deepseek-ai / FlashMLA
File Size

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

Intro
Learn more...
File Size Overall
51% | 0% | 18% | 14% | 16%
Legend:
1001+
501-1000
201-500
101-200
1-100


explore: grouped by folders | grouped by size | sunburst | 3D view
File Size per Extension
1001+
501-1000
201-500
101-200
1-100
cu82% | 0% | 0% | 12% | 4%
py0% | 0% | 75% | 0% | 24%
cpp0% | 0% | 0% | 100% | 0%
h0% | 0% | 0% | 0% | 100%
File Size per Logical Decomposition
primary
1001+
501-1000
201-500
101-200
1-100
csrc67% | 0% | 0% | 19% | 13%
benchmark0% | 0% | 95% | 0% | 4%
ROOT0% | 0% | 0% | 0% | 100%
flash_mla0% | 0% | 0% | 0% | 100%
Longest Files (Top 16)
File# lines# units
splitkv_mla.cu
in csrc/kernels
 1230 -
bench_flash_mla.py
in benchmark
 434 13
mla_combine.cu
in csrc/kernels
 179 -
176 -
traits.h
in csrc/kernels
 87 -
setup.py
in root
 79 2
get_mla_metadata.cu
in csrc/kernels
 73 -
params.h
in csrc/kernels
 46 -
34 2
utils.h
in csrc/kernels
 25 -
visualize.py
in benchmark
 20 1
config.h
in csrc/kernels
 8 -
__init__.py
in flash_mla
 5 -
mla_combine.h
in csrc/kernels
 4 -
splitkv_mla.h
in csrc/kernels
 4 -
get_mla_metadata.h
in csrc/kernels
 3 -
Files With Most Units (Top 4)
File# lines# units
bench_flash_mla.py
in benchmark
 434 13
setup.py
in root
 79 2
34 2
visualize.py
in benchmark
 20 1
Files With Long Lines (Top 6)

There are 6 files with lines longer than 120 characters. In total, there are 85 long lines.

File# lines# units# long lines
splitkv_mla.cu
in csrc/kernels
 1230 - 51
bench_flash_mla.py
in benchmark
 434 13 22
traits.h
in csrc/kernels
 87 - 4
mla_combine.cu
in csrc/kernels
 179 - 3
176 - 3
get_mla_metadata.cu
in csrc/kernels
 73 - 2
Correlations

File Size vs. Commits (all time): 15 points

setup.py x: 8 commits (all time) y: 79 lines of code csrc/kernels/splitkv_mla.cu x: 2 commits (all time) y: 1230 lines of code csrc/kernels/traits.h x: 2 commits (all time) y: 87 lines of code benchmark/bench_flash_mla.py x: 5 commits (all time) y: 434 lines of code csrc/flash_api.cpp x: 6 commits (all time) y: 176 lines of code csrc/kernels/config.h x: 1 commits (all time) y: 8 lines of code csrc/kernels/get_mla_metadata.cu x: 1 commits (all time) y: 73 lines of code csrc/kernels/get_mla_metadata.h x: 1 commits (all time) y: 3 lines of code csrc/kernels/mla_combine.cu x: 1 commits (all time) y: 179 lines of code csrc/kernels/params.h x: 1 commits (all time) y: 46 lines of code csrc/kernels/utils.h x: 1 commits (all time) y: 25 lines of code flash_mla/flash_mla_interface.py x: 3 commits (all time) y: 34 lines of code benchmark/visualize.py x: 4 commits (all time) y: 20 lines of code
1230.0
lines of code
  min: 3.0
  average: 160.13
  25th percentile: 8.0
  median: 46.0
  75th percentile: 176.0
  max: 1230.0
0 8.0
commits (all time)
min: 1.0 | average: 2.53 | 25th percentile: 1.0 | median: 1.0 | 75th percentile: 4.0 | max: 8.0

File Size vs. Contributors (all time): 15 points

setup.py x: 5 contributors (all time) y: 79 lines of code csrc/kernels/splitkv_mla.cu x: 2 contributors (all time) y: 1230 lines of code csrc/kernels/traits.h x: 2 contributors (all time) y: 87 lines of code benchmark/bench_flash_mla.py x: 4 contributors (all time) y: 434 lines of code csrc/flash_api.cpp x: 4 contributors (all time) y: 176 lines of code csrc/kernels/config.h x: 1 contributors (all time) y: 8 lines of code csrc/kernels/get_mla_metadata.cu x: 1 contributors (all time) y: 73 lines of code csrc/kernels/get_mla_metadata.h x: 1 contributors (all time) y: 3 lines of code csrc/kernels/mla_combine.cu x: 1 contributors (all time) y: 179 lines of code csrc/kernels/params.h x: 1 contributors (all time) y: 46 lines of code csrc/kernels/utils.h x: 1 contributors (all time) y: 25 lines of code flash_mla/flash_mla_interface.py x: 3 contributors (all time) y: 34 lines of code benchmark/visualize.py x: 3 contributors (all time) y: 20 lines of code
1230.0
lines of code
  min: 3.0
  average: 160.13
  25th percentile: 8.0
  median: 46.0
  75th percentile: 176.0
  max: 1230.0
0 5.0
contributors (all time)
min: 1.0 | average: 2.07 | 25th percentile: 1.0 | median: 1.0 | 75th percentile: 3.0 | max: 5.0

File Size vs. Commits (30 days): 14 points

setup.py x: 2 commits (30d) y: 79 lines of code csrc/kernels/splitkv_mla.cu x: 2 commits (30d) y: 1230 lines of code csrc/kernels/traits.h x: 2 commits (30d) y: 87 lines of code benchmark/bench_flash_mla.py x: 1 commits (30d) y: 434 lines of code csrc/flash_api.cpp x: 1 commits (30d) y: 176 lines of code csrc/kernels/config.h x: 1 commits (30d) y: 8 lines of code csrc/kernels/get_mla_metadata.cu x: 1 commits (30d) y: 73 lines of code csrc/kernels/get_mla_metadata.h x: 1 commits (30d) y: 3 lines of code csrc/kernels/params.h x: 1 commits (30d) y: 46 lines of code csrc/kernels/utils.h x: 1 commits (30d) y: 25 lines of code flash_mla/flash_mla_interface.py x: 1 commits (30d) y: 34 lines of code
1230.0
lines of code
  min: 3.0
  average: 170.14
  25th percentile: 7.0
  median: 59.5
  75th percentile: 176.75
  max: 1230.0
0 2.0
commits (30d)
min: 1.0 | average: 1.21 | 25th percentile: 1.0 | median: 1.0 | 75th percentile: 1.25 | max: 2.0

File Size vs. Contributors (30 days): 14 points

setup.py x: 2 contributors (30d) y: 79 lines of code csrc/kernels/splitkv_mla.cu x: 2 contributors (30d) y: 1230 lines of code csrc/kernels/traits.h x: 2 contributors (30d) y: 87 lines of code benchmark/bench_flash_mla.py x: 1 contributors (30d) y: 434 lines of code csrc/flash_api.cpp x: 1 contributors (30d) y: 176 lines of code csrc/kernels/config.h x: 1 contributors (30d) y: 8 lines of code csrc/kernels/get_mla_metadata.cu x: 1 contributors (30d) y: 73 lines of code csrc/kernels/get_mla_metadata.h x: 1 contributors (30d) y: 3 lines of code csrc/kernels/params.h x: 1 contributors (30d) y: 46 lines of code csrc/kernels/utils.h x: 1 contributors (30d) y: 25 lines of code flash_mla/flash_mla_interface.py x: 1 contributors (30d) y: 34 lines of code
1230.0
lines of code
  min: 3.0
  average: 170.14
  25th percentile: 7.0
  median: 59.5
  75th percentile: 176.75
  max: 1230.0
0 2.0
contributors (30d)
min: 1.0 | average: 1.21 | 25th percentile: 1.0 | median: 1.0 | 75th percentile: 1.25 | max: 2.0

File Size vs. Commits (90 days): 15 points

setup.py x: 8 commits (90d) y: 79 lines of code csrc/kernels/splitkv_mla.cu x: 2 commits (90d) y: 1230 lines of code csrc/kernels/traits.h x: 2 commits (90d) y: 87 lines of code benchmark/bench_flash_mla.py x: 5 commits (90d) y: 434 lines of code csrc/flash_api.cpp x: 6 commits (90d) y: 176 lines of code csrc/kernels/config.h x: 1 commits (90d) y: 8 lines of code csrc/kernels/get_mla_metadata.cu x: 1 commits (90d) y: 73 lines of code csrc/kernels/get_mla_metadata.h x: 1 commits (90d) y: 3 lines of code csrc/kernels/mla_combine.cu x: 1 commits (90d) y: 179 lines of code csrc/kernels/params.h x: 1 commits (90d) y: 46 lines of code csrc/kernels/utils.h x: 1 commits (90d) y: 25 lines of code flash_mla/flash_mla_interface.py x: 3 commits (90d) y: 34 lines of code benchmark/visualize.py x: 4 commits (90d) y: 20 lines of code
1230.0
lines of code
  min: 3.0
  average: 160.13
  25th percentile: 8.0
  median: 46.0
  75th percentile: 176.0
  max: 1230.0
0 8.0
commits (90d)
min: 1.0 | average: 2.53 | 25th percentile: 1.0 | median: 1.0 | 75th percentile: 4.0 | max: 8.0

File Size vs. Contributors (90 days): 15 points

setup.py x: 5 contributors (90d) y: 79 lines of code csrc/kernels/splitkv_mla.cu x: 2 contributors (90d) y: 1230 lines of code csrc/kernels/traits.h x: 2 contributors (90d) y: 87 lines of code benchmark/bench_flash_mla.py x: 4 contributors (90d) y: 434 lines of code csrc/flash_api.cpp x: 4 contributors (90d) y: 176 lines of code csrc/kernels/config.h x: 1 contributors (90d) y: 8 lines of code csrc/kernels/get_mla_metadata.cu x: 1 contributors (90d) y: 73 lines of code csrc/kernels/get_mla_metadata.h x: 1 contributors (90d) y: 3 lines of code csrc/kernels/mla_combine.cu x: 1 contributors (90d) y: 179 lines of code csrc/kernels/params.h x: 1 contributors (90d) y: 46 lines of code csrc/kernels/utils.h x: 1 contributors (90d) y: 25 lines of code flash_mla/flash_mla_interface.py x: 3 contributors (90d) y: 34 lines of code benchmark/visualize.py x: 3 contributors (90d) y: 20 lines of code
1230.0
lines of code
  min: 3.0
  average: 160.13
  25th percentile: 8.0
  median: 46.0
  75th percentile: 176.0
  max: 1230.0
0 5.0
contributors (90d)
min: 1.0 | average: 2.07 | 25th percentile: 1.0 | median: 1.0 | 75th percentile: 3.0 | max: 5.0