDIUM | xla/python_api/xla_shape.py | 14 | # ====================================== | COMMENT |
| MEDIUM | xla/python_api/xla_shape_test.py | 14 | # ====================================== | COMMENT |
| MEDIUM | …ols/buffer_debug_log/check_thunk_output_consistency.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/tools/buffer_debug_log/checksum_mismatch_report.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | …ools/buffer_debug_log/checksum_mismatch_report_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/tools/benchmarks/json_to_tensorboard.py | 14 | # ============================================================================ | COMMENT |
| MEDIUM | xla/tools/benchmarks/test_json_to_tensorboard.py | 14 | # ============================================================================ | COMMENT |
| MEDIUM | xla/tools/benchmarks/hlo_workload_executor/action.yml | 14 | # ============================================================================ | COMMENT |
| MEDIUM | …a/tools/multihost_hlo_runner/python_hlo_runner_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/python/xla_transform_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/python/xla_client.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/python/xla_compiler_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/python/xla_hlo_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/python/status_casters_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/python/profile_data_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/python/xla_extension.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/python/profiler/profile_data_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/tsl/util/filewrapper_testdata_generator.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/tsl/lib/io/zip_writer_integration_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/tsl/builddata/workspace_status_command.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/gpu/codegen/tools/ncu_rep.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/gpu/codegen/tools/ncu_rep_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/gpu/codegen/tools/ncu_rep_lib.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | …ackends/cpu/codegen/computation_kernel_emitter_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/codegen/scatter_kernel_emitter_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/codegen/fusion_emitter_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/codegen/tiled/tiled_kernel_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/codegen/dot/dot_kernel_emitter_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | …u/codegen/elemental/concatenate_kernel_emitter_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/benchmarks/e2e/keras/setup.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/benchmarks/e2e/keras/benchmark.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/benchmarks/e2e/keras/run.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/benchmarks/e2e/keras/config.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/benchmarks/e2e/keras/cleanup.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/benchmarks/e2e/gemma2/flax_2b/setup.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | …ackends/cpu/benchmarks/e2e/gemma2/flax_2b/benchmark.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/benchmarks/e2e/gemma2/flax_2b/run.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | …a/backends/cpu/benchmarks/e2e/gemma2/flax_2b/config.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | …backends/cpu/benchmarks/e2e/gemma2/pytorch_2b/setup.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | …ends/cpu/benchmarks/e2e/gemma2/pytorch_2b/benchmark.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | …a/backends/cpu/benchmarks/e2e/gemma2/pytorch_2b/run.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | …ackends/cpu/benchmarks/e2e/gemma2/pytorch_2b/config.sh | 15 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/testlib/__init__.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/testlib/kernel_runner_test.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/backends/cpu/testlib/utilities.py | 14 | # ============================================================================== | COMMENT |
| MEDIUM | xla/tests/generate_complex_unary_op_samples.py | 14 | # ============================================================================ | COMMENT |
| MEDIUM⚡ | xla/mlir_hlo/CMakeLists.txt | 37 | #------------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | xla/mlir_hlo/CMakeLists.txt | 39 | #------------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | xla/mlir_hlo/CMakeLists.txt | 43 | #------------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | xla/mlir_hlo/CMakeLists.txt | 45 | #------------------------------------------------------------------------------- | COMMENT |
| MEDIUM | xla/mlir_hlo/CMakeLists.txt | 86 | #------------------------------------------------------------------------------- | COMMENT |
| MEDIUM | xla/mlir_hlo/CMakeLists.txt | 88 | #------------------------------------------------------------------------------- | COMMENT |
| MEDIUM⚡ | xla/mlir_hlo/CMakeLists.txt | 132 | #------------------------------------------------------------------------------- | COMMENT |
| 103 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | configure.py | 89 | def _get_cuda_compute_capabilities_or_die() -> list[str]: | CODE |
| LOW | configure.py | 254 | def get_relevant_paths_and_versions(self, config: "XLAConfigOptions"): | CODE |
| LOW | …ols/buffer_debug_log/check_thunk_output_consistency.py | 33 | def parse_binary_or_text_proto(data: bytes, proto_type): | CODE |
| LOW | xla/tools/buffer_debug_log/checksum_mismatch_report.py | 182 | def _find_inconsistent_output_checksums( | CODE |
| LOW | xla/tools/buffer_debug_log/checksum_mismatch_report.py | 210 | def _find_inconsistent_thunks( | CODE |
| LOW | …ools/buffer_debug_log/checksum_mismatch_report_test.py | 25 | def test_from_protos_loads_metadata(self): | CODE |
| LOW | …ools/buffer_debug_log/checksum_mismatch_report_test.py | 71 | def test_from_protos_finds_mismatches_in_single_proto(self): | STRING |
| LOW | …ools/buffer_debug_log/checksum_mismatch_report_test.py | 129 | def test_from_protos_finds_mismatches_in_multiple_protos(self): | STRING |
| LOW | …ools/buffer_debug_log/checksum_mismatch_report_test.py | 180 | def test_from_protos_does_not_include_consistent_executions(self): | STRING |
| LOW | …a/tools/multihost_hlo_runner/python_hlo_runner_test.py | 28 | def _register_transformer_engine_custom_calls(): | CODE |
| LOW | xla/python/xla_transform_test.py | 23 | def test_register_hlo_xla_transform(self): | CODE |
| LOW | xla/python/xla_transform_test.py | 34 | def test_core_assignment_unimplemented_on_cpu(self): | CODE |
| LOW | xla/python/xla_client.py | 131 | def current_source_info_metadata(op_type=None, op_name=None, skip_frames=1): | CODE |
| LOW | xla/python/xla_client.py | 166 | def window_padding_type_to_pad_values( | CODE |
| LOW | xla/python/xla_client.py | 265 | def make_dot_dimension_numbers( | CODE |
| LOW | xla/python/xla_client.py | 321 | def make_convolution_dimension_numbers( | CODE |
| LOW | xla/python/xla_client.py | 437 | def _make_replica_group_proto(replica_group): | CODE |
| LOW | xla/python/xla_client.py | 454 | def get_backend_config_string(instruction_proto, module_proto=None) -> str: | CODE |
| LOW | xla/python/profile_data_test.py | 22 | def test_find_plane_with_name(self): | CODE |
| LOW | xla/python/profile_data_test.py | 160 | def test_visit_event_missing_metadata(self): | STRING |
| LOW | xla/python/profile_data_test.py | 196 | def test_create_profile_data_from_file(self): | STRING |
| LOW | xla/python/profiler/profile_data_test.py | 23 | def test_find_plane_with_name(self): | CODE |
| LOW | xla/python/profiler/profile_data_test.py | 161 | def test_visit_event_missing_metadata(self): | STRING |
| LOW | xla/python/profiler/profile_data_test.py | 197 | def test_create_profile_data_from_file(self): | STRING |
| LOW | xla/tsl/lib/io/zip_writer_integration_test.py | 27 | def test_generated_zip_validity(self): | CODE |
| LOW | xla/backends/gpu/codegen/tools/ncu_rep_test.py | 23 | def test_get_metrics_by_kernel(self): | CODE |
| LOW | xla/backends/gpu/codegen/tools/ncu_rep_test.py | 47 | def test_aggregate_kernel_metrics(self): | CODE |
| LOW | xla/backends/gpu/codegen/tools/ncu_rep_test.py | 107 | def test_write_metrics_markdown(self): | CODE |
| LOW | …ackends/cpu/codegen/computation_kernel_emitter_test.py | 30 | def create_trivial_add_computation( | CODE |
| LOW | xla/backends/cpu/codegen/scatter_kernel_emitter_test.py | 28 | def create_sclar_add_computation( | CODE |
| LOW | xla/backends/cpu/codegen/scatter_kernel_emitter_test.py | 204 | def test_out_of_bounds_scatter(self): | CODE |
| LOW | xla/backends/cpu/codegen/fusion_emitter_test.py | 206 | def test_constant_with_layout(self): | STRING |
| LOW | xla/backends/cpu/codegen/tiled/tiled_kernel_test.py | 274 | def test_dot_fusion_single_tile(self): | STRING |
| LOW | xla/backends/cpu/codegen/tiled/tiled_kernel_test.py | 412 | def test_reduction_outer_inner(self): | STRING |
| LOW | xla/backends/cpu/codegen/tiled/tiled_kernel_test.py | 445 | def test_reduction_unsigned_add(self): | STRING |
| LOW | xla/backends/cpu/codegen/tiled/tiled_kernel_test.py | 477 | def test_broadcast_in_dim_inner(self): | STRING |
| LOW | xla/backends/cpu/codegen/tiled/tiled_kernel_test.py | 502 | def test_broadcast_in_dim_outer(self): | STRING |
| LOW | xla/backends/cpu/codegen/dot/dot_kernel_emitter_test.py | 131 | def test_matrix_multiplication(self, emitter_type, shapes, dtype): | CODE |
| LOW | …ackends/cpu/benchmarks/e2e/gemma2/flax_2b/benchmark.py | 52 | def benchmark_generation_time(output_token_len): | CODE |
| LOW | …ackends/cpu/benchmarks/e2e/gemma2/flax_2b/benchmark.py | 73 | def display_benchmark_results(timer_list, metric_name): | CODE |
| LOW | …ends/cpu/benchmarks/e2e/gemma2/pytorch_2b/benchmark.py | 60 | def benchmark_generation_time(output_token_len): | CODE |
| LOW | …ends/cpu/benchmarks/e2e/gemma2/pytorch_2b/benchmark.py | 85 | def display_benchmark_results(timer_list, metric_name): | CODE |
| LOW | xla/backends/cpu/testlib/kernel_runner_test.py | 28 | def test_llvm_ir_kernel_runner(self): | CODE |
| LOW⚡ | xla/mlir_hlo/tests/python/attributes.py | 39 | def test_comparison_direction_attr(): | CODE |
| LOW⚡ | xla/mlir_hlo/tests/python/attributes.py | 47 | def test_comparison_type_attr(): | CODE |
| LOW⚡ | xla/mlir_hlo/tests/python/attributes.py | 55 | def test_conv_dimension_numbers(): | CODE |
| LOW | xla/mlir_hlo/tests/python/attributes.py | 81 | def test_dequantize_mode_attr(): | CODE |
| LOW | xla/mlir_hlo/tests/python/attributes.py | 89 | def test_dot_dimension_numbers(): | CODE |
| LOW | xla/mlir_hlo/tests/python/attributes.py | 124 | def test_gather_dimension_numbers(): | CODE |
| LOW | xla/mlir_hlo/tests/python/attributes.py | 148 | def test_output_operand_alias(): | CODE |
| LOW | xla/mlir_hlo/tests/python/attributes.py | 178 | def test_rng_distribution_attr(): | CODE |
| LOW | xla/mlir_hlo/tests/python/attributes.py | 186 | def test_scatter_dimension_numbers(): | CODE |
| LOW | xla/codegen/testlib/kernel_runner_test.py | 55 | def test_output_same_as_input(self): | CODE |
| LOW | xla/hlo/tools/generate_hlo_test_checks.py | 186 | def _format_function_declaration_file_check(line_text: str) -> str: | CODE |
| LOW | xla/hlo/tools/generate_hlo_test_checks.py | 467 | def _prefix_lines_with_check_directives( | CODE |
| LOW | xla/hlo/tools/generate_hlo_test_checks.py | 508 | def _replace_symbol_names_with_regex_captures( | CODE |
| LOW | xla/hlo/tools/generate_hlo_test_checks.py | 1007 | def transform_and_overwrite_file( | CODE |
| LOW | xla/hlo/tools/tests/generate_hlo_test_checks_test.py | 38 | def test_parallel_mode_generates_expected_output_file(self): | CODE |
| LOW | xla/hlo/tools/tests/generate_hlo_test_checks_test.py | 62 | def test_sequential_mode_generates_expected_output_file(self): | CODE |
| LOW | xla/hlo/tools/tests/generate_hlo_test_checks_test.py | 83 | def test_custom_input_file_placeholder_string(self): | CODE |
| 193 more matches not shown… | ||||
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | xla/pjrt/cpu/abstract_cpu_buffer.h | 57 | // robust by using setting the AsyncValue in the destructor. | COMMENT |
| MEDIUM | …/distributed/coordination/coordination_service_agent.h | 181 | // The agent does not need to be connected to utilize the key-value store. | COMMENT |
| MEDIUM | …a/pjrt/distributed/coordination/coordination_service.h | 191 | // connected to utilize the distributed key-value store. | COMMENT |
| MEDIUM | xla/python_api/xla_literal_test.py | 53 | # To facilitate iteration over different test cases, we collect similar array | COMMENT |
| MEDIUM | xla/tools/hlo_dump/hlo_lexer.h | 32 | // 2. The HTML dumper needs to be robust to malformed or partial HLO text | COMMENT |
| MEDIUM | xla/python/ifrt_proxy/server/ifrt_backend.h | 52 | // service to interact with a variety of backend runtime system it can utilize. | COMMENT |
| MEDIUM | xla/python/ifrt/client.h | 177 | // robust cleanup, it is strongly recommended to capture RAII objects in the | COMMENT |
| MEDIUM | xla/python/ifrt/ir/program_interpreter.h | 58 | // 1. It allows us to leverage the static information available in the program | COMMENT |
| MEDIUM | xla/tsl/framework/cancellation.h | 124 | // is robust to `callback` being invoked asynchronously (e.g. from | COMMENT |
| MEDIUM | xla/tsl/lib/gtl/int_type.h | 24 | // other operations from other similar integer-like types. Essentially, the | COMMENT |
| MEDIUM | xla/tsl/lib/gtl/int_type.h | 133 | // The following bad usage is permissible although discouraged. Essentially, it | COMMENT |
| MEDIUM | …uted_runtime/coordination/coordination_service_agent.h | 191 | // The agent does not need to be connected to utilize the key-value store. | COMMENT |
| MEDIUM | …istributed_runtime/coordination/coordination_service.h | 191 | // connected to utilize the distributed key-value store. | COMMENT |
| MEDIUM | xla/backends/gpu/runtime/while_loop.h | 49 | // storage. Essentially we keep a thread local stack of `WhileLoopState` entries | COMMENT |
| MEDIUM | xla/backends/gpu/transforms/conv_fp8_fallback.h | 54 | // rewriting the convolution earlier is not robust because inserted converts | COMMENT |
| MEDIUM | xla/backends/gpu/transforms/splitk_rewriter.h | 29 | // Rewrites dot instructions that don't fully utilize cores but have a long K | COMMENT |
| MEDIUM | xla/hlo/tools/hlo_diff/matchers/gumgraph_matcher.h | 49 | // output types, ensuring seamless integration and compatibility within any | COMMENT |
| MEDIUM | xla/service/compiler.h | 196 | // Note: The default implementation of the API here does not utilize the given | COMMENT |
| MEDIUM | xla/service/hlo_runner_interface.h | 238 | // Note: The default implementation of the API here does not utilize the given | COMMENT |
| MEDIUM | xla/service/hlo_runner_interface.h | 263 | // Note: The default implementation of the API here does not utilize the given | COMMENT |
| MEDIUM | xla/service/batchnorm_expander.h | 30 | // big operation into smaller operations helps leverage our generic fusion | COMMENT |
| MEDIUM | xla/service/local_service.h | 39 | // Service implementation that extends the XLA Service to leverage running | COMMENT |
| MEDIUM | xla/service/hlo_graph_dumper.h | 35 | // Fundamentally all graphs are rendered using the DOT language, but they can be | COMMENT |
| MEDIUM | xla/service/spmd/spmd_partitioner_util_internal.h | 38 | // NOTE: If provided sharding does not utilize all the partitions, we skip | COMMENT |
| MEDIUM | xla/service/heap_simulator/allocation_block.h | 75 | // composed of slices. In such cases, the repacker can utilize | COMMENT |
| LOW | tools/ci_build/gpu_build/parallel_gpu_execute.sh | 41 | # If the file path is already fully specified, simply return it. | COMMENT |
| LOW | tools/ci_build/gpu_build/parallel_gpu_execute.sh | 44 | # If the file exists in the $TEST_SRCDIR then just use it. | COMMENT |
| LOW | build_tools/ci/parallel_gpu_execute.sh | 41 | # If the file path is already fully specified, simply return it. | COMMENT |
| LOW | build_tools/ci/parallel_gpu_execute.sh | 44 | # If the file exists in the $TEST_SRCDIR then just use it. | COMMENT |
| LOW | build_tools/rocm/parallel_gpu_execute.sh | 50 | # If the file path is already fully specified, simply return it. | COMMENT |
| LOW | build_tools/rocm/parallel_gpu_execute.sh | 53 | # If the file exists in the $TEST_SRCDIR then just use it. | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | xla/sh_test_with_runfiles.py | 109 | except FileNotFoundError: | CODE |
| HIGH | xla/sh_test_with_runfiles.py | 51 | except FileExistsError: | CODE |
| HIGH | …ols/buffer_debug_log/check_thunk_output_consistency.py | 37 | except message.DecodeError: | CODE |
| HIGH | …odegen/intrinsic/accuracy/generate_golden_baselines.py | 145 | except ValueError: | CODE |
| HIGH | …a/benchmarks/jax_microbenchmarks/jax_profiler_utils.py | 122 | except ValueError: | CODE |
| HIGH | xla/service/fusion_explorer/viewer_server.py | 201 | except webbrowser.Error: | CODE |
| HIGH | xla/service/fusion_explorer/viewer_server.py | 214 | except KeyboardInterrupt: | CODE |
| HIGH | xla/service/fusion_explorer/viewer_server.py | 66 | except sqlite3.OperationalError: | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | xla/online_topsort.h | 513 | // Step 1: test order: if w is already higher than v in the lexicographical | COMMENT |
| LOW⚡ | xla/online_topsort.h | 523 | // Step 2: search backwards from v, until we either find `w`, which means we | COMMENT |
| LOW | xla/online_topsort.h | 547 | // Step 3: search forwards from w, following outgoing edges only from nodes | COMMENT |
| LOW | xla/online_topsort.h | 565 | // Step 4: update indices. | COMMENT |
| LOW | xla/online_topsort.h | 581 | // Step 5: add the edge. | COMMENT |
| LOW | xla/tsl/platform/test.h | 77 | // NOTE: This function is not thread-safe. | COMMENT |
| LOW | xla/tsl/profiler/utils/xplane_utils.h | 248 | // NOTE: This class assumes that the ancestors are ordered with respect to the | COMMENT |
| LOW | xla/tsl/distributed_runtime/rpc/grpc_call.h | 255 | // NOTE: This method must be called before this call is enqueued on a | COMMENT |
| LOW | xla/hlo/ir/tile_assignment.h | 180 | // NOTE: This class is immutable. | COMMENT |
| LOW | xla/hlo/builder/lib/matrix.h | 118 | // NOTE: This function is meant for testing, there is no need to call it | COMMENT |
| LOW⚡ | xla/service/mapped_ptr_container_sorter.h | 370 | // Step 1: build a set of the ptrs in unordered_container | COMMENT |
| LOW⚡ | xla/service/mapped_ptr_container_sorter.h | 377 | // Step 2: for mapped elements (in unordered_container), create a map from | COMMENT |
| LOW | xla/service/mapped_ptr_container_sorter.h | 399 | // Step 3: create sorted unordered element indices | COMMENT |
| LOW | docs/pjrt/pjrt_integration.md | 8 | ### Step 1: Implement [PJRT C API interface](https://github.com/openxla/xla/blob/71a4e6e6e4e9f0f8b8f25c07a32ad489aff1923 | COMMENT |
| LOW | docs/pjrt/pjrt_integration.md | 38 | ### Step 2: Implement GetPjRtApi | COMMENT |
| LOW | docs/pjrt/pjrt_integration.md | 50 | ### Step 3: Test C API implementations | COMMENT |
| LOW | docs/pjrt/pjrt_integration.md | 56 | ### Step 1: Set up JAX | COMMENT |
| LOW | docs/pjrt/pjrt_integration.md | 76 | ### Step 2: Use jax\_plugins namespace or set up entry\_point | COMMENT |
| LOW | docs/pjrt/pjrt_integration.md | 104 | ### Step 3: Implement an initialize() method | COMMENT |
| LOW | .github/workflows/benchmarks/README.md | 130 | ### Step 1: Prepare Your Benchmark Artifact | COMMENT |
| LOW | .github/workflows/benchmarks/README.md | 148 | ### Step 2: Define the Benchmark Configuration in `benchmark_registry.pbtxt` | COMMENT |
| LOW | .github/workflows/benchmarks/README.md | 163 | ### Step 3: Establish Baselines & Thresholds | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | configure.py | 62 | Finds executable and resolves symlinks or raises RuntimeError. Resolving symlinks is sometimes necessary for finding | STRING |
| HIGH | xla/hlo/tools/generate_hlo_test_checks.py | 219 | Passes `input_stream` through an optimizer; output is another stream. Args: input_stream: An iterator over the li | STRING |
| HIGH | xla/hlo/tools/generate_hlo_test_checks.py | 301 | Returns the next item in the sub-stream corresponding to `target_buffer`. Args: target_buffer: The queue back | STRING |
| HIGH | build_tools/configure/configure.py | 62 | Finds executable and resolves symlinks or raises RuntimeError. Resolving symlinks is sometimes necessary for finding | STRING |
| HIGH | build_tools/ci/clang_tidy_diff.py | 308 | Parses a Bazel BEP JSON file and returns a list of paths to .clang-tidy.yaml files. Args: bep_path: Path to the B | STRING |
| HIGH | build_tools/ci/change_detector.py | 83 | Validates that ref is a safe Git revision string and cannot inject flags. Args: ref: Git commit SHA or ref name t | STRING |
| HIGH | .github/workflows/github_api.py | 39 | Helper method to make a request and raise an HTTPError if one occurred. Arguments: verb: The HTTP verb to use | STRING |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | configure.py | 319 | CODE | |
| LOW | xla/sh_test_with_runfiles.py | 25 | CODE | |
| LOW | xla/python/xla_hlo_test.py | 30 | CODE | |
| LOW | xla/python/xla_hlo_test.py | 246 | CODE | |
| LOW | xla/python/xla_hlo_test.py | 459 | CODE | |
| LOW | xla/python/xla_hlo_test.py | 510 | CODE | |
| LOW | xla/tsl/lib/io/zip_writer_integration_test.py | 27 | CODE | |
| LOW | xla/backends/cpu/benchmarks/e2e/keras/benchmark.py | 110 | CODE | |
| LOW | xla/tests/generate_complex_unary_op_samples.py | 53 | CODE | |
| LOW | xla/hlo/tools/generate_hlo_test_checks.py | 300 | CODE | |
| LOW | xla/benchmarks/benchmark_configs.py | 57 | CODE | |
| LOW | …a/benchmarks/jax_microbenchmarks/jax_profiler_utils.py | 81 | CODE | |
| LOW | xla/benchmarks/core/benchmark.py | 125 | CODE | |
| LOW | xla/benchmarks/core/flag_utils.py | 53 | CODE | |
| LOW | xla/benchmarks/core/flag_utils.py | 118 | CODE | |
| LOW | xla/benchmarks/dma_microbenchmarks/memory_base.py | 66 | CODE | |
| LOW | xla/benchmarks/collectives/collectives_benchmark.py | 146 | CODE | |
| LOW | xla/service/fusion_explorer/viewer_server.py | 159 | CODE | |
| LOW | xla/service/fusion_explorer/viewer_server.py | 43 | CODE | |
| LOW | third_party/py/setup_py_nvidia_dependencies_util.py | 46 | CODE | |
| LOW | third_party/mkl_dnn/gen_gpu_kernel_list.py | 133 | CODE | |
| LOW | third_party/gpus/find_cuda_config.py | 223 | CODE | |
| LOW | build_tools/lint/check_dwyu.py | 165 | CODE | |
| LOW | build_tools/configure/configure.py | 319 | CODE | |
| LOW | build_tools/ci/build.py | 958 | CODE | |
| LOW | build_tools/ci/bazel_diff.py | 304 | CODE | |
| LOW | build_tools/ci/bazel_diff.py | 433 | CODE | |
| LOW | build_tools/ci/clang_tidy_diff.py | 307 | CODE | |
| LOW | build_tools/ci/clang_tidy_diff.py | 351 | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | xla/debug_options_flags.h | 132 | // Example usage: | COMMENT |
| LOW | xla/reference_util.h | 599 | // Example usage: | COMMENT |
| LOW | xla/tsl/util/sorted_range.h | 29 | // Example usage: | COMMENT |
| LOW | xla/tsl/util/sorted_range.h | 49 | // Example usage: | COMMENT |
| LOW | xla/tsl/util/fixed_option_set_flag.h | 47 | // Example usage: | COMMENT |
| LOW | xla/tsl/platform/debug_me_context.h | 47 | // Example usage: | COMMENT |
| LOW | xla/tsl/platform/errors.h | 201 | // Example usage: | COMMENT |
| LOW | xla/tsl/platform/cloud/compute_engine_metadata_client.h | 50 | /// Example usage: | COMMENT |
| LOW | xla/tsl/lib/gtl/value_or_die.h | 21 | // Example usage: | COMMENT |
| LOW | xla/tsl/lib/monitoring/collection_registry.h | 499 | // Example usage: | COMMENT |
| LOW | xla/tests/xla_test_backend_predicates.h | 119 | // Example usage: | COMMENT |
| LOW | xla/hlo/analysis/shape_tracker.h | 42 | // Example usage: | COMMENT |
| LOW | xla/hlo/ir/hlo_computation.h | 180 | // Example usage: | COMMENT |
| LOW | xla/service/pattern_matcher.h | 3244 | // Example usage: | COMMENT |
| LOW | xla/service/cpu/ir_function.h | 46 | // Example usage: | COMMENT |
| LOW | xla/stream_executor/kernel_symbol_registry.h | 85 | // Example usage: | COMMENT |
| LOW | …stream_executor/gpu/scoped_command_buffer_annotation.h | 28 | // Example usage: | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | xla/backends/cpu/benchmarks/e2e/gemma2/flax_2b/setup.sh | 26 | # Create a virtual environment | COMMENT |
| MEDIUM | …backends/cpu/benchmarks/e2e/gemma2/pytorch_2b/setup.sh | 27 | # Create a virtual environment | COMMENT |
| MEDIUM | …arks/dma_microbenchmarks/chip_to_chip_dma_benchmark.py | 35 | # Create a mesh with 2 devices, each on a different chip. | COMMENT |
| MEDIUM | tools/ci_build/gpu_build/parallel_gpu_execute.sh | 27 | # This function is used below in rlocation to check that a path is absolute | COMMENT |
| MEDIUM | build_tools/ci/parallel_gpu_execute.sh | 27 | # This function is used below in rlocation to check that a path is absolute | COMMENT |
| MEDIUM | build_tools/rocm/parallel_gpu_execute.sh | 36 | # This function is used below in rlocation to check that a path is absolute | COMMENT |
| MEDIUM | .github/workflows/benchmarks/prepare_artifact.sh | 26 | # Create the directory HERE, inside this script, right before using it. | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| HIGH | xla/packed_literal_reader.h | 33 | // Reads packed data from a metadata-less file as requested by a user (who must | COMMENT |
| HIGH | docs/contributing.md | 201 | ### AI-generated contributions | COMMENT |
| HIGH | docs/contributing.md | 220 | their PRs in detail regardless of whether it was generated by AI or not. | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW⚡ | xla/online_topsort.h | 513 | // Step 1: test order: if w is already higher than v in the lexicographical | COMMENT |
| LOW⚡ | xla/online_topsort.h | 523 | // Step 2: search backwards from v, until we either find `w`, which means we | COMMENT |
| LOW | xla/online_topsort.h | 547 | // Step 3: search forwards from w, following outgoing edges only from nodes | COMMENT |
| LOW | xla/online_topsort.h | 565 | // Step 4: update indices. | COMMENT |
| LOW | xla/online_topsort.h | 581 | // Step 5: add the edge. | COMMENT |
| LOW⚡ | xla/service/mapped_ptr_container_sorter.h | 370 | // Step 1: build a set of the ptrs in unordered_container | COMMENT |
| LOW⚡ | xla/service/mapped_ptr_container_sorter.h | 377 | // Step 2: for mapped elements (in unordered_container), create a map from | COMMENT |
| LOW | xla/service/mapped_ptr_container_sorter.h | 399 | // Step 3: create sorted unordered element indices | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | …ols/buffer_debug_log/check_thunk_output_consistency.py | 94 | except Exception as e: | CODE |
| LOW | …ols/buffer_debug_log/check_thunk_output_consistency.py | 104 | except Exception as e: | CODE |
| MEDIUM | xla/backends/cpu/benchmarks/e2e/keras/benchmark.py | 74 | print(f"Error: max_len {max_len} should be >= 1") | CODE |
| LOW | …a/benchmarks/jax_microbenchmarks/jax_profiler_utils.py | 75 | except Exception as e: # pylint: disable=broad-except | CODE |
| LOW | …a/benchmarks/jax_microbenchmarks/jax_profiler_utils.py | 134 | except Exception: # pylint: disable=broad-except | CODE |
| LOW⚡ | xla/service/fusion_explorer/viewer_server.py | 177 | except Exception as e: # pylint: disable=broad-exception-caught | CODE |
| MEDIUM⚡ | xla/service/fusion_explorer/viewer_server.py | 178 | print(f"Error extracting database: {e}") | CODE |
| LOW⚡ | xla/service/fusion_explorer/viewer_server.py | 185 | except Exception: # pylint: disable=broad-exception-caught | CODE |
| LOW | xla/service/fusion_explorer/viewer_server.py | 209 | except Exception as e: # pylint: disable=broad-exception-caught | CODE |
| LOW | .github/workflows/pr_size_check.py | 77 | except Exception as e: # pylint: disable=broad-except | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | xla/python/xla_client.py | 24 | CODE | |
| LOW | xla/python/xla_client.py | 34 | CODE | |
| LOW | xla/python/xla_client.py | 36 | CODE | |
| LOW | xla/python/xla_client.py | 37 | CODE | |
| LOW | xla/python/xla_extension.py | 18 | CODE | |
| LOW | xla/python/xla_extension.py | 20 | CODE | |
| LOW | xla/mlir_hlo/tests/python/smoketest.py | 19 | CODE | |
| LOW | xla/mlir_hlo/tests/python/smoketest.py | 20 | CODE | |
| LOW | xla/mlir_hlo/bindings/python/mlir/dialects/mhlo.py | 19 | CODE | |
| LOW | xla/mlir_hlo/bindings/python/mlir/dialects/mhlo.py | 20 | CODE | |
| LOW | xla/hlo/tools/generate_hlo_test_checks.py | 91 | CODE | |
| LOW | build_tools/ci/clang_tidy_diff.py | 25 | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| CRITICAL | xla/backends/gpu/tests/gpu_pjrt_codegen_test.h | 114 | return "void @llvm.nvvm.barrier.cta.sync.aligned.all(i32 0)"; | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| MEDIUM | xla/tools/hlo_isolation/hlo_inf_nan_intent_analyzer.h | 48 | // numerical guardrails) rather than genuine numerical errors. | COMMENT |
| MEDIUM | build_tools/ci/run_clang_tidy.sh | 79 | # TODO(sohaibiftikhar): Move the bash orchestration to python. | COMMENT |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | build_tools/lint/generate_compile_commands.py | 53 | def _setup_external_symlink(xla_src_root: pathlib.Path) -> None: | STRING |
| LOW | build_tools/ci/clang_tidy_diff.py | 174 | def _set_log_level(log_level: str) -> None: | CODE |
| LOW | build_tools/ci/clang_tidy_diff.py | 184 | def _set_level(level: int) -> None: | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | xla/backends/cpu/codegen/dot/dot_kernel_emitter_test.py | 47 | CODE |
| Severity | File | Line | Snippet | Context |
|---|---|---|---|---|
| LOW | xla/python/xla_hlo_test.py | 234 | # Check if we can find expected opcodes in the main computation. | STRING |