azure / AI-benchmarking-guide

Intro

For duplication, we look at places in code where there are 6 or more lines of code that are exactly the same.
Before duplication is calculated, the code is cleaned to remove empty lines, comments, and frequently duplicated constructs such as imports.
You should aim at having as little as possible (<5%) of duplicated code as high-level of duplication can lead to maintenance difficulties, poor factoring, and logical contradictions.

Learn more...

Duplication Overall

33% duplication:

1,382 cleaned lines of cleaned code (without empty lines, comments, and frequently duplicated constructs such as imports)
461 duplicated lines

80 duplicates

dependency graphs: 2D graph | 3D graph | 3D graph (with duplicates)...

Duplication per Extension

Duplication per Component (primary)

Longest Duplicates

The list of 50 longest duplicates.

See data for all 80 duplicates...

Size	#	Folders	Files	Lines	Code
27	x 2	Benchmarks/AMD Benchmarks/NVIDIA	GEMMHipblasLt.py GEMMCublasLt.py	19:47 (28%) 27:55 (21%)	view
26	x 2	Benchmarks/AMD Benchmarks/NVIDIA	HBMBandwidth.py HBMBandwidth.py	70:98 (30%) 108:137 (22%)	view
19	x 2	Benchmarks/AMD Benchmarks/NVIDIA	HBMBandwidth.py CPUStream.py	70:91 (22%) 100:122 (17%)	view
19	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	CPUStream.py HBMBandwidth.py	100:122 (17%) 108:130 (16%)	view
17	x 2	Benchmarks/AMD Benchmarks/AMD	HBMBandwidth.py TransferBench.py	15:35 (19%) 12:32 (35%)	view
15	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	CPUStream.py HBMBandwidth.py	17:35 (14%) 17:35 (13%)	view
15	x 2	Benchmarks/AMD Benchmarks/AMD	GEMMHipblasLt.py RCCLBandwidth.py	49:64 (15%) 34:49 (15%)	view
14	x 2	Benchmarks/AMD Benchmarks/NVIDIA	TransferBench.py HBMBandwidth.py	15:32 (29%) 17:34 (12%)	view
14	x 2	Benchmarks/AMD Benchmarks/NVIDIA	HBMBandwidth.py CPUStream.py	18:35 (16%) 17:34 (13%)	view
14	x 2	Benchmarks/AMD Benchmarks/NVIDIA	HBMBandwidth.py HBMBandwidth.py	18:35 (16%) 17:34 (12%)	view
14	x 2	Benchmarks/AMD Benchmarks/NVIDIA	TransferBench.py CPUStream.py	15:32 (29%) 17:34 (13%)	view
13	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	HBMBandwidth.py NVBandwidth.py	13:28 (11%) 9:23 (26%)	view
12	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	CPUStream.py HBMBandwidth.py	59:73 (11%) 67:81 (10%)	view
12	x 2	Benchmarks/AMD Benchmarks/AMD	HBMBandwidth.py RCCLBandwidth.py	16:29 (13%) 15:28 (12%)	view
12	x 2	Benchmarks/AMD Benchmarks/NVIDIA	RCCLBandwidth.py NCCLBandwidth.py	19:32 (12%) 18:31 (15%)	view
12	x 2	Benchmarks/AMD Benchmarks/AMD	RCCLBandwidth.py TransferBench.py	15:28 (12%) 13:26 (25%)	view
11	x 2	Benchmarks/AMD Benchmarks/NVIDIA	FIO.py FIO.py	11:21 (28%) 11:21 (30%)	view
11	x 2	Benchmarks/AMD Benchmarks/NVIDIA	FIO.py FIO.py	32:43 (28%) 31:42 (30%)	view
10	x 2	Benchmarks/AMD Benchmarks/NVIDIA	TransferBench.py NVBandwidth.py	15:26 (20%) 12:23 (20%)	view
10	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	CPUStream.py HBMBandwidth.py	84:96 (9%) 92:104 (8%)	view
10	x 2	Benchmarks/AMD Benchmarks/NVIDIA	HBMBandwidth.py NVBandwidth.py	18:29 (11%) 12:23 (20%)	view
10	x 2	Benchmarks/AMD Benchmarks/NVIDIA	RCCLBandwidth.py CPUStream.py	17:28 (10%) 17:28 (9%)	view
10	x 2	Benchmarks/AMD Benchmarks/NVIDIA	RCCLBandwidth.py NVBandwidth.py	17:28 (10%) 12:23 (20%)	view
10	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	CPUStream.py NVBandwidth.py	17:28 (9%) 12:23 (20%)	view
10	x 2	Benchmarks/AMD Benchmarks/NVIDIA	RCCLBandwidth.py HBMBandwidth.py	17:28 (10%) 17:28 (8%)	view
9	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	CPUStream.py NCCLBandwidth.py	19:28 (8%) 18:27 (11%)	view
9	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	NCCLBandwidth.py NVBandwidth.py	18:27 (11%) 14:23 (18%)	view
9	x 2	Benchmarks/AMD Benchmarks/NVIDIA	TransferBench.py NCCLBandwidth.py	17:26 (18%) 18:27 (11%)	view
9	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	HBMBandwidth.py NCCLBandwidth.py	19:28 (7%) 18:27 (11%)	view
9	x 2	Benchmarks/AMD Benchmarks/NVIDIA	HBMBandwidth.py NCCLBandwidth.py	20:29 (10%) 18:27 (11%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	GEMMHipblasLt.py CPUStream.py	19:26 (8%) 19:26 (7%)	view
8	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	GEMMCublasLt.py LLMBenchmark.py	27:34 (6%) 20:27 (8%)	view
8	x 2	Benchmarks/AMD Benchmarks/AMD	GEMMHipblasLt.py LLMBenchmark.py	19:26 (8%) 18:25 (9%)	view
8	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	HBMBandwidth.py LLMBenchmark.py	19:26 (6%) 20:27 (8%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	LLMBenchmark.py NCCLBandwidth.py	18:25 (9%) 18:25 (10%)	view
8	x 2	Benchmarks/AMD Benchmarks/AMD	GEMMHipblasLt.py HBMBandwidth.py	19:26 (8%) 20:27 (9%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	RCCLBandwidth.py GEMMCublasLt.py	19:26 (8%) 27:34 (6%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	RCCLBandwidth.py LLMBenchmark.py	19:26 (8%) 20:27 (8%)	view
8	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	LLMBenchmark.py NCCLBandwidth.py	20:27 (8%) 18:25 (10%)	view
8	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	GEMMCublasLt.py NVBandwidth.py	27:34 (6%) 14:21 (16%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	GEMMHipblasLt.py LLMBenchmark.py	19:26 (8%) 20:27 (8%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	LLMBenchmark.py NVBandwidth.py	18:25 (9%) 14:21 (16%)	view
8	x 2	Benchmarks/AMD Benchmarks/AMD	HBMBandwidth.py LLMBenchmark.py	20:27 (9%) 18:25 (9%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	GEMMHipblasLt.py NCCLBandwidth.py	19:26 (8%) 18:25 (10%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	GEMMHipblasLt.py HBMBandwidth.py	19:26 (8%) 19:26 (6%)	view
8	x 2	ROOT ROOT	AMD_runner.py NVIDIA_runner.py	72:81 (9%) 94:103 (6%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	LLMBenchmark.py CPUStream.py	18:25 (9%) 19:26 (7%)	view
8	x 2	Benchmarks/NVIDIA Benchmarks/NVIDIA	GEMMCublasLt.py HBMBandwidth.py	27:34 (6%) 19:26 (6%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	LLMBenchmark.py GEMMCublasLt.py	18:25 (9%) 27:34 (6%)	view
8	x 2	Benchmarks/AMD Benchmarks/NVIDIA	TransferBench.py LLMBenchmark.py	17:24 (16%) 20:27 (8%)	view