Benchmarks

This page compares VarMQ with other Go worker pools, Redis job queues and SQLite job queues. Every row on it is one line of the raw go test -bench output in goptics/benchmarks, and every figure is read from those rows or averaged over them. The run is from 2026-10-25, with varmq v1.10.0, at commit a0d0e0b. The page holds all 1,667 rows of that run. Each result links to its line in the raw file.

Explore the rows

Pick a dataset and narrow it with the selects: each lane is one subject’s mean over the groups in scope, and its last column divides it by VarMQ. Press a square in the map to see one group, with a link to each of its rows in the raw file.

All pools24 CPUs8 CPUs4 CPUsRedisSQLite

Worker pool, GOMAXPROCS 24, 8 and 4

The three pool files read together. A mean here averages the 24, 8 and 4 CPU runs.

Show

ThroughputMemory

Channel baselines †

165 groups in scope · VarMQ leads 91, others 74 · 9 subjects

Tasks per second

tasks/s · higher is better · mean of 165 groups

SubjectRate× VarMQ

1. VarMQwon 91 of 165 4,479,925 1.00×1.00 times VarMQ, the baseline
2. BufferedPool †won 47 of 165 3,694,917 0.82×0.82 times VarMQ, behind VarMQ
3. ErrGroupwon 22 of 165 2,131,294 0.48×0.48 times VarMQ, behind VarMQ
4. Gopoolwon 1 of 165 2,027,999 0.45×0.45 times VarMQ, behind VarMQ
5. Antswon 4 of 165 1,960,521 0.44×0.44 times VarMQ, behind VarMQ
6. GoroutinePool †won 0 of 165 1,625,000 0.36×0.36 times VarMQ, behind VarMQ
7. Concwon 0 of 165 1,326,509 0.30×0.30 times VarMQ, behind VarMQ
8. WorkerPoolwon 0 of 165 1,221,358 0.27×0.27 times VarMQ, behind VarMQ
9. Pondwon 0 of 165 655,940 0.15×0.15 times VarMQ, behind VarMQ

The dashed line marks VarMQ. An amber ratio means the lane is ahead of it. The “×” column divides the lane by VarMQ.

† A baseline, not a library: N goroutines reading one channel. The buffered one holds all 1M tasks.

Who leads each group

one square = one group of rows

A filled green square: a VarMQ subject leads the group. An outlined amber square: another subject leads, and its code is in the square. A dashed square with “=”: a VarMQ subject and another subject share the best value, so neither leads. Press a square to see that group in the lanes above.

	24 CPUs					8 CPUs					4 CPUs
Task · workers	11 producer	100100 producers	1K1K producers	10K10K producers	1M1M producers	11 producer	100100 producers	1K1K producers	10K10K producers	1M1M producers	11 producer	100100 producers	1K1K producers	10K10K producers	1M1M producers
Sleep10ms · 50K	VQ	BP	BP	BP	VQ	BP	BP	BP	BP	EG	BP	BP	BP	BP	EG
Loop10 · 50K	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 50K	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	BP	BP	BP	EG
Sleep10ms · 100K	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	EG	EG	BP	BP	BP	EG
Loop10 · 100K	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 100K	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	EG	VQ	VQ	BP	VQ	EG
Sleep10ms · 300K	EG	VQ	VQ	VQ	EG	EG	VQ	VQ	EG	EG	EG	AN	AN	EG	EG
Loop10 · 300K	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 300K	EG	VQ	VQ	VQ	VQ	EG	VQ	VQ	VQ	EG	EG	AN	AN	EG	EG
Loop10 · 1 worker	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP
Loop10K · 1 worker	GP	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP	BP

Hover or focus a square to read it. Press it to show that group in the lanes.

• VQVarMQ91
• BPBufferedPool †47
• EGErrGroup22
• ANAnts4
• GPGopool1

Where VarMQ does not win

Another subject has the fastest throughput in 74 groups of 165. VarMQ has it in 91.

• BufferedPool † 47 groups
• ErrGroup 22 groups
• Ants 4 groups
• Gopool 1 group

All 74 groups where VarMQ does not win, biggest gap first

Group	Best	VarMQ	Ahead by	Source
Loop10, 1 worker, 1 producer, 24 CPUs	BufferedPool † 26,076,554	5,612,142 3 of 9	4.65×	24cpu L410–418
Loop10, 1 worker, 1K producers, 24 CPUs	BufferedPool † 24,944,261	5,587,181 2 of 9	4.46×	24cpu L428–436
Loop10, 1 worker, 1 producer, 8 CPUs	BufferedPool † 26,148,651	6,326,035 3 of 9	4.13×	8cpu L410–418
Loop10, 1 worker, 1 producer, 4 CPUs	BufferedPool † 26,202,379	6,632,332 3 of 9	3.95×	4cpu L410–418
Loop10, 1 worker, 100 producers, 8 CPUs	BufferedPool † 27,854,937	7,090,918 2 of 9	3.93×	8cpu L419–427
Loop10, 1 worker, 100 producers, 24 CPUs	BufferedPool † 25,921,880	6,768,459 2 of 9	3.83×	24cpu L419–427
Loop10, 1 worker, 100 producers, 4 CPUs	BufferedPool † 27,380,238	7,308,651 2 of 9	3.75×	4cpu L419–427
Loop10, 1 worker, 1K producers, 4 CPUs	BufferedPool † 26,829,003	7,342,679 2 of 9	3.65×	4cpu L428–436
Loop10, 1 worker, 1K producers, 8 CPUs	BufferedPool † 26,084,154	7,190,233 2 of 9	3.63×	8cpu L428–436
Loop10, 1 worker, 10K producers, 8 CPUs	BufferedPool † 24,737,062	7,177,970 2 of 9	3.45×	8cpu L437–445
Loop10, 1 worker, 10K producers, 4 CPUs	BufferedPool † 23,643,298	7,271,060 2 of 9	3.25×	4cpu L437–445
Loop10, 1 worker, 10K producers, 24 CPUs	BufferedPool † 21,230,114	6,584,676 2 of 9	3.22×	24cpu L437–445
Sleep10ms, 50K workers, 1 producer, 4 CPUs	BufferedPool † 3,736,757	2,504,306 2 of 9	1.49×	4cpu L5–13
Sleep10ms, 50K workers, 1K producers, 4 CPUs	BufferedPool † 3,674,940	2,481,416 2 of 9	1.48×	4cpu L23–31
Sleep10ms, 50K workers, 10K producers, 4 CPUs	BufferedPool † 3,662,015	2,491,621 2 of 9	1.47×	4cpu L32–40
Sleep10ms, 50K workers, 100 producers, 4 CPUs	BufferedPool † 3,673,447	2,554,864 2 of 9	1.44×	4cpu L14–22
Sleep10ms, 100K workers, 1M producers, 4 CPUs	ErrGroup 1,967,247	1,373,342 6 of 9	1.43×	4cpu L176–184
Sleep10ms, 300K workers, 1M producers, 4 CPUs	ErrGroup 1,951,801	1,392,523 4 of 9	1.40×	4cpu L311–319
Sleep5msLoop5, 300K workers, 1M producers, 4 CPUs	ErrGroup 2,107,373	1,521,728 4 of 9	1.38×	4cpu L401–409
Sleep10ms, 50K workers, 1M producers, 4 CPUs	ErrGroup 1,968,705	1,436,960 6 of 9	1.37×	4cpu L41–49
Sleep5msLoop5, 100K workers, 1M producers, 4 CPUs	ErrGroup 2,040,959	1,491,073 4 of 9	1.37×	4cpu L266–274
Sleep5msLoop5, 50K workers, 1M producers, 4 CPUs	ErrGroup 2,047,829	1,499,597 6 of 9	1.37×	4cpu L131–139
Sleep10ms, 50K workers, 1 producer, 8 CPUs	BufferedPool † 3,939,669	3,176,393 2 of 9	1.24×	8cpu L5–13
Sleep5msLoop5, 300K workers, 1M producers, 8 CPUs	ErrGroup 3,068,841	2,482,123 2 of 9	1.24×	8cpu L401–409
Sleep5msLoop5, 300K workers, 1 producer, 8 CPUs	ErrGroup 2,997,183	2,459,278 2 of 9	1.22×	8cpu L365–373
Sleep10ms, 50K workers, 10K producers, 8 CPUs	BufferedPool † 3,731,132	3,169,295 2 of 9	1.18×	8cpu L32–40
Loop10, 1 worker, 1M producers, 24 CPUs	BufferedPool † 4,756,005	4,048,978 2 of 9	1.17×	24cpu L446–454
Sleep5msLoop5, 300K workers, 1 producer, 4 CPUs	ErrGroup 2,540,219	2,180,862 3 of 9	1.16×	4cpu L365–373
Sleep10ms, 50K workers, 100 producers, 8 CPUs	BufferedPool † 3,813,389	3,274,819 2 of 9	1.16×	8cpu L14–22
Sleep10ms, 50K workers, 1K producers, 8 CPUs	BufferedPool † 3,800,703	3,266,799 2 of 9	1.16×	8cpu L23–31
Sleep5msLoop5, 50K workers, 100 producers, 4 CPUs	BufferedPool † 3,332,321	2,873,030 2 of 9	1.16×	4cpu L104–112
Sleep10ms, 300K workers, 1 producer, 4 CPUs	ErrGroup 2,282,556	1,972,243 3 of 9	1.16×	4cpu L275–283
Sleep5msLoop5, 50K workers, 10K producers, 4 CPUs	BufferedPool † 3,185,706	2,771,238 2 of 9	1.15×	4cpu L122–130
Sleep10ms, 300K workers, 1 producer, 8 CPUs	ErrGroup 2,784,449	2,455,963 2 of 9	1.13×	8cpu L275–283
Sleep10ms, 50K workers, 10K producers, 24 CPUs	BufferedPool † 2,962,595	2,632,451 2 of 9	1.13×	24cpu L32–40
Sleep10ms, 300K workers, 100 producers, 4 CPUs	Ants 2,132,405	1,900,906 3 of 9	1.12×	4cpu L284–292
Sleep5msLoop5, 300K workers, 1 producer, 24 CPUs	ErrGroup 2,543,927	2,275,746 2 of 9	1.12×	24cpu L365–373
Sleep10ms, 300K workers, 1M producers, 24 CPUs	ErrGroup 2,161,621	1,937,685 2 of 9	1.12×	24cpu L311–319
Loop10, 1 worker, 1M producers, 4 CPUs	BufferedPool † 4,798,241	4,304,050 2 of 9	1.11×	4cpu L446–454
Sleep5msLoop5, 300K workers, 1K producers, 4 CPUs	Ants 2,338,347	2,136,678 2 of 9	1.09×	4cpu L383–391
Sleep10ms, 50K workers, 100 producers, 24 CPUs	BufferedPool † 3,011,339	2,771,948 2 of 9	1.09×	24cpu L14–22
Sleep5msLoop5, 50K workers, 1K producers, 4 CPUs	BufferedPool † 3,101,827	2,860,632 2 of 9	1.08×	4cpu L113–121
Sleep10ms, 300K workers, 1 producer, 24 CPUs	ErrGroup 2,293,517	2,135,964 2 of 9	1.07×	24cpu L275–283
Sleep10ms, 300K workers, 1K producers, 4 CPUs	Ants 2,212,462	2,070,837 2 of 9	1.07×	4cpu L293–301
Sleep10ms, 50K workers, 1K producers, 24 CPUs	BufferedPool † 2,915,777	2,741,551 2 of 9	1.06×	24cpu L23–31
Sleep10ms, 100K workers, 100 producers, 4 CPUs	BufferedPool † 2,572,532	2,423,256 2 of 9	1.06×	4cpu L149–157
Loop10K, 1 worker, 1 producer, 8 CPUs	BufferedPool † 337,903	322,978 3 of 9	1.05×	8cpu L455–463
Sleep5msLoop5, 100K workers, 1M producers, 8 CPUs	ErrGroup 2,879,461	2,757,349 2 of 9	1.04×	8cpu L266–274
Sleep10ms, 50K workers, 1M producers, 8 CPUs	ErrGroup 2,732,450	2,628,276 2 of 9	1.04×	8cpu L41–49
Loop10K, 1 worker, 100 producers, 24 CPUs	BufferedPool † 334,515	322,322 3 of 9	1.04×	24cpu L464–472
Loop10K, 1 worker, 1 producer, 24 CPUs	Gopool 336,425	324,485 3 of 9	1.04×	24cpu L455–463
Loop10K, 1 worker, 10K producers, 8 CPUs	BufferedPool † 337,132	325,199 2 of 9	1.04×	8cpu L482–490
Loop10K, 1 worker, 1K producers, 8 CPUs	BufferedPool † 336,500	325,224 2 of 9	1.03×	8cpu L473–481
Loop10K, 1 worker, 1K producers, 24 CPUs	BufferedPool † 333,606	322,936 2 of 9	1.03×	24cpu L473–481
Loop10K, 1 worker, 1M producers, 24 CPUs	BufferedPool † 322,314	312,975 2 of 9	1.03×	24cpu L491–499
Loop10K, 1 worker, 10K producers, 24 CPUs	BufferedPool † 333,318	323,713 2 of 9	1.03×	24cpu L482–490
Loop10K, 1 worker, 1 producer, 4 CPUs	BufferedPool † 333,826	324,401 3 of 9	1.03×	4cpu L455–463
Loop10K, 1 worker, 1K producers, 4 CPUs	BufferedPool † 334,843	325,729 2 of 9	1.03×	4cpu L473–481
Loop10, 1 worker, 1M producers, 8 CPUs	BufferedPool † 5,320,750	5,181,886 2 of 9	1.03×	8cpu L446–454
Loop10K, 1 worker, 1M producers, 4 CPUs	BufferedPool † 333,400	325,027 3 of 9	1.03×	4cpu L491–499
Loop10K, 1 worker, 100 producers, 8 CPUs	BufferedPool † 334,568	326,262 2 of 9	1.03×	8cpu L464–472
Sleep10ms, 100K workers, 1K producers, 4 CPUs	BufferedPool † 2,526,768	2,464,268 2 of 9	1.03×	4cpu L158–166
Loop10K, 1 worker, 100 producers, 4 CPUs	BufferedPool † 332,585	324,909 2 of 9	1.02×	4cpu L464–472
Loop10K, 1 worker, 10K producers, 4 CPUs	BufferedPool † 333,216	325,666 2 of 9	1.02×	4cpu L482–490
Sleep10ms, 100K workers, 1 producer, 4 CPUs	ErrGroup 2,326,813	2,277,277 2 of 9	1.02×	4cpu L140–148
Sleep5msLoop5, 100K workers, 1K producers, 4 CPUs	BufferedPool † 2,737,841	2,685,087 2 of 9	1.02×	4cpu L248–256
Sleep10ms, 300K workers, 10K producers, 8 CPUs	ErrGroup 2,815,102	2,766,179 2 of 9	1.02×	8cpu L302–310
Sleep10ms, 100K workers, 1M producers, 8 CPUs	ErrGroup 2,675,045	2,629,661 2 of 9	1.02×	8cpu L176–184
Loop10K, 1 worker, 1M producers, 8 CPUs	BufferedPool † 328,842	324,508 2 of 9	1.01×	8cpu L491–499
Sleep5msLoop5, 300K workers, 100 producers, 4 CPUs	Ants 2,276,359	2,261,985 2 of 9	1.01×	4cpu L374–382
Sleep10ms, 100K workers, 10K producers, 4 CPUs	BufferedPool † 2,439,078	2,423,984 2 of 9	1.01×	4cpu L167–175
Sleep10ms, 300K workers, 1M producers, 8 CPUs	ErrGroup 2,669,081	2,660,367 2 of 9	1.00×	8cpu L311–319
Sleep10ms, 300K workers, 10K producers, 4 CPUs	ErrGroup 2,115,244	2,108,599 2 of 9	1.00×	4cpu L302–310
Sleep5msLoop5, 300K workers, 10K producers, 4 CPUs	ErrGroup 2,226,923	2,220,232 2 of 9	1.00×	4cpu L392–400

Every row behind these numbers · 165 groups

Tasks per second (tasks/s) for each group in scope. The best value of each group is marked.

Group	VarMQ	Ants	WorkerPool	Conc	ErrGroup	Gopool	GoroutinePool †	BufferedPool †	Pond	Source
Sleep10ms, 50K workers, 1 producer, 24 CPUs	2,912,550 (best)	1,519,070	656,765	1,429,409	1,794,858	954,564	1,339,549	1,537,452	399,351	24cpu L5–13
Sleep10ms, 50K workers, 100 producers, 24 CPUs	2,771,948	2,035,061	1,015,615	1,619,384	1,582,002	1,361,653	1,017,638	3,011,339 (best)	368,806	24cpu L14–22
Sleep10ms, 50K workers, 1K producers, 24 CPUs	2,741,551	2,077,158	1,156,901	586,431	1,464,832	1,141,372	778,011	2,915,777 (best)	284,263	24cpu L23–31
Sleep10ms, 50K workers, 10K producers, 24 CPUs	2,632,451	1,884,900	973,489	497,178	1,268,658	1,189,586	1,137,502	2,962,595 (best)	415,109	24cpu L32–40
Sleep10ms, 50K workers, 1M producers, 24 CPUs	2,553,298 (best)	1,382,534	864,947	1,251,845	1,758,812	1,403,797	1,319,959	1,261,415	410,218	24cpu L41–49
Loop10, 50K workers, 1 producer, 24 CPUs	7,391,698 (best)	1,990,627	1,811,122	1,293,970	2,330,329	1,342,225	1,604,738	1,567,348	305,817	24cpu L50–58
Loop10, 50K workers, 100 producers, 24 CPUs	11,460,237 (best)	2,480,988	880,365	1,231,462	1,588,223	2,005,832	1,144,077	1,069,505	324,663	24cpu L59–67
Loop10, 50K workers, 1K producers, 24 CPUs	11,943,831 (best)	2,462,577	1,399,805	718,067	1,131,098	2,423,449	2,319,056	2,108,059	627,675	24cpu L68–76
Loop10, 50K workers, 10K producers, 24 CPUs	11,123,489 (best)	2,097,792	1,587,189	716,207	1,802,430	2,256,852	1,073,129	1,048,061	422,508	24cpu L77–85
Loop10, 50K workers, 1M producers, 24 CPUs	3,701,745 (best)	2,066,346	886,019	1,351,280	2,693,246	1,992,786	1,795,731	1,804,391	274,306	24cpu L86–94
Sleep5msLoop5, 50K workers, 1 producer, 24 CPUs	3,702,257 (best)	1,481,031	414,158	1,313,209	2,244,662	941,503	885,899	1,268,019	284,774	24cpu L95–103
Sleep5msLoop5, 50K workers, 100 producers, 24 CPUs	4,141,394 (best)	2,216,414	671,836	829,635	1,698,787	1,540,400	1,748,018	3,077,938	381,410	24cpu L104–112
Sleep5msLoop5, 50K workers, 1K producers, 24 CPUs	4,081,581 (best)	2,186,966	1,192,459	536,919	1,608,759	1,453,572	747,096	3,133,797	266,469	24cpu L113–121
Sleep5msLoop5, 50K workers, 10K producers, 24 CPUs	3,884,958 (best)	1,925,470	1,078,418	539,814	1,550,695	1,387,420	949,423	2,806,689	415,162	24cpu L122–130
Sleep5msLoop5, 50K workers, 1M producers, 24 CPUs	3,013,711 (best)	1,354,235	964,243	1,324,931	1,204,768	935,379	1,452,245	1,445,972	392,536	24cpu L131–139
Sleep10ms, 100K workers, 1 producer, 24 CPUs	3,186,940 (best)	1,232,040	464,330	1,023,228	2,244,373	1,306,482	1,415,210	1,329,567	433,481	24cpu L140–148
Sleep10ms, 100K workers, 100 producers, 24 CPUs	3,099,420 (best)	2,128,994	640,768	629,472	1,410,503	1,520,710	748,027	1,235,336	365,591	24cpu L149–157
Sleep10ms, 100K workers, 1K producers, 24 CPUs	3,070,265 (best)	2,127,243	1,208,889	884,789	1,568,382	1,510,232	873,092	1,788,000	282,140	24cpu L158–166
Sleep10ms, 100K workers, 10K producers, 24 CPUs	3,106,371 (best)	1,916,333	1,113,100	447,415	1,147,200	1,281,530	864,372	2,300,995	258,489	24cpu L167–175
Sleep10ms, 100K workers, 1M producers, 24 CPUs	2,590,761 (best)	1,258,017	938,921	1,241,020	1,822,877	983,668	1,251,392	1,313,986	413,874	24cpu L176–184
Loop10, 100K workers, 1 producer, 24 CPUs	7,655,402 (best)	1,797,014	1,780,325	871,350	2,452,591	650,795	881,705	1,143,310	258,078	24cpu L185–193
Loop10, 100K workers, 100 producers, 24 CPUs	12,148,546 (best)	2,491,122	885,864	622,712	2,484,519	1,798,938	748,174	715,315	303,219	24cpu L194–202
Loop10, 100K workers, 1K producers, 24 CPUs	12,226,055 (best)	2,448,765	1,587,293	698,892	1,221,789	2,527,880	841,209	745,780	302,656	24cpu L203–211
Loop10, 100K workers, 10K producers, 24 CPUs	10,991,630 (best)	2,327,981	1,429,656	763,993	2,076,663	1,501,982	1,228,318	940,196	359,744	24cpu L212–220
Loop10, 100K workers, 1M producers, 24 CPUs	4,945,904 (best)	1,972,167	900,888	1,471,112	2,688,984	1,864,032	1,694,485	1,713,957	195,794	24cpu L221–229
Sleep5msLoop5, 100K workers, 1 producer, 24 CPUs	3,637,486 (best)	1,447,185	546,915	1,295,835	2,216,760	918,703	1,049,195	955,075	245,538	24cpu L230–238
Sleep5msLoop5, 100K workers, 100 producers, 24 CPUs	4,388,983 (best)	2,254,168	667,374	732,438	1,975,484	1,396,246	1,435,188	1,681,656	406,611	24cpu L239–247
Sleep5msLoop5, 100K workers, 1K producers, 24 CPUs	4,142,646 (best)	2,209,361	1,200,520	681,386	1,505,560	1,461,071	994,664	1,964,340	287,311	24cpu L248–256
Sleep5msLoop5, 100K workers, 10K producers, 24 CPUs	4,321,934 (best)	1,998,802	1,095,639	432,165	1,843,065	1,522,098	892,702	2,146,119	350,291	24cpu L257–265
Sleep5msLoop5, 100K workers, 1M producers, 24 CPUs	2,441,000 (best)	1,389,629	974,400	1,240,756	2,252,720	1,114,660	1,309,237	1,300,391	356,724	24cpu L266–274
Sleep10ms, 300K workers, 1 producer, 24 CPUs	2,135,964	1,423,911	674,477	905,784	2,293,517 (best)	1,158,774	1,164,541	1,279,765	404,293	24cpu L275–283
Sleep10ms, 300K workers, 100 producers, 24 CPUs	2,641,759 (best)	2,103,215	730,914	943,261	1,168,031	1,139,678	778,535	711,112	410,141	24cpu L284–292
Sleep10ms, 300K workers, 1K producers, 24 CPUs	2,672,786 (best)	2,092,344	1,242,138	1,149,947	1,517,095	1,482,393	709,468	1,087,562	283,971	24cpu L293–301
Sleep10ms, 300K workers, 10K producers, 24 CPUs	2,581,158 (best)	1,836,557	1,139,974	597,356	1,487,572	1,353,100	695,951	883,328	292,336	24cpu L302–310
Sleep10ms, 300K workers, 1M producers, 24 CPUs	1,937,685	1,273,929	919,127	978,159	2,161,621 (best)	659,276	1,147,985	1,127,933	300,361	24cpu L311–319
Loop10, 300K workers, 1 producer, 24 CPUs	7,061,717 (best)	1,926,912	1,151,327	730,006	2,332,201	774,258	996,672	1,000,133	344,545	24cpu L320–328
Loop10, 300K workers, 100 producers, 24 CPUs	11,120,957 (best)	2,468,320	989,128	568,315	2,783,019	2,042,220	821,431	786,589	590,471	24cpu L329–337
Loop10, 300K workers, 1K producers, 24 CPUs	11,837,012 (best)	2,419,535	1,739,799	778,083	2,158,559	2,376,341	801,351	1,334,788	425,958	24cpu L338–346
Loop10, 300K workers, 10K producers, 24 CPUs	10,647,574 (best)	2,217,089	1,471,641	831,012	3,177,492	2,714,958	1,130,040	945,233	324,284	24cpu L347–355
Loop10, 300K workers, 1M producers, 24 CPUs	4,838,041 (best)	1,989,109	885,105	1,061,057	3,048,488	1,944,313	1,308,876	1,444,907	179,640	24cpu L356–364
Sleep5msLoop5, 300K workers, 1 producer, 24 CPUs	2,275,746	1,519,060	538,180	930,919	2,543,927 (best)	1,160,267	1,017,716	1,242,966	279,136	24cpu L365–373
Sleep5msLoop5, 300K workers, 100 producers, 24 CPUs	2,702,395 (best)	2,213,105	730,379	1,078,856	1,781,723	1,653,042	1,001,925	1,237,635	325,528	24cpu L374–382
Sleep5msLoop5, 300K workers, 1K producers, 24 CPUs	2,617,388 (best)	2,206,429	1,260,367	1,027,524	1,329,192	1,254,891	702,793	1,137,222	382,119	24cpu L383–391
Sleep5msLoop5, 300K workers, 10K producers, 24 CPUs	2,471,812 (best)	1,951,986	1,157,052	582,503	1,349,137	1,408,464	932,752	1,395,710	508,267	24cpu L392–400
Sleep5msLoop5, 300K workers, 1M producers, 24 CPUs	1,800,531 (best)	1,402,851	979,629	997,380	1,169,486	634,759	1,221,871	1,226,829	217,506	24cpu L401–409
Loop10, 1 worker, 1 producer, 24 CPUs	5,612,142	1,582,899	1,473,002	3,423,771	1,487,417	8,177,809	3,920,683	26,076,554 (best)	3,028,169	24cpu L410–418
Loop10, 1 worker, 100 producers, 24 CPUs	6,768,459	2,076,506	543,788	463,815	1,313,642	4,157,265	816,034	25,921,880 (best)	1,041,918	24cpu L419–427
Loop10, 1 worker, 1K producers, 24 CPUs	5,587,181	1,470,470	729,037	504,896	1,116,652	2,782,257	1,000,047	24,944,261 (best)	870,740	24cpu L428–436
Loop10, 1 worker, 10K producers, 24 CPUs	6,584,676	1,516,219	709,503	625,526	2,124,776	3,713,812	1,020,295	21,230,114 (best)	1,185,321	24cpu L437–445
Loop10, 1 worker, 1M producers, 24 CPUs	4,048,978	812,581	1,143,188	737,708	1,174,023	2,731,570	1,941,971	4,756,005 (best)	1,109,138	24cpu L446–454
Loop10K, 1 worker, 1 producer, 24 CPUs	324,485	265,060	170,583	182,923	264,163	336,425 (best)	217,878	335,389	298,315	24cpu L455–463
Loop10K, 1 worker, 100 producers, 24 CPUs	322,322	263,642	154,748	310,680	263,194	316,513	322,404	334,515 (best)	245,228	24cpu L464–472
Loop10K, 1 worker, 1K producers, 24 CPUs	322,936	263,783	153,473	306,176	262,053	309,949	318,883	333,606 (best)	244,325	24cpu L473–481
Loop10K, 1 worker, 10K producers, 24 CPUs	323,713	255,670	171,686	304,454	253,845	308,782	309,248	333,318 (best)	247,535	24cpu L482–490
Loop10K, 1 worker, 1M producers, 24 CPUs	312,975	240,117	191,623	272,352	236,378	285,413	284,754	322,314 (best)	255,940	24cpu L491–499
Sleep10ms, 50K workers, 1 producer, 8 CPUs	3,176,393	1,986,231	679,563	1,849,255	2,640,266	1,839,933	2,073,880	3,939,669 (best)	704,764	8cpu L5–13
Sleep10ms, 50K workers, 100 producers, 8 CPUs	3,274,819	2,327,861	1,569,760	1,833,619	2,447,916	1,738,482	2,131,513	3,813,389 (best)	710,970	8cpu L14–22
Sleep10ms, 50K workers, 1K producers, 8 CPUs	3,266,799	2,284,979	1,708,169	1,615,934	2,263,140	1,599,249	2,049,195	3,800,703 (best)	716,863	8cpu L23–31
Sleep10ms, 50K workers, 10K producers, 8 CPUs	3,169,295	2,112,959	1,537,853	1,402,121	1,838,928	2,264,740	2,099,609	3,731,132 (best)	779,413	8cpu L32–40
Sleep10ms, 50K workers, 1M producers, 8 CPUs	2,628,276	1,790,694	1,166,711	1,461,992	2,732,450 (best)	1,655,538	1,659,281	1,665,211	677,640	8cpu L41–49
Loop10, 50K workers, 1 producer, 8 CPUs	7,671,684 (best)	3,311,130	1,291,498	2,191,790	3,266,709	2,139,690	2,738,622	2,895,918	891,515	8cpu L50–58
Loop10, 50K workers, 100 producers, 8 CPUs	10,048,582 (best)	2,785,465	2,048,257	2,441,882	2,636,984	2,339,683	2,786,947	4,043,435	804,596	8cpu L59–67
Loop10, 50K workers, 1K producers, 8 CPUs	9,631,427 (best)	2,710,156	2,114,193	2,422,538	2,616,806	3,319,312	2,997,697	2,965,094	556,446	8cpu L68–76
Loop10, 50K workers, 10K producers, 8 CPUs	9,342,026 (best)	2,559,600	1,957,654	1,820,891	3,112,282	3,348,069	3,124,697	2,298,151	625,787	8cpu L77–85
Loop10, 50K workers, 1M producers, 8 CPUs	5,183,600 (best)	2,493,160	1,327,188	1,774,230	3,814,425	2,028,430	2,026,601	2,047,610	385,713	8cpu L86–94
Sleep5msLoop5, 50K workers, 1 producer, 8 CPUs	4,080,137 (best)	2,199,664	595,668	1,858,440	2,860,603	1,938,857	2,113,304	2,118,268	724,425	8cpu L95–103
Sleep5msLoop5, 50K workers, 100 producers, 8 CPUs	4,370,254 (best)	2,458,764	1,545,827	1,829,243	2,783,624	1,759,654	2,030,108	3,658,544	765,675	8cpu L104–112
Sleep5msLoop5, 50K workers, 1K producers, 8 CPUs	4,342,675 (best)	2,394,348	1,723,581	1,641,479	2,615,562	1,638,460	1,885,046	3,271,588	695,670	8cpu L113–121
Sleep5msLoop5, 50K workers, 10K producers, 8 CPUs	3,897,961 (best)	2,176,721	1,489,095	1,308,161	2,147,916	2,323,846	2,083,902	3,830,494	836,893	8cpu L122–130
Sleep5msLoop5, 50K workers, 1M producers, 8 CPUs	3,066,865 (best)	1,820,757	1,243,540	1,520,168	2,875,622	1,742,844	1,679,313	1,667,937	633,235	8cpu L131–139
Sleep10ms, 100K workers, 1 producer, 8 CPUs	3,382,564 (best)	2,031,611	797,093	1,684,581	2,771,764	1,886,008	1,961,994	1,985,532	688,822	8cpu L140–148
Sleep10ms, 100K workers, 100 producers, 8 CPUs	3,495,042 (best)	2,325,278	1,510,855	1,596,958	2,653,807	1,798,224	1,785,989	3,026,245	730,926	8cpu L149–157
Sleep10ms, 100K workers, 1K producers, 8 CPUs	3,404,792 (best)	2,298,816	1,611,359	1,619,952	2,655,545	1,609,919	1,809,348	2,940,701	868,513	8cpu L158–166
Sleep10ms, 100K workers, 10K producers, 8 CPUs	3,307,640 (best)	2,084,848	1,482,996	1,378,631	2,489,694	1,609,554	1,913,004	2,660,361	666,634	8cpu L167–175
Sleep10ms, 100K workers, 1M producers, 8 CPUs	2,629,661	1,743,809	1,209,273	1,341,418	2,675,045 (best)	1,496,747	1,527,097	1,492,396	588,528	8cpu L176–184
Loop10, 100K workers, 1 producer, 8 CPUs	8,016,914 (best)	3,318,506	915,582	1,962,675	3,223,144	2,125,860	2,315,982	2,467,976	879,748	8cpu L185–193
Loop10, 100K workers, 100 producers, 8 CPUs	9,973,951 (best)	2,802,805	2,263,989	2,053,640	2,757,581	2,245,171	2,303,811	2,861,536	800,109	8cpu L194–202
Loop10, 100K workers, 1K producers, 8 CPUs	9,616,461 (best)	2,783,820	2,068,040	2,203,184	2,772,634	3,084,433	2,330,652	2,577,287	864,647	8cpu L203–211
Loop10, 100K workers, 10K producers, 8 CPUs	9,303,365 (best)	2,589,295	1,830,866	2,072,131	3,119,879	3,140,786	2,473,921	2,379,015	779,417	8cpu L212–220
Loop10, 100K workers, 1M producers, 8 CPUs	5,040,936 (best)	2,505,392	1,257,497	1,406,473	3,846,960	2,014,759	1,925,627	1,909,667	284,459	8cpu L221–229
Sleep5msLoop5, 100K workers, 1 producer, 8 CPUs	3,734,297 (best)	2,077,719	679,474	1,589,036	2,883,425	1,921,903	2,052,352	1,916,682	730,519	8cpu L230–238
Sleep5msLoop5, 100K workers, 100 producers, 8 CPUs	4,111,318 (best)	2,482,053	1,532,271	1,642,233	2,685,430	1,829,144	1,801,063	2,497,989	778,669	8cpu L239–247
Sleep5msLoop5, 100K workers, 1K producers, 8 CPUs	4,163,080 (best)	2,373,647	1,716,568	1,634,549	2,682,940	1,680,077	1,870,507	2,661,241	753,595	8cpu L248–256
Sleep5msLoop5, 100K workers, 10K producers, 8 CPUs	3,983,207 (best)	2,149,628	1,530,045	1,438,901	2,702,715	1,612,935	1,919,240	2,789,687	684,114	8cpu L257–265
Sleep5msLoop5, 100K workers, 1M producers, 8 CPUs	2,757,349	1,805,562	1,225,296	1,382,658	2,879,461 (best)	1,545,701	1,560,507	1,560,923	630,118	8cpu L266–274
Sleep10ms, 300K workers, 1 producer, 8 CPUs	2,455,963	2,093,774	576,738	1,213,314	2,784,449 (best)	1,855,333	1,547,576	1,621,264	722,264	8cpu L275–283
Sleep10ms, 300K workers, 100 producers, 8 CPUs	2,719,832 (best)	2,306,727	1,484,983	1,251,695	2,617,496	1,759,026	1,476,386	1,688,833	942,273	8cpu L284–292
Sleep10ms, 300K workers, 1K producers, 8 CPUs	2,726,497 (best)	2,213,675	1,633,217	1,302,743	2,680,243	1,628,227	1,468,225	1,746,018	709,496	8cpu L293–301
Sleep10ms, 300K workers, 10K producers, 8 CPUs	2,766,179	2,002,153	1,477,399	1,318,660	2,815,102 (best)	1,632,026	1,519,349	1,715,160	675,031	8cpu L302–310
Sleep10ms, 300K workers, 1M producers, 8 CPUs	2,660,367	1,747,875	1,173,250	1,153,433	2,669,081 (best)	748,295	1,326,906	1,325,233	320,406	8cpu L311–319
Loop10, 300K workers, 1 producer, 8 CPUs	7,731,116 (best)	3,256,171	1,115,385	1,383,162	3,147,900	2,043,314	1,752,500	2,041,223	885,063	8cpu L320–328
Loop10, 300K workers, 100 producers, 8 CPUs	9,990,429 (best)	2,788,718	2,115,681	1,566,953	2,884,260	2,312,304	1,802,248	1,927,430	815,860	8cpu L329–337
Loop10, 300K workers, 1K producers, 8 CPUs	9,748,598 (best)	2,723,620	2,071,194	1,612,649	2,961,010	3,133,647	1,722,204	1,898,363	938,049	8cpu L338–346
Loop10, 300K workers, 10K producers, 8 CPUs	9,514,309 (best)	2,597,026	1,988,261	1,593,171	3,787,789	3,325,401	1,565,705	1,829,661	547,704	8cpu L347–355
Loop10, 300K workers, 1M producers, 8 CPUs	5,112,507 (best)	2,515,607	1,297,076	1,326,960	3,819,515	2,005,256	1,572,745	1,643,865	183,942	8cpu L356–364
Sleep5msLoop5, 300K workers, 1 producer, 8 CPUs	2,459,278	2,248,233	882,246	1,211,796	2,997,183 (best)	1,896,457	1,575,617	1,717,074	763,774	8cpu L365–373
Sleep5msLoop5, 300K workers, 100 producers, 8 CPUs	2,742,270 (best)	2,438,900	1,539,005	1,261,448	2,634,212	1,799,708	1,481,643	1,777,802	802,117	8cpu L374–382
Sleep5msLoop5, 300K workers, 1K producers, 8 CPUs	2,804,855 (best)	2,384,087	1,682,370	1,293,508	2,709,618	1,660,992	1,507,317	1,703,386	769,307	8cpu L383–391
Sleep5msLoop5, 300K workers, 10K producers, 8 CPUs	2,989,275 (best)	2,164,418	1,508,185	1,320,203	2,773,516	1,647,167	1,517,762	1,799,994	673,978	8cpu L392–400
Sleep5msLoop5, 300K workers, 1M producers, 8 CPUs	2,482,123	1,834,465	1,222,747	1,214,876	3,068,841 (best)	1,070,799	1,357,213	1,351,918	357,014	8cpu L401–409
Loop10, 1 worker, 1 producer, 8 CPUs	6,326,035	1,708,634	1,477,781	4,029,748	2,505,977	11,383,821	3,830,596	26,148,651 (best)	2,783,170	8cpu L410–418
Loop10, 1 worker, 100 producers, 8 CPUs	7,090,918	1,829,284	1,257,569	2,485,982	983,882	4,734,976	2,661,374	27,854,937 (best)	1,580,776	8cpu L419–427
Loop10, 1 worker, 1K producers, 8 CPUs	7,190,233	1,164,781	950,948	1,991,543	1,976,966	4,724,164	3,535,456	26,084,154 (best)	1,564,245	8cpu L428–436
Loop10, 1 worker, 10K producers, 8 CPUs	7,177,970	1,587,642	1,608,207	1,347,437	826,160	4,376,133	2,947,091	24,737,062 (best)	1,568,077	8cpu L437–445
Loop10, 1 worker, 1M producers, 8 CPUs	5,181,886	802,451	1,151,548	1,323,368	1,266,037	3,965,171	2,352,358	5,320,750 (best)	1,283,997	8cpu L446–454
Loop10K, 1 worker, 1 producer, 8 CPUs	322,978	262,550	150,606	206,440	264,605	336,341	220,928	337,903 (best)	299,628	8cpu L455–463
Loop10K, 1 worker, 100 producers, 8 CPUs	326,262	263,467	173,781	307,338	268,565	314,035	320,871	334,568 (best)	278,476	8cpu L464–472
Loop10K, 1 worker, 1K producers, 8 CPUs	325,224	263,508	152,058	302,173	262,424	316,598	321,028	336,500 (best)	277,681	8cpu L473–481
Loop10K, 1 worker, 10K producers, 8 CPUs	325,199	256,045	202,793	296,590	255,972	318,011	313,400	337,132 (best)	271,426	8cpu L482–490
Loop10K, 1 worker, 1M producers, 8 CPUs	324,508	268,194	192,389	284,650	241,213	322,427	295,566	328,842 (best)	260,446	8cpu L491–499
Sleep10ms, 50K workers, 1 producer, 4 CPUs	2,504,306	1,913,216	1,183,121	1,883,293	2,253,850	2,215,600	2,024,379	3,736,757 (best)	623,183	4cpu L5–13
Sleep10ms, 50K workers, 100 producers, 4 CPUs	2,554,864	2,177,404	1,619,490	1,725,821	2,110,137	2,021,565	1,921,155	3,673,447 (best)	654,648	4cpu L14–22
Sleep10ms, 50K workers, 1K producers, 4 CPUs	2,481,416	2,237,604	1,591,348	1,644,058	2,066,129	1,712,781	1,915,121	3,674,940 (best)	664,580	4cpu L23–31
Sleep10ms, 50K workers, 10K producers, 4 CPUs	2,491,621	1,920,782	1,452,288	1,454,461	1,913,117	2,175,690	2,113,921	3,662,015 (best)	750,254	4cpu L32–40
Sleep10ms, 50K workers, 1M producers, 4 CPUs	1,436,960	1,517,066	1,042,072	1,398,267	1,968,705 (best)	1,883,409	1,510,171	1,544,148	653,792	4cpu L41–49
Loop10, 50K workers, 1 producer, 4 CPUs	8,974,684 (best)	3,663,215	1,557,553	2,427,842	3,588,268	2,604,365	2,880,121	2,932,268	854,865	4cpu L50–58
Loop10, 50K workers, 100 producers, 4 CPUs	9,414,707 (best)	2,909,906	2,280,660	2,440,272	2,610,244	2,293,829	2,663,600	4,692,294	882,400	4cpu L59–67
Loop10, 50K workers, 1K producers, 4 CPUs	9,242,992 (best)	2,884,879	2,111,816	2,562,457	2,723,893	2,553,615	2,954,328	3,428,164	821,661	4cpu L68–76
Loop10, 50K workers, 10K producers, 4 CPUs	9,100,058 (best)	2,713,562	1,927,611	2,772,204	3,016,296	2,652,457	3,405,348	3,249,114	766,849	4cpu L77–85
Loop10, 50K workers, 1M producers, 4 CPUs	4,078,666 (best)	2,728,615	1,365,879	1,726,126	3,757,427	2,399,119	2,076,888	2,168,784	707,310	4cpu L86–94
Sleep5msLoop5, 50K workers, 1 producer, 4 CPUs	2,683,120 (best)	2,188,831	1,397,144	1,827,467	2,548,192	2,230,850	1,993,757	2,020,323	703,813	4cpu L95–103
Sleep5msLoop5, 50K workers, 100 producers, 4 CPUs	2,873,030	2,365,318	1,601,606	1,628,028	2,082,071	1,916,477	1,803,299	3,332,321 (best)	724,189	4cpu L104–112
Sleep5msLoop5, 50K workers, 1K producers, 4 CPUs	2,860,632	2,316,929	1,595,246	1,677,868	2,129,081	1,529,943	1,871,692	3,101,827 (best)	750,682	4cpu L113–121
Sleep5msLoop5, 50K workers, 10K producers, 4 CPUs	2,771,238	1,979,868	1,393,430	1,437,549	2,034,452	2,113,058	2,032,782	3,185,706 (best)	792,167	4cpu L122–130
Sleep5msLoop5, 50K workers, 1M producers, 4 CPUs	1,499,597	1,631,679	1,029,438	1,386,109	2,047,829 (best)	1,957,175	1,543,432	1,542,083	649,209	4cpu L131–139
Sleep10ms, 100K workers, 1 producer, 4 CPUs	2,277,277	1,929,536	1,344,598	1,635,295	2,326,813 (best)	2,213,291	1,752,724	1,778,211	621,073	4cpu L140–148
Sleep10ms, 100K workers, 100 producers, 4 CPUs	2,423,256	2,235,957	1,544,855	1,439,483	1,902,991	2,016,200	1,538,496	2,572,532 (best)	676,464	4cpu L149–157
Sleep10ms, 100K workers, 1K producers, 4 CPUs	2,464,268	2,161,787	1,573,492	1,526,274	1,990,995	1,593,568	1,629,140	2,526,768 (best)	667,401	4cpu L158–166
Sleep10ms, 100K workers, 10K producers, 4 CPUs	2,423,984	1,901,625	1,393,122	1,427,641	2,150,416	1,599,449	1,757,459	2,439,078 (best)	813,016	4cpu L167–175
Sleep10ms, 100K workers, 1M producers, 4 CPUs	1,373,342	1,482,758	1,004,084	1,289,624	1,967,247 (best)	1,870,184	1,414,272	1,411,968	591,249	4cpu L176–184
Loop10, 100K workers, 1 producer, 4 CPUs	8,822,850 (best)	3,695,054	999,547	2,203,406	3,614,516	2,526,563	2,594,506	2,748,426	887,493	4cpu L185–193
Loop10, 100K workers, 100 producers, 4 CPUs	9,073,531 (best)	2,850,110	2,254,006	2,176,856	2,595,711	2,334,207	2,352,562	2,934,572	889,293	4cpu L194–202
Loop10, 100K workers, 1K producers, 4 CPUs	8,963,968 (best)	2,823,795	2,092,593	2,162,966	2,860,459	2,605,763	2,411,816	2,778,263	819,885	4cpu L203–211
Loop10, 100K workers, 10K producers, 4 CPUs	8,741,077 (best)	2,712,149	1,878,664	2,321,116	3,242,404	2,619,015	2,967,802	2,785,151	769,006	4cpu L212–220
Loop10, 100K workers, 1M producers, 4 CPUs	4,085,856 (best)	2,712,580	1,367,833	1,491,690	3,796,375	2,379,867	1,962,092	2,047,578	445,405	4cpu L221–229
Sleep5msLoop5, 100K workers, 1 producer, 4 CPUs	2,588,982 (best)	2,177,372	1,266,818	1,678,743	2,534,752	2,282,109	1,830,325	1,874,953	713,349	4cpu L230–238
Sleep5msLoop5, 100K workers, 100 producers, 4 CPUs	2,675,082 (best)	2,369,545	1,615,656	1,476,984	1,919,244	1,911,984	1,558,296	2,363,042	731,775	4cpu L239–247
Sleep5msLoop5, 100K workers, 1K producers, 4 CPUs	2,685,087	2,325,682	1,589,778	1,586,390	2,003,671	1,785,487	1,631,038	2,737,841 (best)	698,138	4cpu L248–256
Sleep5msLoop5, 100K workers, 10K producers, 4 CPUs	2,612,762 (best)	1,962,630	1,441,348	1,462,821	2,156,860	1,675,246	1,804,550	2,345,062	842,579	4cpu L257–265
Sleep5msLoop5, 100K workers, 1M producers, 4 CPUs	1,491,073	1,672,673	1,036,474	1,342,939	2,040,959 (best)	1,947,089	1,410,187	1,459,386	578,672	4cpu L266–274
Sleep10ms, 300K workers, 1 producer, 4 CPUs	1,972,243	1,895,103	1,400,181	1,259,592	2,282,556 (best)	2,114,410	1,449,094	1,528,653	619,592	4cpu L275–283
Sleep10ms, 300K workers, 100 producers, 4 CPUs	1,900,906	2,132,405 (best)	1,549,604	1,165,948	1,843,322	1,816,248	1,295,588	2,019,173	681,027	4cpu L284–292
Sleep10ms, 300K workers, 1K producers, 4 CPUs	2,070,837	2,212,462 (best)	1,531,333	1,222,419	1,940,576	1,344,606	1,309,023	1,682,078	728,195	4cpu L293–301
Sleep10ms, 300K workers, 10K producers, 4 CPUs	2,108,599	1,899,876	1,378,367	1,262,676	2,115,244 (best)	1,526,651	1,455,983	1,683,873	609,106	4cpu L302–310
Sleep10ms, 300K workers, 1M producers, 4 CPUs	1,392,523	1,483,417	1,017,560	1,082,728	1,951,801 (best)	1,875,449	1,210,427	1,254,215	400,028	4cpu L311–319
Loop10, 300K workers, 1 producer, 4 CPUs	8,904,451 (best)	3,691,593	876,263	1,587,828	3,559,990	2,486,249	2,030,857	2,167,295	880,499	4cpu L320–328
Loop10, 300K workers, 100 producers, 4 CPUs	9,540,875 (best)	2,884,675	2,303,545	1,512,610	2,725,520	2,314,553	1,763,438	2,057,987	877,699	4cpu L329–337
Loop10, 300K workers, 1K producers, 4 CPUs	8,864,308 (best)	2,880,763	2,139,906	1,525,242	2,817,087	2,411,791	1,891,427	1,864,496	821,228	4cpu L338–346
Loop10, 300K workers, 10K producers, 4 CPUs	8,823,550 (best)	2,649,041	1,937,796	1,600,822	3,558,489	2,536,054	1,900,095	1,914,892	772,312	4cpu L347–355
Loop10, 300K workers, 1M producers, 4 CPUs	4,006,494 (best)	2,774,204	1,380,812	1,501,498	3,906,769	2,394,671	1,659,267	1,691,046	404,836	4cpu L356–364
Sleep5msLoop5, 300K workers, 1 producer, 4 CPUs	2,180,862	2,158,372	1,469,828	1,300,435	2,540,219 (best)	2,267,699	1,485,305	1,568,345	697,295	4cpu L365–373
Sleep5msLoop5, 300K workers, 100 producers, 4 CPUs	2,261,985	2,276,359 (best)	1,601,331	1,169,340	1,899,621	1,968,919	1,286,994	1,614,006	727,407	4cpu L374–382
Sleep5msLoop5, 300K workers, 1K producers, 4 CPUs	2,136,678	2,338,347 (best)	1,562,446	1,195,559	1,985,445	1,721,649	1,337,528	1,588,406	700,550	4cpu L383–391
Sleep5msLoop5, 300K workers, 10K producers, 4 CPUs	2,220,232	1,939,939	1,425,925	1,234,456	2,226,923 (best)	1,448,610	1,441,574	1,684,125	637,865	4cpu L392–400
Sleep5msLoop5, 300K workers, 1M producers, 4 CPUs	1,521,728	1,645,270	1,037,657	1,156,707	2,107,373 (best)	1,896,843	1,204,219	1,270,588	372,732	4cpu L401–409
Loop10, 1 worker, 1 producer, 4 CPUs	6,632,332	1,242,226	1,769,691	2,090,601	1,341,601	11,813,021	3,692,385	26,202,379 (best)	2,773,548	4cpu L410–418
Loop10, 1 worker, 100 producers, 4 CPUs	7,308,651	1,696,633	1,712,354	3,028,265	2,482,547	4,781,570	5,785,343	27,380,238 (best)	1,733,181	4cpu L419–427
Loop10, 1 worker, 1K producers, 4 CPUs	7,342,679	1,461,392	1,494,450	2,709,995	2,371,571	4,789,004	3,910,772	26,829,003 (best)	1,686,809	4cpu L428–436
Loop10, 1 worker, 10K producers, 4 CPUs	7,271,060	1,402,085	1,337,644	2,115,881	1,302,625	4,927,431	3,558,888	23,643,298 (best)	1,705,779	4cpu L437–445
Loop10, 1 worker, 1M producers, 4 CPUs	4,304,050	895,593	1,349,366	1,352,913	1,366,016	3,808,882	2,519,020	4,798,241 (best)	1,530,277	4cpu L446–454
Loop10K, 1 worker, 1 producer, 4 CPUs	324,401	262,078	179,912	217,314	268,953	330,298	205,916	333,826 (best)	300,071	4cpu L455–463
Loop10K, 1 worker, 100 producers, 4 CPUs	324,909	265,614	186,457	317,155	262,930	318,468	322,479	332,585 (best)	282,164	4cpu L464–472
Loop10K, 1 worker, 1K producers, 4 CPUs	325,729	263,156	161,846	313,304	264,692	310,645	321,535	334,843 (best)	280,022	4cpu L473–481
Loop10K, 1 worker, 10K producers, 4 CPUs	325,666	253,377	228,233	305,421	252,112	312,945	311,607	333,216 (best)	280,109	4cpu L482–490
Loop10K, 1 worker, 1M producers, 4 CPUs	325,027	261,287	186,733	283,844	250,315	329,648	304,900	333,400 (best)	282,891	4cpu L491–499

Run detailsgo test command, machine, versions

shcd pool && go test -run xxx -bench 'Tasks' -cpu 24,8,4 -benchtime=2x -benchmem -timeout 0

The command runs the groups in scope on your machine, from a clone of the benchmarks repository.

Run

2026-10-25

Commit

a0d0e0b

varmq

v1.10.0

Go

1.26

CPU

Intel i7-13700

GOMAXPROCS

24, 8, 4

Raw

pool-24cpu.txt, pool-8cpu.txt, pool-4cpu.txt

Charts

benchmarks.goptics.org

Intel Core i7-13700 (24 threads), linux/amd64, Go 1.26, SQLite databases on ext4 over NVMe.

The pool suite sets GOMAXPROCS to 24, 8 and 4 with no core pinning. The SQLite suite runs on all 24 threads. Redis 7.2 (Debian image) is pinned to CPU 14 and the Redis suite's client to CPUs 0-13 and 16-23 (22 threads), so Redis has a whole physical core.

CPU line of the raw file: 13th Gen Intel(R) Core(TM) i7-13700. Copied from the benchmarks README and the header of each raw file.

Library versions for this suite

Library	Version
varmq	v1.10.0
ants	v2.12.1
workerpool	v1.2.2
conc	v0.3.0
x/sync (errgroup)	v0.24.0
bytedance/gopkg (gopool)	v0.1.4
pond	v2.7.2

Worker pool, GOMAXPROCS 24

Nine pools, 1M tasks per run.

Show

ThroughputMemory

Channel baselines †

55 groups in scope · VarMQ leads 39, others 16 · 9 subjects

Tasks per second

tasks/s · higher is better · mean of 55 groups

SubjectRate× VarMQ

1. VarMQwon 39 of 55 4,719,281 1.00×1.00 times VarMQ, the baseline
2. BufferedPool †won 12 of 55 3,158,583 0.67×0.67 times VarMQ, behind VarMQ
3. Antswon 0 of 55 1,744,569 0.37×0.37 times VarMQ, behind VarMQ
4. ErrGroupwon 3 of 55 1,698,742 0.36×0.36 times VarMQ, behind VarMQ
5. Gopoolwon 1 of 55 1,610,290 0.34×0.34 times VarMQ, behind VarMQ
6. GoroutinePool †won 0 of 55 1,093,148 0.23×0.23 times VarMQ, behind VarMQ
7. WorkerPoolwon 0 of 55 946,147 0.20×0.20 times VarMQ, behind VarMQ
8. Concwon 0 of 55 883,068 0.19×0.19 times VarMQ, behind VarMQ
9. Pondwon 0 of 55 439,778 0.09×0.09 times VarMQ, behind VarMQ

The dashed line marks VarMQ. An amber ratio means the lane is ahead of it. The “×” column divides the lane by VarMQ.

† A baseline, not a library: N goroutines reading one channel. The buffered one holds all 1M tasks.

Who leads each group

one square = one group of rows

A filled green square: a VarMQ subject leads the group. An outlined amber square: another subject leads, and its code is in the square. A dashed square with “=”: a VarMQ subject and another subject share the best value, so neither leads. Press a square to see that group in the lanes above.

Task · workers	11 producer	100100 producers	1K1K producers	10K10K producers	1M1M producers
Sleep10ms · 50K	VQ	BP	BP	BP	VQ
Loop10 · 50K	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 50K	VQ	VQ	VQ	VQ	VQ
Sleep10ms · 100K	VQ	VQ	VQ	VQ	VQ
Loop10 · 100K	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 100K	VQ	VQ	VQ	VQ	VQ
Sleep10ms · 300K	EG	VQ	VQ	VQ	EG
Loop10 · 300K	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 300K	EG	VQ	VQ	VQ	VQ
Loop10 · 1 worker	BP	BP	BP	BP	BP
Loop10K · 1 worker	GP	BP	BP	BP	BP

Hover or focus a square to read it. Press it to show that group in the lanes.

• VQVarMQ39
• BPBufferedPool †12
• EGErrGroup3
• GPGopool1

Where VarMQ does not win

Another subject has the fastest throughput in 16 groups of 55. VarMQ has it in 39.

• BufferedPool † 12 groups
• ErrGroup 3 groups
• Gopool 1 group

All 16 groups where VarMQ does not win, biggest gap first

Group	Best	VarMQ	Ahead by	Source
Loop10, 1 worker, 1 producer	BufferedPool † 26,076,554	5,612,142 3 of 9	4.65×	L410–418
Loop10, 1 worker, 1K producers	BufferedPool † 24,944,261	5,587,181 2 of 9	4.46×	L428–436
Loop10, 1 worker, 100 producers	BufferedPool † 25,921,880	6,768,459 2 of 9	3.83×	L419–427
Loop10, 1 worker, 10K producers	BufferedPool † 21,230,114	6,584,676 2 of 9	3.22×	L437–445
Loop10, 1 worker, 1M producers	BufferedPool † 4,756,005	4,048,978 2 of 9	1.17×	L446–454
Sleep10ms, 50K workers, 10K producers	BufferedPool † 2,962,595	2,632,451 2 of 9	1.13×	L32–40
Sleep5msLoop5, 300K workers, 1 producer	ErrGroup 2,543,927	2,275,746 2 of 9	1.12×	L365–373
Sleep10ms, 300K workers, 1M producers	ErrGroup 2,161,621	1,937,685 2 of 9	1.12×	L311–319
Sleep10ms, 50K workers, 100 producers	BufferedPool † 3,011,339	2,771,948 2 of 9	1.09×	L14–22
Sleep10ms, 300K workers, 1 producer	ErrGroup 2,293,517	2,135,964 2 of 9	1.07×	L275–283
Sleep10ms, 50K workers, 1K producers	BufferedPool † 2,915,777	2,741,551 2 of 9	1.06×	L23–31
Loop10K, 1 worker, 100 producers	BufferedPool † 334,515	322,322 3 of 9	1.04×	L464–472
Loop10K, 1 worker, 1 producer	Gopool 336,425	324,485 3 of 9	1.04×	L455–463
Loop10K, 1 worker, 1K producers	BufferedPool † 333,606	322,936 2 of 9	1.03×	L473–481
Loop10K, 1 worker, 1M producers	BufferedPool † 322,314	312,975 2 of 9	1.03×	L491–499
Loop10K, 1 worker, 10K producers	BufferedPool † 333,318	323,713 2 of 9	1.03×	L482–490

Every row behind these numbers · 55 groups

Tasks per second (tasks/s) for each group in scope. The best value of each group is marked.

Group	VarMQ	Ants	WorkerPool	Conc	ErrGroup	Gopool	GoroutinePool †	BufferedPool †	Pond	Source
Sleep10ms, 50K workers, 1 producer	2,912,550 (best)	1,519,070	656,765	1,429,409	1,794,858	954,564	1,339,549	1,537,452	399,351	L5–13
Sleep10ms, 50K workers, 100 producers	2,771,948	2,035,061	1,015,615	1,619,384	1,582,002	1,361,653	1,017,638	3,011,339 (best)	368,806	L14–22
Sleep10ms, 50K workers, 1K producers	2,741,551	2,077,158	1,156,901	586,431	1,464,832	1,141,372	778,011	2,915,777 (best)	284,263	L23–31
Sleep10ms, 50K workers, 10K producers	2,632,451	1,884,900	973,489	497,178	1,268,658	1,189,586	1,137,502	2,962,595 (best)	415,109	L32–40
Sleep10ms, 50K workers, 1M producers	2,553,298 (best)	1,382,534	864,947	1,251,845	1,758,812	1,403,797	1,319,959	1,261,415	410,218	L41–49
Loop10, 50K workers, 1 producer	7,391,698 (best)	1,990,627	1,811,122	1,293,970	2,330,329	1,342,225	1,604,738	1,567,348	305,817	L50–58
Loop10, 50K workers, 100 producers	11,460,237 (best)	2,480,988	880,365	1,231,462	1,588,223	2,005,832	1,144,077	1,069,505	324,663	L59–67
Loop10, 50K workers, 1K producers	11,943,831 (best)	2,462,577	1,399,805	718,067	1,131,098	2,423,449	2,319,056	2,108,059	627,675	L68–76
Loop10, 50K workers, 10K producers	11,123,489 (best)	2,097,792	1,587,189	716,207	1,802,430	2,256,852	1,073,129	1,048,061	422,508	L77–85
Loop10, 50K workers, 1M producers	3,701,745 (best)	2,066,346	886,019	1,351,280	2,693,246	1,992,786	1,795,731	1,804,391	274,306	L86–94
Sleep5msLoop5, 50K workers, 1 producer	3,702,257 (best)	1,481,031	414,158	1,313,209	2,244,662	941,503	885,899	1,268,019	284,774	L95–103
Sleep5msLoop5, 50K workers, 100 producers	4,141,394 (best)	2,216,414	671,836	829,635	1,698,787	1,540,400	1,748,018	3,077,938	381,410	L104–112
Sleep5msLoop5, 50K workers, 1K producers	4,081,581 (best)	2,186,966	1,192,459	536,919	1,608,759	1,453,572	747,096	3,133,797	266,469	L113–121
Sleep5msLoop5, 50K workers, 10K producers	3,884,958 (best)	1,925,470	1,078,418	539,814	1,550,695	1,387,420	949,423	2,806,689	415,162	L122–130
Sleep5msLoop5, 50K workers, 1M producers	3,013,711 (best)	1,354,235	964,243	1,324,931	1,204,768	935,379	1,452,245	1,445,972	392,536	L131–139
Sleep10ms, 100K workers, 1 producer	3,186,940 (best)	1,232,040	464,330	1,023,228	2,244,373	1,306,482	1,415,210	1,329,567	433,481	L140–148
Sleep10ms, 100K workers, 100 producers	3,099,420 (best)	2,128,994	640,768	629,472	1,410,503	1,520,710	748,027	1,235,336	365,591	L149–157
Sleep10ms, 100K workers, 1K producers	3,070,265 (best)	2,127,243	1,208,889	884,789	1,568,382	1,510,232	873,092	1,788,000	282,140	L158–166
Sleep10ms, 100K workers, 10K producers	3,106,371 (best)	1,916,333	1,113,100	447,415	1,147,200	1,281,530	864,372	2,300,995	258,489	L167–175
Sleep10ms, 100K workers, 1M producers	2,590,761 (best)	1,258,017	938,921	1,241,020	1,822,877	983,668	1,251,392	1,313,986	413,874	L176–184
Loop10, 100K workers, 1 producer	7,655,402 (best)	1,797,014	1,780,325	871,350	2,452,591	650,795	881,705	1,143,310	258,078	L185–193
Loop10, 100K workers, 100 producers	12,148,546 (best)	2,491,122	885,864	622,712	2,484,519	1,798,938	748,174	715,315	303,219	L194–202
Loop10, 100K workers, 1K producers	12,226,055 (best)	2,448,765	1,587,293	698,892	1,221,789	2,527,880	841,209	745,780	302,656	L203–211
Loop10, 100K workers, 10K producers	10,991,630 (best)	2,327,981	1,429,656	763,993	2,076,663	1,501,982	1,228,318	940,196	359,744	L212–220
Loop10, 100K workers, 1M producers	4,945,904 (best)	1,972,167	900,888	1,471,112	2,688,984	1,864,032	1,694,485	1,713,957	195,794	L221–229
Sleep5msLoop5, 100K workers, 1 producer	3,637,486 (best)	1,447,185	546,915	1,295,835	2,216,760	918,703	1,049,195	955,075	245,538	L230–238
Sleep5msLoop5, 100K workers, 100 producers	4,388,983 (best)	2,254,168	667,374	732,438	1,975,484	1,396,246	1,435,188	1,681,656	406,611	L239–247
Sleep5msLoop5, 100K workers, 1K producers	4,142,646 (best)	2,209,361	1,200,520	681,386	1,505,560	1,461,071	994,664	1,964,340	287,311	L248–256
Sleep5msLoop5, 100K workers, 10K producers	4,321,934 (best)	1,998,802	1,095,639	432,165	1,843,065	1,522,098	892,702	2,146,119	350,291	L257–265
Sleep5msLoop5, 100K workers, 1M producers	2,441,000 (best)	1,389,629	974,400	1,240,756	2,252,720	1,114,660	1,309,237	1,300,391	356,724	L266–274
Sleep10ms, 300K workers, 1 producer	2,135,964	1,423,911	674,477	905,784	2,293,517 (best)	1,158,774	1,164,541	1,279,765	404,293	L275–283
Sleep10ms, 300K workers, 100 producers	2,641,759 (best)	2,103,215	730,914	943,261	1,168,031	1,139,678	778,535	711,112	410,141	L284–292
Sleep10ms, 300K workers, 1K producers	2,672,786 (best)	2,092,344	1,242,138	1,149,947	1,517,095	1,482,393	709,468	1,087,562	283,971	L293–301
Sleep10ms, 300K workers, 10K producers	2,581,158 (best)	1,836,557	1,139,974	597,356	1,487,572	1,353,100	695,951	883,328	292,336	L302–310
Sleep10ms, 300K workers, 1M producers	1,937,685	1,273,929	919,127	978,159	2,161,621 (best)	659,276	1,147,985	1,127,933	300,361	L311–319
Loop10, 300K workers, 1 producer	7,061,717 (best)	1,926,912	1,151,327	730,006	2,332,201	774,258	996,672	1,000,133	344,545	L320–328
Loop10, 300K workers, 100 producers	11,120,957 (best)	2,468,320	989,128	568,315	2,783,019	2,042,220	821,431	786,589	590,471	L329–337
Loop10, 300K workers, 1K producers	11,837,012 (best)	2,419,535	1,739,799	778,083	2,158,559	2,376,341	801,351	1,334,788	425,958	L338–346
Loop10, 300K workers, 10K producers	10,647,574 (best)	2,217,089	1,471,641	831,012	3,177,492	2,714,958	1,130,040	945,233	324,284	L347–355
Loop10, 300K workers, 1M producers	4,838,041 (best)	1,989,109	885,105	1,061,057	3,048,488	1,944,313	1,308,876	1,444,907	179,640	L356–364
Sleep5msLoop5, 300K workers, 1 producer	2,275,746	1,519,060	538,180	930,919	2,543,927 (best)	1,160,267	1,017,716	1,242,966	279,136	L365–373
Sleep5msLoop5, 300K workers, 100 producers	2,702,395 (best)	2,213,105	730,379	1,078,856	1,781,723	1,653,042	1,001,925	1,237,635	325,528	L374–382
Sleep5msLoop5, 300K workers, 1K producers	2,617,388 (best)	2,206,429	1,260,367	1,027,524	1,329,192	1,254,891	702,793	1,137,222	382,119	L383–391
Sleep5msLoop5, 300K workers, 10K producers	2,471,812 (best)	1,951,986	1,157,052	582,503	1,349,137	1,408,464	932,752	1,395,710	508,267	L392–400
Sleep5msLoop5, 300K workers, 1M producers	1,800,531 (best)	1,402,851	979,629	997,380	1,169,486	634,759	1,221,871	1,226,829	217,506	L401–409
Loop10, 1 worker, 1 producer	5,612,142	1,582,899	1,473,002	3,423,771	1,487,417	8,177,809	3,920,683	26,076,554 (best)	3,028,169	L410–418
Loop10, 1 worker, 100 producers	6,768,459	2,076,506	543,788	463,815	1,313,642	4,157,265	816,034	25,921,880 (best)	1,041,918	L419–427
Loop10, 1 worker, 1K producers	5,587,181	1,470,470	729,037	504,896	1,116,652	2,782,257	1,000,047	24,944,261 (best)	870,740	L428–436
Loop10, 1 worker, 10K producers	6,584,676	1,516,219	709,503	625,526	2,124,776	3,713,812	1,020,295	21,230,114 (best)	1,185,321	L437–445
Loop10, 1 worker, 1M producers	4,048,978	812,581	1,143,188	737,708	1,174,023	2,731,570	1,941,971	4,756,005 (best)	1,109,138	L446–454
Loop10K, 1 worker, 1 producer	324,485	265,060	170,583	182,923	264,163	336,425 (best)	217,878	335,389	298,315	L455–463
Loop10K, 1 worker, 100 producers	322,322	263,642	154,748	310,680	263,194	316,513	322,404	334,515 (best)	245,228	L464–472
Loop10K, 1 worker, 1K producers	322,936	263,783	153,473	306,176	262,053	309,949	318,883	333,606 (best)	244,325	L473–481
Loop10K, 1 worker, 10K producers	323,713	255,670	171,686	304,454	253,845	308,782	309,248	333,318 (best)	247,535	L482–490
Loop10K, 1 worker, 1M producers	312,975	240,117	191,623	272,352	236,378	285,413	284,754	322,314 (best)	255,940	L491–499

Run detailsgo test command, machine, versions

shcd pool && go test -run xxx -bench 'Tasks' -cpu 24 -benchtime=2x -benchmem -timeout 0

The command runs the groups in scope on your machine, from a clone of the benchmarks repository.

Run

2026-10-25

Commit

a0d0e0b

varmq

v1.10.0

Go

1.26

CPU

Intel i7-13700

GOMAXPROCS

24

Raw

pool-24cpu.txt

Charts

benchmarks.goptics.org

Intel Core i7-13700 (24 threads), linux/amd64, Go 1.26, SQLite databases on ext4 over NVMe.

The pool suite sets GOMAXPROCS to 24, 8 and 4 with no core pinning. The SQLite suite runs on all 24 threads. Redis 7.2 (Debian image) is pinned to CPU 14 and the Redis suite's client to CPUs 0-13 and 16-23 (22 threads), so Redis has a whole physical core.

CPU line of the raw file: 13th Gen Intel(R) Core(TM) i7-13700. Copied from the benchmarks README and the header of each raw file.

Library versions for this suite

Library	Version
varmq	v1.10.0
ants	v2.12.1
workerpool	v1.2.2
conc	v0.3.0
x/sync (errgroup)	v0.24.0
bytedance/gopkg (gopool)	v0.1.4
pond	v2.7.2

Worker pool, GOMAXPROCS 8

The same matrix with GOMAXPROCS 8.

Show

ThroughputMemory

Channel baselines †

55 groups in scope · VarMQ leads 33, others 22 · 9 subjects

Tasks per second

tasks/s · higher is better · mean of 55 groups

SubjectRate× VarMQ

1. VarMQwon 33 of 55 4,683,021 1.00×1.00 times VarMQ, the baseline
2. BufferedPool †won 14 of 55 3,983,311 0.85×0.85 times VarMQ, behind VarMQ
3. ErrGroupwon 8 of 55 2,484,641 0.53×0.53 times VarMQ, behind VarMQ
4. Gopoolwon 0 of 55 2,192,502 0.47×0.47 times VarMQ, behind VarMQ
5. Antswon 0 of 55 2,081,951 0.44×0.44 times VarMQ, behind VarMQ
6. GoroutinePool †won 0 of 55 1,873,143 0.40×0.40 times VarMQ, behind VarMQ
7. Concwon 0 of 55 1,537,793 0.33×0.33 times VarMQ, behind VarMQ
8. WorkerPoolwon 0 of 55 1,322,094 0.28×0.28 times VarMQ, behind VarMQ
9. Pondwon 0 of 55 754,775 0.16×0.16 times VarMQ, behind VarMQ

The dashed line marks VarMQ. An amber ratio means the lane is ahead of it. The “×” column divides the lane by VarMQ.

† A baseline, not a library: N goroutines reading one channel. The buffered one holds all 1M tasks.

Who leads each group

one square = one group of rows

A filled green square: a VarMQ subject leads the group. An outlined amber square: another subject leads, and its code is in the square. A dashed square with “=”: a VarMQ subject and another subject share the best value, so neither leads. Press a square to see that group in the lanes above.

Task · workers	11 producer	100100 producers	1K1K producers	10K10K producers	1M1M producers
Sleep10ms · 50K	BP	BP	BP	BP	EG
Loop10 · 50K	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 50K	VQ	VQ	VQ	VQ	VQ
Sleep10ms · 100K	VQ	VQ	VQ	VQ	EG
Loop10 · 100K	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 100K	VQ	VQ	VQ	VQ	EG
Sleep10ms · 300K	EG	VQ	VQ	EG	EG
Loop10 · 300K	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 300K	EG	VQ	VQ	VQ	EG
Loop10 · 1 worker	BP	BP	BP	BP	BP
Loop10K · 1 worker	BP	BP	BP	BP	BP

Hover or focus a square to read it. Press it to show that group in the lanes.

• VQVarMQ33
• BPBufferedPool †14
• EGErrGroup8

Where VarMQ does not win

Another subject has the fastest throughput in 22 groups of 55. VarMQ has it in 33.

• BufferedPool † 14 groups
• ErrGroup 8 groups

All 22 groups where VarMQ does not win, biggest gap first

Group	Best	VarMQ	Ahead by	Source
Loop10, 1 worker, 1 producer	BufferedPool † 26,148,651	6,326,035 3 of 9	4.13×	L410–418
Loop10, 1 worker, 100 producers	BufferedPool † 27,854,937	7,090,918 2 of 9	3.93×	L419–427
Loop10, 1 worker, 1K producers	BufferedPool † 26,084,154	7,190,233 2 of 9	3.63×	L428–436
Loop10, 1 worker, 10K producers	BufferedPool † 24,737,062	7,177,970 2 of 9	3.45×	L437–445
Sleep10ms, 50K workers, 1 producer	BufferedPool † 3,939,669	3,176,393 2 of 9	1.24×	L5–13
Sleep5msLoop5, 300K workers, 1M producers	ErrGroup 3,068,841	2,482,123 2 of 9	1.24×	L401–409
Sleep5msLoop5, 300K workers, 1 producer	ErrGroup 2,997,183	2,459,278 2 of 9	1.22×	L365–373
Sleep10ms, 50K workers, 10K producers	BufferedPool † 3,731,132	3,169,295 2 of 9	1.18×	L32–40
Sleep10ms, 50K workers, 100 producers	BufferedPool † 3,813,389	3,274,819 2 of 9	1.16×	L14–22
Sleep10ms, 50K workers, 1K producers	BufferedPool † 3,800,703	3,266,799 2 of 9	1.16×	L23–31
Sleep10ms, 300K workers, 1 producer	ErrGroup 2,784,449	2,455,963 2 of 9	1.13×	L275–283
Loop10K, 1 worker, 1 producer	BufferedPool † 337,903	322,978 3 of 9	1.05×	L455–463
Sleep5msLoop5, 100K workers, 1M producers	ErrGroup 2,879,461	2,757,349 2 of 9	1.04×	L266–274
Sleep10ms, 50K workers, 1M producers	ErrGroup 2,732,450	2,628,276 2 of 9	1.04×	L41–49
Loop10K, 1 worker, 10K producers	BufferedPool † 337,132	325,199 2 of 9	1.04×	L482–490
Loop10K, 1 worker, 1K producers	BufferedPool † 336,500	325,224 2 of 9	1.03×	L473–481
Loop10, 1 worker, 1M producers	BufferedPool † 5,320,750	5,181,886 2 of 9	1.03×	L446–454
Loop10K, 1 worker, 100 producers	BufferedPool † 334,568	326,262 2 of 9	1.03×	L464–472
Sleep10ms, 300K workers, 10K producers	ErrGroup 2,815,102	2,766,179 2 of 9	1.02×	L302–310
Sleep10ms, 100K workers, 1M producers	ErrGroup 2,675,045	2,629,661 2 of 9	1.02×	L176–184
Loop10K, 1 worker, 1M producers	BufferedPool † 328,842	324,508 2 of 9	1.01×	L491–499
Sleep10ms, 300K workers, 1M producers	ErrGroup 2,669,081	2,660,367 2 of 9	1.00×	L311–319

Every row behind these numbers · 55 groups

Tasks per second (tasks/s) for each group in scope. The best value of each group is marked.

Group	VarMQ	Ants	WorkerPool	Conc	ErrGroup	Gopool	GoroutinePool †	BufferedPool †	Pond	Source
Sleep10ms, 50K workers, 1 producer	3,176,393	1,986,231	679,563	1,849,255	2,640,266	1,839,933	2,073,880	3,939,669 (best)	704,764	L5–13
Sleep10ms, 50K workers, 100 producers	3,274,819	2,327,861	1,569,760	1,833,619	2,447,916	1,738,482	2,131,513	3,813,389 (best)	710,970	L14–22
Sleep10ms, 50K workers, 1K producers	3,266,799	2,284,979	1,708,169	1,615,934	2,263,140	1,599,249	2,049,195	3,800,703 (best)	716,863	L23–31
Sleep10ms, 50K workers, 10K producers	3,169,295	2,112,959	1,537,853	1,402,121	1,838,928	2,264,740	2,099,609	3,731,132 (best)	779,413	L32–40
Sleep10ms, 50K workers, 1M producers	2,628,276	1,790,694	1,166,711	1,461,992	2,732,450 (best)	1,655,538	1,659,281	1,665,211	677,640	L41–49
Loop10, 50K workers, 1 producer	7,671,684 (best)	3,311,130	1,291,498	2,191,790	3,266,709	2,139,690	2,738,622	2,895,918	891,515	L50–58
Loop10, 50K workers, 100 producers	10,048,582 (best)	2,785,465	2,048,257	2,441,882	2,636,984	2,339,683	2,786,947	4,043,435	804,596	L59–67
Loop10, 50K workers, 1K producers	9,631,427 (best)	2,710,156	2,114,193	2,422,538	2,616,806	3,319,312	2,997,697	2,965,094	556,446	L68–76
Loop10, 50K workers, 10K producers	9,342,026 (best)	2,559,600	1,957,654	1,820,891	3,112,282	3,348,069	3,124,697	2,298,151	625,787	L77–85
Loop10, 50K workers, 1M producers	5,183,600 (best)	2,493,160	1,327,188	1,774,230	3,814,425	2,028,430	2,026,601	2,047,610	385,713	L86–94
Sleep5msLoop5, 50K workers, 1 producer	4,080,137 (best)	2,199,664	595,668	1,858,440	2,860,603	1,938,857	2,113,304	2,118,268	724,425	L95–103
Sleep5msLoop5, 50K workers, 100 producers	4,370,254 (best)	2,458,764	1,545,827	1,829,243	2,783,624	1,759,654	2,030,108	3,658,544	765,675	L104–112
Sleep5msLoop5, 50K workers, 1K producers	4,342,675 (best)	2,394,348	1,723,581	1,641,479	2,615,562	1,638,460	1,885,046	3,271,588	695,670	L113–121
Sleep5msLoop5, 50K workers, 10K producers	3,897,961 (best)	2,176,721	1,489,095	1,308,161	2,147,916	2,323,846	2,083,902	3,830,494	836,893	L122–130
Sleep5msLoop5, 50K workers, 1M producers	3,066,865 (best)	1,820,757	1,243,540	1,520,168	2,875,622	1,742,844	1,679,313	1,667,937	633,235	L131–139
Sleep10ms, 100K workers, 1 producer	3,382,564 (best)	2,031,611	797,093	1,684,581	2,771,764	1,886,008	1,961,994	1,985,532	688,822	L140–148
Sleep10ms, 100K workers, 100 producers	3,495,042 (best)	2,325,278	1,510,855	1,596,958	2,653,807	1,798,224	1,785,989	3,026,245	730,926	L149–157
Sleep10ms, 100K workers, 1K producers	3,404,792 (best)	2,298,816	1,611,359	1,619,952	2,655,545	1,609,919	1,809,348	2,940,701	868,513	L158–166
Sleep10ms, 100K workers, 10K producers	3,307,640 (best)	2,084,848	1,482,996	1,378,631	2,489,694	1,609,554	1,913,004	2,660,361	666,634	L167–175
Sleep10ms, 100K workers, 1M producers	2,629,661	1,743,809	1,209,273	1,341,418	2,675,045 (best)	1,496,747	1,527,097	1,492,396	588,528	L176–184
Loop10, 100K workers, 1 producer	8,016,914 (best)	3,318,506	915,582	1,962,675	3,223,144	2,125,860	2,315,982	2,467,976	879,748	L185–193
Loop10, 100K workers, 100 producers	9,973,951 (best)	2,802,805	2,263,989	2,053,640	2,757,581	2,245,171	2,303,811	2,861,536	800,109	L194–202
Loop10, 100K workers, 1K producers	9,616,461 (best)	2,783,820	2,068,040	2,203,184	2,772,634	3,084,433	2,330,652	2,577,287	864,647	L203–211
Loop10, 100K workers, 10K producers	9,303,365 (best)	2,589,295	1,830,866	2,072,131	3,119,879	3,140,786	2,473,921	2,379,015	779,417	L212–220
Loop10, 100K workers, 1M producers	5,040,936 (best)	2,505,392	1,257,497	1,406,473	3,846,960	2,014,759	1,925,627	1,909,667	284,459	L221–229
Sleep5msLoop5, 100K workers, 1 producer	3,734,297 (best)	2,077,719	679,474	1,589,036	2,883,425	1,921,903	2,052,352	1,916,682	730,519	L230–238
Sleep5msLoop5, 100K workers, 100 producers	4,111,318 (best)	2,482,053	1,532,271	1,642,233	2,685,430	1,829,144	1,801,063	2,497,989	778,669	L239–247
Sleep5msLoop5, 100K workers, 1K producers	4,163,080 (best)	2,373,647	1,716,568	1,634,549	2,682,940	1,680,077	1,870,507	2,661,241	753,595	L248–256
Sleep5msLoop5, 100K workers, 10K producers	3,983,207 (best)	2,149,628	1,530,045	1,438,901	2,702,715	1,612,935	1,919,240	2,789,687	684,114	L257–265
Sleep5msLoop5, 100K workers, 1M producers	2,757,349	1,805,562	1,225,296	1,382,658	2,879,461 (best)	1,545,701	1,560,507	1,560,923	630,118	L266–274
Sleep10ms, 300K workers, 1 producer	2,455,963	2,093,774	576,738	1,213,314	2,784,449 (best)	1,855,333	1,547,576	1,621,264	722,264	L275–283
Sleep10ms, 300K workers, 100 producers	2,719,832 (best)	2,306,727	1,484,983	1,251,695	2,617,496	1,759,026	1,476,386	1,688,833	942,273	L284–292
Sleep10ms, 300K workers, 1K producers	2,726,497 (best)	2,213,675	1,633,217	1,302,743	2,680,243	1,628,227	1,468,225	1,746,018	709,496	L293–301
Sleep10ms, 300K workers, 10K producers	2,766,179	2,002,153	1,477,399	1,318,660	2,815,102 (best)	1,632,026	1,519,349	1,715,160	675,031	L302–310
Sleep10ms, 300K workers, 1M producers	2,660,367	1,747,875	1,173,250	1,153,433	2,669,081 (best)	748,295	1,326,906	1,325,233	320,406	L311–319
Loop10, 300K workers, 1 producer	7,731,116 (best)	3,256,171	1,115,385	1,383,162	3,147,900	2,043,314	1,752,500	2,041,223	885,063	L320–328
Loop10, 300K workers, 100 producers	9,990,429 (best)	2,788,718	2,115,681	1,566,953	2,884,260	2,312,304	1,802,248	1,927,430	815,860	L329–337
Loop10, 300K workers, 1K producers	9,748,598 (best)	2,723,620	2,071,194	1,612,649	2,961,010	3,133,647	1,722,204	1,898,363	938,049	L338–346
Loop10, 300K workers, 10K producers	9,514,309 (best)	2,597,026	1,988,261	1,593,171	3,787,789	3,325,401	1,565,705	1,829,661	547,704	L347–355
Loop10, 300K workers, 1M producers	5,112,507 (best)	2,515,607	1,297,076	1,326,960	3,819,515	2,005,256	1,572,745	1,643,865	183,942	L356–364
Sleep5msLoop5, 300K workers, 1 producer	2,459,278	2,248,233	882,246	1,211,796	2,997,183 (best)	1,896,457	1,575,617	1,717,074	763,774	L365–373
Sleep5msLoop5, 300K workers, 100 producers	2,742,270 (best)	2,438,900	1,539,005	1,261,448	2,634,212	1,799,708	1,481,643	1,777,802	802,117	L374–382
Sleep5msLoop5, 300K workers, 1K producers	2,804,855 (best)	2,384,087	1,682,370	1,293,508	2,709,618	1,660,992	1,507,317	1,703,386	769,307	L383–391
Sleep5msLoop5, 300K workers, 10K producers	2,989,275 (best)	2,164,418	1,508,185	1,320,203	2,773,516	1,647,167	1,517,762	1,799,994	673,978	L392–400
Sleep5msLoop5, 300K workers, 1M producers	2,482,123	1,834,465	1,222,747	1,214,876	3,068,841 (best)	1,070,799	1,357,213	1,351,918	357,014	L401–409
Loop10, 1 worker, 1 producer	6,326,035	1,708,634	1,477,781	4,029,748	2,505,977	11,383,821	3,830,596	26,148,651 (best)	2,783,170	L410–418
Loop10, 1 worker, 100 producers	7,090,918	1,829,284	1,257,569	2,485,982	983,882	4,734,976	2,661,374	27,854,937 (best)	1,580,776	L419–427
Loop10, 1 worker, 1K producers	7,190,233	1,164,781	950,948	1,991,543	1,976,966	4,724,164	3,535,456	26,084,154 (best)	1,564,245	L428–436
Loop10, 1 worker, 10K producers	7,177,970	1,587,642	1,608,207	1,347,437	826,160	4,376,133	2,947,091	24,737,062 (best)	1,568,077	L437–445
Loop10, 1 worker, 1M producers	5,181,886	802,451	1,151,548	1,323,368	1,266,037	3,965,171	2,352,358	5,320,750 (best)	1,283,997	L446–454
Loop10K, 1 worker, 1 producer	322,978	262,550	150,606	206,440	264,605	336,341	220,928	337,903 (best)	299,628	L455–463
Loop10K, 1 worker, 100 producers	326,262	263,467	173,781	307,338	268,565	314,035	320,871	334,568 (best)	278,476	L464–472
Loop10K, 1 worker, 1K producers	325,224	263,508	152,058	302,173	262,424	316,598	321,028	336,500 (best)	277,681	L473–481
Loop10K, 1 worker, 10K producers	325,199	256,045	202,793	296,590	255,972	318,011	313,400	337,132 (best)	271,426	L482–490
Loop10K, 1 worker, 1M producers	324,508	268,194	192,389	284,650	241,213	322,427	295,566	328,842 (best)	260,446	L491–499

Run detailsgo test command, machine, versions

shcd pool && go test -run xxx -bench 'Tasks' -cpu 8 -benchtime=2x -benchmem -timeout 0

The command runs the groups in scope on your machine, from a clone of the benchmarks repository.

Run

2026-10-25

Commit

a0d0e0b

varmq

v1.10.0

Go

1.26

CPU

Intel i7-13700

GOMAXPROCS

8

Raw

pool-8cpu.txt

Charts

benchmarks.goptics.org

Intel Core i7-13700 (24 threads), linux/amd64, Go 1.26, SQLite databases on ext4 over NVMe.

The pool suite sets GOMAXPROCS to 24, 8 and 4 with no core pinning. The SQLite suite runs on all 24 threads. Redis 7.2 (Debian image) is pinned to CPU 14 and the Redis suite's client to CPUs 0-13 and 16-23 (22 threads), so Redis has a whole physical core.

CPU line of the raw file: 13th Gen Intel(R) Core(TM) i7-13700. Copied from the benchmarks README and the header of each raw file.

Library versions for this suite

Library	Version
varmq	v1.10.0
ants	v2.12.1
workerpool	v1.2.2
conc	v0.3.0
x/sync (errgroup)	v0.24.0
bytedance/gopkg (gopool)	v0.1.4
pond	v2.7.2

Worker pool, GOMAXPROCS 4

The same matrix with GOMAXPROCS 4.

Show

ThroughputMemory

Channel baselines †

55 groups in scope · VarMQ leads 19, others 36 · 9 subjects

Tasks per second

tasks/s · higher is better · mean of 55 groups

SubjectRate× VarMQ

1. VarMQwon 19 of 55 4,037,474 1.00×1.00 times VarMQ, the baseline
2. BufferedPool †won 21 of 55 3,942,859 0.98×0.98 times VarMQ, behind VarMQ
3. Gopoolwon 0 of 55 2,281,205 0.57×0.57 times VarMQ, behind VarMQ
4. ErrGroupwon 11 of 55 2,210,498 0.55×0.55 times VarMQ, behind VarMQ
5. Antswon 4 of 55 2,055,044 0.51×0.51 times VarMQ, behind VarMQ
6. GoroutinePool †won 0 of 55 1,908,708 0.47×0.47 times VarMQ, behind VarMQ
7. Concwon 0 of 55 1,558,666 0.39×0.39 times VarMQ, behind VarMQ
8. WorkerPoolwon 0 of 55 1,395,833 0.35×0.35 times VarMQ, behind VarMQ
9. Pondwon 0 of 55 773,266 0.19×0.19 times VarMQ, behind VarMQ

The dashed line marks VarMQ. An amber ratio means the lane is ahead of it. The “×” column divides the lane by VarMQ.

† A baseline, not a library: N goroutines reading one channel. The buffered one holds all 1M tasks.

Who leads each group

one square = one group of rows

A filled green square: a VarMQ subject leads the group. An outlined amber square: another subject leads, and its code is in the square. A dashed square with “=”: a VarMQ subject and another subject share the best value, so neither leads. Press a square to see that group in the lanes above.

Task · workers	11 producer	100100 producers	1K1K producers	10K10K producers	1M1M producers
Sleep10ms · 50K	BP	BP	BP	BP	EG
Loop10 · 50K	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 50K	VQ	BP	BP	BP	EG
Sleep10ms · 100K	EG	BP	BP	BP	EG
Loop10 · 100K	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 100K	VQ	VQ	BP	VQ	EG
Sleep10ms · 300K	EG	AN	AN	EG	EG
Loop10 · 300K	VQ	VQ	VQ	VQ	VQ
Sleep5msLoop5 · 300K	EG	AN	AN	EG	EG
Loop10 · 1 worker	BP	BP	BP	BP	BP
Loop10K · 1 worker	BP	BP	BP	BP	BP

Hover or focus a square to read it. Press it to show that group in the lanes.

• VQVarMQ19
• BPBufferedPool †21
• EGErrGroup11
• ANAnts4

Where VarMQ does not win

Another subject has the fastest throughput in 36 groups of 55. VarMQ has it in 19.

• BufferedPool † 21 groups
• ErrGroup 11 groups
• Ants 4 groups

All 36 groups where VarMQ does not win, biggest gap first

Group	Best	VarMQ	Ahead by	Source
Loop10, 1 worker, 1 producer	BufferedPool † 26,202,379	6,632,332 3 of 9	3.95×	L410–418
Loop10, 1 worker, 100 producers	BufferedPool † 27,380,238	7,308,651 2 of 9	3.75×	L419–427
Loop10, 1 worker, 1K producers	BufferedPool † 26,829,003	7,342,679 2 of 9	3.65×	L428–436
Loop10, 1 worker, 10K producers	BufferedPool † 23,643,298	7,271,060 2 of 9	3.25×	L437–445
Sleep10ms, 50K workers, 1 producer	BufferedPool † 3,736,757	2,504,306 2 of 9	1.49×	L5–13
Sleep10ms, 50K workers, 1K producers	BufferedPool † 3,674,940	2,481,416 2 of 9	1.48×	L23–31
Sleep10ms, 50K workers, 10K producers	BufferedPool † 3,662,015	2,491,621 2 of 9	1.47×	L32–40
Sleep10ms, 50K workers, 100 producers	BufferedPool † 3,673,447	2,554,864 2 of 9	1.44×	L14–22
Sleep10ms, 100K workers, 1M producers	ErrGroup 1,967,247	1,373,342 6 of 9	1.43×	L176–184
Sleep10ms, 300K workers, 1M producers	ErrGroup 1,951,801	1,392,523 4 of 9	1.40×	L311–319
Sleep5msLoop5, 300K workers, 1M producers	ErrGroup 2,107,373	1,521,728 4 of 9	1.38×	L401–409
Sleep10ms, 50K workers, 1M producers	ErrGroup 1,968,705	1,436,960 6 of 9	1.37×	L41–49
Sleep5msLoop5, 100K workers, 1M producers	ErrGroup 2,040,959	1,491,073 4 of 9	1.37×	L266–274
Sleep5msLoop5, 50K workers, 1M producers	ErrGroup 2,047,829	1,499,597 6 of 9	1.37×	L131–139
Sleep5msLoop5, 300K workers, 1 producer	ErrGroup 2,540,219	2,180,862 3 of 9	1.16×	L365–373
Sleep5msLoop5, 50K workers, 100 producers	BufferedPool † 3,332,321	2,873,030 2 of 9	1.16×	L104–112
Sleep10ms, 300K workers, 1 producer	ErrGroup 2,282,556	1,972,243 3 of 9	1.16×	L275–283
Sleep5msLoop5, 50K workers, 10K producers	BufferedPool † 3,185,706	2,771,238 2 of 9	1.15×	L122–130
Sleep10ms, 300K workers, 100 producers	Ants 2,132,405	1,900,906 3 of 9	1.12×	L284–292
Loop10, 1 worker, 1M producers	BufferedPool † 4,798,241	4,304,050 2 of 9	1.11×	L446–454
Sleep5msLoop5, 300K workers, 1K producers	Ants 2,338,347	2,136,678 2 of 9	1.09×	L383–391
Sleep5msLoop5, 50K workers, 1K producers	BufferedPool † 3,101,827	2,860,632 2 of 9	1.08×	L113–121
Sleep10ms, 300K workers, 1K producers	Ants 2,212,462	2,070,837 2 of 9	1.07×	L293–301
Sleep10ms, 100K workers, 100 producers	BufferedPool † 2,572,532	2,423,256 2 of 9	1.06×	L149–157
Loop10K, 1 worker, 1 producer	BufferedPool † 333,826	324,401 3 of 9	1.03×	L455–463
Loop10K, 1 worker, 1K producers	BufferedPool † 334,843	325,729 2 of 9	1.03×	L473–481
Loop10K, 1 worker, 1M producers	BufferedPool † 333,400	325,027 3 of 9	1.03×	L491–499
Sleep10ms, 100K workers, 1K producers	BufferedPool † 2,526,768	2,464,268 2 of 9	1.03×	L158–166
Loop10K, 1 worker, 100 producers	BufferedPool † 332,585	324,909 2 of 9	1.02×	L464–472
Loop10K, 1 worker, 10K producers	BufferedPool † 333,216	325,666 2 of 9	1.02×	L482–490
Sleep10ms, 100K workers, 1 producer	ErrGroup 2,326,813	2,277,277 2 of 9	1.02×	L140–148
Sleep5msLoop5, 100K workers, 1K producers	BufferedPool † 2,737,841	2,685,087 2 of 9	1.02×	L248–256
Sleep5msLoop5, 300K workers, 100 producers	Ants 2,276,359	2,261,985 2 of 9	1.01×	L374–382
Sleep10ms, 100K workers, 10K producers	BufferedPool † 2,439,078	2,423,984 2 of 9	1.01×	L167–175
Sleep10ms, 300K workers, 10K producers	ErrGroup 2,115,244	2,108,599 2 of 9	1.00×	L302–310
Sleep5msLoop5, 300K workers, 10K producers	ErrGroup 2,226,923	2,220,232 2 of 9	1.00×	L392–400

Every row behind these numbers · 55 groups

Tasks per second (tasks/s) for each group in scope. The best value of each group is marked.

Group	VarMQ	Ants	WorkerPool	Conc	ErrGroup	Gopool	GoroutinePool †	BufferedPool †	Pond	Source
Sleep10ms, 50K workers, 1 producer	2,504,306	1,913,216	1,183,121	1,883,293	2,253,850	2,215,600	2,024,379	3,736,757 (best)	623,183	L5–13
Sleep10ms, 50K workers, 100 producers	2,554,864	2,177,404	1,619,490	1,725,821	2,110,137	2,021,565	1,921,155	3,673,447 (best)	654,648	L14–22
Sleep10ms, 50K workers, 1K producers	2,481,416	2,237,604	1,591,348	1,644,058	2,066,129	1,712,781	1,915,121	3,674,940 (best)	664,580	L23–31
Sleep10ms, 50K workers, 10K producers	2,491,621	1,920,782	1,452,288	1,454,461	1,913,117	2,175,690	2,113,921	3,662,015 (best)	750,254	L32–40
Sleep10ms, 50K workers, 1M producers	1,436,960	1,517,066	1,042,072	1,398,267	1,968,705 (best)	1,883,409	1,510,171	1,544,148	653,792	L41–49
Loop10, 50K workers, 1 producer	8,974,684 (best)	3,663,215	1,557,553	2,427,842	3,588,268	2,604,365	2,880,121	2,932,268	854,865	L50–58
Loop10, 50K workers, 100 producers	9,414,707 (best)	2,909,906	2,280,660	2,440,272	2,610,244	2,293,829	2,663,600	4,692,294	882,400	L59–67
Loop10, 50K workers, 1K producers	9,242,992 (best)	2,884,879	2,111,816	2,562,457	2,723,893	2,553,615	2,954,328	3,428,164	821,661	L68–76
Loop10, 50K workers, 10K producers	9,100,058 (best)	2,713,562	1,927,611	2,772,204	3,016,296	2,652,457	3,405,348	3,249,114	766,849	L77–85
Loop10, 50K workers, 1M producers	4,078,666 (best)	2,728,615	1,365,879	1,726,126	3,757,427	2,399,119	2,076,888	2,168,784	707,310	L86–94
Sleep5msLoop5, 50K workers, 1 producer	2,683,120 (best)	2,188,831	1,397,144	1,827,467	2,548,192	2,230,850	1,993,757	2,020,323	703,813	L95–103
Sleep5msLoop5, 50K workers, 100 producers	2,873,030	2,365,318	1,601,606	1,628,028	2,082,071	1,916,477	1,803,299	3,332,321 (best)	724,189	L104–112
Sleep5msLoop5, 50K workers, 1K producers	2,860,632	2,316,929	1,595,246	1,677,868	2,129,081	1,529,943	1,871,692	3,101,827 (best)	750,682	L113–121
Sleep5msLoop5, 50K workers, 10K producers	2,771,238	1,979,868	1,393,430	1,437,549	2,034,452	2,113,058	2,032,782	3,185,706 (best)	792,167	L122–130
Sleep5msLoop5, 50K workers, 1M producers	1,499,597	1,631,679	1,029,438	1,386,109	2,047,829 (best)	1,957,175	1,543,432	1,542,083	649,209	L131–139
Sleep10ms, 100K workers, 1 producer	2,277,277	1,929,536	1,344,598	1,635,295	2,326,813 (best)	2,213,291	1,752,724	1,778,211	621,073	L140–148
Sleep10ms, 100K workers, 100 producers	2,423,256	2,235,957	1,544,855	1,439,483	1,902,991	2,016,200	1,538,496	2,572,532 (best)	676,464	L149–157
Sleep10ms, 100K workers, 1K producers	2,464,268	2,161,787	1,573,492	1,526,274	1,990,995	1,593,568	1,629,140	2,526,768 (best)	667,401	L158–166
Sleep10ms, 100K workers, 10K producers	2,423,984	1,901,625	1,393,122	1,427,641	2,150,416	1,599,449	1,757,459	2,439,078 (best)	813,016	L167–175
Sleep10ms, 100K workers, 1M producers	1,373,342	1,482,758	1,004,084	1,289,624	1,967,247 (best)	1,870,184	1,414,272	1,411,968	591,249	L176–184
Loop10, 100K workers, 1 producer	8,822,850 (best)	3,695,054	999,547	2,203,406	3,614,516	2,526,563	2,594,506	2,748,426	887,493	L185–193
Loop10, 100K workers, 100 producers	9,073,531 (best)	2,850,110	2,254,006	2,176,856	2,595,711	2,334,207	2,352,562	2,934,572	889,293	L194–202
Loop10, 100K workers, 1K producers	8,963,968 (best)	2,823,795	2,092,593	2,162,966	2,860,459	2,605,763	2,411,816	2,778,263	819,885	L203–211
Loop10, 100K workers, 10K producers	8,741,077 (best)	2,712,149	1,878,664	2,321,116	3,242,404	2,619,015	2,967,802	2,785,151	769,006	L212–220
Loop10, 100K workers, 1M producers	4,085,856 (best)	2,712,580	1,367,833	1,491,690	3,796,375	2,379,867	1,962,092	2,047,578	445,405	L221–229
Sleep5msLoop5, 100K workers, 1 producer	2,588,982 (best)	2,177,372	1,266,818	1,678,743	2,534,752	2,282,109	1,830,325	1,874,953	713,349	L230–238
Sleep5msLoop5, 100K workers, 100 producers	2,675,082 (best)	2,369,545	1,615,656	1,476,984	1,919,244	1,911,984	1,558,296	2,363,042	731,775	L239–247
Sleep5msLoop5, 100K workers, 1K producers	2,685,087	2,325,682	1,589,778	1,586,390	2,003,671	1,785,487	1,631,038	2,737,841 (best)	698,138	L248–256
Sleep5msLoop5, 100K workers, 10K producers	2,612,762 (best)	1,962,630	1,441,348	1,462,821	2,156,860	1,675,246	1,804,550	2,345,062	842,579	L257–265
Sleep5msLoop5, 100K workers, 1M producers	1,491,073	1,672,673	1,036,474	1,342,939	2,040,959 (best)	1,947,089	1,410,187	1,459,386	578,672	L266–274
Sleep10ms, 300K workers, 1 producer	1,972,243	1,895,103	1,400,181	1,259,592	2,282,556 (best)	2,114,410	1,449,094	1,528,653	619,592	L275–283
Sleep10ms, 300K workers, 100 producers	1,900,906	2,132,405 (best)	1,549,604	1,165,948	1,843,322	1,816,248	1,295,588	2,019,173	681,027	L284–292
Sleep10ms, 300K workers, 1K producers	2,070,837	2,212,462 (best)	1,531,333	1,222,419	1,940,576	1,344,606	1,309,023	1,682,078	728,195	L293–301
Sleep10ms, 300K workers, 10K producers	2,108,599	1,899,876	1,378,367	1,262,676	2,115,244 (best)	1,526,651	1,455,983	1,683,873	609,106	L302–310
Sleep10ms, 300K workers, 1M producers	1,392,523	1,483,417	1,017,560	1,082,728	1,951,801 (best)	1,875,449	1,210,427	1,254,215	400,028	L311–319
Loop10, 300K workers, 1 producer	8,904,451 (best)	3,691,593	876,263	1,587,828	3,559,990	2,486,249	2,030,857	2,167,295	880,499	L320–328
Loop10, 300K workers, 100 producers	9,540,875 (best)	2,884,675	2,303,545	1,512,610	2,725,520	2,314,553	1,763,438	2,057,987	877,699	L329–337
Loop10, 300K workers, 1K producers	8,864,308 (best)	2,880,763	2,139,906	1,525,242	2,817,087	2,411,791	1,891,427	1,864,496	821,228	L338–346
Loop10, 300K workers, 10K producers	8,823,550 (best)	2,649,041	1,937,796	1,600,822	3,558,489	2,536,054	1,900,095	1,914,892	772,312	L347–355
Loop10, 300K workers, 1M producers	4,006,494 (best)	2,774,204	1,380,812	1,501,498	3,906,769	2,394,671	1,659,267	1,691,046	404,836	L356–364
Sleep5msLoop5, 300K workers, 1 producer	2,180,862	2,158,372	1,469,828	1,300,435	2,540,219 (best)	2,267,699	1,485,305	1,568,345	697,295	L365–373
Sleep5msLoop5, 300K workers, 100 producers	2,261,985	2,276,359 (best)	1,601,331	1,169,340	1,899,621	1,968,919	1,286,994	1,614,006	727,407	L374–382
Sleep5msLoop5, 300K workers, 1K producers	2,136,678	2,338,347 (best)	1,562,446	1,195,559	1,985,445	1,721,649	1,337,528	1,588,406	700,550	L383–391
Sleep5msLoop5, 300K workers, 10K producers	2,220,232	1,939,939	1,425,925	1,234,456	2,226,923 (best)	1,448,610	1,441,574	1,684,125	637,865	L392–400
Sleep5msLoop5, 300K workers, 1M producers	1,521,728	1,645,270	1,037,657	1,156,707	2,107,373 (best)	1,896,843	1,204,219	1,270,588	372,732	L401–409
Loop10, 1 worker, 1 producer	6,632,332	1,242,226	1,769,691	2,090,601	1,341,601	11,813,021	3,692,385	26,202,379 (best)	2,773,548	L410–418
Loop10, 1 worker, 100 producers	7,308,651	1,696,633	1,712,354	3,028,265	2,482,547	4,781,570	5,785,343	27,380,238 (best)	1,733,181	L419–427
Loop10, 1 worker, 1K producers	7,342,679	1,461,392	1,494,450	2,709,995	2,371,571	4,789,004	3,910,772	26,829,003 (best)	1,686,809	L428–436
Loop10, 1 worker, 10K producers	7,271,060	1,402,085	1,337,644	2,115,881	1,302,625	4,927,431	3,558,888	23,643,298 (best)	1,705,779	L437–445
Loop10, 1 worker, 1M producers	4,304,050	895,593	1,349,366	1,352,913	1,366,016	3,808,882	2,519,020	4,798,241 (best)	1,530,277	L446–454
Loop10K, 1 worker, 1 producer	324,401	262,078	179,912	217,314	268,953	330,298	205,916	333,826 (best)	300,071	L455–463
Loop10K, 1 worker, 100 producers	324,909	265,614	186,457	317,155	262,930	318,468	322,479	332,585 (best)	282,164	L464–472
Loop10K, 1 worker, 1K producers	325,729	263,156	161,846	313,304	264,692	310,645	321,535	334,843 (best)	280,022	L473–481
Loop10K, 1 worker, 10K producers	325,666	253,377	228,233	305,421	252,112	312,945	311,607	333,216 (best)	280,109	L482–490
Loop10K, 1 worker, 1M producers	325,027	261,287	186,733	283,844	250,315	329,648	304,900	333,400 (best)	282,891	L491–499

Run detailsgo test command, machine, versions

shcd pool && go test -run xxx -bench 'Tasks' -cpu 4 -benchtime=2x -benchmem -timeout 0

The command runs the groups in scope on your machine, from a clone of the benchmarks repository.

Run

2026-10-25

Commit

a0d0e0b

varmq

v1.10.0

Go

1.26

CPU

Intel i7-13700

GOMAXPROCS

4

Raw

pool-4cpu.txt

Charts

benchmarks.goptics.org

Intel Core i7-13700 (24 threads), linux/amd64, Go 1.26, SQLite databases on ext4 over NVMe.

The pool suite sets GOMAXPROCS to 24, 8 and 4 with no core pinning. The SQLite suite runs on all 24 threads. Redis 7.2 (Debian image) is pinned to CPU 14 and the Redis suite's client to CPUs 0-13 and 16-23 (22 threads), so Redis has a whole physical core.

CPU line of the raw file: 13th Gen Intel(R) Core(TM) i7-13700. Copied from the benchmarks README and the header of each raw file.

Library versions for this suite

Library	Version
varmq	v1.10.0
ants	v2.12.1
workerpool	v1.2.2
conc	v0.3.0
x/sync (errgroup)	v0.24.0
bytedance/gopkg (gopool)	v0.1.4
pond	v2.7.2

Message queues on Redis

Six subjects from five libraries on Redis 7.2.

Show

ThroughputMemory

12 groups of 24 in scope · VarMQ leads 4, others 8 · 6 subjects

Jobs per second

jobs/s · higher is better · mean of 12 groups

SubjectRate× VarMQ

1. VarMQwon 4 of 12 33,005 1.00×1.00 times VarMQ, the baseline
2. VarMQDistributedwon 0 of 12 31,044 0.94×0.94 times VarMQ, behind VarMQ
3. GocraftWorkwon 1 of 12 29,200 0.88×0.88 times VarMQ, behind VarMQ
4. Rmqwon 7 of 12 26,476 0.80×0.80 times VarMQ, behind VarMQ
5. Taskqwon 0 of 12 20,835 0.63×0.63 times VarMQ, behind VarMQ
6. Asynqwon 0 of 12 14,804 0.45×0.45 times VarMQ, behind VarMQ

The dashed line marks VarMQ. An amber ratio means the lane is ahead of it. The “×” column divides the lane by VarMQ.

Who leads each group

one square = one group of rows

A filled green square: a VarMQ subject leads the group. An outlined amber square: another subject leads, and its code is in the square. A dashed square with “=”: a VarMQ subject and another subject share the best value, so neither leads. Press a square to see that group in the lanes above.

Task · workers · instances	11 producer	100100 producers
Noop · 10 · C1	RM	GW
Noop · 100 · C1	RM	VQ
Noop · 1000 · C1	RM	VQ
Sleep1ms · 10 · C1	RM	RM
Sleep1ms · 100 · C1	RM	VQ
Sleep1ms · 1000 · C1	RM	VQ

Hover or focus a square to read it. Press it to show that group in the lanes.

• VQVarMQ4
• RMRmq7
• GWGocraftWork1

Where VarMQ does not win

Another subject has the fastest throughput in 8 groups of 12. VarMQ has it in 4.

• Rmq 7 groups
• GocraftWork 1 group

All 8 groups where VarMQ does not win, biggest gap first

Group	Best	VarMQ	Ahead by	Source
Noop, 100 workers, C1, 1 producer	Rmq 35,708	VarMQDistributed 21,426 3 of 6	1.67×	L27–32
Noop, 1000 workers, C1, 1 producer	Rmq 34,621	VarMQDistributed 21,425 3 of 6	1.62×	L49–54
Noop, 10 workers, C1, 1 producer	Rmq 35,129	21,968 3 of 6	1.60×	L5–10
Sleep 1 ms, 100 workers, C1, 1 producer	Rmq 28,837	VarMQDistributed 21,177 3 of 6	1.36×	L93–98
Sleep 1 ms, 1000 workers, C1, 1 producer	Rmq 28,898	21,243 3 of 6	1.36×	L115–120
Sleep 1 ms, 10 workers, C1, 100 producers	Rmq 6,314	5,077 4 of 6	1.24×	L77–82
Sleep 1 ms, 10 workers, C1, 1 producer	Rmq 7,032	5,695 4 of 6	1.23×	L71–76
Noop, 10 workers, C1, 100 producers	GocraftWork 48,917	41,816 2 of 6	1.17×	L11–16

Every row behind these numbers · 12 groups

Jobs per second (jobs/s) for each group in scope. The best value of each group is marked.

Group	VarMQ	Asynq	VarMQDistributed	Rmq	Taskq	GocraftWork	Source
Noop, 10 workers, C1, 1 producer	21,968	15,806	21,664	35,129 (best)	19,031	25,428	L5–10
Noop, 10 workers, C1, 100 producers	41,816	17,823	38,909	29,616	34,434	48,917 (best)	L11–16
Noop, 100 workers, C1, 1 producer	20,812	15,096	21,426	35,708 (best)	18,448	25,731	L27–32
Noop, 100 workers, C1, 100 producers	59,898 (best)	17,744	54,031	28,521	34,499	44,775	L33–38
Noop, 1000 workers, C1, 1 producer	21,157	17,710	21,425	34,621 (best)	16,351	25,147	L49–54
Noop, 1000 workers, C1, 100 producers	58,308 (best)	18,954	54,007	30,010	27,197	40,255	L55–60
Sleep 1 ms, 10 workers, C1, 1 producer	5,695	5,332	5,514	7,032 (best)	5,745	6,144	L71–76
Sleep 1 ms, 10 workers, C1, 100 producers	5,077	4,907	4,866	6,314 (best)	5,227	5,528	L77–82
Sleep 1 ms, 100 workers, C1, 1 producer	21,017	14,703	21,177	28,837 (best)	18,810	25,093	L93–98
Sleep 1 ms, 100 workers, C1, 100 producers	45,894 (best)	17,315	42,158	26,208	27,743	44,324	L99–104
Sleep 1 ms, 1000 workers, C1, 1 producer	21,243	16,264	21,197	28,898 (best)	16,517	22,712	L115–120
Sleep 1 ms, 1000 workers, C1, 100 producers	73,175 (best)	15,998	66,152	26,821	26,015	36,343	L121–126

Run detailsgo test command, machine, versions

shtask redis:upcd redis && go test -run xxx -bench 'Durable/^(Noop|Sleep1ms)W\d+C1$' -benchtime=2x -benchmem -timeout 0

The command runs the groups in scope on your machine, from a clone of the benchmarks repository.

Run

2026-10-25

Commit

a0d0e0b

varmq

v1.10.0

Go

1.26

CPU

Intel i7-13700

Threads

22

Raw

redis.txt

Charts

benchmarks.goptics.org

Intel Core i7-13700 (24 threads), linux/amd64, Go 1.26, SQLite databases on ext4 over NVMe.

The pool suite sets GOMAXPROCS to 24, 8 and 4 with no core pinning. The SQLite suite runs on all 24 threads. Redis 7.2 (Debian image) is pinned to CPU 14 and the Redis suite's client to CPUs 0-13 and 16-23 (22 threads), so Redis has a whole physical core.

CPU line of the raw file: 13th Gen Intel(R) Core(TM) i7-13700. Copied from the benchmarks README and the header of each raw file.

Library versions for this suite

Library	Version
varmq	v1.10.0
redisq	v1.5.0
Asynq	v0.26.0
rmq	v5.2.0
taskq	v3.2.9
gocraft/work	v0.5.1
go-redis	v9.23.0 (taskq uses v8.11.5)
redigo	v1.9.3

Message queues on SQLite

Five subjects from four libraries, WAL checkpoint at 1000 and 10000 pages.

Show

ThroughputWAL size

8 groups in scope · VarMQ leads 8, others 0 · 5 subjects

Jobs per second

jobs/s · higher is better · mean of 8 groups

SubjectRate× VarMQ

1. VarMQwon 8 of 8 28,628 1.00×1.00 times VarMQ, the baseline
2. Goqitewon 0 of 8 4,064 0.14×0.14 times VarMQ, behind VarMQ
3. Riverwon 0 of 8 2,614 0.09×0.09 times VarMQ, behind VarMQ

The dashed line marks VarMQ. An amber ratio means the lane is ahead of it. The “×” column divides the lane by VarMQ.

Not in the lanes: Backlite (6 of 8 groups), VarMQSingleFile (2 of 8 groups). A mean over unequal groups would not compare like with like. Narrow the filters to see them.

Who leads each group

one square = one group of rows

A filled green square: a VarMQ subject leads the group. An outlined amber square: another subject leads, and its code is in the square. A dashed square with “=”: a VarMQ subject and another subject share the best value, so neither leads. Press a square to see that group in the lanes above.

Scenario	10001000 pages	1000010000 pages
Mixed1Q	VQ	VQ
Drain1Q	VQ	VQ
DrainPrio	VQ	VQ
Mixed10Q	VQ	VQ

Hover or focus a square to read it. Press it to show that group in the lanes.

• VQVarMQ8

Where VarMQ does not win

VarMQ has the fastest throughput in every one of the 8 groups in this scope.

Every row behind these numbers · 8 groups

Jobs per second (jobs/s) for each group in scope. The best value of each group is marked.

Group	VarMQ	Goqite	Backlite	River	VarMQSingleFile	Source
Mixed1Q, 1000 pages	8,575 (best)	3,032	924	1,375	–not run	L5–8
Mixed1Q, 10000 pages	11,860 (best)	7,091	294	2,115	–not run	L9–12
Drain1Q, 1000 pages	14,725 (best)	2,933	1,119	3,230	–not run	L13–16
Drain1Q, 10000 pages	21,225 (best)	4,921	1,016	3,719	–not run	L17–20
DrainPrio, 1000 pages	12,937 (best)	2,908	–not run	3,240	–not run	L21–23
DrainPrio, 10000 pages	25,608 (best)	4,828	–not run	3,625	–not run	L24–26
Mixed10Q, 1000 pages	29,624 (best)	2,131	5,326	1,431	4,553	L27–31
Mixed10Q, 10000 pages	104,471 (best)	4,668	5,419	2,175	14,110	L32–36

Run detailsgo test command, machine, versions

shcd sqlite && go test -run xxx -bench 'SQLite' -benchtime=2x -benchmem -timeout 0

The command runs the groups in scope on your machine, from a clone of the benchmarks repository.

Run

2026-10-25

Commit

a0d0e0b

varmq

v1.10.0

Go

1.26

CPU

Intel i7-13700

Threads

24

Raw

sqlite.txt

Charts

benchmarks.goptics.org

Intel Core i7-13700 (24 threads), linux/amd64, Go 1.26, SQLite databases on ext4 over NVMe.

The pool suite sets GOMAXPROCS to 24, 8 and 4 with no core pinning. The SQLite suite runs on all 24 threads. Redis 7.2 (Debian image) is pinned to CPU 14 and the Redis suite's client to CPUs 0-13 and 16-23 (22 threads), so Redis has a whole physical core.

CPU line of the raw file: 13th Gen Intel(R) Core(TM) i7-13700. Copied from the benchmarks README and the header of each raw file.

Library versions for this suite

Library	Version
varmq	v1.10.0
sqliteq	v0.3.0
goqite	v0.4.0
backlite	v0.8.0
River with its SQLite driver	v0.49.0
go-sqlite3	v1.14.28

How to read the numbers

• A row is one line of output. A group is the rows that run one case with different subjects, for example Sleep10ms with 50K workers and one producer. A subject leads a group when its row is the best one. When a VarMQ subject and another subject share the best value, the group is tied: it counts for neither side, and the map marks it with “=”.
• Two iterations per row. The files use -benchtime=2x, so each row covers two iterations, with no warm-up iteration. Absolute numbers change from run to run, so compare ratios inside one run.
• A lane is a mean. When several groups are in scope, a lane is the arithmetic mean of that subject over them. A subject that misses some of the groups is left out, because its mean would cover other rows. Narrow the filters to see it.
• MB means 10^6 bytes. Memory is the B/op column divided by one million. The charts at benchmarks.goptics.org divide by 2^20 and still print MB, so their memory figures are 4.6 % smaller.
• Pool memory covers the whole process. It counts every allocation during a run of 1M tasks. VarMQ’s Add never blocks, so waiting tasks sit in an unbounded queue and use memory.
• † marks a baseline. GoroutinePool and BufferedPool are N goroutines that read one channel. They are not libraries. Switch them off with the Baselines control.
• Redis has two VarMQ subjects. VarMQ runs with one instance. VarMQDistributed runs with one and with four. Where VarMQ does not run, the × column divides by VarMQDistributed. The charts leave VarMQDistributed out of the one-instance, enqueue and latency groups. This page keeps it, because it is in the raw file.
• Latency is enqueue to handler. One job at a time, 1,000 jobs, 10 workers, one instance. Lower is better.

For the full method of each suite, read the READMEs in the benchmarks repository: pool, Redis and SQLite.

Where the data comes from

The page reads a committed snapshot, src/components/bench/data.json. A script builds it from the raw .txt files at one commit of the benchmarks repository. The script keeps the line number of every row, which is how the page links each figure to its line. The docs build does not use the network.

The live charts at benchmarks.goptics.org do not send a CORS header. A browser on this site cannot read their dataset, so the page cannot refresh itself. To update the snapshot, run these commands in the docs/ folder:

node scripts/bench-data.mjs --ref main --check-live
pnpm test

--check-live compares every row with the published chart data and fails on a difference. The tests check the figures against the home page and check that each go test command selects exactly the groups in scope.

Run it yourself

Install Go 1.26 or later and Task, then clone goptics/benchmarks. The Redis suite needs Docker. The SQLite suite needs a C compiler.

task pool:bench                      # one iteration of every pool benchmark
task redis:up redis:bench redis:down
task sqlite:bench

Run details, at the foot of the explorer, shows the command for the groups in scope. The full run takes hours: the committed pool files took 1282 s, 917 s and 904 s.

See also the comparison of features and the bench log.