fulmine, fulmine-tuned: publish under the package name, fulmine.js

[nigrosimone]

The package is published as fulmine.js on npm and the repository is fulmine.js, but the two entries are on the board as fulmine and fulmine-tuned, so the board, the badge and the per-framework page use a name nobody searches for. This changes the display_name in the two meta.json and nothing else about the entries: same directories, same Dockerfiles, same /benchmark -f fulmine.

The rename moves the slug, so site/data/results/fulmine.json becomes fulmine.js.json and the framework field of every row goes with it. That keeps the published history attached to the entry instead of leaving it to be purged as a stale name. scripts/gen_leaderboard_data.py runs clean after it (no unknown-name warning, the two framework pages are written) and scripts/check_badge_parity.js passes, with site/leaderboard/data.js regenerated here.

On both entries the two POST routes now read the body with the framework's parsers, express.text({ type: () => true }) on /baseline11 and express.raw({ type: () => true }) on /echo, instead of collecting chunks on 'data'. The parser takes the whole body off uWS in one piece, chunked or not, and on a laptop it costs 5 to 7 us less per POST. The type function is there because the load generator sends no Content-Type. Checked with the validator's shapes: fragmented, lower-cased names, chunked, echo 1 B to 100 KB, chunked echo, empty body, on 8080 and 8081.

The saved run measured the parser change: baseline +11.9%, short-lived +12.6%, 8gbit 72.0 -> 68.4 us/req and p99.9 6.8 -> 4.3 ms, the other rows flat. The results commit also brought back fulmine.json and fulmine-tuned.json with the old rows, because the compare step copies main's site/data over the branch and never removes the files the branch renamed; they are dropped again in the last commit and data.js is regenerated.

[nigrosimone]

/benchmark-multiple -f fulmine,fulmine-tuned --save

[github-actions]

👋 Benchmark request received. A collaborator will review and approve the run.

[github-actions]

⚠️ /benchmark --save cannot start: main has diverged and cannot be auto-merged into this branch. Please merge or rebase main manually, push, and re-run /benchmark --save.

[nigrosimone]

/benchmark-multiple -f fulmine,fulmine-tuned --save

[github-actions]

👋 Benchmark request received. A collaborator will review and approve the run.

[github-actions]

Benchmark Results

Frameworks: 2 | Test: all tests

✅ fulmine

Test	Conn	RPS	Rate	CPU/req	p99	p99.9	CPU	Mem	Δ RPS	Δ Mem
baseline	4096	1,646,820	—	—	—	—	6791.8%	2.6GiB	+11.9%	-3.7%
pipelined	4096	32,530,268	—	—	—	—	6420.8%	2.1GiB	+2.1%	~0%
limited-conn	4096	1,485,268	—	—	—	—	6611.2%	2.7GiB	+12.6%	~0%
json-comp	4096	478,013	—	—	—	—	6373.7%	7.6GiB	+0.7%	+1.3%
json-comp	16384	494,728	—	—	—	—	6222.3%	8.5GiB	+0.8%	~0%
json-tls	4096	1,022,233	—	—	—	—	6410.9%	7.1GiB	+0.2%	-13.4%
8gbit	512	49,377	0.9875	68.39us	159.0us	4270.0us	330.3%	2.7GiB	+0.1%	~0%
static-tls	1024	594,289	—	—	—	—	6594.5%	3.2GiB	-2.0%	-3.0%
async-db	1024	233,634	—	—	—	—	5770.7%	4.3GiB	+1.9%	+16.2%
fortunes	1024	68,675	—	—	—	—	6497.8%	4.2GiB	+0.7%	~0%
gateway-64	512	279,500	—	—	—	—	5185.7%	4.8GiB	+18.6%	-4.0%
gateway-64	1024	264,856	—	—	—	—	5391.7%	5.5GiB	+13.4%	-3.5%
gateway-h3	64	90,254	—	—	—	—	3337.2%	3.5GiB	-1.0%	~0%
gateway-h3	256	61,282	—	—	—	—	2408.6%	4.6GiB	+0.1%	+2.2%
production-stack	256	116,894	—	—	—	—	2769.4%	3.6GiB	-12.3%	+2.9%
production-stack	1024	105,309	—	—	—	—	2641.7%	4.3GiB	-4.0%	~0%
echo-ws	512	3,689,829	—	—	—	—	6512.5%	2.3GiB	-0.9%	~0%
echo-ws	4096	3,777,226	—	—	—	—	6276.7%	2.4GiB	-0.6%	~0%
echo-ws	16384	3,523,105	—	—	—	—	6131.7%	2.5GiB	-0.6%	~0%
echo-ws-pipeline	512	24,189,686	—	—	—	—	6726.6%	2.3GiB	-0.4%	~0%
echo-ws-pipeline	4096	24,296,604	—	—	—	—	6538.9%	2.4GiB	+0.3%	~0%
echo-ws-pipeline	16384	23,202,560	—	—	—	—	6313.0%	2.6GiB	+0.2%	~0%
echo-ws-limited	512	2,076,963	—	—	—	—	5635.9%	2.3GiB	-0.2%	~0%
echo-ws-limited	4096	2,551,279	—	—	—	—	6264.8%	2.4GiB	-0.4%	~0%
latency-1m	1024	997,869	0.9979	33.66us	263.0us	1646.0us	3911.7%	2.5GiB	~0%	~0%
latency-10k	1024	9,982	0.9982	39.08us	121.0us	373.0us	37.2%	2.1GiB	~0%	~0%
async	32000	2,076,210	—	—	—	—	6403.9%	5.7GiB	-0.6%	~0%

✅ fulmine-tuned

Test	Conn	RPS	Rate	CPU/req	p99	p99.9	CPU	Mem	Δ RPS	Δ Mem
baseline	4096	1,595,557	—	—	—	—	6551.1%	2.6GiB	+10.0%	-3.7%
pipelined	4096	32,750,279	—	—	—	—	6349.1%	2.1GiB	-0.7%	~0%
limited-conn	4096	1,450,810	—	—	—	—	6354.8%	2.6GiB	+9.0%	-7.1%
json-comp	4096	474,140	—	—	—	—	6357.3%	7.6GiB	-0.6%	+1.3%
json-comp	16384	491,530	—	—	—	—	6245.6%	8.5GiB	-0.2%	+1.2%
json-tls	4096	1,030,277	—	—	—	—	6575.2%	8.2GiB	-1.6%	+18.8%
8gbit	512	49,364	0.9873	69.09us	162.0us	5786.0us	318.9%	2.7GiB	~0%	~0%
static-tls	1024	595,686	—	—	—	—	6585.4%	3.3GiB	-2.2%	+3.1%
async-db	1024	239,935	—	—	—	—	5813.2%	6.2GiB	-4.6%	~0%
fortunes	1024	70,705	—	—	—	—	6506.8%	6.1GiB	-0.2%	~0%
gateway-64	512	314,541	—	—	—	—	5473.9%	5.1GiB	-0.7%	-3.8%
gateway-64	1024	311,169	—	—	—	—	6032.4%	6.5GiB	+0.7%	~0%
gateway-h3	64	90,285	—	—	—	—	3252.8%	3.4GiB	-0.2%	-5.6%
gateway-h3	256	62,553	—	—	—	—	2434.9%	4.7GiB	+1.7%	+23.7%
production-stack	256	120,770	—	—	—	—	2878.3%	3.7GiB	-16.0%	~0%
production-stack	1024	105,395	—	—	—	—	2639.5%	4.4GiB	-3.3%	~0%
echo-ws	512	3,701,735	—	—	—	—	6393.0%	2.3GiB	+0.5%	~0%
echo-ws	4096	3,772,036	—	—	—	—	6279.7%	2.4GiB	-0.6%	+4.3%
echo-ws	16384	3,538,165	—	—	—	—	6173.2%	2.5GiB	-0.2%	~0%
echo-ws-pipeline	512	23,989,939	—	—	—	—	6721.9%	2.3GiB	-1.0%	~0%
echo-ws-pipeline	4096	24,204,617	—	—	—	—	6533.3%	2.4GiB	-0.3%	~0%
echo-ws-pipeline	16384	22,945,078	—	—	—	—	6261.9%	2.6GiB	-0.6%	+4.0%
echo-ws-limited	512	2,084,011	—	—	—	—	5672.1%	2.4GiB	-0.3%	+4.3%
echo-ws-limited	4096	2,553,620	—	—	—	—	6262.5%	2.4GiB	-0.4%	~0%
latency-1m	1024	998,214	0.9982	33.36us	263.0us	792.0us	3670.3%	2.5GiB	~0%	~0%
latency-10k	1024	9,982	0.9982	39.76us	122.0us	210.0us	37.7%	2.1GiB	~0%	~0%
async	32000	2,110,747	—	—	—	—	6367.9%	5.7GiB	-0.7%	~0%

[nigrosimone]

Okay, the results are good for me.