pytorch / FBGEMM
File Size

The distribution of size of files (measured in lines of code).

Intro
  • File size measurements show the distribution of size of files.
  • Files are classified in four categories based on their size (lines of code): 1-100 (very small files), 101-200 (small files), 201-500 (medium size files), 501-1000 (long files), 1001+(very long files).
  • It is a good practice to keep files small. Long files may become "bloaters", code that have increased to such gargantuan proportions that they are hard to work with.
Learn more...
File Size Overall
  • There are 234 files with 70,952 lines of code.
    • 14 very long files (24,815 lines of code)
    • 18 long files (13,278 lines of code)
    • 70 medium size files (22,274 lines of codeclsfd_ftr_w_mp_ins)
    • 47 small files (6,957 lines of code)
    • 85 very small files (3,628 lines of code)
34% | 18% | 31% | 9% | 5%
Legend:
1001+
501-1000
201-500
101-200
1-100


explore: zoomable circles | sunburst | 3D view
File Size per Extension
1001+
501-1000
201-500
101-200
1-100
cc31% | 22% | 34% | 8% | 2%
cu60% | 15% | 20% | 2% | <1%
py58% | 7% | 17% | 9% | 7%
cuh78% | 0% | 0% | 9% | 11%
cpp22% | 19% | 37% | 16% | 3%
h0% | 18% | 45% | 17% | 17%
cmake0% | 0% | 96% | 0% | 3%
bzl0% | 0% | 0% | 100% | 0%
toml0% | 0% | 0% | 0% | 100%
File Size per Logical Decomposition
primary
1001+
501-1000
201-500
101-200
1-100
src34% | 26% | 29% | 5% | 3%
fbgemm_gpu/src59% | 14% | 17% | 6% | 2%
fbgemm_gpu/fbgemm_gpu83% | 0% | 0% | 6% | 9%
fbgemm_gpu/include/fbgemm_gpu57% | 0% | 15% | 11% | 15%
fbgemm_gpu/bench50% | 0% | 37% | 8% | 3%
fbgemm_gpu/codegen15% | 29% | 38% | 12% | 3%
include/fbgemm0% | 23% | 36% | 26% | 14%
bench0% | 0% | 68% | 24% | 6%
cmake/modules0% | 0% | 96% | 0% | 3%
fbgemm_gpu0% | 0% | 0% | 88% | 11%
ROOT0% | 0% | 0% | 100% | 0%
fbgemm_gpu/bench/scripts0% | 0% | 0% | 0% | 100%
Longest Files (Top 50)
File# lines# units
3130 14
2247 8
1963 48
sparse_ops.cu
in fbgemm_gpu/src
 1953 -
fbgemm_cuda_utils.cuh
in fbgemm_gpu/include/fbgemm_gpu
 1948 -
1916 21
1803 -
1761 15
1755 31
sparse_ops_cpu.cpp
in fbgemm_gpu/src
 1683 33
1251 1
1170 1
jagged_tensor_ops.cu
in fbgemm_gpu/src
 1150 -
1085 -
989 5
jagged_tensor_ops_cpu.cpp
in fbgemm_gpu/src
 967 28
938 12
917 -
886 6
845 9
842 4
823 1
Fbgemm.h
in include/fbgemm
 804 46
706 24
quantize_ops.cu
in fbgemm_gpu/src
 666 -
656 4
645 4
531 34
embedding_forward_split_cpu.cpp
in fbgemm_gpu/codegen
 526 5
521 29
Fbgemm.cc
in src
 510 3
506 6
491 2
487 2
486 -
483 2
467 2
461 4
450 -
449 2
448 15
440 10
434 6
BenchUtils.h
in bench
 415 6
413 2
412 3
409 3
Utils.cc
in src
 397 19
388 -
380 19
Files With Most Units (Top 50)
File# lines# units
1963 48
Fbgemm.h
in include/fbgemm
 804 46
PackingTraits-inl.h
in include/fbgemm
 298 35
531 34
sparse_ops_cpu.cpp
in fbgemm_gpu/src
 1683 33
1755 31
521 29
jagged_tensor_ops_cpu.cpp
in fbgemm_gpu/src
 967 28
706 24
1916 21
380 19
Utils.cc
in src
 397 19
FbgemmPackMatrixB.h
in include/fbgemm
 200 18
448 15
sparse_ops_utils.h
in fbgemm_gpu/include/fbgemm_gpu
 214 15
1761 15
167 15
111 15
quantize_ops_cpu.cpp
in fbgemm_gpu/src
 303 14
3130 14
182 13
938 12
440 10
QuantUtils.h
in include/fbgemm
 224 10
92 9
845 9
242 8
2247 8
split_embedding_configs.py
in fbgemm_gpu/fbgemm_gpu
 88 7
341 7
343 7
264 7
86 7
BenchUtils.h
in bench
 415 6
145 6
359 6
setup.py
in fbgemm_gpu
 144 6
sparse_ops_gpu.cpp
in fbgemm_gpu/src
 157 6
506 6
886 6
189 6
434 6
253 6
batched_unary_embeddings_ops.py
in fbgemm_gpu/fbgemm_gpu
 61 5
363 5
embedding_forward_split_cpu.cpp
in fbgemm_gpu/codegen
 526 5
input_combine_cpu.cpp
in fbgemm_gpu/src
 278 5
132 5
989 5
42 5
Files With Long Lines (Top 23)

There are 23 files with lines longer than 120 characters. In total, there are 121 long lines.

File# lines# units# long lines
917 - 22
1085 - 16
sparse_ops_cpu.cpp
in fbgemm_gpu/src
 1683 33 12
1916 21 10
440 10 7
359 6 7
467 2 5
101 - 5
jagged_tensor_ops_cpu.cpp
in fbgemm_gpu/src
 967 28 5
quantize_ops_benchmark.py
in fbgemm_gpu/bench
 186 3 4
274 - 4
450 - 4
1963 48 3
200 2 3
380 19 3
input_combine_cpu.cpp
in fbgemm_gpu/src
 278 5 3
132 5 2
448 15 1
embedding_bounds_check.cu
in fbgemm_gpu/codegen
 138 - 1
98 1 1
164 3 1
63 1 1
341 7 1
Correlations

File Size vs. Commits (all time): 233 points

fbgemm_gpu/codegen/embedding_forward_quantized_cpu_template.cpp x: 31 commits (all time) y: 359 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host.cpp x: 24 commits (all time) y: 180 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host_cpu.cpp x: 28 commits (all time) y: 380 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_split_template.cu x: 38 commits (all time) y: 917 lines of code fbgemm_gpu/fbgemm_gpu/split_table_batched_embeddings_ops.py x: 93 commits (all time) y: 1963 lines of code fbgemm_gpu/src/sparse_ops.cu x: 52 commits (all time) y: 1953 lines of code fbgemm_gpu/fbgemm_gpu/permute_pooled_embedding_modules.py x: 8 commits (all time) y: 51 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embedding_ops.h x: 8 commits (all time) y: 16 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embs_function.h x: 3 commits (all time) y: 1 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops_gpu.cpp x: 11 commits (all time) y: 132 lines of code fbgemm_gpu/src/jagged_tensor_ops.cu x: 22 commits (all time) y: 1150 lines of code fbgemm_gpu/src/jagged_tensor_ops_cpu.cpp x: 14 commits (all time) y: 967 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops.h x: 40 commits (all time) y: 317 lines of code fbgemm_gpu/src/sparse_ops_cpu.cpp x: 51 commits (all time) y: 1683 lines of code fbgemm_gpu/src/sparse_ops_gpu.cpp x: 35 commits (all time) y: 157 lines of code fbgemm_gpu/src/split_embeddings_cache_cuda.cu x: 27 commits (all time) y: 1803 lines of code fbgemm_gpu/include/fbgemm_gpu/embedding_common.h x: 5 commits (all time) y: 54 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops_utils.h x: 10 commits (all time) y: 214 lines of code fbgemm_gpu/src/split_table_batched_embeddings.cpp x: 17 commits (all time) y: 101 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host.cpp x: 17 commits (all time) y: 363 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host_cpu.cpp x: 17 commits (all time) y: 164 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_cpu_template.cpp x: 18 commits (all time) y: 200 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_template.cpp x: 24 commits (all time) y: 467 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host.cpp x: 11 commits (all time) y: 16 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host_cpu.cpp x: 16 commits (all time) y: 98 lines of code fbgemm_gpu/src/cumem_utils_host.cpp x: 15 commits (all time) y: 34 lines of code fbgemm_gpu/bench/split_table_batched_embeddings_benchmark.py x: 64 commits (all time) y: 1916 lines of code fbgemm_gpu/pyproject.toml x: 2 commits (all time) y: 18 lines of code fbgemm_gpu/bench/sparse_ops_benchmark.py x: 1 commits (all time) y: 52 lines of code fbgemm_gpu/codegen/embedding_backward_code_generator.py x: 25 commits (all time) y: 531 lines of code fbgemm_gpu/bench/batched_unary_embeddings_benchmark.py x: 2 commits (all time) y: 137 lines of code fbgemm_gpu/bench/quantize_ops_benchmark.py x: 9 commits (all time) y: 186 lines of code fbgemm_gpu/include/fbgemm_gpu/input_combine.h x: 6 commits (all time) y: 16 lines of code fbgemm_gpu/src/input_combine_cpu.cpp x: 13 commits (all time) y: 278 lines of code fbgemm_gpu/include/fbgemm_gpu/quantize_ops_gpu.h x: 1 commits (all time) y: 14 lines of code fbgemm_gpu/src/quantize_ops.cu x: 9 commits (all time) y: 666 lines of code fbgemm_gpu/src/quantize_ops_cpu.cpp x: 13 commits (all time) y: 303 lines of code fbgemm_gpu/src/quantize_ops_gpu.cpp x: 11 commits (all time) y: 35 lines of code bench/GEMMsBenchmark.cc x: 11 commits (all time) y: 260 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops.cuh x: 20 commits (all time) y: 35 lines of code fbgemm_gpu/include/fbgemm_gpu/fbgemm_cuda_utils.cuh x: 25 commits (all time) y: 1948 lines of code fbgemm_gpu/bench/bench_utils.py x: 1 commits (all time) y: 34 lines of code fbgemm_gpu/bench/merge_embeddings_benchmark.py x: 17 commits (all time) y: 440 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_approx_template.cpp x: 20 commits (all time) y: 191 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_template.cpp x: 30 commits (all time) y: 340 lines of code fbgemm_gpu/codegen/embedding_backward_split_indice_weights_template.cu x: 12 commits (all time) y: 274 lines of code fbgemm_gpu/codegen/embedding_backward_split_template.cu x: 25 commits (all time) y: 1085 lines of code fbgemm_gpu/codegen/embedding_bounds_check.cu x: 15 commits (all time) y: 138 lines of code fbgemm_gpu/codegen/embedding_forward_split_cpu.cpp x: 38 commits (all time) y: 526 lines of code fbgemm_gpu/codegen/embedding_forward_split_template.cu x: 18 commits (all time) y: 450 lines of code fbgemm_gpu/src/histogram_binning_calibration_ops.cu x: 8 commits (all time) y: 388 lines of code fbgemm_gpu/src/layout_transform_ops_cpu.cpp x: 9 commits (all time) y: 63 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops.cu x: 8 commits (all time) y: 74 lines of code fbgemm_gpu/src/split_embeddings_utils.cu x: 6 commits (all time) y: 284 lines of code src/FbgemmConv.cc x: 27 commits (all time) y: 434 lines of code fbgemm_gpu/include/fbgemm_gpu/cpu_utils.h x: 9 commits (all time) y: 42 lines of code include/fbgemm/FbgemmEmbedding.h x: 42 commits (all time) y: 234 lines of code src/GroupwiseConv.cc x: 18 commits (all time) y: 938 lines of code src/TransposeUtils.cc x: 9 commits (all time) y: 74 lines of code fbgemm_gpu/bench/split_embeddings_cache_benchmark.py x: 5 commits (all time) y: 448 lines of code fbgemm_gpu/src/merge_pooled_embeddings_cpu.cpp x: 10 commits (all time) y: 33 lines of code fbgemm_gpu/src/merge_pooled_embeddings_gpu.cpp x: 16 commits (all time) y: 341 lines of code include/fbgemm/FbgemmFPCommon.h x: 8 commits (all time) y: 190 lines of code include/fbgemm/QuantUtils.h x: 29 commits (all time) y: 224 lines of code include/fbgemm/Utils.h x: 48 commits (all time) y: 214 lines of code src/ExecuteKernelU8S8.cc x: 38 commits (all time) y: 486 lines of code src/Fbgemm.cc x: 52 commits (all time) y: 510 lines of code src/FbgemmI8Depthwise2DAvx2-inl.h x: 18 commits (all time) y: 656 lines of code src/FbgemmI8Depthwise3DAvx2.cc x: 23 commits (all time) y: 989 lines of code src/QuantUtils.cc x: 31 commits (all time) y: 706 lines of code src/Utils.cc x: 47 commits (all time) y: 397 lines of code defs.bzl x: 26 commits (all time) y: 118 lines of code include/fbgemm/Fbgemm.h x: 69 commits (all time) y: 804 lines of code src/DirectConv.h x: 5 commits (all time) y: 148 lines of code src/GenerateKernelDirectConvU8S8S32ACC32.cc x: 6 commits (all time) y: 491 lines of code src/PackWeightsForConv.cc x: 16 commits (all time) y: 172 lines of code src/PackWeightsForDirectConv.cc x: 1 commits (all time) y: 412 lines of code fbgemm_gpu/setup.py x: 36 commits (all time) y: 144 lines of code fbgemm_gpu/version.py x: 1 commits (all time) y: 1 lines of code fbgemm_gpu/src/cumem_utils.cu x: 12 commits (all time) y: 305 lines of code src/CodeCache.h x: 12 commits (all time) y: 96 lines of code src/EmbeddingSpMDM.cc x: 36 commits (all time) y: 1251 lines of code src/EmbeddingSpMDMNBit.cc x: 27 commits (all time) y: 1170 lines of code fbgemm_gpu/include/fbgemm_gpu/embedding_backward_template_helpers.cuh x: 6 commits (all time) y: 98 lines of code fbgemm_gpu/include/fbgemm_gpu/split_embeddings_utils.cuh x: 5 commits (all time) y: 139 lines of code fbgemm_gpu/include/fbgemm_gpu/dispatch_macros.h x: 8 commits (all time) y: 141 lines of code include/fbgemm/QuantUtilsAvx2.h x: 22 commits (all time) y: 111 lines of code src/QuantUtilsAvx2.cc x: 41 commits (all time) y: 1761 lines of code bench/BenchUtils.cc x: 20 commits (all time) y: 182 lines of code bench/BenchUtils.h x: 21 commits (all time) y: 415 lines of code bench/EmbeddingIndexRemappingBenchmark.cc x: 4 commits (all time) y: 134 lines of code bench/EmbeddingSpMDM8BitBenchmark.cc x: 19 commits (all time) y: 282 lines of code bench/EmbeddingSpMDMBenchmark.cc x: 16 commits (all time) y: 293 lines of code bench/EmbeddingSpMDMNBitBenchmark.cc x: 10 commits (all time) y: 278 lines of code bench/EmbeddingSpMDMNBitRowWiseSparseBenchmark.cc x: 10 commits (all time) y: 309 lines of code bench/GEMMsTunableBenchmark.cc x: 9 commits (all time) y: 286 lines of code bench/GroupwiseConvRequantizeBenchmark.cc x: 10 commits (all time) y: 413 lines of code bench/Im2ColFusedRequantizeBenchmark.cc x: 7 commits (all time) y: 296 lines of code bench/PackedFloatInOutBenchmark.cc x: 17 commits (all time) y: 242 lines of code bench/PackedRequantizeAcc32Benchmark.cc x: 17 commits (all time) y: 262 lines of code bench/RowwiseAdagradBenchmark.cc x: 13 commits (all time) y: 189 lines of code bench/RowwiseAdagradFusedBenchmark.cc x: 7 commits (all time) y: 162 lines of code bench/SparseDenseMMFP32Benchmark.cc x: 5 commits (all time) y: 93 lines of code bench/SparseDenseMMInt8Benchmark.cc x: 5 commits (all time) y: 112 lines of code include/fbgemm/Types.h x: 9 commits (all time) y: 116 lines of code src/FbgemmSparseDenseVectorInt8Avx512.cc x: 6 commits (all time) y: 217 lines of code src/GenerateKernelU8S8S32ACC16.cc x: 29 commits (all time) y: 262 lines of code src/GenerateKernelU8S8S32ACC16Avx512.cc x: 34 commits (all time) y: 256 lines of code src/GenerateKernelU8S8S32ACC32.cc x: 37 commits (all time) y: 308 lines of code src/GenerateKernelU8S8S32ACC32Avx512VNNI.cc x: 25 commits (all time) y: 271 lines of code src/FbgemmI8DepthwiseAvx2.cc x: 23 commits (all time) y: 158 lines of code src/FbgemmFloat16Convert.cc x: 16 commits (all time) y: 74 lines of code include/fbgemm/FbgemmI8DepthwiseAvx2.h x: 17 commits (all time) y: 63 lines of code src/FbgemmFP16UKernelsAvx2.cc x: 19 commits (all time) y: 886 lines of code src/FbgemmFP16UKernelsAvx512.cc x: 16 commits (all time) y: 3130 lines of code src/FbgemmFP16UKernelsAvx512_256.cc x: 11 commits (all time) y: 2247 lines of code src/codegen_fp16fp32.cc x: 22 commits (all time) y: 461 lines of code bench/ConvUnifiedBenchmark.cc x: 17 commits (all time) y: 449 lines of code bench/ConvertBenchmark.cc x: 4 commits (all time) y: 51 lines of code bench/Depthwise3DBenchmark.cc x: 18 commits (all time) y: 188 lines of code bench/DepthwiseBenchmark.cc x: 20 commits (all time) y: 305 lines of code bench/FP16Benchmark.cc x: 29 commits (all time) y: 40 lines of code bench/I64Benchmark.cc x: 6 commits (all time) y: 115 lines of code bench/I8SpmdmBenchmark.cc x: 8 commits (all time) y: 172 lines of code bench/RequantizeBenchmark.cc x: 7 commits (all time) y: 123 lines of code bench/RowOffsetBenchmark.cc x: 6 commits (all time) y: 37 lines of code fbgemm_gpu/bench/histogram_binning_calibration_benchmark.py x: 6 commits (all time) y: 242 lines of code fbgemm_gpu/bench/scripts/batch_benchmark_run.py x: 4 commits (all time) y: 73 lines of code fbgemm_gpu/codegen/embedding_forward_split_cpu.h x: 13 commits (all time) y: 58 lines of code fbgemm_gpu/codegen/embedding_forward_template_helpers.cuh x: 13 commits (all time) y: 29 lines of code fbgemm_gpu/codegen/lookup_args.py x: 7 commits (all time) y: 39 lines of code fbgemm_gpu/fbgemm_gpu/__init__.py x: 4 commits (all time) y: 4 lines of code fbgemm_gpu/fbgemm_gpu/batched_unary_embeddings_ops.py x: 6 commits (all time) y: 61 lines of code fbgemm_gpu/fbgemm_gpu/split_embedding_inference_converter.py x: 16 commits (all time) y: 145 lines of code fbgemm_gpu/fbgemm_gpu/uvm.py x: 4 commits (all time) y: 19 lines of code fbgemm_gpu/include/fbgemm_gpu/bench_utils.cuh x: 6 commits (all time) y: 65 lines of code fbgemm_gpu/include/fbgemm_gpu/cuda_utils.cuh x: 3 commits (all time) y: 14 lines of code fbgemm_gpu/include/fbgemm_gpu/layout_transform_ops.cuh x: 5 commits (all time) y: 105 lines of code fbgemm_gpu/include/fbgemm_gpu/quantize_ops.cuh x: 18 commits (all time) y: 27 lines of code fbgemm_gpu/src/cpu_utils.cpp x: 4 commits (all time) y: 145 lines of code fbgemm_gpu/src/layout_transform_ops_gpu.cpp x: 7 commits (all time) y: 16 lines of code include/fbgemm/ConvUtils.h x: 12 commits (all time) y: 161 lines of code include/fbgemm/FbgemmBuild.h x: 10 commits (all time) y: 60 lines of code include/fbgemm/FbgemmConvert.h x: 9 commits (all time) y: 56 lines of code include/fbgemm/FbgemmFP16.h x: 32 commits (all time) y: 40 lines of code include/fbgemm/FbgemmI8Spmdm.h x: 7 commits (all time) y: 77 lines of code include/fbgemm/FbgemmPackMatrixB.h x: 9 commits (all time) y: 200 lines of code include/fbgemm/FbgemmSparse.h x: 3 commits (all time) y: 135 lines of code include/fbgemm/OutputProcessing-inl.h x: 14 commits (all time) y: 282 lines of code include/fbgemm/PackingTraits-inl.h x: 18 commits (all time) y: 298 lines of code include/fbgemm/QuantUtilsAvx512.h x: 5 commits (all time) y: 21 lines of code include/fbgemm/spmmUtils.h x: 3 commits (all time) y: 44 lines of code include/fbgemm/spmmUtilsAvx2.h x: 3 commits (all time) y: 28 lines of code src/EmbeddingSpMDMAvx2.cc x: 9 commits (all time) y: 132 lines of code src/EmbeddingSpMDMAvx512.cc x: 7 commits (all time) y: 521 lines of code src/ExecuteKernel.cc x: 6 commits (all time) y: 1 lines of code src/FbgemmBfloat16Convert.cc x: 12 commits (all time) y: 56 lines of code src/FbgemmBfloat16ConvertAvx2.cc x: 5 commits (all time) y: 42 lines of code src/FbgemmFP16.cc x: 34 commits (all time) y: 119 lines of code src/FbgemmFP16UKernelsAvx2.h x: 16 commits (all time) y: 14 lines of code src/FbgemmFP16UKernelsAvx512.h x: 14 commits (all time) y: 22 lines of code src/FbgemmFP16UKernelsAvx512_256.h x: 9 commits (all time) y: 16 lines of code src/FbgemmFP16UKernelsIntrinsicAvx2.cc x: 4 commits (all time) y: 86 lines of code src/FbgemmFP16UKernelsIntrinsicAvx512.cc x: 4 commits (all time) y: 111 lines of code src/FbgemmFP16UKernelsIntrinsicAvx512_256.cc x: 4 commits (all time) y: 92 lines of code src/FbgemmFPCommon.cc x: 6 commits (all time) y: 260 lines of code src/FbgemmI64.cc x: 11 commits (all time) y: 409 lines of code src/FbgemmI8DepthwiseAvx2-inl.h x: 10 commits (all time) y: 487 lines of code src/FbgemmI8DepthwisePerChannelQuantAvx2.cc x: 10 commits (all time) y: 91 lines of code src/FbgemmI8Spmdm.cc x: 14 commits (all time) y: 296 lines of code src/FbgemmSparseDense.cc x: 6 commits (all time) y: 253 lines of code src/FbgemmSparseDenseAvx512.cc x: 5 commits (all time) y: 130 lines of code src/FbgemmSparseDenseInt8Avx2.cc x: 5 commits (all time) y: 221 lines of code src/FbgemmSparseDenseInt8Avx512.cc x: 4 commits (all time) y: 506 lines of code src/GenerateI8Depthwise.cc x: 12 commits (all time) y: 483 lines of code src/GenerateI8Depthwise.h x: 6 commits (all time) y: 30 lines of code src/GenerateKernel.h x: 27 commits (all time) y: 99 lines of code src/GenerateKernelU8S8S32ACC16Avx512VNNI.cc x: 9 commits (all time) y: 27 lines of code src/GroupwiseConv.h x: 31 commits (all time) y: 256 lines of code src/GroupwiseConvAcc32Avx2.cc x: 52 commits (all time) y: 204 lines of code src/OptimizedKernelsAvx2.cc x: 8 commits (all time) y: 299 lines of code src/PackAMatrix.cc x: 22 commits (all time) y: 153 lines of code src/PackAWithIm2Col.cc x: 33 commits (all time) y: 645 lines of code src/PackAWithQuantRowOffset.cc x: 26 commits (all time) y: 217 lines of code src/PackAWithRowOffset.cc x: 22 commits (all time) y: 201 lines of code src/PackBMatrix.cc x: 31 commits (all time) y: 264 lines of code src/PackDepthwiseConvMatrixAvx2.cc x: 13 commits (all time) y: 113 lines of code src/PackWeightMatrixForGConv.cc x: 29 commits (all time) y: 189 lines of code src/QuantUtilsAvx512.cc x: 7 commits (all time) y: 356 lines of code src/RefImplementations.cc x: 65 commits (all time) y: 1755 lines of code src/RefImplementations.h x: 41 commits (all time) y: 259 lines of code src/RowWiseSparseAdagradFused.cc x: 15 commits (all time) y: 823 lines of code src/SparseAdagrad.cc x: 23 commits (all time) y: 842 lines of code src/TransposeUtilsAvx2.h x: 7 commits (all time) y: 343 lines of code src/UtilsAvx512.cc x: 13 commits (all time) y: 845 lines of code src/spmmUtils.cc x: 4 commits (all time) y: 292 lines of code src/spmmUtilsAvx2.cc x: 4 commits (all time) y: 271 lines of code bench/AlignedVec.h x: 3 commits (all time) y: 86 lines of code
3130.0
lines of code
  min: 1.0
  average: 303.43
  25th percentile: 54.5
  median: 162.0
  75th percentile: 313.0
  max: 3130.0
0 93.0
commits (all time)
min: 1.0 | average: 15.36 | 25th percentile: 6.0 | median: 11.0 | 75th percentile: 20.0 | max: 93.0

File Size vs. Contributors (all time): 233 points

fbgemm_gpu/codegen/embedding_forward_quantized_cpu_template.cpp x: 12 contributors (all time) y: 359 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host.cpp x: 7 contributors (all time) y: 180 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host_cpu.cpp x: 8 contributors (all time) y: 380 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_split_template.cu x: 9 contributors (all time) y: 917 lines of code fbgemm_gpu/fbgemm_gpu/split_table_batched_embeddings_ops.py x: 28 contributors (all time) y: 1963 lines of code fbgemm_gpu/src/sparse_ops.cu x: 14 contributors (all time) y: 1953 lines of code fbgemm_gpu/fbgemm_gpu/permute_pooled_embedding_modules.py x: 6 contributors (all time) y: 51 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embedding_ops.h x: 5 contributors (all time) y: 16 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embs_function.h x: 2 contributors (all time) y: 1 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops_gpu.cpp x: 7 contributors (all time) y: 132 lines of code fbgemm_gpu/src/jagged_tensor_ops.cu x: 4 contributors (all time) y: 1150 lines of code fbgemm_gpu/src/jagged_tensor_ops_cpu.cpp x: 2 contributors (all time) y: 967 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops.h x: 10 contributors (all time) y: 317 lines of code fbgemm_gpu/src/sparse_ops_cpu.cpp x: 12 contributors (all time) y: 1683 lines of code fbgemm_gpu/src/sparse_ops_gpu.cpp x: 10 contributors (all time) y: 157 lines of code fbgemm_gpu/src/split_embeddings_cache_cuda.cu x: 10 contributors (all time) y: 1803 lines of code fbgemm_gpu/include/fbgemm_gpu/embedding_common.h x: 4 contributors (all time) y: 54 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops_utils.h x: 5 contributors (all time) y: 214 lines of code fbgemm_gpu/src/split_table_batched_embeddings.cpp x: 5 contributors (all time) y: 101 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host.cpp x: 4 contributors (all time) y: 363 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host_cpu.cpp x: 7 contributors (all time) y: 164 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_cpu_template.cpp x: 8 contributors (all time) y: 200 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_template.cpp x: 6 contributors (all time) y: 467 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host.cpp x: 4 contributors (all time) y: 16 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host_cpu.cpp x: 6 contributors (all time) y: 98 lines of code fbgemm_gpu/src/cumem_utils_host.cpp x: 5 contributors (all time) y: 34 lines of code fbgemm_gpu/bench/split_table_batched_embeddings_benchmark.py x: 14 contributors (all time) y: 1916 lines of code fbgemm_gpu/pyproject.toml x: 2 contributors (all time) y: 18 lines of code fbgemm_gpu/bench/sparse_ops_benchmark.py x: 1 contributors (all time) y: 52 lines of code fbgemm_gpu/codegen/embedding_backward_code_generator.py x: 11 contributors (all time) y: 531 lines of code fbgemm_gpu/bench/batched_unary_embeddings_benchmark.py x: 1 contributors (all time) y: 137 lines of code fbgemm_gpu/bench/quantize_ops_benchmark.py x: 6 contributors (all time) y: 186 lines of code fbgemm_gpu/src/input_combine_cpu.cpp x: 8 contributors (all time) y: 278 lines of code fbgemm_gpu/include/fbgemm_gpu/quantize_ops_gpu.h x: 1 contributors (all time) y: 14 lines of code fbgemm_gpu/src/quantize_ops.cu x: 6 contributors (all time) y: 666 lines of code fbgemm_gpu/src/quantize_ops_cpu.cpp x: 7 contributors (all time) y: 303 lines of code bench/GEMMsBenchmark.cc x: 6 contributors (all time) y: 260 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops.cuh x: 8 contributors (all time) y: 35 lines of code fbgemm_gpu/include/fbgemm_gpu/fbgemm_cuda_utils.cuh x: 7 contributors (all time) y: 1948 lines of code fbgemm_gpu/bench/bench_utils.py x: 1 contributors (all time) y: 34 lines of code fbgemm_gpu/bench/merge_embeddings_benchmark.py x: 8 contributors (all time) y: 440 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_template.cpp x: 8 contributors (all time) y: 340 lines of code fbgemm_gpu/codegen/embedding_backward_split_indice_weights_template.cu x: 5 contributors (all time) y: 274 lines of code fbgemm_gpu/codegen/embedding_backward_split_template.cu x: 10 contributors (all time) y: 1085 lines of code fbgemm_gpu/codegen/embedding_bounds_check.cu x: 6 contributors (all time) y: 138 lines of code fbgemm_gpu/codegen/embedding_forward_split_cpu.cpp x: 9 contributors (all time) y: 526 lines of code fbgemm_gpu/codegen/embedding_forward_split_template.cu x: 6 contributors (all time) y: 450 lines of code fbgemm_gpu/src/histogram_binning_calibration_ops.cu x: 3 contributors (all time) y: 388 lines of code fbgemm_gpu/src/layout_transform_ops.cu x: 6 contributors (all time) y: 129 lines of code fbgemm_gpu/src/layout_transform_ops_cpu.cpp x: 6 contributors (all time) y: 63 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops.cu x: 4 contributors (all time) y: 74 lines of code fbgemm_gpu/src/split_embeddings_utils.cu x: 3 contributors (all time) y: 284 lines of code src/FbgemmConv.cc x: 7 contributors (all time) y: 434 lines of code fbgemm_gpu/include/fbgemm_gpu/cpu_utils.h x: 4 contributors (all time) y: 42 lines of code include/fbgemm/FbgemmEmbedding.h x: 7 contributors (all time) y: 234 lines of code src/GroupwiseConv.cc x: 5 contributors (all time) y: 938 lines of code fbgemm_gpu/bench/split_embeddings_cache_benchmark.py x: 1 contributors (all time) y: 448 lines of code fbgemm_gpu/src/merge_pooled_embeddings_gpu.cpp x: 7 contributors (all time) y: 341 lines of code include/fbgemm/FbgemmFPCommon.h x: 5 contributors (all time) y: 190 lines of code include/fbgemm/QuantUtils.h x: 9 contributors (all time) y: 224 lines of code include/fbgemm/Utils.h x: 11 contributors (all time) y: 214 lines of code src/ExecuteKernelU8S8.cc x: 9 contributors (all time) y: 486 lines of code src/Fbgemm.cc x: 7 contributors (all time) y: 510 lines of code src/FbgemmI8Depthwise2DAvx2-inl.h x: 4 contributors (all time) y: 656 lines of code src/FbgemmI8Depthwise3DAvx2.cc x: 4 contributors (all time) y: 989 lines of code src/QuantUtils.cc x: 11 contributors (all time) y: 706 lines of code src/Utils.cc x: 10 contributors (all time) y: 397 lines of code defs.bzl x: 8 contributors (all time) y: 118 lines of code include/fbgemm/Fbgemm.h x: 10 contributors (all time) y: 804 lines of code src/DirectConv.h x: 2 contributors (all time) y: 148 lines of code src/GenerateKernelDirectConvU8S8S32ACC32.cc x: 3 contributors (all time) y: 491 lines of code src/PackWeightsForConv.cc x: 6 contributors (all time) y: 172 lines of code src/PackWeightsForDirectConv.cc x: 1 contributors (all time) y: 412 lines of code fbgemm_gpu/setup.py x: 10 contributors (all time) y: 144 lines of code fbgemm_gpu/version.py x: 1 contributors (all time) y: 1 lines of code fbgemm_gpu/src/cumem_utils.cu x: 4 contributors (all time) y: 305 lines of code fbgemm_gpu/src/cumem_utils.h x: 3 contributors (all time) y: 18 lines of code src/CodeCache.h x: 5 contributors (all time) y: 96 lines of code src/EmbeddingSpMDM.cc x: 8 contributors (all time) y: 1251 lines of code src/EmbeddingSpMDMNBit.cc x: 5 contributors (all time) y: 1170 lines of code fbgemm_gpu/include/fbgemm_gpu/embedding_backward_template_helpers.cuh x: 3 contributors (all time) y: 98 lines of code fbgemm_gpu/include/fbgemm_gpu/split_embeddings_utils.cuh x: 3 contributors (all time) y: 139 lines of code src/QuantUtilsAvx2.cc x: 10 contributors (all time) y: 1761 lines of code bench/BenchUtils.cc x: 5 contributors (all time) y: 182 lines of code bench/BenchUtils.h x: 4 contributors (all time) y: 415 lines of code bench/EmbeddingIndexRemappingBenchmark.cc x: 3 contributors (all time) y: 134 lines of code bench/EmbeddingSpMDMBenchmark.cc x: 4 contributors (all time) y: 293 lines of code bench/EmbeddingSpMDMNBitBenchmark.cc x: 2 contributors (all time) y: 278 lines of code bench/EmbeddingSpMDMNBitRowWiseSparseBenchmark.cc x: 2 contributors (all time) y: 309 lines of code bench/GroupwiseConvRequantizeBenchmark.cc x: 3 contributors (all time) y: 413 lines of code bench/Im2ColFusedRequantizeBenchmark.cc x: 3 contributors (all time) y: 296 lines of code bench/PackedFloatInOutBenchmark.cc x: 4 contributors (all time) y: 242 lines of code bench/PackedRequantizeAcc32Benchmark.cc x: 4 contributors (all time) y: 262 lines of code bench/RowwiseAdagradFusedBenchmark.cc x: 2 contributors (all time) y: 162 lines of code bench/SparseDenseMMInt8Benchmark.cc x: 3 contributors (all time) y: 112 lines of code include/fbgemm/Types.h x: 4 contributors (all time) y: 116 lines of code src/CodeGenHelpers.h x: 5 contributors (all time) y: 167 lines of code src/FbgemmSparseDenseVectorInt8Avx512.cc x: 3 contributors (all time) y: 217 lines of code src/GenerateKernelU8S8S32ACC16.cc x: 10 contributors (all time) y: 262 lines of code src/GenerateKernelU8S8S32ACC16Avx512.cc x: 10 contributors (all time) y: 256 lines of code src/GenerateKernelU8S8S32ACC32.cc x: 12 contributors (all time) y: 308 lines of code src/GenerateKernelU8S8S32ACC32Avx512VNNI.cc x: 9 contributors (all time) y: 271 lines of code src/FbgemmI8DepthwiseAvx2.cc x: 5 contributors (all time) y: 158 lines of code src/FbgemmFloat16Convert.cc x: 8 contributors (all time) y: 74 lines of code include/fbgemm/FbgemmI8DepthwiseAvx2.h x: 3 contributors (all time) y: 63 lines of code src/FbgemmFP16UKernelsAvx2.cc x: 7 contributors (all time) y: 886 lines of code src/FbgemmFP16UKernelsAvx512.cc x: 6 contributors (all time) y: 3130 lines of code src/FbgemmFP16UKernelsAvx512_256.cc x: 6 contributors (all time) y: 2247 lines of code bench/ConvUnifiedBenchmark.cc x: 7 contributors (all time) y: 449 lines of code bench/Depthwise3DBenchmark.cc x: 4 contributors (all time) y: 188 lines of code bench/FP16Benchmark.cc x: 8 contributors (all time) y: 40 lines of code bench/RowOffsetBenchmark.cc x: 3 contributors (all time) y: 37 lines of code fbgemm_gpu/bench/histogram_binning_calibration_benchmark.py x: 2 contributors (all time) y: 242 lines of code fbgemm_gpu/bench/verify_fp16_stochastic_benchmark.cu x: 2 contributors (all time) y: 287 lines of code fbgemm_gpu/codegen/embedding_forward_split_cpu.h x: 7 contributors (all time) y: 58 lines of code fbgemm_gpu/codegen/embedding_forward_template_helpers.cuh x: 7 contributors (all time) y: 29 lines of code fbgemm_gpu/codegen/lookup_args.py x: 2 contributors (all time) y: 39 lines of code fbgemm_gpu/fbgemm_gpu/batched_unary_embeddings_ops.py x: 3 contributors (all time) y: 61 lines of code fbgemm_gpu/fbgemm_gpu/split_embedding_configs.py x: 8 contributors (all time) y: 88 lines of code fbgemm_gpu/fbgemm_gpu/split_embedding_inference_converter.py x: 9 contributors (all time) y: 145 lines of code fbgemm_gpu/include/fbgemm_gpu/bench_utils.cuh x: 1 contributors (all time) y: 65 lines of code fbgemm_gpu/include/fbgemm_gpu/enum_utils.h x: 2 contributors (all time) y: 57 lines of code fbgemm_gpu/include/fbgemm_gpu/layout_transform_ops.cuh x: 1 contributors (all time) y: 105 lines of code include/fbgemm/FbgemmBuild.h x: 5 contributors (all time) y: 60 lines of code include/fbgemm/FbgemmFP16.h x: 9 contributors (all time) y: 40 lines of code include/fbgemm/FbgemmI8Spmdm.h x: 3 contributors (all time) y: 77 lines of code include/fbgemm/FbgemmPackMatrixB.h x: 5 contributors (all time) y: 200 lines of code include/fbgemm/FbgemmSparse.h x: 2 contributors (all time) y: 135 lines of code include/fbgemm/OutputProcessing-inl.h x: 3 contributors (all time) y: 282 lines of code include/fbgemm/PackingTraits-inl.h x: 5 contributors (all time) y: 298 lines of code include/fbgemm/spmmUtilsAvx2.h x: 2 contributors (all time) y: 28 lines of code src/EmbeddingSpMDMAvx2.cc x: 4 contributors (all time) y: 132 lines of code src/EmbeddingSpMDMAvx512.cc x: 3 contributors (all time) y: 521 lines of code src/ExecuteKernel.cc x: 3 contributors (all time) y: 1 lines of code src/FbgemmFP16UKernelsIntrinsicAvx2.cc x: 2 contributors (all time) y: 86 lines of code src/FbgemmFP16UKernelsIntrinsicAvx512.cc x: 2 contributors (all time) y: 111 lines of code src/FbgemmFP16UKernelsIntrinsicAvx512_256.cc x: 2 contributors (all time) y: 92 lines of code src/FbgemmFPCommon.cc x: 3 contributors (all time) y: 260 lines of code src/FbgemmI64.cc x: 5 contributors (all time) y: 409 lines of code src/FbgemmSparseDense.cc x: 3 contributors (all time) y: 253 lines of code src/FbgemmSparseDenseInt8Avx512.cc x: 2 contributors (all time) y: 506 lines of code src/GenerateI8Depthwise.cc x: 4 contributors (all time) y: 483 lines of code src/GenerateKernel.h x: 10 contributors (all time) y: 99 lines of code src/GenerateKernelU8S8S32ACC16Avx512VNNI.cc x: 3 contributors (all time) y: 27 lines of code src/GroupwiseConv.h x: 6 contributors (all time) y: 256 lines of code src/GroupwiseConvAcc32Avx512.cc x: 2 contributors (all time) y: 206 lines of code src/PackAWithIm2Col.cc x: 7 contributors (all time) y: 645 lines of code src/PackAWithQuantRowOffset.cc x: 6 contributors (all time) y: 217 lines of code src/PackMatrix.cc x: 9 contributors (all time) y: 116 lines of code src/QuantUtilsAvx512.cc x: 3 contributors (all time) y: 356 lines of code src/RefImplementations.cc x: 12 contributors (all time) y: 1755 lines of code src/RowWiseSparseAdagradFused.cc x: 6 contributors (all time) y: 823 lines of code src/SparseAdagrad.cc x: 8 contributors (all time) y: 842 lines of code src/TransposeUtilsAvx2.h x: 4 contributors (all time) y: 343 lines of code src/UtilsAvx2.cc x: 5 contributors (all time) y: 269 lines of code src/UtilsAvx512.cc x: 5 contributors (all time) y: 845 lines of code src/spmmUtilsAvx2.cc x: 2 contributors (all time) y: 271 lines of code
3130.0
lines of code
  min: 1.0
  average: 303.43
  25th percentile: 54.5
  median: 162.0
  75th percentile: 313.0
  max: 3130.0
0 28.0
contributors (all time)
min: 1.0 | average: 4.97 | 25th percentile: 3.0 | median: 4.0 | 75th percentile: 7.0 | max: 28.0

File Size vs. Commits (30 days): 61 points

fbgemm_gpu/codegen/embedding_forward_quantized_cpu_template.cpp x: 3 commits (30d) y: 359 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host.cpp x: 2 commits (30d) y: 180 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host_cpu.cpp x: 2 commits (30d) y: 380 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_split_template.cu x: 3 commits (30d) y: 917 lines of code fbgemm_gpu/fbgemm_gpu/split_table_batched_embeddings_ops.py x: 7 commits (30d) y: 1963 lines of code fbgemm_gpu/src/sparse_ops.cu x: 8 commits (30d) y: 1953 lines of code fbgemm_gpu/fbgemm_gpu/permute_pooled_embedding_modules.py x: 2 commits (30d) y: 51 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embedding_ops.h x: 2 commits (30d) y: 16 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embs_function.h x: 3 commits (30d) y: 1 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops_gpu.cpp x: 2 commits (30d) y: 132 lines of code fbgemm_gpu/src/jagged_tensor_ops.cu x: 16 commits (30d) y: 1150 lines of code fbgemm_gpu/src/jagged_tensor_ops_cpu.cpp x: 14 commits (30d) y: 967 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops.h x: 5 commits (30d) y: 317 lines of code fbgemm_gpu/src/sparse_ops_cpu.cpp x: 8 commits (30d) y: 1683 lines of code fbgemm_gpu/src/sparse_ops_gpu.cpp x: 5 commits (30d) y: 157 lines of code fbgemm_gpu/src/split_embeddings_cache_cuda.cu x: 2 commits (30d) y: 1803 lines of code fbgemm_gpu/include/fbgemm_gpu/embedding_common.h x: 1 commits (30d) y: 54 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops_utils.h x: 3 commits (30d) y: 214 lines of code fbgemm_gpu/src/split_table_batched_embeddings.cpp x: 4 commits (30d) y: 101 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host.cpp x: 1 commits (30d) y: 363 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host_cpu.cpp x: 1 commits (30d) y: 164 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_cpu_template.cpp x: 1 commits (30d) y: 200 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_template.cpp x: 1 commits (30d) y: 467 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host.cpp x: 1 commits (30d) y: 16 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host_cpu.cpp x: 2 commits (30d) y: 98 lines of code fbgemm_gpu/src/cumem_utils_host.cpp x: 1 commits (30d) y: 34 lines of code fbgemm_gpu/bench/split_table_batched_embeddings_benchmark.py x: 1 commits (30d) y: 1916 lines of code fbgemm_gpu/codegen/embedding_backward_code_generator.py x: 1 commits (30d) y: 531 lines of code fbgemm_gpu/bench/quantize_ops_benchmark.py x: 3 commits (30d) y: 186 lines of code fbgemm_gpu/src/input_combine_cpu.cpp x: 2 commits (30d) y: 278 lines of code fbgemm_gpu/include/fbgemm_gpu/quantize_ops_gpu.h x: 1 commits (30d) y: 14 lines of code fbgemm_gpu/include/fbgemm_gpu/quantize_ops_utils.h x: 1 commits (30d) y: 55 lines of code fbgemm_gpu/src/quantize_ops.cu x: 2 commits (30d) y: 666 lines of code fbgemm_gpu/src/quantize_ops_cpu.cpp x: 1 commits (30d) y: 303 lines of code bench/GEMMsBenchmark.cc x: 1 commits (30d) y: 260 lines of code fbgemm_gpu/include/fbgemm_gpu/fbgemm_cuda_utils.cuh x: 1 commits (30d) y: 1948 lines of code fbgemm_gpu/bench/merge_embeddings_benchmark.py x: 2 commits (30d) y: 440 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_approx_template.cpp x: 1 commits (30d) y: 191 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_template.cpp x: 2 commits (30d) y: 340 lines of code fbgemm_gpu/codegen/embedding_backward_split_indice_weights_template.cu x: 1 commits (30d) y: 274 lines of code fbgemm_gpu/codegen/embedding_backward_split_template.cu x: 1 commits (30d) y: 1085 lines of code fbgemm_gpu/codegen/embedding_bounds_check.cu x: 1 commits (30d) y: 138 lines of code fbgemm_gpu/codegen/embedding_forward_split_cpu.cpp x: 2 commits (30d) y: 526 lines of code fbgemm_gpu/codegen/embedding_forward_split_template.cu x: 1 commits (30d) y: 450 lines of code fbgemm_gpu/src/histogram_binning_calibration_ops.cu x: 1 commits (30d) y: 388 lines of code fbgemm_gpu/src/layout_transform_ops.cu x: 1 commits (30d) y: 129 lines of code fbgemm_gpu/src/layout_transform_ops_cpu.cpp x: 1 commits (30d) y: 63 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops.cu x: 1 commits (30d) y: 74 lines of code fbgemm_gpu/src/split_embeddings_utils.cu x: 1 commits (30d) y: 284 lines of code src/FbgemmConv.cc x: 1 commits (30d) y: 434 lines of code fbgemm_gpu/include/fbgemm_gpu/cpu_utils.h x: 1 commits (30d) y: 42 lines of code include/fbgemm/FbgemmEmbedding.h x: 1 commits (30d) y: 234 lines of code
1963.0
lines of code
  min: 1.0
  average: 415.98
  25th percentile: 54.5
  median: 200.0
  75th percentile: 445.0
  max: 1963.0
0 16.0
commits (30d)
min: 1.0 | average: 2.38 | 25th percentile: 1.0 | median: 1.0 | 75th percentile: 2.0 | max: 16.0

File Size vs. Contributors (30 days): 61 points

fbgemm_gpu/codegen/embedding_forward_quantized_cpu_template.cpp x: 3 contributors (30d) y: 359 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host.cpp x: 2 contributors (30d) y: 180 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host_cpu.cpp x: 2 contributors (30d) y: 380 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_split_template.cu x: 3 contributors (30d) y: 917 lines of code fbgemm_gpu/fbgemm_gpu/split_table_batched_embeddings_ops.py x: 4 contributors (30d) y: 1963 lines of code fbgemm_gpu/src/sparse_ops.cu x: 4 contributors (30d) y: 1953 lines of code fbgemm_gpu/fbgemm_gpu/permute_pooled_embedding_modules.py x: 2 contributors (30d) y: 51 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embedding_ops.h x: 2 contributors (30d) y: 16 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embs_function.h x: 2 contributors (30d) y: 1 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops_gpu.cpp x: 2 contributors (30d) y: 132 lines of code fbgemm_gpu/src/jagged_tensor_ops.cu x: 2 contributors (30d) y: 1150 lines of code fbgemm_gpu/src/jagged_tensor_ops_cpu.cpp x: 2 contributors (30d) y: 967 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops.h x: 3 contributors (30d) y: 317 lines of code fbgemm_gpu/src/sparse_ops_cpu.cpp x: 4 contributors (30d) y: 1683 lines of code fbgemm_gpu/src/sparse_ops_gpu.cpp x: 3 contributors (30d) y: 157 lines of code fbgemm_gpu/src/split_embeddings_cache_cuda.cu x: 2 contributors (30d) y: 1803 lines of code fbgemm_gpu/include/fbgemm_gpu/embedding_common.h x: 1 contributors (30d) y: 54 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops_utils.h x: 2 contributors (30d) y: 214 lines of code fbgemm_gpu/src/split_table_batched_embeddings.cpp x: 2 contributors (30d) y: 101 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host.cpp x: 1 contributors (30d) y: 363 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host_cpu.cpp x: 1 contributors (30d) y: 164 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_cpu_template.cpp x: 1 contributors (30d) y: 200 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_template.cpp x: 1 contributors (30d) y: 467 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host.cpp x: 1 contributors (30d) y: 16 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host_cpu.cpp x: 2 contributors (30d) y: 98 lines of code fbgemm_gpu/src/cumem_utils_host.cpp x: 1 contributors (30d) y: 34 lines of code fbgemm_gpu/bench/split_table_batched_embeddings_benchmark.py x: 1 contributors (30d) y: 1916 lines of code fbgemm_gpu/codegen/embedding_backward_code_generator.py x: 1 contributors (30d) y: 531 lines of code fbgemm_gpu/bench/batched_unary_embeddings_benchmark.py x: 1 contributors (30d) y: 137 lines of code fbgemm_gpu/bench/quantize_ops_benchmark.py x: 3 contributors (30d) y: 186 lines of code fbgemm_gpu/src/input_combine_cpu.cpp x: 2 contributors (30d) y: 278 lines of code fbgemm_gpu/include/fbgemm_gpu/quantize_ops_gpu.h x: 1 contributors (30d) y: 14 lines of code fbgemm_gpu/include/fbgemm_gpu/quantize_ops_utils.h x: 1 contributors (30d) y: 55 lines of code fbgemm_gpu/src/quantize_ops.cu x: 2 contributors (30d) y: 666 lines of code fbgemm_gpu/src/quantize_ops_cpu.cpp x: 1 contributors (30d) y: 303 lines of code bench/GEMMsBenchmark.cc x: 1 contributors (30d) y: 260 lines of code fbgemm_gpu/include/fbgemm_gpu/fbgemm_cuda_utils.cuh x: 1 contributors (30d) y: 1948 lines of code fbgemm_gpu/bench/merge_embeddings_benchmark.py x: 2 contributors (30d) y: 440 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_approx_template.cpp x: 1 contributors (30d) y: 191 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_template.cpp x: 2 contributors (30d) y: 340 lines of code fbgemm_gpu/codegen/embedding_backward_split_indice_weights_template.cu x: 1 contributors (30d) y: 274 lines of code fbgemm_gpu/codegen/embedding_backward_split_template.cu x: 1 contributors (30d) y: 1085 lines of code fbgemm_gpu/codegen/embedding_forward_split_cpu.cpp x: 2 contributors (30d) y: 526 lines of code fbgemm_gpu/codegen/embedding_forward_split_template.cu x: 1 contributors (30d) y: 450 lines of code fbgemm_gpu/src/histogram_binning_calibration_ops.cu x: 1 contributors (30d) y: 388 lines of code fbgemm_gpu/src/layout_transform_ops.cu x: 1 contributors (30d) y: 129 lines of code fbgemm_gpu/src/layout_transform_ops_cpu.cpp x: 1 contributors (30d) y: 63 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops.cu x: 1 contributors (30d) y: 74 lines of code fbgemm_gpu/src/split_embeddings_utils.cu x: 1 contributors (30d) y: 284 lines of code src/FbgemmConv.cc x: 1 contributors (30d) y: 434 lines of code fbgemm_gpu/include/fbgemm_gpu/cpu_utils.h x: 1 contributors (30d) y: 42 lines of code include/fbgemm/FbgemmEmbedding.h x: 1 contributors (30d) y: 234 lines of code
1963.0
lines of code
  min: 1.0
  average: 415.98
  25th percentile: 54.5
  median: 200.0
  75th percentile: 445.0
  max: 1963.0
0 4.0
contributors (30d)
min: 1.0 | average: 1.61 | 25th percentile: 1.0 | median: 1.0 | 75th percentile: 2.0 | max: 4.0

File Size vs. Commits (90 days): 231 points

fbgemm_gpu/codegen/embedding_forward_quantized_cpu_template.cpp x: 10 commits (90d) y: 359 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host.cpp x: 9 commits (90d) y: 180 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host_cpu.cpp x: 9 commits (90d) y: 380 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_split_template.cu x: 8 commits (90d) y: 917 lines of code fbgemm_gpu/fbgemm_gpu/split_table_batched_embeddings_ops.py x: 14 commits (90d) y: 1963 lines of code fbgemm_gpu/src/sparse_ops.cu x: 16 commits (90d) y: 1953 lines of code fbgemm_gpu/fbgemm_gpu/permute_pooled_embedding_modules.py x: 4 commits (90d) y: 51 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embedding_ops.h x: 4 commits (90d) y: 16 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embs_function.h x: 3 commits (90d) y: 1 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops_gpu.cpp x: 6 commits (90d) y: 132 lines of code fbgemm_gpu/src/jagged_tensor_ops.cu x: 20 commits (90d) y: 1150 lines of code fbgemm_gpu/src/jagged_tensor_ops_cpu.cpp x: 14 commits (90d) y: 967 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops.h x: 13 commits (90d) y: 317 lines of code fbgemm_gpu/src/sparse_ops_cpu.cpp x: 21 commits (90d) y: 1683 lines of code fbgemm_gpu/src/sparse_ops_gpu.cpp x: 13 commits (90d) y: 157 lines of code fbgemm_gpu/src/split_embeddings_cache_cuda.cu x: 12 commits (90d) y: 1803 lines of code fbgemm_gpu/include/fbgemm_gpu/embedding_common.h x: 5 commits (90d) y: 54 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops_utils.h x: 6 commits (90d) y: 214 lines of code fbgemm_gpu/src/split_table_batched_embeddings.cpp x: 10 commits (90d) y: 101 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host.cpp x: 9 commits (90d) y: 363 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host_cpu.cpp x: 9 commits (90d) y: 164 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_cpu_template.cpp x: 10 commits (90d) y: 200 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_template.cpp x: 8 commits (90d) y: 467 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host.cpp x: 7 commits (90d) y: 16 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host_cpu.cpp x: 9 commits (90d) y: 98 lines of code fbgemm_gpu/src/cumem_utils_host.cpp x: 6 commits (90d) y: 34 lines of code fbgemm_gpu/bench/split_table_batched_embeddings_benchmark.py x: 5 commits (90d) y: 1916 lines of code fbgemm_gpu/pyproject.toml x: 1 commits (90d) y: 18 lines of code fbgemm_gpu/bench/sparse_ops_benchmark.py x: 1 commits (90d) y: 52 lines of code fbgemm_gpu/codegen/embedding_backward_code_generator.py x: 6 commits (90d) y: 531 lines of code fbgemm_gpu/bench/batched_unary_embeddings_benchmark.py x: 2 commits (90d) y: 137 lines of code fbgemm_gpu/bench/quantize_ops_benchmark.py x: 6 commits (90d) y: 186 lines of code fbgemm_gpu/include/fbgemm_gpu/input_combine.h x: 3 commits (90d) y: 16 lines of code fbgemm_gpu/src/input_combine_cpu.cpp x: 5 commits (90d) y: 278 lines of code fbgemm_gpu/src/quantize_ops.cu x: 8 commits (90d) y: 666 lines of code fbgemm_gpu/src/quantize_ops_cpu.cpp x: 6 commits (90d) y: 303 lines of code fbgemm_gpu/src/quantize_ops_gpu.cpp x: 5 commits (90d) y: 35 lines of code bench/GEMMsBenchmark.cc x: 4 commits (90d) y: 260 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops.cuh x: 7 commits (90d) y: 35 lines of code fbgemm_gpu/include/fbgemm_gpu/fbgemm_cuda_utils.cuh x: 8 commits (90d) y: 1948 lines of code fbgemm_gpu/bench/bench_utils.py x: 1 commits (90d) y: 34 lines of code fbgemm_gpu/bench/merge_embeddings_benchmark.py x: 9 commits (90d) y: 440 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_approx_template.cpp x: 7 commits (90d) y: 191 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_template.cpp x: 7 commits (90d) y: 340 lines of code fbgemm_gpu/codegen/embedding_backward_split_template.cu x: 7 commits (90d) y: 1085 lines of code fbgemm_gpu/codegen/embedding_bounds_check.cu x: 5 commits (90d) y: 138 lines of code fbgemm_gpu/codegen/embedding_forward_split_cpu.cpp x: 8 commits (90d) y: 526 lines of code fbgemm_gpu/codegen/embedding_forward_split_template.cu x: 5 commits (90d) y: 450 lines of code fbgemm_gpu/src/histogram_binning_calibration_ops.cu x: 6 commits (90d) y: 388 lines of code fbgemm_gpu/src/layout_transform_ops.cu x: 4 commits (90d) y: 129 lines of code fbgemm_gpu/src/layout_transform_ops_cpu.cpp x: 4 commits (90d) y: 63 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops.cu x: 3 commits (90d) y: 74 lines of code fbgemm_gpu/src/split_embeddings_utils.cu x: 6 commits (90d) y: 284 lines of code src/FbgemmConv.cc x: 5 commits (90d) y: 434 lines of code fbgemm_gpu/include/fbgemm_gpu/cpu_utils.h x: 4 commits (90d) y: 42 lines of code include/fbgemm/FbgemmEmbedding.h x: 5 commits (90d) y: 234 lines of code src/GroupwiseConv.cc x: 5 commits (90d) y: 938 lines of code fbgemm_gpu/src/merge_pooled_embeddings_cpu.cpp x: 4 commits (90d) y: 33 lines of code fbgemm_gpu/src/merge_pooled_embeddings_gpu.cpp x: 6 commits (90d) y: 341 lines of code include/fbgemm/FbgemmFPCommon.h x: 4 commits (90d) y: 190 lines of code include/fbgemm/QuantUtils.h x: 1 commits (90d) y: 224 lines of code include/fbgemm/Utils.h x: 4 commits (90d) y: 214 lines of code src/ExecuteKernelU8S8.cc x: 3 commits (90d) y: 486 lines of code src/Fbgemm.cc x: 3 commits (90d) y: 510 lines of code src/FbgemmI8Depthwise2DAvx2-inl.h x: 4 commits (90d) y: 656 lines of code src/FbgemmI8Depthwise3DAvx2.cc x: 5 commits (90d) y: 989 lines of code src/QuantUtils.cc x: 1 commits (90d) y: 706 lines of code src/Utils.cc x: 4 commits (90d) y: 397 lines of code defs.bzl x: 3 commits (90d) y: 118 lines of code include/fbgemm/Fbgemm.h x: 5 commits (90d) y: 804 lines of code src/DirectConv.h x: 4 commits (90d) y: 148 lines of code src/GenerateKernelDirectConvU8S8S32ACC32.cc x: 5 commits (90d) y: 491 lines of code src/PackWeightsForConv.cc x: 3 commits (90d) y: 172 lines of code src/PackWeightsForDirectConv.cc x: 1 commits (90d) y: 412 lines of code fbgemm_gpu/setup.py x: 6 commits (90d) y: 144 lines of code fbgemm_gpu/version.py x: 1 commits (90d) y: 1 lines of code fbgemm_gpu/src/cumem_utils.cu x: 4 commits (90d) y: 305 lines of code src/CodeCache.h x: 3 commits (90d) y: 96 lines of code src/EmbeddingSpMDM.cc x: 5 commits (90d) y: 1251 lines of code src/EmbeddingSpMDMNBit.cc x: 5 commits (90d) y: 1170 lines of code fbgemm_gpu/include/fbgemm_gpu/embedding_backward_template_helpers.cuh x: 6 commits (90d) y: 98 lines of code fbgemm_gpu/include/fbgemm_gpu/dispatch_macros.h x: 4 commits (90d) y: 141 lines of code include/fbgemm/QuantUtilsAvx2.h x: 1 commits (90d) y: 111 lines of code src/QuantUtilsAvx2.cc x: 3 commits (90d) y: 1761 lines of code bench/BenchUtils.cc x: 3 commits (90d) y: 182 lines of code bench/BenchUtils.h x: 3 commits (90d) y: 415 lines of code bench/EmbeddingIndexRemappingBenchmark.cc x: 3 commits (90d) y: 134 lines of code bench/EmbeddingSpMDM8BitBenchmark.cc x: 3 commits (90d) y: 282 lines of code bench/EmbeddingSpMDMBenchmark.cc x: 3 commits (90d) y: 293 lines of code bench/EmbeddingSpMDMNBitRowWiseSparseBenchmark.cc x: 3 commits (90d) y: 309 lines of code bench/Im2ColFusedRequantizeBenchmark.cc x: 3 commits (90d) y: 296 lines of code bench/PackedFloatInOutBenchmark.cc x: 3 commits (90d) y: 242 lines of code bench/PackedRequantizeAcc16Benchmark.cc x: 3 commits (90d) y: 359 lines of code bench/PackedRequantizeAcc32Benchmark.cc x: 3 commits (90d) y: 262 lines of code bench/RowwiseAdagradBenchmark.cc x: 3 commits (90d) y: 189 lines of code src/FbgemmSparseDenseVectorInt8Avx512.cc x: 3 commits (90d) y: 217 lines of code src/GenerateKernelU8S8S32ACC16Avx512.cc x: 3 commits (90d) y: 256 lines of code include/fbgemm/FbgemmI8DepthwiseAvx2.h x: 3 commits (90d) y: 63 lines of code src/FbgemmFP16UKernelsAvx2.cc x: 3 commits (90d) y: 886 lines of code src/FbgemmFP16UKernelsAvx512.cc x: 3 commits (90d) y: 3130 lines of code src/FbgemmFP16UKernelsAvx512_256.cc x: 3 commits (90d) y: 2247 lines of code src/codegen_fp16fp32.cc x: 3 commits (90d) y: 461 lines of code bench/ConvUnifiedBenchmark.cc x: 2 commits (90d) y: 449 lines of code bench/ConvertBenchmark.cc x: 2 commits (90d) y: 51 lines of code bench/Depthwise3DBenchmark.cc x: 2 commits (90d) y: 188 lines of code bench/DepthwiseBenchmark.cc x: 2 commits (90d) y: 305 lines of code bench/EmbeddingQuantizeBenchmark.cc x: 2 commits (90d) y: 89 lines of code bench/FP16Benchmark.cc x: 2 commits (90d) y: 40 lines of code bench/I64Benchmark.cc x: 2 commits (90d) y: 115 lines of code bench/I8SpmdmBenchmark.cc x: 2 commits (90d) y: 172 lines of code fbgemm_gpu/bench/histogram_binning_calibration_benchmark.py x: 2 commits (90d) y: 242 lines of code fbgemm_gpu/bench/scripts/batch_benchmark_run.py x: 2 commits (90d) y: 73 lines of code fbgemm_gpu/bench/verify_fp16_stochastic_benchmark.cu x: 2 commits (90d) y: 287 lines of code fbgemm_gpu/codegen/embedding_forward_split_cpu.h x: 3 commits (90d) y: 58 lines of code fbgemm_gpu/fbgemm_gpu/__init__.py x: 2 commits (90d) y: 4 lines of code fbgemm_gpu/fbgemm_gpu/uvm.py x: 2 commits (90d) y: 19 lines of code fbgemm_gpu/include/fbgemm_gpu/layout_transform_ops.cuh x: 2 commits (90d) y: 105 lines of code fbgemm_gpu/include/fbgemm_gpu/quantize_ops.cuh x: 2 commits (90d) y: 27 lines of code include/fbgemm/FbgemmI8Spmdm.h x: 2 commits (90d) y: 77 lines of code include/fbgemm/FbgemmPackMatrixB.h x: 2 commits (90d) y: 200 lines of code include/fbgemm/FbgemmSparse.h x: 2 commits (90d) y: 135 lines of code include/fbgemm/OutputProcessing-inl.h x: 2 commits (90d) y: 282 lines of code src/EmbeddingSpMDMAvx512.cc x: 2 commits (90d) y: 521 lines of code src/FbgemmFPCommon.cc x: 2 commits (90d) y: 260 lines of code src/FbgemmI64.cc x: 2 commits (90d) y: 409 lines of code src/FbgemmI8DepthwiseAvx2-inl.h x: 2 commits (90d) y: 487 lines of code src/FbgemmSparseDense.cc x: 2 commits (90d) y: 253 lines of code src/FbgemmSparseDenseInt8Avx2.cc x: 2 commits (90d) y: 221 lines of code src/FbgemmSparseDenseInt8Avx512.cc x: 2 commits (90d) y: 506 lines of code src/PackAMatrix.cc x: 2 commits (90d) y: 153 lines of code src/PackAWithIm2Col.cc x: 2 commits (90d) y: 645 lines of code src/QuantUtilsAvx512.cc x: 2 commits (90d) y: 356 lines of code src/RowWiseSparseAdagradFused.cc x: 2 commits (90d) y: 823 lines of code src/SparseAdagrad.cc x: 2 commits (90d) y: 842 lines of code src/TransposeUtilsAvx2.h x: 2 commits (90d) y: 343 lines of code
3130.0
lines of code
  min: 1.0
  average: 305.65
  25th percentile: 55.0
  median: 164.0
  75th percentile: 317.0
  max: 3130.0
0 21.0
commits (90d)
min: 1.0 | average: 3.69 | 25th percentile: 2.0 | median: 3.0 | 75th percentile: 4.0 | max: 21.0

File Size vs. Contributors (90 days): 231 points

fbgemm_gpu/codegen/embedding_forward_quantized_cpu_template.cpp x: 3 contributors (90d) y: 359 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host.cpp x: 3 contributors (90d) y: 180 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_host_cpu.cpp x: 4 contributors (90d) y: 380 lines of code fbgemm_gpu/codegen/embedding_forward_quantized_split_template.cu x: 3 contributors (90d) y: 917 lines of code fbgemm_gpu/fbgemm_gpu/split_table_batched_embeddings_ops.py x: 5 contributors (90d) y: 1963 lines of code fbgemm_gpu/src/sparse_ops.cu x: 6 contributors (90d) y: 1953 lines of code fbgemm_gpu/fbgemm_gpu/permute_pooled_embedding_modules.py x: 3 contributors (90d) y: 51 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embedding_ops.h x: 3 contributors (90d) y: 16 lines of code fbgemm_gpu/include/fbgemm_gpu/permute_pooled_embs_function.h x: 2 contributors (90d) y: 1 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops_gpu.cpp x: 4 contributors (90d) y: 132 lines of code fbgemm_gpu/src/jagged_tensor_ops.cu x: 4 contributors (90d) y: 1150 lines of code fbgemm_gpu/src/jagged_tensor_ops_cpu.cpp x: 2 contributors (90d) y: 967 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops.h x: 8 contributors (90d) y: 317 lines of code fbgemm_gpu/src/sparse_ops_cpu.cpp x: 7 contributors (90d) y: 1683 lines of code fbgemm_gpu/src/sparse_ops_gpu.cpp x: 6 contributors (90d) y: 157 lines of code fbgemm_gpu/src/split_embeddings_cache_cuda.cu x: 6 contributors (90d) y: 1803 lines of code fbgemm_gpu/include/fbgemm_gpu/embedding_common.h x: 4 contributors (90d) y: 54 lines of code fbgemm_gpu/include/fbgemm_gpu/sparse_ops_utils.h x: 3 contributors (90d) y: 214 lines of code fbgemm_gpu/src/split_table_batched_embeddings.cpp x: 3 contributors (90d) y: 101 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host.cpp x: 2 contributors (90d) y: 363 lines of code fbgemm_gpu/codegen/embedding_backward_dense_host_cpu.cpp x: 4 contributors (90d) y: 164 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_cpu_template.cpp x: 5 contributors (90d) y: 200 lines of code fbgemm_gpu/codegen/embedding_backward_split_host_template.cpp x: 2 contributors (90d) y: 467 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host.cpp x: 2 contributors (90d) y: 16 lines of code fbgemm_gpu/codegen/embedding_bounds_check_host_cpu.cpp x: 3 contributors (90d) y: 98 lines of code fbgemm_gpu/src/cumem_utils_host.cpp x: 3 contributors (90d) y: 34 lines of code fbgemm_gpu/bench/split_table_batched_embeddings_benchmark.py x: 2 contributors (90d) y: 1916 lines of code fbgemm_gpu/pyproject.toml x: 1 contributors (90d) y: 18 lines of code fbgemm_gpu/bench/sparse_ops_benchmark.py x: 1 contributors (90d) y: 52 lines of code fbgemm_gpu/codegen/embedding_backward_code_generator.py x: 3 contributors (90d) y: 531 lines of code fbgemm_gpu/bench/batched_unary_embeddings_benchmark.py x: 1 contributors (90d) y: 137 lines of code fbgemm_gpu/bench/quantize_ops_benchmark.py x: 4 contributors (90d) y: 186 lines of code fbgemm_gpu/src/input_combine_cpu.cpp x: 4 contributors (90d) y: 278 lines of code fbgemm_gpu/src/quantize_ops.cu x: 6 contributors (90d) y: 666 lines of code fbgemm_gpu/src/quantize_ops_cpu.cpp x: 5 contributors (90d) y: 303 lines of code fbgemm_gpu/src/quantize_ops_gpu.cpp x: 4 contributors (90d) y: 35 lines of code bench/GEMMsBenchmark.cc x: 3 contributors (90d) y: 260 lines of code fbgemm_gpu/include/fbgemm_gpu/fbgemm_cuda_utils.cuh x: 3 contributors (90d) y: 1948 lines of code fbgemm_gpu/bench/bench_utils.py x: 1 contributors (90d) y: 34 lines of code fbgemm_gpu/bench/merge_embeddings_benchmark.py x: 4 contributors (90d) y: 440 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_approx_template.cpp x: 3 contributors (90d) y: 191 lines of code fbgemm_gpu/codegen/embedding_backward_split_cpu_template.cpp x: 4 contributors (90d) y: 340 lines of code fbgemm_gpu/codegen/embedding_backward_split_indice_weights_template.cu x: 3 contributors (90d) y: 274 lines of code fbgemm_gpu/codegen/embedding_backward_split_template.cu x: 3 contributors (90d) y: 1085 lines of code fbgemm_gpu/codegen/embedding_bounds_check.cu x: 2 contributors (90d) y: 138 lines of code fbgemm_gpu/codegen/embedding_forward_split_cpu.cpp x: 4 contributors (90d) y: 526 lines of code fbgemm_gpu/codegen/embedding_forward_split_template.cu x: 3 contributors (90d) y: 450 lines of code fbgemm_gpu/src/histogram_binning_calibration_ops.cu x: 3 contributors (90d) y: 388 lines of code fbgemm_gpu/src/layout_transform_ops.cu x: 3 contributors (90d) y: 129 lines of code fbgemm_gpu/src/layout_transform_ops_cpu.cpp x: 3 contributors (90d) y: 63 lines of code fbgemm_gpu/src/permute_pooled_embedding_ops.cu x: 2 contributors (90d) y: 74 lines of code fbgemm_gpu/src/split_embeddings_utils.cu x: 3 contributors (90d) y: 284 lines of code src/FbgemmConv.cc x: 2 contributors (90d) y: 434 lines of code fbgemm_gpu/include/fbgemm_gpu/cpu_utils.h x: 2 contributors (90d) y: 42 lines of code include/fbgemm/FbgemmEmbedding.h x: 2 contributors (90d) y: 234 lines of code src/GroupwiseConv.cc x: 2 contributors (90d) y: 938 lines of code fbgemm_gpu/bench/split_embeddings_cache_benchmark.py x: 1 contributors (90d) y: 448 lines of code include/fbgemm/FbgemmFPCommon.h x: 2 contributors (90d) y: 190 lines of code include/fbgemm/QuantUtils.h x: 1 contributors (90d) y: 224 lines of code src/ExecuteKernelU8S8.cc x: 2 contributors (90d) y: 486 lines of code src/Fbgemm.cc x: 2 contributors (90d) y: 510 lines of code src/FbgemmI8Depthwise2DAvx2-inl.h x: 2 contributors (90d) y: 656 lines of code src/FbgemmI8Depthwise3DAvx2.cc x: 2 contributors (90d) y: 989 lines of code src/QuantUtils.cc x: 1 contributors (90d) y: 706 lines of code src/Utils.cc x: 2 contributors (90d) y: 397 lines of code defs.bzl x: 2 contributors (90d) y: 118 lines of code include/fbgemm/Fbgemm.h x: 3 contributors (90d) y: 804 lines of code src/DirectConv.h x: 2 contributors (90d) y: 148 lines of code src/GenerateKernelDirectConvU8S8S32ACC32.cc x: 3 contributors (90d) y: 491 lines of code src/PackWeightsForConv.cc x: 2 contributors (90d) y: 172 lines of code src/PackWeightsForDirectConv.cc x: 1 contributors (90d) y: 412 lines of code fbgemm_gpu/setup.py x: 3 contributors (90d) y: 144 lines of code fbgemm_gpu/version.py x: 1 contributors (90d) y: 1 lines of code fbgemm_gpu/src/cumem_utils.cu x: 2 contributors (90d) y: 305 lines of code src/CodeCache.h x: 2 contributors (90d) y: 96 lines of code src/EmbeddingSpMDM.cc x: 2 contributors (90d) y: 1251 lines of code src/EmbeddingSpMDMNBit.cc x: 2 contributors (90d) y: 1170 lines of code include/fbgemm/QuantUtilsAvx2.h x: 1 contributors (90d) y: 111 lines of code src/QuantUtilsAvx2.cc x: 2 contributors (90d) y: 1761 lines of code bench/BenchUtils.cc x: 2 contributors (90d) y: 182 lines of code bench/BenchUtils.h x: 2 contributors (90d) y: 415 lines of code bench/EmbeddingIndexRemappingBenchmark.cc x: 2 contributors (90d) y: 134 lines of code bench/EmbeddingSpMDM8BitBenchmark.cc x: 2 contributors (90d) y: 282 lines of code bench/EmbeddingSpMDMBenchmark.cc x: 2 contributors (90d) y: 293 lines of code bench/EmbeddingSpMDMNBitRowWiseSparseBenchmark.cc x: 2 contributors (90d) y: 309 lines of code bench/PackedFloatInOutBenchmark.cc x: 2 contributors (90d) y: 242 lines of code bench/PackedRequantizeAcc32Benchmark.cc x: 2 contributors (90d) y: 262 lines of code src/FbgemmSparseDenseVectorInt8Avx512.cc x: 2 contributors (90d) y: 217 lines of code src/GenerateKernelU8S8S32ACC16Avx512.cc x: 2 contributors (90d) y: 256 lines of code include/fbgemm/FbgemmI8DepthwiseAvx2.h x: 2 contributors (90d) y: 63 lines of code src/FbgemmFP16UKernelsAvx2.cc x: 2 contributors (90d) y: 886 lines of code src/FbgemmFP16UKernelsAvx512.cc x: 2 contributors (90d) y: 3130 lines of code src/FbgemmFP16UKernelsAvx512_256.cc x: 2 contributors (90d) y: 2247 lines of code bench/Depthwise3DBenchmark.cc x: 1 contributors (90d) y: 188 lines of code bench/DepthwiseBenchmark.cc x: 1 contributors (90d) y: 305 lines of code bench/EmbeddingQuantizeBenchmark.cc x: 1 contributors (90d) y: 89 lines of code bench/FP16Benchmark.cc x: 1 contributors (90d) y: 40 lines of code bench/I8SpmdmBenchmark.cc x: 1 contributors (90d) y: 172 lines of code fbgemm_gpu/bench/histogram_binning_calibration_benchmark.py x: 1 contributors (90d) y: 242 lines of code fbgemm_gpu/bench/scripts/batch_benchmark_run.py x: 1 contributors (90d) y: 73 lines of code fbgemm_gpu/bench/verify_fp16_stochastic_benchmark.cu x: 1 contributors (90d) y: 287 lines of code fbgemm_gpu/codegen/embedding_forward_split_cpu.h x: 2 contributors (90d) y: 58 lines of code fbgemm_gpu/include/fbgemm_gpu/layout_transform_ops.cuh x: 1 contributors (90d) y: 105 lines of code include/fbgemm/FbgemmI8Spmdm.h x: 1 contributors (90d) y: 77 lines of code include/fbgemm/FbgemmPackMatrixB.h x: 1 contributors (90d) y: 200 lines of code include/fbgemm/FbgemmSparse.h x: 1 contributors (90d) y: 135 lines of code include/fbgemm/OutputProcessing-inl.h x: 1 contributors (90d) y: 282 lines of code src/EmbeddingSpMDMAvx512.cc x: 1 contributors (90d) y: 521 lines of code src/FbgemmFPCommon.cc x: 1 contributors (90d) y: 260 lines of code src/FbgemmI8DepthwiseAvx2-inl.h x: 1 contributors (90d) y: 487 lines of code src/FbgemmSparseDense.cc x: 1 contributors (90d) y: 253 lines of code src/FbgemmSparseDenseInt8Avx2.cc x: 1 contributors (90d) y: 221 lines of code src/FbgemmSparseDenseInt8Avx512.cc x: 1 contributors (90d) y: 506 lines of code src/PackAMatrix.cc x: 1 contributors (90d) y: 153 lines of code src/PackAWithIm2Col.cc x: 1 contributors (90d) y: 645 lines of code src/QuantUtilsAvx512.cc x: 1 contributors (90d) y: 356 lines of code src/RowWiseSparseAdagradFused.cc x: 1 contributors (90d) y: 823 lines of code src/SparseAdagrad.cc x: 1 contributors (90d) y: 842 lines of code src/TransposeUtilsAvx2.h x: 1 contributors (90d) y: 343 lines of code
3130.0
lines of code
  min: 1.0
  average: 305.65
  25th percentile: 55.0
  median: 164.0
  75th percentile: 317.0
  max: 3130.0
0 8.0
contributors (90d)
min: 1.0 | average: 1.9 | 25th percentile: 1.0 | median: 2.0 | 75th percentile: 2.0 | max: 8.0