')
- print(f"
Hostname: {socket.gethostname()}
")
- print(
- f'
Operating System: {platform.system()} {platform.release()}
'
- )
- print(f'
Kernel Version : {platform.version()}
')
- with open("/proc/cpuinfo") as f:
- print(
- f'
CPU: {next(line.strip().split(": ")[1] for line in f if "model name" in line)} ({psutil.cpu_count(logical=False)} cores)
'
- )
- print(f'
Memory: {get_mem_info()} GB
')
- print(f"
GPU: {get_first_gpu_info()}
")
- print(f"
CUDA Version: {get_cuda_version()}
")
-
-
-if __name__ == "__main__":
- logs_dir = pathlib.Path("logs")
-
- dataset_patt = re.compile(".*ds=([\w-]+).*")
- backend_patt = re.compile(".*backend=(\w+).*")
- k_patt = re.compile(".*k=(\d+).*")
-
- # Organize all benchmark runs by the following hierarchy: algo -> backend -> dataset
- benchmarks = get_all_benchmark_info()
-
- # dump HTML table
- ordered_datasets = [
- "netscience",
- "email_Eu_core",
- "amazon0302",
- "cit-patents",
- "hollywood",
- "soc-livejournal1",
- ]
- # dataset, # Node, # Edge, Directed info
- dataset_meta = {
- "netscience": ["1,461", "5,484", "Yes"],
- "email_Eu_core": ["1,005", "25,571", "Yes"],
- "amazon0302": ["262,111", "1,234,877", "Yes"],
- "cit-patents": ["3,774,768", "16,518,948", "Yes"],
- "hollywood": ["1,139,905", "57,515,616", "No"],
- "soc-livejournal1": ["4,847,571", "68,993,773", "Yes"],
- }
-
- print(
- """
-
-
-
-
-
-
-
- Dataset
Nodes
Edges
Directed | """
- )
- for ds in ordered_datasets:
- print(
- f" {ds}
{dataset_meta[ds][0]}
{dataset_meta[ds][1]}
{dataset_meta[ds][2]}
| "
- )
- print(
- """
-
-
- """
- )
- for algo_name in sorted(benchmarks):
- algo_runs = benchmarks[algo_name]
- print(" ")
- print(f" | {algo_name} | ")
- # Proceed only if any results are present for both cugraph and NX
- if "cugraph" in algo_runs and "networkx" in algo_runs:
- cugraph_algo_runs = algo_runs["cugraph"]
- networkx_algo_runs = algo_runs["networkx"]
- datasets_in_both = set(cugraph_algo_runs).intersection(networkx_algo_runs)
-
- # populate the table with speedup results for each dataset in the order
- # specified in ordered_datasets. If results for a run using a dataset
- # are not present for both cugraph and NX, output an empty cell.
- for dataset in ordered_datasets:
- if dataset in datasets_in_both:
- cugraph_runtime = cugraph_algo_runs[dataset]
- networkx_runtime = networkx_algo_runs[dataset]
- (speedup, runtime_delta) = compute_perf_vals(
- cugraph_runtime=cugraph_runtime,
- networkx_runtime=networkx_runtime,
- )
- nx_formatted = get_formatted_time_value(networkx_runtime)
- cg_formatted = get_formatted_time_value(cugraph_runtime)
- print(
- f" {nx_formatted} / {cg_formatted}
{speedup}
{runtime_delta} | "
- )
- else:
- print(f" | ")
-
- # If a comparison between cugraph and NX cannot be made, output empty cells
- # for each dataset
- else:
- for _ in range(len(ordered_datasets)):
- print(" | ")
- print("
")
- print(
- """
- \n
-
-
-
Table Format:
-
- - NetworkX time / nx-cugraph time
- - Speed-up of using nx-cugraph
- - Time-delta
-
-
"""
- )
- get_system_info()
- print("""
\n