mirror of
https://github.com/luxfi/bench.git
synced 2026-10-30 17:01:57 +00:00
zeekay
65a56203c3
github.com/luxfi/bench existed only as loose files: no .git, no history, no
remote. An rm -rf would have destroyed it with nothing to recover from, and it
was invisible to every repo sweep that looks for a .git directory.
Includes the dependency repair that makes it build again. It previously failed
with:
reading github.com/luxfi/geth/go.mod at revision v1.16.98: unknown revision
The v1.16.x line was DELETED from github.com/luxfi/geth. The requirement came
from github.com/luxfi/crypto, which pins the dead tag AND excludes the
alternative in every version through v1.20.0; v1.20.2 is the first that moved
to geth v1.20.1, so bumping crypto is what escapes it. geth then drops out of
go.mod entirely.
The compiled luxbench binary is ignored rather than committed.
Verified: go build ./... exits 0.
NOTE: no remote — luxfi/bench does not exist on GitHub. This is still one copy
in one place.
57 lines
2.6 KiB
Python
57 lines
2.6 KiB
Python
#!/usr/bin/env python3
|
|
"""Parse the vs-ava raw benchmark outputs into median tables with provenance.
|
|
Usage: python3 parse_results.py [results_dir]"""
|
|
import os, re, sys, statistics, json
|
|
|
|
RD = sys.argv[1] if len(sys.argv) > 1 else os.path.join(os.path.dirname(__file__), "results")
|
|
|
|
def bench_rows(path):
|
|
"""name -> list of (ns_per_op, extra_tx_per_s or None)."""
|
|
rows = {}
|
|
if not os.path.exists(path): return rows
|
|
for ln in open(path, errors="ignore"):
|
|
m = re.match(r"^(Benchmark\S+?)-\d+\s+\d+\s+([\d.]+)\s+ns/op(.*)$", ln)
|
|
if not m: continue
|
|
name = re.sub(r"-\d+$", "", m.group(1))
|
|
ns = float(m.group(2))
|
|
txm = re.search(r"([\d.]+)\s+tx/s", m.group(3))
|
|
rows.setdefault(name, []).append((ns, float(txm.group(1)) if txm else None))
|
|
return rows
|
|
|
|
def med(xs): return statistics.median(xs) if xs else float("nan")
|
|
|
|
def summarize(path, label):
|
|
rows = bench_rows(path)
|
|
print(f"\n== {label} ({os.path.basename(path)}) ==")
|
|
out = {}
|
|
for name in sorted(rows):
|
|
nss = [r[0] for r in rows[name]]
|
|
txs = [r[1] for r in rows[name] if r[1] is not None]
|
|
n = len(nss); mns = med(nss)
|
|
line = f" {name:<46} n={n:2d} median {mns:>12.0f} ns/op"
|
|
if txs: line += f" median {med(txs):>14.0f} tx/s"
|
|
else: line += f" = {1e9/mns:>10.1f} op/s"
|
|
print(line)
|
|
out[name] = {"n": n, "median_ns_per_op": round(mns, 1),
|
|
"median_tx_per_s": round(med(txs), 1) if txs else None,
|
|
"op_per_s": round(1e9/mns, 2) if mns == mns else None}
|
|
return out
|
|
|
|
result = {}
|
|
result["insertchain_lux"] = summarize(os.path.join(RD, "insertchain_lux_raw.txt"), "InsertChain — luxfi/evm (CGO=0)")
|
|
result["insertchain_coreth"] = summarize(os.path.join(RD, "insertchain_coreth_raw.txt"), "InsertChain — ava/coreth (CGO=1)")
|
|
result["blockstm_lux"] = summarize(os.path.join(RD, "blockstm_lux_raw.txt"), "Block-STM parallel — luxfi/evm (CGO=0)")
|
|
|
|
# head-to-head ratio table (coreth as baseline) for the memdb rows both ran
|
|
lux, cor = result["insertchain_lux"], result["insertchain_coreth"]
|
|
if lux and cor:
|
|
print("\n== InsertChain head-to-head (ns/op median; ratio = lux/coreth) ==")
|
|
for b in ["empty", "valueTx", "ring200", "ring1000"]:
|
|
k = f"BenchmarkInsertChain_{b}_memdb"
|
|
if k in lux and k in cor:
|
|
lv, cv = lux[k]["median_ns_per_op"], cor[k]["median_ns_per_op"]
|
|
print(f" {b:<10} lux {lv:>12.0f} coreth {cv:>12.0f} ratio {lv/cv:5.2f}x")
|
|
|
|
open(os.path.join(RD, "summary.json"), "w").write(json.dumps(result, indent=2))
|
|
print(f"\nwrote {os.path.join(RD, 'summary.json')}")
|