mirror of
https://github.com/zap-proto/bench.git
synced 2026-10-30 15:06:06 +00:00
zeekayandHanzo Dev
713be1506c
The throughput columns in the last stored run were taken at a one-minute load average between 122 and 190 on ten cores. At that load the same 5,000-request loop returned 10,408, 796 and 4,018 req/s in three consecutive repetitions, and ZAP-HTTP against fasthttp came out 1.32x, 0.93x and 1.28x at 256 B. Those numbers describe a busy CPU. No reading of the transport can be taken out of them, including the cautious reading that the two are hard to tell apart. run.sh now reads the load average before it starts. The allocation and wire sections run unconditionally, because counts do not move with load: across four runs between load 99 and load 208, every allocation figure repeated to the tenth of a byte. The concurrent and per-op sections run only below two fifths of the core count, four on this box, and otherwise the file carries NOT MEASURED and the load that caused it in place of the table. A figure missing from bench-results.txt was not measured; none is ever estimated. The stored run is regenerated under that rule, so it now carries the five- workload allocation table and the wire sizes and no rate table at all. Co-authored-by: Hanzo Dev <[email protected]>
272 lines
20 KiB
Plaintext
272 lines
20 KiB
Plaintext
=== environment ===
|
|
go version go1.26.5 darwin/arm64
|
|
Darwin 25.6.0
|
|
Apple M1 Max
|
|
10
|
|
20:21 up 7 days, 8:08, 1 user, load averages: 102.21 128.14 138.31
|
|
quiet enough for rates: no (load 102.21, 10 cores, threshold 4.0)
|
|
github.com/zap-proto/bench
|
|
github.com/valyala/fasthttp v1.70.0
|
|
github.com/zap-proto/go v1.8.3
|
|
github.com/zap-proto/http v0.3.9
|
|
|
|
=== allocations, 3 repetitions ===
|
|
=== RUN TestMemoryPressure
|
|
=== RUN TestMemoryPressure/tiny
|
|
bench_test.go:472:
|
|
workload: tiny (body=16 B, headers=4)
|
|
net/http 5000 reqs in 4.922s ( 1016 req/s) | 10409.7 bytes/req | 115.4 allocs/req | 33 GCs | heap=3928 KiB
|
|
fasthttp 5000 reqs in 1.008s ( 4963 req/s) | 6.7 bytes/req | 0.0 allocs/req | 0 GCs | heap=2480 KiB
|
|
ZAP-HTTP 5000 reqs in 1.3s ( 3845 req/s) | 2.6 bytes/req | 0.0 allocs/req | 0 GCs | heap=2520 KiB
|
|
native ZAP 5000 reqs in 1.131s ( 4423 req/s) | 224.0 bytes/req | 6.0 allocs/req | 0 GCs | heap=3520 KiB
|
|
floor 5000 reqs in 752ms ( 6648 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2440 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:2.57x allocs:1.31x throughput:0.77x
|
|
ZAP-HTTP vs net/http bytes:3979.25x allocs:8613.34x throughput:3.79x
|
|
native ZAP vs fasthttp bytes:0.03x allocs:0.00x throughput:0.89x
|
|
native ZAP vs net/http bytes:46.47x allocs:19.23x throughput:4.35x
|
|
=== RUN TestMemoryPressure/small
|
|
bench_test.go:472:
|
|
workload: small (body=256 B, headers=8)
|
|
net/http 5000 reqs in 3.229s ( 1548 req/s) | 10531.4 bytes/req | 122.4 allocs/req | 34 GCs | heap=4336 KiB
|
|
fasthttp 5000 reqs in 1.338s ( 3737 req/s) | 5.3 bytes/req | 0.0 allocs/req | 0 GCs | heap=2568 KiB
|
|
ZAP-HTTP 5000 reqs in 2.13s ( 2347 req/s) | 3.2 bytes/req | 0.0 allocs/req | 0 GCs | heap=2624 KiB
|
|
native ZAP 5000 reqs in 1.084s ( 4612 req/s) | 704.0 bytes/req | 6.0 allocs/req | 1 GCs | heap=4120 KiB
|
|
floor 5000 reqs in 935ms ( 5347 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2528 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:1.63x allocs:1.07x throughput:0.63x
|
|
ZAP-HTTP vs net/http bytes:3260.09x allocs:7373.08x throughput:1.52x
|
|
native ZAP vs fasthttp bytes:0.01x allocs:0.00x throughput:1.23x
|
|
native ZAP vs net/http bytes:14.96x allocs:20.40x throughput:2.98x
|
|
=== RUN TestMemoryPressure/chat
|
|
bench_test.go:472:
|
|
workload: chat (body=824 B, headers=12)
|
|
net/http 5000 reqs in 2.586s ( 1933 req/s) | 14680.2 bytes/req | 151.5 allocs/req | 50 GCs | heap=3448 KiB
|
|
fasthttp 5000 reqs in 1.48s ( 3379 req/s) | 7.5 bytes/req | 0.0 allocs/req | 0 GCs | heap=2696 KiB
|
|
ZAP-HTTP 5000 reqs in 1.271s ( 3934 req/s) | 4.4 bytes/req | 0.0 allocs/req | 0 GCs | heap=2696 KiB
|
|
native ZAP 5000 reqs in 1.105s ( 4526 req/s) | 1984.0 bytes/req | 6.0 allocs/req | 3 GCs | heap=4960 KiB
|
|
floor 5000 reqs in 258ms ( 19343 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2576 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:1.72x allocs:1.06x throughput:1.16x
|
|
ZAP-HTTP vs net/http bytes:3369.50x allocs:6588.45x throughput:2.04x
|
|
native ZAP vs fasthttp bytes:0.00x allocs:0.00x throughput:1.34x
|
|
native ZAP vs net/http bytes:7.40x allocs:25.25x throughput:2.34x
|
|
=== RUN TestMemoryPressure/medium
|
|
bench_test.go:472:
|
|
workload: medium (body=4096 B, headers=12)
|
|
net/http 5000 reqs in 4.38s ( 1142 req/s) | 24747.7 bytes/req | 158.9 allocs/req | 91 GCs | heap=3184 KiB
|
|
fasthttp 5000 reqs in 2.32s ( 2155 req/s) | 11.7 bytes/req | 0.0 allocs/req | 0 GCs | heap=2696 KiB
|
|
ZAP-HTTP 5000 reqs in 1.368s ( 3654 req/s) | 6.9 bytes/req | 0.0 allocs/req | 0 GCs | heap=2744 KiB
|
|
native ZAP 5000 reqs in 888ms ( 5628 req/s) | 8384.1 bytes/req | 6.0 allocs/req | 12 GCs | heap=16360 KiB
|
|
floor 5000 reqs in 243ms ( 20590 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2632 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:1.69x allocs:1.21x throughput:1.70x
|
|
ZAP-HTTP vs net/http bytes:3596.21x allocs:7095.76x throughput:3.20x
|
|
native ZAP vs fasthttp bytes:0.00x allocs:0.00x throughput:2.61x
|
|
native ZAP vs net/http bytes:2.95x allocs:26.49x throughput:4.93x
|
|
=== RUN TestMemoryPressure/large
|
|
bench_test.go:472:
|
|
workload: large (body=65536 B, headers=8)
|
|
net/http 5000 reqs in 7.044s ( 710 req/s) | 193830.8 bytes/req | 144.2 allocs/req | 669 GCs | heap=4568 KiB
|
|
fasthttp 5000 reqs in 1.45s ( 3449 req/s) | 58.5 bytes/req | 0.0 allocs/req | 0 GCs | heap=3000 KiB
|
|
ZAP-HTTP 5000 reqs in 2.141s ( 2335 req/s) | 68.5 bytes/req | 0.0 allocs/req | 0 GCs | heap=3528 KiB
|
|
native ZAP 5000 reqs in 2.693s ( 1857 req/s) | 131268.9 bytes/req | 6.0 allocs/req | 254 GCs | heap=3640 KiB
|
|
floor 5000 reqs in 2.283s ( 2190 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2864 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:0.85x allocs:1.16x throughput:0.68x
|
|
ZAP-HTTP vs net/http bytes:2830.74x allocs:8383.02x throughput:3.29x
|
|
native ZAP vs fasthttp bytes:0.00x allocs:0.00x throughput:0.54x
|
|
native ZAP vs net/http bytes:1.48x allocs:23.85x throughput:2.62x
|
|
--- PASS: TestMemoryPressure (52.56s)
|
|
--- PASS: TestMemoryPressure/tiny (9.98s)
|
|
--- PASS: TestMemoryPressure/small (9.50s)
|
|
--- PASS: TestMemoryPressure/chat (7.32s)
|
|
--- PASS: TestMemoryPressure/medium (9.62s)
|
|
--- PASS: TestMemoryPressure/large (16.13s)
|
|
=== RUN TestMemoryPressure
|
|
=== RUN TestMemoryPressure/tiny
|
|
bench_test.go:472:
|
|
workload: tiny (body=16 B, headers=4)
|
|
net/http 5000 reqs in 2.887s ( 1732 req/s) | 10399.1 bytes/req | 115.4 allocs/req | 36 GCs | heap=3624 KiB
|
|
fasthttp 5000 reqs in 1.316s ( 3801 req/s) | 6.7 bytes/req | 0.0 allocs/req | 0 GCs | heap=2544 KiB
|
|
ZAP-HTTP 5000 reqs in 1.493s ( 3348 req/s) | 2.8 bytes/req | 0.0 allocs/req | 0 GCs | heap=2576 KiB
|
|
native ZAP 5000 reqs in 1.116s ( 4480 req/s) | 224.0 bytes/req | 6.0 allocs/req | 0 GCs | heap=3576 KiB
|
|
floor 5000 reqs in 798ms ( 6262 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2504 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:2.39x allocs:1.18x throughput:0.88x
|
|
ZAP-HTTP vs net/http bytes:3705.48x allocs:7904.74x throughput:1.93x
|
|
native ZAP vs fasthttp bytes:0.03x allocs:0.00x throughput:1.18x
|
|
native ZAP vs net/http bytes:46.42x allocs:19.23x throughput:2.59x
|
|
=== RUN TestMemoryPressure/small
|
|
bench_test.go:472:
|
|
workload: small (body=256 B, headers=8)
|
|
net/http 5000 reqs in 2.814s ( 1777 req/s) | 10448.9 bytes/req | 122.4 allocs/req | 33 GCs | heap=3776 KiB
|
|
fasthttp 5000 reqs in 1.119s ( 4469 req/s) | 6.1 bytes/req | 0.0 allocs/req | 0 GCs | heap=2616 KiB
|
|
ZAP-HTTP 5000 reqs in 1.393s ( 3589 req/s) | 3.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2672 KiB
|
|
native ZAP 5000 reqs in 874ms ( 5723 req/s) | 704.3 bytes/req | 6.0 allocs/req | 1 GCs | heap=3848 KiB
|
|
floor 5000 reqs in 428ms ( 11670 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2568 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:2.08x allocs:1.12x throughput:0.80x
|
|
ZAP-HTTP vs net/http bytes:3541.53x allocs:7461.32x throughput:2.02x
|
|
native ZAP vs fasthttp bytes:0.01x allocs:0.00x throughput:1.28x
|
|
native ZAP vs net/http bytes:14.84x allocs:20.39x throughput:3.22x
|
|
=== RUN TestMemoryPressure/chat
|
|
bench_test.go:472:
|
|
workload: chat (body=824 B, headers=12)
|
|
net/http 5000 reqs in 2.136s ( 2341 req/s) | 14752.3 bytes/req | 151.6 allocs/req | 55 GCs | heap=4056 KiB
|
|
fasthttp 5000 reqs in 2.363s ( 2116 req/s) | 8.5 bytes/req | 0.0 allocs/req | 0 GCs | heap=2776 KiB
|
|
ZAP-HTTP 5000 reqs in 2.373s ( 2107 req/s) | 4.7 bytes/req | 0.0 allocs/req | 0 GCs | heap=2856 KiB
|
|
native ZAP 5000 reqs in 2.297s ( 2177 req/s) | 1984.1 bytes/req | 6.0 allocs/req | 4 GCs | heap=4120 KiB
|
|
floor 5000 reqs in 616ms ( 8119 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2664 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:1.81x allocs:1.15x throughput:1.00x
|
|
ZAP-HTTP vs net/http bytes:3132.91x allocs:6532.55x throughput:0.90x
|
|
native ZAP vs fasthttp bytes:0.00x allocs:0.00x throughput:1.03x
|
|
native ZAP vs net/http bytes:7.44x allocs:25.25x throughput:0.93x
|
|
=== RUN TestMemoryPressure/medium
|
|
bench_test.go:472:
|
|
workload: medium (body=4096 B, headers=12)
|
|
net/http 5000 reqs in 3.9s ( 1282 req/s) | 24746.8 bytes/req | 158.9 allocs/req | 88 GCs | heap=4736 KiB
|
|
fasthttp 5000 reqs in 1.268s ( 3943 req/s) | 10.8 bytes/req | 0.0 allocs/req | 0 GCs | heap=2784 KiB
|
|
ZAP-HTTP 5000 reqs in 1.567s ( 3190 req/s) | 6.6 bytes/req | 0.0 allocs/req | 0 GCs | heap=2824 KiB
|
|
native ZAP 5000 reqs in 1.743s ( 2869 req/s) | 8384.0 bytes/req | 6.0 allocs/req | 17 GCs | heap=3448 KiB
|
|
floor 5000 reqs in 1.345s ( 3717 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2712 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:1.63x allocs:1.11x throughput:0.81x
|
|
ZAP-HTTP vs net/http bytes:3723.34x allocs:6969.84x throughput:2.49x
|
|
native ZAP vs fasthttp bytes:0.00x allocs:0.00x throughput:0.73x
|
|
native ZAP vs net/http bytes:2.95x allocs:26.48x throughput:2.24x
|
|
=== RUN TestMemoryPressure/large
|
|
bench_test.go:472:
|
|
workload: large (body=65536 B, headers=8)
|
|
net/http 5000 reqs in 12.36s ( 405 req/s) | 191015.1 bytes/req | 143.1 allocs/req | 513 GCs | heap=4352 KiB
|
|
fasthttp 5000 reqs in 3.486s ( 1434 req/s) | 73.4 bytes/req | 0.0 allocs/req | 0 GCs | heap=3160 KiB
|
|
ZAP-HTTP 5000 reqs in 3.573s ( 1399 req/s) | 68.5 bytes/req | 0.0 allocs/req | 0 GCs | heap=3584 KiB
|
|
native ZAP 5000 reqs in 9.106s ( 549 req/s) | 131267.7 bytes/req | 6.0 allocs/req | 112 GCs | heap=4664 KiB
|
|
floor 5000 reqs in 7.372s ( 678 req/s) | 0.1 bytes/req | 0.0 allocs/req | 0 GCs | heap=2736 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:1.07x allocs:1.27x throughput:0.98x
|
|
ZAP-HTTP vs net/http bytes:2788.44x allocs:7693.23x throughput:3.46x
|
|
native ZAP vs fasthttp bytes:0.00x allocs:0.00x throughput:0.38x
|
|
native ZAP vs net/http bytes:1.46x allocs:23.74x throughput:1.36x
|
|
--- PASS: TestMemoryPressure (74.10s)
|
|
--- PASS: TestMemoryPressure/tiny (8.27s)
|
|
--- PASS: TestMemoryPressure/small (7.33s)
|
|
--- PASS: TestMemoryPressure/chat (10.64s)
|
|
--- PASS: TestMemoryPressure/medium (10.43s)
|
|
--- PASS: TestMemoryPressure/large (37.40s)
|
|
=== RUN TestMemoryPressure
|
|
=== RUN TestMemoryPressure/tiny
|
|
bench_test.go:472:
|
|
workload: tiny (body=16 B, headers=4)
|
|
net/http 5000 reqs in 9.109s ( 549 req/s) | 10253.3 bytes/req | 115.3 allocs/req | 29 GCs | heap=4848 KiB
|
|
fasthttp 5000 reqs in 3.924s ( 1274 req/s) | 6.7 bytes/req | 0.0 allocs/req | 0 GCs | heap=2672 KiB
|
|
ZAP-HTTP 5000 reqs in 2.382s ( 2099 req/s) | 2.8 bytes/req | 0.0 allocs/req | 0 GCs | heap=2720 KiB
|
|
native ZAP 5000 reqs in 2.157s ( 2318 req/s) | 224.0 bytes/req | 6.0 allocs/req | 0 GCs | heap=3648 KiB
|
|
floor 5000 reqs in 2.524s ( 1981 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2608 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:2.38x allocs:1.28x throughput:1.65x
|
|
ZAP-HTTP vs net/http bytes:3641.10x allocs:7689.11x throughput:3.82x
|
|
native ZAP vs fasthttp bytes:0.03x allocs:0.00x throughput:1.82x
|
|
native ZAP vs net/http bytes:45.77x allocs:19.22x throughput:4.22x
|
|
=== RUN TestMemoryPressure/small
|
|
bench_test.go:472:
|
|
workload: small (body=256 B, headers=8)
|
|
net/http 5000 reqs in 6.77s ( 739 req/s) | 10446.0 bytes/req | 122.4 allocs/req | 34 GCs | heap=4000 KiB
|
|
fasthttp 5000 reqs in 3.309s ( 1511 req/s) | 7.1 bytes/req | 0.0 allocs/req | 0 GCs | heap=2840 KiB
|
|
ZAP-HTTP 5000 reqs in 2.373s ( 2107 req/s) | 3.2 bytes/req | 0.0 allocs/req | 0 GCs | heap=2864 KiB
|
|
native ZAP 5000 reqs in 2.204s ( 2269 req/s) | 704.1 bytes/req | 6.0 allocs/req | 1 GCs | heap=4072 KiB
|
|
floor 5000 reqs in 1.841s ( 2715 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2776 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:2.22x allocs:1.27x throughput:1.39x
|
|
ZAP-HTTP vs net/http bytes:3274.19x allocs:7116.20x throughput:2.85x
|
|
native ZAP vs fasthttp bytes:0.01x allocs:0.00x throughput:1.50x
|
|
native ZAP vs net/http bytes:14.84x allocs:20.39x throughput:3.07x
|
|
=== RUN TestMemoryPressure/chat
|
|
bench_test.go:472:
|
|
workload: chat (body=824 B, headers=12)
|
|
net/http 5000 reqs in 11.094s ( 451 req/s) | 14681.3 bytes/req | 151.5 allocs/req | 46 GCs | heap=3920 KiB
|
|
fasthttp 5000 reqs in 3.378s ( 1480 req/s) | 9.2 bytes/req | 0.0 allocs/req | 0 GCs | heap=2800 KiB
|
|
ZAP-HTTP 5000 reqs in 2.489s ( 2009 req/s) | 4.1 bytes/req | 0.0 allocs/req | 0 GCs | heap=2824 KiB
|
|
native ZAP 5000 reqs in 2.685s ( 1862 req/s) | 1984.1 bytes/req | 6.0 allocs/req | 4 GCs | heap=4504 KiB
|
|
floor 5000 reqs in 1.306s ( 3830 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2672 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:2.25x allocs:1.21x throughput:1.36x
|
|
ZAP-HTTP vs net/http bytes:3598.35x allocs:6704.65x throughput:4.46x
|
|
native ZAP vs fasthttp bytes:0.00x allocs:0.00x throughput:1.26x
|
|
native ZAP vs net/http bytes:7.40x allocs:25.24x throughput:4.13x
|
|
=== RUN TestMemoryPressure/medium
|
|
bench_test.go:472:
|
|
workload: medium (body=4096 B, headers=12)
|
|
net/http 5000 reqs in 6.602s ( 757 req/s) | 24985.4 bytes/req | 159.0 allocs/req | 91 GCs | heap=4072 KiB
|
|
fasthttp 5000 reqs in 1.311s ( 3814 req/s) | 11.6 bytes/req | 0.0 allocs/req | 0 GCs | heap=2792 KiB
|
|
ZAP-HTTP 5000 reqs in 2.429s ( 2059 req/s) | 7.6 bytes/req | 0.0 allocs/req | 0 GCs | heap=2856 KiB
|
|
native ZAP 5000 reqs in 2.315s ( 2160 req/s) | 8384.2 bytes/req | 6.0 allocs/req | 12 GCs | heap=4512 KiB
|
|
floor 5000 reqs in 964ms ( 5187 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2744 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:1.52x allocs:1.12x throughput:0.54x
|
|
ZAP-HTTP vs net/http bytes:3270.35x allocs:6854.51x throughput:2.72x
|
|
native ZAP vs fasthttp bytes:0.00x allocs:0.00x throughput:0.57x
|
|
native ZAP vs net/http bytes:2.98x allocs:26.49x throughput:2.85x
|
|
=== RUN TestMemoryPressure/large
|
|
bench_test.go:472:
|
|
workload: large (body=65536 B, headers=8)
|
|
net/http 5000 reqs in 13.029s ( 384 req/s) | 188602.8 bytes/req | 142.3 allocs/req | 396 GCs | heap=4512 KiB
|
|
fasthttp 5000 reqs in 3.627s ( 1379 req/s) | 72.5 bytes/req | 0.0 allocs/req | 0 GCs | heap=3272 KiB
|
|
ZAP-HTTP 5000 reqs in 4.099s ( 1220 req/s) | 55.2 bytes/req | 0.0 allocs/req | 0 GCs | heap=3648 KiB
|
|
native ZAP 5000 reqs in 5.853s ( 854 req/s) | 131268.3 bytes/req | 6.0 allocs/req | 150 GCs | heap=5944 KiB
|
|
floor 5000 reqs in 2.279s ( 2194 req/s) | 0.0 bytes/req | 0.0 allocs/req | 0 GCs | heap=2864 KiB
|
|
ZAP-HTTP vs fasthttp (wire only) bytes:1.31x allocs:1.29x throughput:0.88x
|
|
ZAP-HTTP vs net/http bytes:3416.52x allocs:7992.57x throughput:3.18x
|
|
native ZAP vs fasthttp bytes:0.00x allocs:0.00x throughput:0.62x
|
|
native ZAP vs net/http bytes:1.44x allocs:23.59x throughput:2.23x
|
|
--- PASS: TestMemoryPressure (107.80s)
|
|
--- PASS: TestMemoryPressure/tiny (22.72s)
|
|
--- PASS: TestMemoryPressure/small (18.11s)
|
|
--- PASS: TestMemoryPressure/chat (22.54s)
|
|
--- PASS: TestMemoryPressure/medium (14.75s)
|
|
--- PASS: TestMemoryPressure/large (29.69s)
|
|
PASS
|
|
ok github.com/zap-proto/bench 234.984s
|
|
|
|
=== wire bytes ===
|
|
=== RUN TestWireBytes
|
|
=== RUN TestWireBytes/tiny
|
|
bench_test.go:505:
|
|
workload: tiny (body=16 B, headers=4)
|
|
HTTP/1.1 wire: 243 bytes
|
|
ZAP-HTTP wire: 272 bytes (1.12x HTTP, 11.9% larger)
|
|
native ZAP wire: 212 bytes (0.87x HTTP, 12.8% smaller)
|
|
=== RUN TestWireBytes/small
|
|
bench_test.go:505:
|
|
workload: small (body=256 B, headers=8)
|
|
HTTP/1.1 wire: 562 bytes
|
|
ZAP-HTTP wire: 638 bytes (1.14x HTTP, 13.5% larger)
|
|
native ZAP wire: 580 bytes (1.03x HTTP, 3.2% larger)
|
|
=== RUN TestWireBytes/chat
|
|
bench_test.go:505:
|
|
workload: chat (body=824 B, headers=12)
|
|
HTTP/1.1 wire: 1340 bytes
|
|
ZAP-HTTP wire: 1432 bytes (1.07x HTTP, 6.9% larger)
|
|
native ZAP wire: 1372 bytes (1.02x HTTP, 2.4% larger)
|
|
=== RUN TestWireBytes/medium
|
|
bench_test.go:505:
|
|
workload: medium (body=4096 B, headers=12)
|
|
HTTP/1.1 wire: 4613 bytes
|
|
ZAP-HTTP wire: 4704 bytes (1.02x HTTP, 2.0% larger)
|
|
native ZAP wire: 4644 bytes (1.01x HTTP, 0.7% larger)
|
|
=== RUN TestWireBytes/large
|
|
bench_test.go:505:
|
|
workload: large (body=65536 B, headers=8)
|
|
HTTP/1.1 wire: 65844 bytes
|
|
ZAP-HTTP wire: 65918 bytes (1.00x HTTP, 0.1% larger)
|
|
native ZAP wire: 65860 bytes (1.00x HTTP, 0.0% larger)
|
|
--- PASS: TestWireBytes (0.00s)
|
|
--- PASS: TestWireBytes/tiny (0.00s)
|
|
--- PASS: TestWireBytes/small (0.00s)
|
|
--- PASS: TestWireBytes/chat (0.00s)
|
|
--- PASS: TestWireBytes/medium (0.00s)
|
|
--- PASS: TestWireBytes/large (0.00s)
|
|
PASS
|
|
ok github.com/zap-proto/bench 0.383s
|
|
|
|
=== concurrent throughput: NOT MEASURED ===
|
|
Load average was 102.21 on 10 cores when this run started. A rate
|
|
taken there is scheduler contention, not transport cost. Re-run on a
|
|
quiet machine to fill this section in.
|
|
|
|
=== per-op benchmarks: NOT MEASURED ===
|
|
Same reason. The B/op and allocs/op columns these would report are
|
|
already covered by the allocation section above, which is valid at
|
|
any load; only the ns/op column needs a quiet box, so the whole
|
|
section waits for one.
|
|
|
|
=== environment after ===
|
|
20:25 up 7 days, 8:12, 1 user, load averages: 102.13 109.55 126.89
|