deepseek-ai/TileKernels
File Churn
File change frequency (churn) shows the distribution of file updates (days with at least one commit).
Overview
File Change Frequency Overall
  • There are 138 files with 16,188 lines of code.
    • 0 files changed more than 100 times (0 lines of code)
    • 0 files changed 51-100 times (0 lines of code)
    • 0 files changed 21-50 times (0 lines of code)
    • 0 files changed 6-20 times (0 lines of code)
    • 138 files changed 1-5 times (16,188 lines of code)
0% | 0% | 0% | 0% | 100%
Legend:
101+
51-100
21-50
6-20
1-5

explore: grouped by folders | grouped by update frequency | data
Contributors Count Frequency Overall
  • There are 138 files with 16,188 lines of code.
    • 0 files changed by more than 25 contributors (0 lines of code)
    • 0 files changed by 11-25 contributors (0 lines of code)
    • 0 files changed by 6-10 contributors (0 lines of code)
    • 138 files changed by 2-5 contributors (16,188 lines of code)
    • 0 files changed by 1 contributor (0 lines of code)
0% | 0% | 0% | 100% | 0%
Legend:
26+
11-25
6-10
2-5
1

explore: grouped by folders | grouped by contributors count | data
File Change Frequency per File Extension
py, txt, gitignore, md, editorconfig, toml, yaml
File Change Frequency per Extension
The number of recorded file updates
101+
51-100
21-50
6-20
1-5
py0% | 0% | 0% | 0% | 100%
yaml0% | 0% | 0% | 0% | 100%
File Change Frequency per Logical Decomposition
primary
primary (file change frequency)
The number of recorded file updates
101+
51-100
21-50
6-20
1-5
tile_kernels0% | 0% | 0% | 0% | 100%
ROOT0% | 0% | 0% | 0% | 100%
Most Frequently Changed Files (Top 50)

See data for all files...

File# lines# unitscreatedlast modified# changes
(days)		# contributorsline
churn		first
contributor		latest
contributor
__init__.py
in tile_kernels/transform
 1 - 2026-09-30 2026-09-30 1 2 +2 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/testing/pytest
 1 - 2026-09-30 2026-09-30 1 2 +46 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/modeling
 1 - 2026-09-30 2026-09-30 1 2 +0 / -2 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/modeling/engram
 1 - 2026-09-30 2026-09-30 1 2 +2 / -2 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/testing
 2 - 2026-09-30 2026-09-30 1 2 +2 / -2 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/rand
 3 - 2026-09-30 2026-09-30 1 2 +8 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/moe
 4 - 2026-09-30 2026-09-30 1 2 +4 / -18 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/engram
 11 - 2026-09-30 2026-09-30 1 2 +16 / -2 122071158+bucket-xv@users.n... xch@deepseek.com
12 - 2026-09-30 2026-09-30 1 2 +32 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/quant
 13 - 2026-09-30 2026-09-30 1 2 +14 / -10 122071158+bucket-xv@users.n... xch@deepseek.com
seed.py
in tile_kernels/rand
 14 2 2026-09-30 2026-09-30 1 2 +40 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
random.py
in tile_kernels/testing/pytest
 14 2 2026-09-30 2026-09-30 1 2 +42 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
sinkhorn_kernel.py
in tile_kernels/mhc
 15 2 2026-09-30 2026-09-30 1 2 +30 / -318 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/mhc
 17 - 2026-09-30 2026-09-30 1 2 +34 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
topk_gate_kernel.py
in tile_kernels/moe
 18 1 2026-09-30 2026-09-30 1 2 +20 / -114 122071158+bucket-xv@users.n... xch@deepseek.com
normalize_weight_kernel.py
in tile_kernels/moe
 18 1 2026-09-30 2026-09-30 1 2 +24 / -90 122071158+bucket-xv@users.n... xch@deepseek.com
moe.py
in tile_kernels/torch
 21 2 2026-09-30 2026-09-30 1 2 +2 / -140 122071158+bucket-xv@users.n... xch@deepseek.com
engram_fused_weight_kernel.py
in tile_kernels/engram
 21 1 2026-09-30 2026-09-30 1 2 +32 / -76 122071158+bucket-xv@users.n... xch@deepseek.com
utils.py
in tile_kernels
 24 6 2026-09-30 2026-09-30 1 2 +56 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
common.py
in tile_kernels/testing/pytest
 25 4 2026-09-30 2026-09-30 1 2 +142 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/torch
 25 - 2026-09-30 2026-09-30 1 2 +48 / -18 122071158+bucket-xv@users.n... xch@deepseek.com
moe.py
in tile_kernels/testing
 27 1 2026-09-30 2026-09-30 1 2 +64 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
precompile.py
in tile_kernels/testing/pytest
 27 2 2026-09-30 2026-09-30 1 2 +76 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_grad_w_reduce_kernel.py
in tile_kernels/engram
 31 1 2026-09-30 2026-09-30 1 2 +42 / -128 122071158+bucket-xv@users.n... xch@deepseek.com
randn_asc.py
in tile_kernels/rand
 32 1 2026-09-30 2026-09-30 1 2 +84 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm.py
in tile_kernels/torch
 32 2 2026-09-30 2026-09-30 1 2 +88 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_hash_kernel.py
in tile_kernels/engram
 34 1 2026-09-30 2026-09-30 1 2 +46 / -128 122071158+bucket-xv@users.n... xch@deepseek.com
tilelang_warning.py
in tile_kernels/testing/pytest
 35 3 2026-09-30 2026-09-30 1 2 +140 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
expand_kernel.py
in tile_kernels/mhc
 36 4 2026-09-30 2026-09-30 1 2 +84 / -106 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels
 36 - 2026-09-30 2026-09-30 1 2 +52 / -4 122071158+bucket-xv@users.n... xch@deepseek.com
quant.py
in tile_kernels/testing
 37 3 2026-09-30 2026-09-30 1 2 +100 / -4 122071158+bucket-xv@users.n... xch@deepseek.com
engram_fused_weight_cuda.py
in tile_kernels/engram
 37 2 2026-09-30 2026-09-30 1 2 +84 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
cast_kernel.py
in tile_kernels/quant
 37 1 2026-09-30 2026-09-30 1 2 +118 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
normalize_weight_asc.py
in tile_kernels/moe
 38 1 2026-09-30 2026-09-30 1 2 +96 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_channel_cast_kernel.py
in tile_kernels/quant
 39 1 2026-09-30 2026-09-30 1 2 +104 / -26 122071158+bucket-xv@users.n... xch@deepseek.com
randn_kernel.py
in tile_kernels/rand
 40 3 2026-09-30 2026-09-30 1 2 +110 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
normalize_weight_cuda.py
in tile_kernels/moe
 41 1 2026-09-30 2026-09-30 1 2 +100 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
41 1 2026-09-30 2026-09-30 1 2 +122 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
rope.py
in tile_kernels/torch
 42 1 2026-09-30 2026-09-30 1 2 +98 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
pre_apply_mix_kernel.py
in tile_kernels/mhc
 44 2 2026-09-30 2026-09-30 1 2 +86 / -204 122071158+bucket-xv@users.n... xch@deepseek.com
expand_cuda.py
in tile_kernels/mhc
 45 2 2026-09-30 2026-09-30 1 2 +106 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
cast_back_kernel.py
in tile_kernels/quant
 48 2 2026-09-30 2026-09-30 1 2 +42 / -160 122071158+bucket-xv@users.n... xch@deepseek.com
engram_fused_weight_asc.py
in tile_kernels/engram
 49 2 2026-09-30 2026-09-30 1 2 +114 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
49 1 2026-09-30 2026-09-30 1 2 +40 / -308 122071158+bucket-xv@users.n... xch@deepseek.com
numeric.py
in tile_kernels/testing
 52 4 2026-09-30 2026-09-30 1 2 +60 / -42 122071158+bucket-xv@users.n... xch@deepseek.com
topk_gate_cuda.py
in tile_kernels/moe
 56 1 2026-09-30 2026-09-30 1 2 +136 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_grad_w_reduce_cuda.py
in tile_kernels/engram
 56 2 2026-09-30 2026-09-30 1 2 +134 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
59 1 2026-09-30 2026-09-30 1 2 +190 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
xdist_failfast.py
in tile_kernels/testing/pytest
 59 6 2026-09-30 2026-09-30 1 2 +194 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate.py
in tile_kernels/modeling/engram
 61 2 2026-09-30 2026-09-30 1 2 +62 / -42 122071158+bucket-xv@users.n... xch@deepseek.com
Files With Most Churn (Top 50)
Based on the number of lines added and deleted across all commits.

See data for all files...

File# lines# unitscreatedlast modified# changes
(days)		# contributorsline
churn		first
contributor		latest
contributor
benchmark.py
in tile_kernels/testing/pytest
 656 45 2026-09-30 2026-09-30 1 2 +1886 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_backward_asc.py
in tile_kernels/quant
 445 8 2026-09-30 2026-09-30 1 2 +1000 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
pre_split_mixes_asc.py
in tile_kernels/mhc
 414 5 2026-09-30 2026-09-30 1 2 +956 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
gpu_mem.py
in tile_kernels/testing/pytest
 319 21 2026-09-30 2026-09-30 1 2 +874 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_fn_asc.py
in tile_kernels/mhc
 369 5 2026-09-30 2026-09-30 1 2 +856 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
pre_big_fuse_asc.py
in tile_kernels/mhc
 354 4 2026-09-30 2026-09-30 1 2 +846 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_forward_asc.py
in tile_kernels/quant
 379 11 2026-09-30 2026-09-30 1 2 +840 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_fn_kernel.py
in tile_kernels/mhc
 117 4 2026-09-30 2026-09-30 1 2 +240 / -550 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_forward_asc.py
in tile_kernels/quant
 347 12 2026-09-30 2026-09-30 1 2 +790 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
sinkhorn_asc.py
in tile_kernels/mhc
 245 5 2026-09-30 2026-09-30 1 2 +778 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
330 17 2026-09-30 2026-09-30 1 2 +770 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_token_cast_asc.py
in tile_kernels/quant
 336 11 2026-09-30 2026-09-30 1 2 +740 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_bwd_cuda.py
in tile_kernels/engram
 274 8 2026-09-30 2026-09-30 1 2 +680 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_forward_cuda.py
in tile_kernels/quant
 260 2 2026-09-30 2026-09-30 1 2 +646 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_backward_kernel.py
in tile_kernels/quant
 180 3 2026-09-30 2026-09-30 1 2 +622 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_channel_cast_asc.py
in tile_kernels/quant
 262 5 2026-09-30 2026-09-30 1 2 +614 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_sinkhorn_asc.py
in tile_kernels/engram
 252 8 2026-09-30 2026-09-30 1 2 +588 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
220 3 2026-09-30 2026-09-30 1 2 +586 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
moe_topk_gate_forward_asc.py
in tile_kernels/moe
 265 2 2026-09-30 2026-09-30 1 2 +562 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_token_cast_kernel.py
in tile_kernels/quant
 149 4 2026-09-30 2026-09-30 1 2 +194 / -360 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu.py
in tile_kernels/torch
 143 5 2026-09-30 2026-09-30 1 2 +306 / -242 122071158+bucket-xv@users.n... xch@deepseek.com
common.py
in tile_kernels/quant
 261 24 2026-09-30 2026-09-30 1 2 +346 / -202 122071158+bucket-xv@users.n... xch@deepseek.com
norm_forward_kernel.py
in tile_kernels/quant
 194 4 2026-09-30 2026-09-30 1 2 +538 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_backward_cuda.py
in tile_kernels/quant
 211 2 2026-09-30 2026-09-30 1 2 +524 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_fwd_asc.py
in tile_kernels/engram
 233 2 2026-09-30 2026-09-30 1 2 +506 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
pre_split_mixes_kernel.py
in tile_kernels/mhc
 108 2 2026-09-30 2026-09-30 1 2 +204 / -300 122071158+bucket-xv@users.n... xch@deepseek.com
engram_sinkhorn_cuda.py
in tile_kernels/engram
 213 11 2026-09-30 2026-09-30 1 2 +500 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_lossless_asc.py
in tile_kernels/quant
 218 11 2026-09-30 2026-09-30 1 2 +490 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_forward_cuda.py
in tile_kernels/quant
 217 3 2026-09-30 2026-09-30 1 2 +488 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_kernel.py
in tile_kernels/quant
 134 4 2026-09-30 2026-09-30 1 2 +158 / -298 122071158+bucket-xv@users.n... xch@deepseek.com
norm_backward_cuda.py
in tile_kernels/quant
 192 3 2026-09-30 2026-09-30 1 2 +454 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_token_cast_cuda.py
in tile_kernels/quant
 179 4 2026-09-30 2026-09-30 1 2 +442 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_asc.py
in tile_kernels/quant
 195 8 2026-09-30 2026-09-30 1 2 +442 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
pre_split_mixes_cuda.py
in tile_kernels/mhc
 177 3 2026-09-30 2026-09-30 1 2 +420 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
cast_back_asc.py
in tile_kernels/quant
 190 6 2026-09-30 2026-09-30 1 2 +416 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
post_kernel.py
in tile_kernels/mhc
 103 2 2026-09-30 2026-09-30 1 2 +102 / -308 122071158+bucket-xv@users.n... xch@deepseek.com
post_asc.py
in tile_kernels/mhc
 164 2 2026-09-30 2026-09-30 1 2 +408 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_bwd_asc.py
in tile_kernels/engram
 181 2 2026-09-30 2026-09-30 1 2 +402 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
rope_asc.py
in tile_kernels/transform
 185 5 2026-09-30 2026-09-30 1 2 +400 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_fwd_cuda.py
in tile_kernels/engram
 163 4 2026-09-30 2026-09-30 1 2 +398 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
head_compute_mix_kernel.py
in tile_kernels/mhc
 94 3 2026-09-30 2026-09-30 1 2 +210 / -166 122071158+bucket-xv@users.n... xch@deepseek.com
norm_backward_asc.py
in tile_kernels/quant
 172 5 2026-09-30 2026-09-30 1 2 +372 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_sinkhorn_kernel.py
in tile_kernels/engram
 112 4 2026-09-30 2026-09-30 1 2 +358 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
topk.py
in tile_kernels/torch
 141 4 2026-09-30 2026-09-30 1 2 +164 / -190 122071158+bucket-xv@users.n... xch@deepseek.com
pre_big_fuse_kernel.py
in tile_kernels/mhc
 67 1 2026-09-30 2026-09-30 1 2 +118 / -234 122071158+bucket-xv@users.n... xch@deepseek.com
sinkhorn_kernel.py
in tile_kernels/mhc
 15 2 2026-09-30 2026-09-30 1 2 +30 / -318 122071158+bucket-xv@users.n... xch@deepseek.com
49 1 2026-09-30 2026-09-30 1 2 +40 / -308 122071158+bucket-xv@users.n... xch@deepseek.com
148 1 2026-09-30 2026-09-30 1 2 +332 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
generator.py
in tile_kernels/testing
 131 11 2026-09-30 2026-09-30 1 2 +236 / -94 122071158+bucket-xv@users.n... xch@deepseek.com
cast.py
in tile_kernels/torch
 255 8 2026-09-30 2026-09-30 1 2 +258 / -64 122071158+bucket-xv@users.n... xch@deepseek.com
Files With Most Contributors (Top 50)
Based on the number of unique email addresses found in commits.

See data for all files...

File# lines# unitscreatedlast modified# changes
(days)		# contributorsline
churn		first
contributor		latest
contributor
12 - 2026-09-30 2026-09-30 1 2 +32 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
49 1 2026-09-30 2026-09-30 1 2 +40 / -308 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_backward_kernel.py
in tile_kernels/quant
 180 3 2026-09-30 2026-09-30 1 2 +622 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_forward_kernel.py
in tile_kernels/quant
 120 3 2026-09-30 2026-09-30 1 2 +282 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_cuda.py
in tile_kernels/quant
 122 4 2026-09-30 2026-09-30 1 2 +282 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_backward_cuda.py
in tile_kernels/quant
 192 3 2026-09-30 2026-09-30 1 2 +454 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_channel_cast_asc.py
in tile_kernels/quant
 262 5 2026-09-30 2026-09-30 1 2 +614 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
rstd.py
in tile_kernels/quant
 65 5 2026-09-30 2026-09-30 1 2 +196 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_forward_asc.py
in tile_kernels/quant
 379 11 2026-09-30 2026-09-30 1 2 +840 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
330 17 2026-09-30 2026-09-30 1 2 +770 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_lossless_cuda.py
in tile_kernels/quant
 108 4 2026-09-30 2026-09-30 1 2 +320 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_forward_cuda.py
in tile_kernels/quant
 260 2 2026-09-30 2026-09-30 1 2 +646 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_forward_kernel.py
in tile_kernels/quant
 194 4 2026-09-30 2026-09-30 1 2 +538 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_channel_cast_kernel.py
in tile_kernels/quant
 39 1 2026-09-30 2026-09-30 1 2 +104 / -26 122071158+bucket-xv@users.n... xch@deepseek.com
norm_backward_asc.py
in tile_kernels/quant
 172 5 2026-09-30 2026-09-30 1 2 +372 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
cast_back_asc.py
in tile_kernels/quant
 190 6 2026-09-30 2026-09-30 1 2 +416 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_asc.py
in tile_kernels/quant
 195 8 2026-09-30 2026-09-30 1 2 +442 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_channel_cast_cuda.py
in tile_kernels/quant
 136 2 2026-09-30 2026-09-30 1 2 +318 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_backward_cuda.py
in tile_kernels/quant
 211 2 2026-09-30 2026-09-30 1 2 +524 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_token_cast_kernel.py
in tile_kernels/quant
 149 4 2026-09-30 2026-09-30 1 2 +194 / -360 122071158+bucket-xv@users.n... xch@deepseek.com
per_token_cast_cuda.py
in tile_kernels/quant
 179 4 2026-09-30 2026-09-30 1 2 +442 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
41 1 2026-09-30 2026-09-30 1 2 +122 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
102 1 2026-09-30 2026-09-30 1 2 +232 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/quant
 13 - 2026-09-30 2026-09-30 1 2 +14 / -10 122071158+bucket-xv@users.n... xch@deepseek.com
78 1 2026-09-30 2026-09-30 1 2 +180 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
cast_cuda.py
in tile_kernels/quant
 70 1 2026-09-30 2026-09-30 1 2 +174 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
cast_back_kernel.py
in tile_kernels/quant
 48 2 2026-09-30 2026-09-30 1 2 +42 / -160 122071158+bucket-xv@users.n... xch@deepseek.com
common.py
in tile_kernels/quant
 261 24 2026-09-30 2026-09-30 1 2 +346 / -202 122071158+bucket-xv@users.n... xch@deepseek.com
cast_kernel.py
in tile_kernels/quant
 37 1 2026-09-30 2026-09-30 1 2 +118 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_token_cast_asc.py
in tile_kernels/quant
 336 11 2026-09-30 2026-09-30 1 2 +740 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_backward_asc.py
in tile_kernels/quant
 445 8 2026-09-30 2026-09-30 1 2 +1000 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
cast_back_cuda.py
in tile_kernels/quant
 64 2 2026-09-30 2026-09-30 1 2 +172 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_lossless_asc.py
in tile_kernels/quant
 218 11 2026-09-30 2026-09-30 1 2 +490 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
cast_asc.py
in tile_kernels/quant
 107 7 2026-09-30 2026-09-30 1 2 +276 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
72 1 2026-09-30 2026-09-30 1 2 +194 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
68 3 2026-09-30 2026-09-30 1 2 +174 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_forward_cuda.py
in tile_kernels/quant
 217 3 2026-09-30 2026-09-30 1 2 +488 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_kernel.py
in tile_kernels/quant
 134 4 2026-09-30 2026-09-30 1 2 +158 / -298 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_forward_asc.py
in tile_kernels/quant
 347 12 2026-09-30 2026-09-30 1 2 +790 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_backward_kernel.py
in tile_kernels/quant
 62 1 2026-09-30 2026-09-30 1 2 +174 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_fwd_kernel.py
in tile_kernels/engram
 62 1 2026-09-30 2026-09-30 1 2 +170 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_fused_weight_kernel.py
in tile_kernels/engram
 21 1 2026-09-30 2026-09-30 1 2 +32 / -76 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_bwd_asc.py
in tile_kernels/engram
 181 2 2026-09-30 2026-09-30 1 2 +402 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_hash_cuda.py
in tile_kernels/engram
 62 2 2026-09-30 2026-09-30 1 2 +138 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_hash_asc.py
in tile_kernels/engram
 77 2 2026-09-30 2026-09-30 1 2 +176 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
__init__.py
in tile_kernels/engram
 11 - 2026-09-30 2026-09-30 1 2 +16 / -2 122071158+bucket-xv@users.n... xch@deepseek.com
engram_fused_weight_asc.py
in tile_kernels/engram
 49 2 2026-09-30 2026-09-30 1 2 +114 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_sinkhorn_asc.py
in tile_kernels/engram
 252 8 2026-09-30 2026-09-30 1 2 +588 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_fwd_cuda.py
in tile_kernels/engram
 163 4 2026-09-30 2026-09-30 1 2 +398 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_grad_w_reduce_cuda.py
in tile_kernels/engram
 56 2 2026-09-30 2026-09-30 1 2 +134 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
Files With Least Contributors (Top 50)
Based on the number of unique email addresses found in commits.

See data for all files...

File# lines# unitscreatedlast modified# changes
(days)		# contributorsline
churn		first
contributor		latest
contributor
reduce_fused.py
in tile_kernels/torch
 41 2
types.py
in tile_kernels/quant
 2 -
benchmark.py
in tile_kernels/testing/pytest
 656 45 2026-09-30 2026-09-30 1 2 +1886 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_backward_asc.py
in tile_kernels/quant
 445 8 2026-09-30 2026-09-30 1 2 +1000 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
pre_split_mixes_asc.py
in tile_kernels/mhc
 414 5 2026-09-30 2026-09-30 1 2 +956 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_forward_asc.py
in tile_kernels/quant
 379 11 2026-09-30 2026-09-30 1 2 +840 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_fn_asc.py
in tile_kernels/mhc
 369 5 2026-09-30 2026-09-30 1 2 +856 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
pre_big_fuse_asc.py
in tile_kernels/mhc
 354 4 2026-09-30 2026-09-30 1 2 +846 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_forward_asc.py
in tile_kernels/quant
 347 12 2026-09-30 2026-09-30 1 2 +790 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_token_cast_asc.py
in tile_kernels/quant
 336 11 2026-09-30 2026-09-30 1 2 +740 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
330 17 2026-09-30 2026-09-30 1 2 +770 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
gpu_mem.py
in tile_kernels/testing/pytest
 319 21 2026-09-30 2026-09-30 1 2 +874 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_bwd_cuda.py
in tile_kernels/engram
 274 8 2026-09-30 2026-09-30 1 2 +680 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
moe_topk_gate_forward_asc.py
in tile_kernels/moe
 265 2 2026-09-30 2026-09-30 1 2 +562 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_channel_cast_asc.py
in tile_kernels/quant
 262 5 2026-09-30 2026-09-30 1 2 +614 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
common.py
in tile_kernels/quant
 261 24 2026-09-30 2026-09-30 1 2 +346 / -202 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_forward_cuda.py
in tile_kernels/quant
 260 2 2026-09-30 2026-09-30 1 2 +646 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
cast.py
in tile_kernels/torch
 255 8 2026-09-30 2026-09-30 1 2 +258 / -64 122071158+bucket-xv@users.n... xch@deepseek.com
engram_sinkhorn_asc.py
in tile_kernels/engram
 252 8 2026-09-30 2026-09-30 1 2 +588 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
sinkhorn_asc.py
in tile_kernels/mhc
 245 5 2026-09-30 2026-09-30 1 2 +778 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_fwd_asc.py
in tile_kernels/engram
 233 2 2026-09-30 2026-09-30 1 2 +506 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
220 3 2026-09-30 2026-09-30 1 2 +586 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_lossless_asc.py
in tile_kernels/quant
 218 11 2026-09-30 2026-09-30 1 2 +490 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_forward_cuda.py
in tile_kernels/quant
 217 3 2026-09-30 2026-09-30 1 2 +488 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_sinkhorn_cuda.py
in tile_kernels/engram
 213 11 2026-09-30 2026-09-30 1 2 +500 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_backward_cuda.py
in tile_kernels/quant
 211 2 2026-09-30 2026-09-30 1 2 +524 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_asc.py
in tile_kernels/quant
 195 8 2026-09-30 2026-09-30 1 2 +442 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_forward_kernel.py
in tile_kernels/quant
 194 4 2026-09-30 2026-09-30 1 2 +538 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_backward_cuda.py
in tile_kernels/quant
 192 3 2026-09-30 2026-09-30 1 2 +454 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
cast_back_asc.py
in tile_kernels/quant
 190 6 2026-09-30 2026-09-30 1 2 +416 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
rope_asc.py
in tile_kernels/transform
 185 5 2026-09-30 2026-09-30 1 2 +400 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_bwd_asc.py
in tile_kernels/engram
 181 2 2026-09-30 2026-09-30 1 2 +402 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu_backward_kernel.py
in tile_kernels/quant
 180 3 2026-09-30 2026-09-30 1 2 +622 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_token_cast_cuda.py
in tile_kernels/quant
 179 4 2026-09-30 2026-09-30 1 2 +442 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
pre_split_mixes_cuda.py
in tile_kernels/mhc
 177 3 2026-09-30 2026-09-30 1 2 +420 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
norm_backward_asc.py
in tile_kernels/quant
 172 5 2026-09-30 2026-09-30 1 2 +372 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
post_asc.py
in tile_kernels/mhc
 164 2 2026-09-30 2026-09-30 1 2 +408 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
engram_gate_fwd_cuda.py
in tile_kernels/engram
 163 4 2026-09-30 2026-09-30 1 2 +398 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_token_cast_kernel.py
in tile_kernels/quant
 149 4 2026-09-30 2026-09-30 1 2 +194 / -360 122071158+bucket-xv@users.n... xch@deepseek.com
148 1 2026-09-30 2026-09-30 1 2 +332 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
swiglu.py
in tile_kernels/torch
 143 5 2026-09-30 2026-09-30 1 2 +306 / -242 122071158+bucket-xv@users.n... xch@deepseek.com
topk.py
in tile_kernels/torch
 141 4 2026-09-30 2026-09-30 1 2 +164 / -190 122071158+bucket-xv@users.n... xch@deepseek.com
multilayer_recompute_asc.py
in tile_kernels/mhc
 138 1 2026-09-30 2026-09-30 1 2 +314 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_channel_cast_cuda.py
in tile_kernels/quant
 136 2 2026-09-30 2026-09-30 1 2 +318 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_kernel.py
in tile_kernels/quant
 134 4 2026-09-30 2026-09-30 1 2 +158 / -298 122071158+bucket-xv@users.n... xch@deepseek.com
generator.py
in tile_kernels/testing
 131 11 2026-09-30 2026-09-30 1 2 +236 / -94 122071158+bucket-xv@users.n... xch@deepseek.com
norm_fn_cuda.py
in tile_kernels/mhc
 131 4 2026-09-30 2026-09-30 1 2 +310 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
sinkhorn_cuda.py
in tile_kernels/mhc
 130 2 2026-09-30 2026-09-30 1 2 +310 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
post_cuda.py
in tile_kernels/mhc
 127 2 2026-09-30 2026-09-30 1 2 +314 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
per_block_cast_cuda.py
in tile_kernels/quant
 122 4 2026-09-30 2026-09-30 1 2 +282 / -0 122071158+bucket-xv@users.n... xch@deepseek.com
Correlations

File Size vs. Number of Changes: 138 points

.pre-commit-config.yaml x: 12 lines of code y: 1 # changes tile_kernels/__init__.py x: 36 lines of code y: 1 # changes tile_kernels/config.py x: 90 lines of code y: 1 # changes tile_kernels/engram/__init__.py x: 11 lines of code y: 1 # changes tile_kernels/engram/engram_fused_weight_asc.py x: 49 lines of code y: 1 # changes tile_kernels/engram/engram_fused_weight_cuda.py x: 37 lines of code y: 1 # changes tile_kernels/engram/engram_fused_weight_kernel.py x: 21 lines of code y: 1 # changes tile_kernels/engram/engram_gate_bwd_asc.py x: 181 lines of code y: 1 # changes tile_kernels/engram/engram_gate_bwd_cuda.py x: 274 lines of code y: 1 # changes tile_kernels/engram/engram_gate_bwd_kernel.py x: 61 lines of code y: 1 # changes tile_kernels/engram/engram_gate_fwd_asc.py x: 233 lines of code y: 1 # changes tile_kernels/engram/engram_gate_fwd_cuda.py x: 163 lines of code y: 1 # changes tile_kernels/engram/engram_gate_fwd_kernel.py x: 62 lines of code y: 1 # changes tile_kernels/engram/engram_grad_w_reduce_asc.py x: 66 lines of code y: 1 # changes tile_kernels/engram/engram_grad_w_reduce_cuda.py x: 56 lines of code y: 1 # changes tile_kernels/engram/engram_grad_w_reduce_kernel.py x: 31 lines of code y: 1 # changes tile_kernels/engram/engram_hash_asc.py x: 77 lines of code y: 1 # changes tile_kernels/engram/engram_hash_kernel.py x: 34 lines of code y: 1 # changes tile_kernels/engram/engram_sinkhorn_asc.py x: 252 lines of code y: 1 # changes tile_kernels/engram/engram_sinkhorn_cuda.py x: 213 lines of code y: 1 # changes tile_kernels/engram/engram_sinkhorn_kernel.py x: 112 lines of code y: 1 # changes tile_kernels/mhc/__init__.py x: 17 lines of code y: 1 # changes tile_kernels/mhc/expand_cuda.py x: 45 lines of code y: 1 # changes tile_kernels/mhc/head_compute_mix_asc.py x: 85 lines of code y: 1 # changes tile_kernels/mhc/head_compute_mix_kernel.py x: 94 lines of code y: 1 # changes tile_kernels/mhc/multilayer_recompute_asc.py x: 138 lines of code y: 1 # changes tile_kernels/mhc/multilayer_recompute_cuda.py x: 103 lines of code y: 1 # changes tile_kernels/mhc/norm_fn_asc.py x: 369 lines of code y: 1 # changes tile_kernels/mhc/norm_fn_cuda.py x: 131 lines of code y: 1 # changes tile_kernels/mhc/norm_fn_kernel.py x: 117 lines of code y: 1 # changes tile_kernels/mhc/post_asc.py x: 164 lines of code y: 1 # changes tile_kernels/mhc/post_cuda.py x: 127 lines of code y: 1 # changes tile_kernels/mhc/pre_apply_mix_asc.py x: 99 lines of code y: 1 # changes tile_kernels/mhc/pre_apply_mix_cuda.py x: 92 lines of code y: 1 # changes tile_kernels/mhc/pre_apply_mix_kernel.py x: 44 lines of code y: 1 # changes tile_kernels/mhc/pre_big_fuse_asc.py x: 354 lines of code y: 1 # changes tile_kernels/mhc/pre_big_fuse_cuda.py x: 106 lines of code y: 1 # changes tile_kernels/mhc/pre_big_fuse_kernel.py x: 67 lines of code y: 1 # changes tile_kernels/mhc/pre_split_mixes_asc.py x: 414 lines of code y: 1 # changes tile_kernels/mhc/pre_split_mixes_cuda.py x: 177 lines of code y: 1 # changes tile_kernels/mhc/pre_split_mixes_kernel.py x: 108 lines of code y: 1 # changes tile_kernels/mhc/sinkhorn_asc.py x: 245 lines of code y: 1 # changes tile_kernels/mhc/sinkhorn_cuda.py x: 130 lines of code y: 1 # changes tile_kernels/mhc/sinkhorn_kernel.py x: 15 lines of code y: 1 # changes tile_kernels/modeling/__init__.py x: 1 lines of code y: 1 # changes tile_kernels/moe/__init__.py x: 4 lines of code y: 1 # changes tile_kernels/moe/moe_topk_gate_backward_asc.py x: 148 lines of code y: 1 # changes tile_kernels/moe/moe_topk_gate_backward_kernel.py x: 59 lines of code y: 1 # changes tile_kernels/moe/moe_topk_gate_forward_asc.py x: 265 lines of code y: 1 # changes tile_kernels/moe/moe_topk_gate_forward_cuda.py x: 220 lines of code y: 1 # changes tile_kernels/moe/moe_topk_gate_forward_kernel.py x: 110 lines of code y: 1 # changes tile_kernels/moe/normalize_weight_asc.py x: 38 lines of code y: 1 # changes tile_kernels/moe/normalize_weight_cuda.py x: 41 lines of code y: 1 # changes tile_kernels/moe/normalize_weight_kernel.py x: 18 lines of code y: 1 # changes tile_kernels/quant/batched_transpose_weight_sf_kernel.py x: 72 lines of code y: 1 # changes tile_kernels/quant/cast_asc.py x: 107 lines of code y: 1 # changes tile_kernels/quant/cast_back_asc.py x: 190 lines of code y: 1 # changes tile_kernels/quant/cast_back_cuda.py x: 64 lines of code y: 1 # changes tile_kernels/quant/cast_back_kernel.py x: 48 lines of code y: 1 # changes tile_kernels/quant/cast_cuda.py x: 70 lines of code y: 1 # changes tile_kernels/quant/common.py x: 261 lines of code y: 1 # changes tile_kernels/quant/norm_backward_asc.py x: 172 lines of code y: 1 # changes tile_kernels/quant/norm_backward_cuda.py x: 192 lines of code y: 1 # changes tile_kernels/quant/norm_forward_asc.py x: 379 lines of code y: 1 # changes tile_kernels/quant/norm_forward_cuda.py x: 217 lines of code y: 1 # changes tile_kernels/quant/norm_forward_kernel.py x: 194 lines of code y: 1 # changes tile_kernels/quant/per_block_cast_asc.py x: 195 lines of code y: 1 # changes tile_kernels/quant/per_block_cast_cuda.py x: 122 lines of code y: 1 # changes tile_kernels/quant/per_block_cast_kernel.py x: 134 lines of code y: 1 # changes tile_kernels/quant/per_block_cast_lossless_asc.py x: 218 lines of code y: 1 # changes tile_kernels/quant/per_channel_cast_asc.py x: 262 lines of code y: 1 # changes tile_kernels/quant/per_channel_cast_cuda.py x: 136 lines of code y: 1 # changes tile_kernels/quant/per_token_cast_and_cast_back_asc.py x: 330 lines of code y: 1 # changes tile_kernels/quant/per_token_cast_and_cast_back_cuda.py x: 68 lines of code y: 1 # changes tile_kernels/quant/per_token_cast_asc.py x: 336 lines of code y: 1 # changes tile_kernels/quant/per_token_cast_cuda.py x: 179 lines of code y: 1 # changes tile_kernels/quant/per_token_cast_kernel.py x: 149 lines of code y: 1 # changes tile_kernels/quant/swiglu_backward_asc.py x: 445 lines of code y: 1 # changes tile_kernels/quant/swiglu_backward_cuda.py x: 211 lines of code y: 1 # changes tile_kernels/quant/swiglu_forward_asc.py x: 347 lines of code y: 1 # changes tile_kernels/quant/swiglu_forward_cuda.py x: 260 lines of code y: 1 # changes tile_kernels/quant/swiglu_forward_kernel.py x: 120 lines of code y: 1 # changes tile_kernels/rand/__init__.py x: 3 lines of code y: 1 # changes tile_kernels/rand/randn_asc.py x: 32 lines of code y: 1 # changes tile_kernels/rand/randn_kernel.py x: 40 lines of code y: 1 # changes tile_kernels/rand/seed.py x: 14 lines of code y: 1 # changes tile_kernels/testing/__init__.py x: 2 lines of code y: 1 # changes tile_kernels/testing/bench.py x: 84 lines of code y: 1 # changes tile_kernels/testing/moe.py x: 27 lines of code y: 1 # changes tile_kernels/testing/numeric.py x: 52 lines of code y: 1 # changes tile_kernels/testing/pytest/benchmark.py x: 656 lines of code y: 1 # changes tile_kernels/testing/pytest/common.py x: 25 lines of code y: 1 # changes tile_kernels/testing/pytest/gpu_mem.py x: 319 lines of code y: 1 # changes tile_kernels/testing/pytest/tilelang_warning.py x: 35 lines of code y: 1 # changes tile_kernels/torch/cast.py x: 255 lines of code y: 1 # changes tile_kernels/torch/mhc.py x: 86 lines of code y: 1 # changes tile_kernels/torch/rope.py x: 42 lines of code y: 1 # changes tile_kernels/torch/swiglu.py x: 143 lines of code y: 1 # changes tile_kernels/torch/topk.py x: 141 lines of code y: 1 # changes tile_kernels/transform/rope_asc.py x: 185 lines of code y: 1 # changes tile_kernels/transform/rope_cuda.py x: 115 lines of code y: 1 # changes tile_kernels/utils.py x: 24 lines of code y: 1 # changes
1.0
# changes
  min: 1.0
  average: 1.0
  25th percentile: 1.0
  median: 1.0
  75th percentile: 1.0
  max: 1.0
0 656.0
lines of code
min: 1.0 | average: 117.3 | 25th percentile: 38.75 | median: 85.5 | 75th percentile: 166.0 | max: 656.0

Number of Contributors vs. Number of Changes: 138 points

.pre-commit-config.yaml x: 2 # contributors y: 1 # changes
1.0
# changes
  min: 1.0
  average: 1.0
  25th percentile: 1.0
  median: 1.0
  75th percentile: 1.0
  max: 1.0
0 2.0
# contributors
min: 2.0 | average: 2.0 | 25th percentile: 2.0 | median: 2.0 | 75th percentile: 2.0 | max: 2.0

Number of Contributors vs. File Size: 138 points

.pre-commit-config.yaml x: 2 # contributors y: 12 lines of code tile_kernels/__init__.py x: 2 # contributors y: 36 lines of code tile_kernels/config.py x: 2 # contributors y: 90 lines of code tile_kernels/engram/engram_fused_weight_asc.py x: 2 # contributors y: 49 lines of code tile_kernels/engram/engram_fused_weight_cuda.py x: 2 # contributors y: 37 lines of code tile_kernels/engram/engram_fused_weight_kernel.py x: 2 # contributors y: 21 lines of code tile_kernels/engram/engram_gate_bwd_asc.py x: 2 # contributors y: 181 lines of code tile_kernels/engram/engram_gate_bwd_cuda.py x: 2 # contributors y: 274 lines of code tile_kernels/engram/engram_gate_bwd_kernel.py x: 2 # contributors y: 61 lines of code tile_kernels/engram/engram_gate_fwd_asc.py x: 2 # contributors y: 233 lines of code tile_kernels/engram/engram_gate_fwd_cuda.py x: 2 # contributors y: 163 lines of code tile_kernels/engram/engram_gate_fwd_kernel.py x: 2 # contributors y: 62 lines of code tile_kernels/engram/engram_grad_w_reduce_asc.py x: 2 # contributors y: 66 lines of code tile_kernels/engram/engram_grad_w_reduce_cuda.py x: 2 # contributors y: 56 lines of code tile_kernels/engram/engram_grad_w_reduce_kernel.py x: 2 # contributors y: 31 lines of code tile_kernels/engram/engram_hash_asc.py x: 2 # contributors y: 77 lines of code tile_kernels/engram/engram_sinkhorn_asc.py x: 2 # contributors y: 252 lines of code tile_kernels/engram/engram_sinkhorn_cuda.py x: 2 # contributors y: 213 lines of code tile_kernels/engram/engram_sinkhorn_kernel.py x: 2 # contributors y: 112 lines of code tile_kernels/mhc/__init__.py x: 2 # contributors y: 17 lines of code tile_kernels/mhc/expand_cuda.py x: 2 # contributors y: 45 lines of code tile_kernels/mhc/head_compute_mix_asc.py x: 2 # contributors y: 85 lines of code tile_kernels/mhc/head_compute_mix_kernel.py x: 2 # contributors y: 94 lines of code tile_kernels/mhc/multilayer_recompute_asc.py x: 2 # contributors y: 138 lines of code tile_kernels/mhc/multilayer_recompute_cuda.py x: 2 # contributors y: 103 lines of code tile_kernels/mhc/norm_fn_asc.py x: 2 # contributors y: 369 lines of code tile_kernels/mhc/norm_fn_cuda.py x: 2 # contributors y: 131 lines of code tile_kernels/mhc/norm_fn_kernel.py x: 2 # contributors y: 117 lines of code tile_kernels/mhc/post_cuda.py x: 2 # contributors y: 127 lines of code tile_kernels/mhc/pre_apply_mix_asc.py x: 2 # contributors y: 99 lines of code tile_kernels/mhc/pre_apply_mix_cuda.py x: 2 # contributors y: 92 lines of code tile_kernels/mhc/pre_big_fuse_asc.py x: 2 # contributors y: 354 lines of code tile_kernels/mhc/pre_big_fuse_cuda.py x: 2 # contributors y: 106 lines of code tile_kernels/mhc/pre_split_mixes_asc.py x: 2 # contributors y: 414 lines of code tile_kernels/mhc/pre_split_mixes_cuda.py x: 2 # contributors y: 177 lines of code tile_kernels/mhc/sinkhorn_asc.py x: 2 # contributors y: 245 lines of code tile_kernels/mhc/sinkhorn_kernel.py x: 2 # contributors y: 15 lines of code tile_kernels/modeling/__init__.py x: 2 # contributors y: 1 lines of code tile_kernels/moe/__init__.py x: 2 # contributors y: 4 lines of code tile_kernels/moe/moe_topk_gate_backward_asc.py x: 2 # contributors y: 148 lines of code tile_kernels/moe/moe_topk_gate_backward_kernel.py x: 2 # contributors y: 59 lines of code tile_kernels/moe/moe_topk_gate_forward_asc.py x: 2 # contributors y: 265 lines of code tile_kernels/moe/moe_topk_gate_forward_cuda.py x: 2 # contributors y: 220 lines of code tile_kernels/moe/moe_topk_gate_forward_kernel.py x: 2 # contributors y: 110 lines of code tile_kernels/moe/normalize_weight_cuda.py x: 2 # contributors y: 41 lines of code tile_kernels/quant/batched_transpose_weight_sf_cuda.py x: 2 # contributors y: 78 lines of code tile_kernels/quant/batched_transpose_weight_sf_kernel.py x: 2 # contributors y: 72 lines of code tile_kernels/quant/cast_back_asc.py x: 2 # contributors y: 190 lines of code tile_kernels/quant/common.py x: 2 # contributors y: 261 lines of code tile_kernels/quant/norm_backward_asc.py x: 2 # contributors y: 172 lines of code tile_kernels/quant/norm_backward_cuda.py x: 2 # contributors y: 192 lines of code tile_kernels/quant/norm_forward_asc.py x: 2 # contributors y: 379 lines of code tile_kernels/quant/norm_forward_cuda.py x: 2 # contributors y: 217 lines of code tile_kernels/quant/norm_forward_kernel.py x: 2 # contributors y: 194 lines of code tile_kernels/quant/per_block_cast_cuda.py x: 2 # contributors y: 122 lines of code tile_kernels/quant/per_block_cast_kernel.py x: 2 # contributors y: 134 lines of code tile_kernels/quant/per_channel_cast_cuda.py x: 2 # contributors y: 136 lines of code tile_kernels/quant/per_token_cast_and_cast_back_asc.py x: 2 # contributors y: 330 lines of code tile_kernels/quant/per_token_cast_and_cast_back_cuda.py x: 2 # contributors y: 68 lines of code tile_kernels/quant/per_token_cast_asc.py x: 2 # contributors y: 336 lines of code tile_kernels/quant/per_token_cast_cuda.py x: 2 # contributors y: 179 lines of code tile_kernels/quant/swiglu_backward_asc.py x: 2 # contributors y: 445 lines of code tile_kernels/quant/swiglu_backward_cuda.py x: 2 # contributors y: 211 lines of code tile_kernels/quant/swiglu_forward_asc.py x: 2 # contributors y: 347 lines of code tile_kernels/quant/swiglu_forward_cuda.py x: 2 # contributors y: 260 lines of code tile_kernels/quant/swiglu_forward_kernel.py x: 2 # contributors y: 120 lines of code tile_kernels/testing/moe.py x: 2 # contributors y: 27 lines of code tile_kernels/testing/numeric.py x: 2 # contributors y: 52 lines of code tile_kernels/testing/pytest/benchmark.py x: 2 # contributors y: 656 lines of code tile_kernels/testing/pytest/common.py x: 2 # contributors y: 25 lines of code tile_kernels/testing/pytest/gpu_mem.py x: 2 # contributors y: 319 lines of code tile_kernels/torch/cast.py x: 2 # contributors y: 255 lines of code tile_kernels/torch/mhc.py x: 2 # contributors y: 86 lines of code tile_kernels/torch/rope.py x: 2 # contributors y: 42 lines of code tile_kernels/torch/swiglu.py x: 2 # contributors y: 143 lines of code tile_kernels/torch/topk.py x: 2 # contributors y: 141 lines of code tile_kernels/transform/rope_asc.py x: 2 # contributors y: 185 lines of code tile_kernels/transform/rope_cuda.py x: 2 # contributors y: 115 lines of code
656.0
lines of code
  min: 1.0
  average: 117.3
  25th percentile: 38.75
  median: 85.5
  75th percentile: 166.0
  max: 656.0
0 2.0
# contributors
min: 2.0 | average: 2.0 | 25th percentile: 2.0 | median: 2.0 | 75th percentile: 2.0 | max: 2.0

Number of Changes vs. Line Churn: 230 points

.gitignore x: 1 # changes y: 6 line churn .pre-commit-config.yaml x: 1 # changes y: 32 line churn README.md x: 1 # changes y: 120 line churn pyproject.toml x: 1 # changes y: 84 line churn tests/conftest.py x: 1 # changes y: 314 line churn tests/engram/test_engram_fused_weight.py x: 1 # changes y: 48 line churn tests/engram/test_engram_gate_bwd.py x: 1 # changes y: 268 line churn tests/engram/test_engram_gate_fwd.py x: 1 # changes y: 306 line churn tests/engram/test_engram_grad_w_reduce.py x: 1 # changes y: 108 line churn tests/engram/test_engram_hash.py x: 1 # changes y: 140 line churn tests/engram/test_engram_sinkhorn.py x: 1 # changes y: 410 line churn tests/mhc/test_expand.py x: 1 # changes y: 250 line churn tests/mhc/test_head_compute_mix.py x: 1 # changes y: 422 line churn tests/mhc/test_norm_fn.py x: 1 # changes y: 540 line churn tests/mhc/test_post.py x: 1 # changes y: 326 line churn tests/mhc/test_pre_apply_mix.py x: 1 # changes y: 280 line churn tests/mhc/test_pre_big_fuse.py x: 1 # changes y: 510 line churn tests/mhc/test_pre_split_mixes.py x: 1 # changes y: 622 line churn tests/mhc/test_sinkhorn.py x: 1 # changes y: 332 line churn tests/moe/test_aux_fi.py x: 1 # changes y: 130 line churn tests/moe/test_expand_to_fused.py x: 1 # changes y: 300 line churn tests/moe/test_get_fused_mapping.py x: 1 # changes y: 206 line churn tests/moe/test_inplace_unique_group_indices.py x: 1 # changes y: 158 line churn tests/moe/test_moe_topk_gate_backward.py x: 1 # changes y: 456 line churn tests/moe/test_moe_topk_gate_forward.py x: 1 # changes y: 634 line churn tests/moe/test_normalize_weight.py x: 1 # changes y: 42 line churn tests/moe/test_reduce_fused.py x: 1 # changes y: 224 line churn tests/moe/test_top2_sum_gate.py x: 1 # changes y: 718 line churn tests/moe/test_topk_gate.py x: 1 # changes y: 58 line churn tests/moe/test_topk_sum_and_topk_idx.py x: 1 # changes y: 172 line churn tests/pytest_benchmark_plugin.py x: 1 # changes y: 954 line churn tests/quant/test_cast.py x: 1 # changes y: 292 line churn tests/quant/test_cast_back.py x: 1 # changes y: 164 line churn tests/quant/test_cast_back_e5m6.py x: 1 # changes y: 240 line churn tests/quant/test_norm_forward.py x: 1 # changes y: 444 line churn tests/quant/test_per_block_cast.py x: 1 # changes y: 80 line churn tests/quant/test_per_block_cast_lossless.py x: 1 # changes y: 74 line churn tests/quant/test_per_channel_cast_and_transpose.py x: 1 # changes y: 154 line churn tests/quant/test_per_channel_cast_fused.py x: 1 # changes y: 208 line churn tests/quant/test_per_token_cast.py x: 1 # changes y: 378 line churn tests/quant/test_per_token_cast_to_e5m6.py x: 1 # changes y: 236 line churn tests/quant/test_swiglu_backward.py x: 1 # changes y: 500 line churn tests/quant/test_swiglu_forward.py x: 1 # changes y: 406 line churn tests/quant/test_swiglu_forward_and_per_token_cast.py x: 1 # changes y: 358 line churn tests/rand/test_randn.py x: 1 # changes y: 184 line churn tile_kernels/engram/__init__.py x: 1 # changes y: 18 line churn tile_kernels/engram/engram_fused_weight_asc.py x: 1 # changes y: 114 line churn tile_kernels/engram/engram_gate_bwd_cuda.py x: 1 # changes y: 680 line churn tile_kernels/engram/engram_gate_fwd_cuda.py x: 1 # changes y: 398 line churn tile_kernels/engram/engram_gate_kernel.py x: 2 # changes y: 1148 line churn tile_kernels/engram/engram_sinkhorn_asc.py x: 1 # changes y: 588 line churn tile_kernels/mhc/expand_kernel.py x: 1 # changes y: 190 line churn tile_kernels/mhc/multilayer_recompute_cuda.py x: 1 # changes y: 254 line churn tile_kernels/mhc/norm_fn_asc.py x: 1 # changes y: 856 line churn tile_kernels/mhc/norm_fn_kernel.py x: 1 # changes y: 790 line churn tile_kernels/mhc/pre_big_fuse_asc.py x: 1 # changes y: 846 line churn tile_kernels/mhc/pre_big_fuse_kernel.py x: 1 # changes y: 352 line churn tile_kernels/mhc/sinkhorn_asc.py x: 1 # changes y: 778 line churn tile_kernels/mhc/sinkhorn_kernel.py x: 1 # changes y: 348 line churn tile_kernels/modeling/mhc/ops/sinkhorn.py x: 1 # changes y: 64 line churn tile_kernels/moe/aux_fi_kernel.py x: 1 # changes y: 144 line churn tile_kernels/moe/moe_topk_gate_forward_asc.py x: 1 # changes y: 562 line churn tile_kernels/moe/moe_topk_gate_forward_cuda.py x: 1 # changes y: 586 line churn tile_kernels/moe/normalize_weight_asc.py x: 1 # changes y: 96 line churn tile_kernels/moe/normalize_weight_cuda.py x: 1 # changes y: 100 line churn tile_kernels/moe/sqrt_softplus.py x: 1 # changes y: 288 line churn tile_kernels/moe/top2_sum_gate_kernel.py x: 1 # changes y: 848 line churn tile_kernels/quant/__init__.py x: 1 # changes y: 24 line churn tile_kernels/quant/batched_transpose_weight_sf_kernel.py x: 1 # changes y: 194 line churn tile_kernels/quant/common.py x: 1 # changes y: 548 line churn tile_kernels/quant/norm_forward_cuda.py x: 1 # changes y: 488 line churn tile_kernels/quant/per_channel_cast_asc.py x: 1 # changes y: 614 line churn tile_kernels/quant/per_token_cast_and_cast_back_asc.py x: 1 # changes y: 770 line churn tile_kernels/quant/per_token_cast_asc.py x: 1 # changes y: 740 line churn tile_kernels/quant/per_token_cast_kernel.py x: 1 # changes y: 554 line churn tile_kernels/quant/per_token_cast_to_e5m6_kernel.py x: 1 # changes y: 430 line churn tile_kernels/quant/swiglu_backward_and_per_token_cast_kernel.py x: 1 # changes y: 474 line churn tile_kernels/quant/swiglu_backward_asc.py x: 1 # changes y: 1000 line churn tile_kernels/quant/swiglu_backward_cuda.py x: 1 # changes y: 524 line churn tile_kernels/quant/swiglu_forward_and_per_channel_cast_and_transpose_kernel.py x: 1 # changes y: 462 line churn tile_kernels/quant/swiglu_forward_cuda.py x: 1 # changes y: 646 line churn tile_kernels/rand/__init__.py x: 1 # changes y: 8 line churn tile_kernels/testing/pytest/benchmark.py x: 1 # changes y: 1886 line churn tile_kernels/testing/pytest/gpu_mem.py x: 1 # changes y: 874 line churn tile_kernels/torch/cast_e5m6.py x: 1 # changes y: 516 line churn tile_kernels/transform/rope_cuda.py x: 1 # changes y: 264 line churn
1886.0
line churn
  min: 2.0
  average: 292.73
  25th percentile: 121.5
  median: 235.0
  75th percentile: 406.0
  max: 1886.0
0 2.0
# changes
min: 1.0 | average: 1.0 | 25th percentile: 1.0 | median: 1.0 | 75th percentile: 1.0 | max: 2.0

Number of Contributors vs. Line Churn: 138 points

.pre-commit-config.yaml x: 2 # contributors y: 32 line churn tile_kernels/__init__.py x: 2 # contributors y: 56 line churn tile_kernels/config.py x: 2 # contributors y: 214 line churn tile_kernels/engram/__init__.py x: 2 # contributors y: 18 line churn tile_kernels/engram/engram_fused_weight_asc.py x: 2 # contributors y: 114 line churn tile_kernels/engram/engram_fused_weight_cuda.py x: 2 # contributors y: 84 line churn tile_kernels/engram/engram_fused_weight_kernel.py x: 2 # contributors y: 108 line churn tile_kernels/engram/engram_gate_bwd_asc.py x: 2 # contributors y: 402 line churn tile_kernels/engram/engram_gate_bwd_cuda.py x: 2 # contributors y: 680 line churn tile_kernels/engram/engram_gate_bwd_kernel.py x: 2 # contributors y: 172 line churn tile_kernels/engram/engram_gate_fwd_asc.py x: 2 # contributors y: 506 line churn tile_kernels/engram/engram_gate_fwd_cuda.py x: 2 # contributors y: 398 line churn tile_kernels/engram/engram_gate_fwd_kernel.py x: 2 # contributors y: 170 line churn tile_kernels/engram/engram_grad_w_reduce_asc.py x: 2 # contributors y: 152 line churn tile_kernels/engram/engram_grad_w_reduce_cuda.py x: 2 # contributors y: 134 line churn tile_kernels/engram/engram_sinkhorn_asc.py x: 2 # contributors y: 588 line churn tile_kernels/engram/engram_sinkhorn_cuda.py x: 2 # contributors y: 500 line churn tile_kernels/engram/engram_sinkhorn_kernel.py x: 2 # contributors y: 358 line churn tile_kernels/mhc/expand_kernel.py x: 2 # contributors y: 190 line churn tile_kernels/mhc/head_compute_mix_kernel.py x: 2 # contributors y: 376 line churn tile_kernels/mhc/multilayer_recompute_asc.py x: 2 # contributors y: 314 line churn tile_kernels/mhc/multilayer_recompute_cuda.py x: 2 # contributors y: 254 line churn tile_kernels/mhc/multilayer_recompute_kernel.py x: 2 # contributors y: 296 line churn tile_kernels/mhc/norm_fn_asc.py x: 2 # contributors y: 856 line churn tile_kernels/mhc/norm_fn_cuda.py x: 2 # contributors y: 310 line churn tile_kernels/mhc/norm_fn_kernel.py x: 2 # contributors y: 790 line churn tile_kernels/mhc/post_kernel.py x: 2 # contributors y: 410 line churn tile_kernels/mhc/pre_apply_mix_cuda.py x: 2 # contributors y: 228 line churn tile_kernels/mhc/pre_big_fuse_asc.py x: 2 # contributors y: 846 line churn tile_kernels/mhc/pre_big_fuse_cuda.py x: 2 # contributors y: 272 line churn tile_kernels/mhc/pre_big_fuse_kernel.py x: 2 # contributors y: 352 line churn tile_kernels/mhc/pre_split_mixes_asc.py x: 2 # contributors y: 956 line churn tile_kernels/mhc/pre_split_mixes_cuda.py x: 2 # contributors y: 420 line churn tile_kernels/mhc/sinkhorn_asc.py x: 2 # contributors y: 778 line churn tile_kernels/mhc/sinkhorn_kernel.py x: 2 # contributors y: 348 line churn tile_kernels/modeling/__init__.py x: 2 # contributors y: 2 line churn tile_kernels/moe/moe_topk_gate_backward_asc.py x: 2 # contributors y: 332 line churn tile_kernels/moe/moe_topk_gate_forward_asc.py x: 2 # contributors y: 562 line churn tile_kernels/moe/moe_topk_gate_forward_cuda.py x: 2 # contributors y: 586 line churn tile_kernels/moe/normalize_weight_asc.py x: 2 # contributors y: 96 line churn tile_kernels/moe/normalize_weight_cuda.py x: 2 # contributors y: 100 line churn tile_kernels/moe/sqrt_softplus.py x: 2 # contributors y: 288 line churn tile_kernels/quant/__init__.py x: 2 # contributors y: 24 line churn tile_kernels/quant/batched_transpose_weight_sf_asc.py x: 2 # contributors y: 232 line churn tile_kernels/quant/batched_transpose_weight_sf_cuda.py x: 2 # contributors y: 180 line churn tile_kernels/quant/batched_transpose_weight_sf_kernel.py x: 2 # contributors y: 194 line churn tile_kernels/quant/cast_asc.py x: 2 # contributors y: 276 line churn tile_kernels/quant/cast_back_kernel.py x: 2 # contributors y: 202 line churn tile_kernels/quant/common.py x: 2 # contributors y: 548 line churn tile_kernels/quant/norm_backward_cuda.py x: 2 # contributors y: 454 line churn tile_kernels/quant/norm_forward_cuda.py x: 2 # contributors y: 488 line churn tile_kernels/quant/norm_forward_kernel.py x: 2 # contributors y: 538 line churn tile_kernels/quant/per_block_cast_asc.py x: 2 # contributors y: 442 line churn tile_kernels/quant/per_block_cast_lossless_cuda.py x: 2 # contributors y: 320 line churn tile_kernels/quant/per_channel_cast_asc.py x: 2 # contributors y: 614 line churn tile_kernels/quant/per_channel_cast_kernel.py x: 2 # contributors y: 130 line churn tile_kernels/quant/per_token_cast_and_cast_back_asc.py x: 2 # contributors y: 770 line churn tile_kernels/quant/per_token_cast_and_cast_back_kernel.py x: 2 # contributors y: 122 line churn tile_kernels/quant/per_token_cast_asc.py x: 2 # contributors y: 740 line churn tile_kernels/quant/per_token_cast_kernel.py x: 2 # contributors y: 554 line churn tile_kernels/quant/swiglu_backward_asc.py x: 2 # contributors y: 1000 line churn tile_kernels/quant/swiglu_backward_cuda.py x: 2 # contributors y: 524 line churn tile_kernels/quant/swiglu_backward_kernel.py x: 2 # contributors y: 622 line churn tile_kernels/quant/swiglu_forward_cuda.py x: 2 # contributors y: 646 line churn tile_kernels/rand/__init__.py x: 2 # contributors y: 8 line churn tile_kernels/rand/seed.py x: 2 # contributors y: 40 line churn tile_kernels/testing/moe.py x: 2 # contributors y: 64 line churn tile_kernels/testing/pytest/__init__.py x: 2 # contributors y: 46 line churn tile_kernels/testing/pytest/benchmark.py x: 2 # contributors y: 1886 line churn tile_kernels/testing/pytest/common.py x: 2 # contributors y: 142 line churn tile_kernels/testing/pytest/gpu_mem.py x: 2 # contributors y: 874 line churn tile_kernels/testing/pytest/precompile.py x: 2 # contributors y: 76 line churn tile_kernels/transform/rope_cuda.py x: 2 # contributors y: 264 line churn
1886.0
line churn
  min: 2.0
  average: 308.65
  25th percentile: 121.0
  median: 245.0
  75th percentile: 417.0
  max: 1886.0
0 2.0
# contributors
min: 2.0 | average: 2.0 | 25th percentile: 2.0 | median: 2.0 | 75th percentile: 2.0 | max: 2.0

File Size vs. Line Churn: 138 points

.pre-commit-config.yaml x: 12 lines of code y: 32 line churn tile_kernels/__init__.py x: 36 lines of code y: 56 line churn tile_kernels/config.py x: 90 lines of code y: 214 line churn tile_kernels/engram/__init__.py x: 11 lines of code y: 18 line churn tile_kernels/engram/engram_fused_weight_asc.py x: 49 lines of code y: 114 line churn tile_kernels/engram/engram_fused_weight_cuda.py x: 37 lines of code y: 84 line churn tile_kernels/engram/engram_fused_weight_kernel.py x: 21 lines of code y: 108 line churn tile_kernels/engram/engram_gate_bwd_asc.py x: 181 lines of code y: 402 line churn tile_kernels/engram/engram_gate_bwd_cuda.py x: 274 lines of code y: 680 line churn tile_kernels/engram/engram_gate_bwd_kernel.py x: 61 lines of code y: 172 line churn tile_kernels/engram/engram_gate_fwd_asc.py x: 233 lines of code y: 506 line churn tile_kernels/engram/engram_gate_fwd_cuda.py x: 163 lines of code y: 398 line churn tile_kernels/engram/engram_gate_fwd_kernel.py x: 62 lines of code y: 170 line churn tile_kernels/engram/engram_grad_w_reduce_asc.py x: 66 lines of code y: 152 line churn tile_kernels/engram/engram_grad_w_reduce_cuda.py x: 56 lines of code y: 134 line churn tile_kernels/engram/engram_grad_w_reduce_kernel.py x: 31 lines of code y: 170 line churn tile_kernels/engram/engram_hash_asc.py x: 77 lines of code y: 176 line churn tile_kernels/engram/engram_hash_cuda.py x: 62 lines of code y: 138 line churn tile_kernels/engram/engram_hash_kernel.py x: 34 lines of code y: 174 line churn tile_kernels/engram/engram_sinkhorn_asc.py x: 252 lines of code y: 588 line churn tile_kernels/engram/engram_sinkhorn_cuda.py x: 213 lines of code y: 500 line churn tile_kernels/engram/engram_sinkhorn_kernel.py x: 112 lines of code y: 358 line churn tile_kernels/mhc/__init__.py x: 17 lines of code y: 34 line churn tile_kernels/mhc/expand_asc.py x: 62 lines of code y: 150 line churn tile_kernels/mhc/expand_cuda.py x: 45 lines of code y: 106 line churn tile_kernels/mhc/expand_kernel.py x: 36 lines of code y: 190 line churn tile_kernels/mhc/head_compute_mix_asc.py x: 85 lines of code y: 188 line churn tile_kernels/mhc/head_compute_mix_kernel.py x: 94 lines of code y: 376 line churn tile_kernels/mhc/multilayer_recompute_asc.py x: 138 lines of code y: 314 line churn tile_kernels/mhc/multilayer_recompute_cuda.py x: 103 lines of code y: 254 line churn tile_kernels/mhc/multilayer_recompute_kernel.py x: 85 lines of code y: 296 line churn tile_kernels/mhc/norm_fn_asc.py x: 369 lines of code y: 856 line churn tile_kernels/mhc/norm_fn_cuda.py x: 131 lines of code y: 310 line churn tile_kernels/mhc/norm_fn_kernel.py x: 117 lines of code y: 790 line churn tile_kernels/mhc/post_asc.py x: 164 lines of code y: 408 line churn tile_kernels/mhc/post_cuda.py x: 127 lines of code y: 314 line churn tile_kernels/mhc/post_kernel.py x: 103 lines of code y: 410 line churn tile_kernels/mhc/pre_apply_mix_asc.py x: 99 lines of code y: 254 line churn tile_kernels/mhc/pre_apply_mix_cuda.py x: 92 lines of code y: 228 line churn tile_kernels/mhc/pre_apply_mix_kernel.py x: 44 lines of code y: 290 line churn tile_kernels/mhc/pre_big_fuse_asc.py x: 354 lines of code y: 846 line churn tile_kernels/mhc/pre_big_fuse_cuda.py x: 106 lines of code y: 272 line churn tile_kernels/mhc/pre_big_fuse_kernel.py x: 67 lines of code y: 352 line churn tile_kernels/mhc/pre_split_mixes_asc.py x: 414 lines of code y: 956 line churn tile_kernels/mhc/pre_split_mixes_cuda.py x: 177 lines of code y: 420 line churn tile_kernels/mhc/pre_split_mixes_kernel.py x: 108 lines of code y: 504 line churn tile_kernels/mhc/sinkhorn_asc.py x: 245 lines of code y: 778 line churn tile_kernels/mhc/sinkhorn_cuda.py x: 130 lines of code y: 310 line churn tile_kernels/mhc/sinkhorn_kernel.py x: 15 lines of code y: 348 line churn tile_kernels/modeling/__init__.py x: 1 lines of code y: 2 line churn tile_kernels/modeling/engram/engram_gate.py x: 61 lines of code y: 104 line churn tile_kernels/moe/__init__.py x: 4 lines of code y: 22 line churn tile_kernels/moe/moe_topk_gate_backward_asc.py x: 148 lines of code y: 332 line churn tile_kernels/moe/moe_topk_gate_backward_cuda.py x: 108 lines of code y: 270 line churn tile_kernels/moe/moe_topk_gate_backward_kernel.py x: 59 lines of code y: 190 line churn tile_kernels/moe/moe_topk_gate_forward_asc.py x: 265 lines of code y: 562 line churn tile_kernels/moe/moe_topk_gate_forward_cuda.py x: 220 lines of code y: 586 line churn tile_kernels/moe/moe_topk_gate_forward_kernel.py x: 110 lines of code y: 312 line churn tile_kernels/moe/normalize_weight_asc.py x: 38 lines of code y: 96 line churn tile_kernels/moe/normalize_weight_cuda.py x: 41 lines of code y: 100 line churn tile_kernels/moe/normalize_weight_kernel.py x: 18 lines of code y: 114 line churn tile_kernels/moe/sqrt_softplus.py x: 106 lines of code y: 288 line churn tile_kernels/moe/topk_gate_asc.py x: 117 lines of code y: 314 line churn tile_kernels/moe/topk_gate_kernel.py x: 18 lines of code y: 134 line churn tile_kernels/quant/__init__.py x: 13 lines of code y: 24 line churn tile_kernels/quant/batched_transpose_weight_sf_asc.py x: 102 lines of code y: 232 line churn tile_kernels/quant/batched_transpose_weight_sf_cuda.py x: 78 lines of code y: 180 line churn tile_kernels/quant/batched_transpose_weight_sf_kernel.py x: 72 lines of code y: 194 line churn tile_kernels/quant/cast_asc.py x: 107 lines of code y: 276 line churn tile_kernels/quant/cast_back_asc.py x: 190 lines of code y: 416 line churn tile_kernels/quant/cast_back_cuda.py x: 64 lines of code y: 172 line churn tile_kernels/quant/cast_back_kernel.py x: 48 lines of code y: 202 line churn tile_kernels/quant/cast_cuda.py x: 70 lines of code y: 174 line churn tile_kernels/quant/cast_kernel.py x: 37 lines of code y: 118 line churn tile_kernels/quant/common.py x: 261 lines of code y: 548 line churn tile_kernels/quant/norm_backward_asc.py x: 172 lines of code y: 372 line churn tile_kernels/quant/norm_backward_cuda.py x: 192 lines of code y: 454 line churn tile_kernels/quant/norm_backward_kernel.py x: 62 lines of code y: 174 line churn tile_kernels/quant/norm_forward_asc.py x: 379 lines of code y: 840 line churn tile_kernels/quant/norm_forward_cuda.py x: 217 lines of code y: 488 line churn tile_kernels/quant/norm_forward_kernel.py x: 194 lines of code y: 538 line churn tile_kernels/quant/per_block_cast_asc.py x: 195 lines of code y: 442 line churn tile_kernels/quant/per_block_cast_cuda.py x: 122 lines of code y: 282 line churn tile_kernels/quant/per_block_cast_kernel.py x: 134 lines of code y: 456 line churn tile_kernels/quant/per_block_cast_lossless_asc.py x: 218 lines of code y: 490 line churn tile_kernels/quant/per_block_cast_lossless_cuda.py x: 108 lines of code y: 320 line churn tile_kernels/quant/per_block_cast_lossless_kernel.py x: 49 lines of code y: 348 line churn tile_kernels/quant/per_channel_cast_asc.py x: 262 lines of code y: 614 line churn tile_kernels/quant/per_channel_cast_cuda.py x: 136 lines of code y: 318 line churn tile_kernels/quant/per_channel_cast_kernel.py x: 39 lines of code y: 130 line churn tile_kernels/quant/per_token_cast_and_cast_back_asc.py x: 330 lines of code y: 770 line churn tile_kernels/quant/per_token_cast_and_cast_back_cuda.py x: 68 lines of code y: 174 line churn tile_kernels/quant/per_token_cast_and_cast_back_kernel.py x: 41 lines of code y: 122 line churn tile_kernels/quant/per_token_cast_asc.py x: 336 lines of code y: 740 line churn tile_kernels/quant/per_token_cast_cuda.py x: 179 lines of code y: 442 line churn tile_kernels/quant/per_token_cast_kernel.py x: 149 lines of code y: 554 line churn tile_kernels/quant/rstd.py x: 65 lines of code y: 196 line churn tile_kernels/quant/swiglu_backward_asc.py x: 445 lines of code y: 1000 line churn tile_kernels/quant/swiglu_backward_cuda.py x: 211 lines of code y: 524 line churn tile_kernels/quant/swiglu_backward_kernel.py x: 180 lines of code y: 622 line churn tile_kernels/quant/swiglu_forward_asc.py x: 347 lines of code y: 790 line churn tile_kernels/quant/swiglu_forward_cuda.py x: 260 lines of code y: 646 line churn tile_kernels/quant/swiglu_forward_kernel.py x: 120 lines of code y: 282 line churn tile_kernels/rand/__init__.py x: 3 lines of code y: 8 line churn tile_kernels/rand/randn_asc.py x: 32 lines of code y: 84 line churn tile_kernels/rand/randn_kernel.py x: 40 lines of code y: 110 line churn tile_kernels/rand/seed.py x: 14 lines of code y: 40 line churn tile_kernels/testing/__init__.py x: 2 lines of code y: 4 line churn tile_kernels/testing/bench.py x: 84 lines of code y: 174 line churn tile_kernels/testing/generator.py x: 131 lines of code y: 330 line churn tile_kernels/testing/moe.py x: 27 lines of code y: 64 line churn tile_kernels/testing/numeric.py x: 52 lines of code y: 102 line churn tile_kernels/testing/pytest/__init__.py x: 1 lines of code y: 46 line churn tile_kernels/testing/pytest/benchmark.py x: 656 lines of code y: 1886 line churn tile_kernels/testing/pytest/common.py x: 25 lines of code y: 142 line churn tile_kernels/testing/pytest/gpu_mem.py x: 319 lines of code y: 874 line churn tile_kernels/testing/pytest/precompile.py x: 27 lines of code y: 76 line churn tile_kernels/testing/pytest/tilelang_warning.py x: 35 lines of code y: 140 line churn tile_kernels/testing/pytest/xdist_failfast.py x: 59 lines of code y: 194 line churn tile_kernels/testing/quant.py x: 37 lines of code y: 104 line churn tile_kernels/torch/__init__.py x: 25 lines of code y: 66 line churn tile_kernels/torch/cast.py x: 255 lines of code y: 322 line churn tile_kernels/torch/engram.py x: 108 lines of code y: 236 line churn tile_kernels/torch/mhc.py x: 86 lines of code y: 52 line churn tile_kernels/torch/moe.py x: 21 lines of code y: 142 line churn tile_kernels/torch/rope.py x: 42 lines of code y: 98 line churn tile_kernels/torch/swiglu.py x: 143 lines of code y: 548 line churn tile_kernels/torch/topk.py x: 141 lines of code y: 354 line churn tile_kernels/transform/rope_asc.py x: 185 lines of code y: 400 line churn tile_kernels/transform/rope_cuda.py x: 115 lines of code y: 264 line churn tile_kernels/transform/rope_kernel.py x: 84 lines of code y: 230 line churn tile_kernels/utils.py x: 24 lines of code y: 56 line churn
1886.0
line churn
  min: 2.0
  average: 308.65
  25th percentile: 121.0
  median: 245.0
  75th percentile: 417.0
  max: 1886.0
0 656.0
lines of code
min: 1.0 | average: 117.3 | 25th percentile: 38.75 | median: 85.5 | 75th percentile: 166.0 | max: 656.0