• 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.

• There are 110 files with 17,308 lines of code.
• 3 very long files (3,697 lines of code)
• 4 long files (2,196 lines of code)
• 17 medium size files (5,773 lines of code)
• 17 small files (2,617 lines of code)
• 69 very small files (3,025 lines of code)

There are 27 files with lines longer than 120 characters. In total, there are 221 long lines.

File	# lines	# units	# long lines
metal-kernels.c in gpt_oss/metal/source	1454	27	78
model.c in gpt_oss/metal/source	514	9	28
jsonl cases.jsonl in compatibility-test	30	-	21
context.c in gpt_oss/metal/source	1010	12	19
end-to-end-threadgroup.cc in gpt_oss/metal/benchmark	502	14	14
uuid.h in gpt_oss/metal/source/include/internal	103	4	14
metal sdpa.metal in gpt_oss/metal/source	264	-	9
healthbench_eval.py in gpt_oss/evals	463	15	5
metal moematmul.metal in gpt_oss/metal/source	627	-	4
end-to-end.cc in gpt_oss/metal/benchmark	198	2	3
create-local-model.py in gpt_oss/metal/scripts	288	9	3
generate.c in gpt_oss/metal/source	285	7	3
chat.py in gpt_oss	326	2	3
weights.py in gpt_oss/torch	99	5	2
moe.py in gpt_oss/triton	49	3	2
metal matmul.metal in gpt_oss/metal/source	491	-	2
token_generator.py in gpt_oss/vllm	44	2	1
attention.py in gpt_oss/triton	175	4	1
model.py in gpt_oss/triton	453	21	1
metal sample.metal in gpt_oss/metal/source	188	-	1
metal.m in gpt_oss/metal/source	423	19	1
context.c in gpt_oss/metal/python	236	2	1
tokenizer.c in gpt_oss/metal/python	154	1	1
simple_browser_tool.py in gpt_oss/tools/simple_browser	553	29	1
generate.py in gpt_oss	83	1	1
runCase.ts in compatibility-test	284	6	1
tools.ts in compatibility-test	153	2	1