comparing benchmarks for some golang IP routing table implementations:
github.com/tailscale/art
github.com/gaissmai/bart
github.com/k-sone/critbitgo
github.com/Asphaltt/lpmtrie
github.com/gaissmai/cidrtree
github.com/yl2chen/cidranger
The ~1_000_000 Tier1 prefix test records (IPv4 and IPv6 routes) are from a full routing table with typical ISP prefix distribution.
In comparison, the prefix lengths for the random test sets are equally distributed between 1-32 for IPv4 and 1-128 bits for IPv6, the randomly generated default-routes with prefix length 0 have been sorted out, they distorts the lookup times and there is no lookup miss at all.
The RandomPrefixes without IP versions labeling are composed of a distribution of 4 parts IPv4 to 1 part IPv6 prefixes, which is approximately the current ratio in the Internet backbone routers.
$ make dep
$ make -B all
The memory consumption of the multibit trie art with more than
100_000 randomly distributed IPv6 prefixes brings the OOM killer in action.
bart has a factor of 10 lower memory consumption compared to art, but is by
a factor of 2 slower in lookup times.
cidrtree is the most economical in terms of memory consumption,
but this is also not a trie but a binary search tree and
slower by a magnitude than the other algorithms.
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/size.bm │ art/size.bm │
│ bytes │ bytes vs base │
Tier1PfxSize/100 42.82Ki ± 4% 802.97Ki ± 0% +1775.21% (p=0.002 n=6)
Tier1PfxSize/1_000 288.9Ki ± 1% 7420.4Ki ± 0% +2468.17% (p=0.002 n=6)
Tier1PfxSize/10_000 1.822Mi ± 0% 47.172Mi ± 0% +2488.91% (p=0.002 n=6)
Tier1PfxSize/100_000 7.545Mi ± 0% 160.300Mi ± 0% +2024.67% (p=0.002 n=6)
Tier1PfxSize/1_000_000 34.33Mi ± 0% 378.23Mi ± 0% +1001.89% (p=0.002 n=6)
RandomPfx4Size/100 43.33Ki ± 4% 706.59Ki ± 0% +1530.80% (p=0.002 n=6)
RandomPfx4Size/1_000 301.1Ki ± 1% 7286.5Ki ± 0% +2319.96% (p=0.002 n=6)
RandomPfx4Size/10_000 2.422Mi ± 0% 57.886Mi ± 0% +2290.43% (p=0.002 n=6)
RandomPfx4Size/100_000 18.27Mi ± 0% 523.23Mi ± 0% +2764.14% (p=0.002 n=6)
RandomPfx4Size/1_000_000 109.8Mi ± 0%
RandomPfx6Size/100 150.2Ki ± 1% 700.2Ki ± 0% +366.29% (p=0.002 n=6)
RandomPfx6Size/1_000 1.319Mi ± 0% 7.477Mi ± 0% +466.92% (p=0.002 n=6)
RandomPfx6Size/10_000 12.29Mi ± 0% 65.51Mi ± 0% +433.02% (p=0.002 n=6)
RandomPfx6Size/100_000 115.3Mi ± 0% 749.0Mi ± 0% +549.46% (p=0.002 n=6)
RandomPfx6Size/1_000_000 1.049Gi ± 0%
RandomPfxSize/100 68.22Ki ± 2% 694.25Ki ± 0% +917.68% (p=0.002 n=6)
RandomPfxSize/1_000 521.2Ki ± 0% 7439.5Ki ± 0% +1327.38% (p=0.002 n=6)
RandomPfxSize/10_000 4.400Mi ± 0% 59.641Mi ± 0% +1255.38% (p=0.002 n=6)
RandomPfxSize/100_000 39.02Mi ± 0% 553.83Mi ± 0% +1319.37% (p=0.002 n=6)
RandomPfxSize/1_000_000 309.1Mi ± 0%
geomean 3.652Mi 22.52Mi +1266.37% ¹
¹ benchmark set differs from baseline; geomeans may not be comparable
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/size.bm │ cidrtree/size.bm │
│ bytes │ bytes vs base │
Tier1PfxSize/100 42.82Ki ± 4% 12.48Ki ± 14% -70.86% (p=0.002 n=6)
Tier1PfxSize/1_000 288.94Ki ± 1% 81.51Ki ± 2% -71.79% (p=0.002 n=6)
Tier1PfxSize/10_000 1865.8Ki ± 0% 784.6Ki ± 0% -57.95% (p=0.002 n=6)
Tier1PfxSize/100_000 7.545Mi ± 0% 7.633Mi ± 0% +1.17% (p=0.002 n=6)
Tier1PfxSize/1_000_000 34.33Mi ± 0% 76.30Mi ± 0% +122.27% (p=0.002 n=6)
RandomPfx4Size/100 43.33Ki ± 4% 11.58Ki ± 14% -73.28% (p=0.002 n=6)
RandomPfx4Size/1_000 301.10Ki ± 1% 81.51Ki ± 2% -72.93% (p=0.002 n=6)
RandomPfx4Size/10_000 2479.7Ki ± 0% 784.6Ki ± 0% -68.36% (p=0.002 n=6)
RandomPfx4Size/100_000 18.268Mi ± 0% 7.633Mi ± 0% -58.22% (p=0.002 n=6)
RandomPfx4Size/1_000_000 109.79Mi ± 0% 76.30Mi ± 0% -30.51% (p=0.002 n=6)
RandomPfx6Size/100 150.16Ki ± 1% 11.58Ki ± 14% -92.29% (p=0.002 n=6)
RandomPfx6Size/1_000 1350.49Ki ± 0% 81.51Ki ± 2% -93.96% (p=0.002 n=6)
RandomPfx6Size/10_000 12585.8Ki ± 0% 784.6Ki ± 0% -93.77% (p=0.002 n=6)
RandomPfx6Size/100_000 115.320Mi ± 0% 7.633Mi ± 0% -93.38% (p=0.002 n=6)
RandomPfx6Size/1_000_000 1074.54Mi ± 0% 76.30Mi ± 0% -92.90% (p=0.002 n=6)
RandomPfxSize/100 68.22Ki ± 2% 11.58Ki ± 14% -83.03% (p=0.002 n=6)
RandomPfxSize/1_000 521.20Ki ± 0% 81.51Ki ± 2% -84.36% (p=0.002 n=6)
RandomPfxSize/10_000 4505.9Ki ± 0% 784.6Ki ± 0% -82.59% (p=0.002 n=6)
RandomPfxSize/100_000 39.019Mi ± 0% 7.633Mi ± 0% -80.44% (p=0.002 n=6)
RandomPfxSize/1_000_000 309.15Mi ± 0% 76.30Mi ± 0% -75.32% (p=0.002 n=6)
geomean 3.652Mi 856.4Ki -77.10%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/size.bm │ critbitgo/size.bm │
│ bytes │ bytes vs base │
Tier1PfxSize/100 42.82Ki ± 4% 15.63Ki ± 11% -63.49% (p=0.002 n=6)
Tier1PfxSize/1_000 288.9Ki ± 1% 115.2Ki ± 1% -60.12% (p=0.002 n=6)
Tier1PfxSize/10_000 1.822Mi ± 0% 1.094Mi ± 0% -39.98% (p=0.002 n=6)
Tier1PfxSize/100_000 7.545Mi ± 0% 10.913Mi ± 0% +44.64% (p=0.002 n=6)
Tier1PfxSize/1_000_000 34.33Mi ± 0% 109.12Mi ± 0% +217.88% (p=0.002 n=6)
RandomPfx4Size/100 43.33Ki ± 4% 14.67Ki ± 11% -66.14% (p=0.002 n=6)
RandomPfx4Size/1_000 301.1Ki ± 1% 112.7Ki ± 1% -62.56% (p=0.002 n=6)
RandomPfx4Size/10_000 2.422Mi ± 0% 1.071Mi ± 0% -55.76% (p=0.002 n=6)
RandomPfx4Size/100_000 18.27Mi ± 0% 10.68Mi ± 0% -41.51% (p=0.002 n=6)
RandomPfx4Size/1_000_000 109.8Mi ± 0% 106.8Mi ± 0% -2.71% (p=0.002 n=6)
RandomPfx6Size/100 150.16Ki ± 1% 16.22Ki ± 10% -89.20% (p=0.002 n=6)
RandomPfx6Size/1_000 1350.5Ki ± 0% 128.3Ki ± 1% -90.50% (p=0.002 n=6)
RandomPfx6Size/10_000 12.291Mi ± 0% 1.224Mi ± 0% -90.04% (p=0.002 n=6)
RandomPfx6Size/100_000 115.32Mi ± 0% 12.21Mi ± 0% -89.41% (p=0.002 n=6)
RandomPfx6Size/1_000_000 1074.5Mi ± 0% 122.1Mi ± 0% -88.64% (p=0.002 n=6)
RandomPfxSize/100 68.22Ki ± 2% 14.94Ki ± 11% -78.10% (p=0.002 n=6)
RandomPfxSize/1_000 521.2Ki ± 0% 115.8Ki ± 1% -77.79% (p=0.002 n=6)
RandomPfxSize/10_000 4.400Mi ± 0% 1.102Mi ± 0% -74.96% (p=0.002 n=6)
RandomPfxSize/100_000 39.02Mi ± 0% 10.99Mi ± 0% -71.83% (p=0.002 n=6)
RandomPfxSize/1_000_000 309.1Mi ± 0% 109.9Mi ± 0% -64.46% (p=0.002 n=6)
geomean 3.652Mi 1.193Mi -67.33%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/size.bm │ lpmtrie/size.bm │
│ bytes │ bytes vs base │
Tier1PfxSize/100 42.82Ki ± 4% 24.25Ki ± 9% -43.37% (p=0.002 n=6)
Tier1PfxSize/1_000 288.9Ki ± 1% 202.3Ki ± 4% -29.97% (p=0.002 n=6)
Tier1PfxSize/10_000 1.822Mi ± 0% 1.942Mi ± 3% +6.58% (p=0.002 n=6)
Tier1PfxSize/100_000 7.545Mi ± 0% 19.330Mi ± 3% +156.21% (p=0.002 n=6)
Tier1PfxSize/1_000_000 34.33Mi ± 0% 189.92Mi ± 6% +453.28% (p=0.002 n=6)
RandomPfx4Size/100 43.33Ki ± 4% 23.20Ki ± 9% -46.45% (p=0.002 n=6)
RandomPfx4Size/1_000 301.1Ki ± 1% 198.5Ki ± 3% -34.07% (p=0.002 n=6)
RandomPfx4Size/10_000 2.422Mi ± 0% 1.898Mi ± 3% -21.64% (p=0.002 n=6)
RandomPfx4Size/100_000 18.27Mi ± 0% 18.28Mi ± 5% ~ (p=0.374 n=6)
RandomPfx4Size/1_000_000 109.8Mi ± 0% 163.5Mi ± 8% +48.93% (p=0.002 n=6)
RandomPfx6Size/100 150.16Ki ± 1% 25.53Ki ± 12% -83.00% (p=0.002 n=6)
RandomPfx6Size/1_000 1350.5Ki ± 0% 222.0Ki ± 8% -83.56% (p=0.002 n=6)
RandomPfx6Size/10_000 12.291Mi ± 0% 2.129Mi ± 7% -82.68% (p=0.002 n=6)
RandomPfx6Size/100_000 115.32Mi ± 0% 20.93Mi ± 8% -81.85% (p=0.002 n=6)
RandomPfx6Size/1_000_000 1074.5Mi ± 0% 204.0Mi ± 8% -81.02% (p=0.002 n=6)
RandomPfxSize/100 68.22Ki ± 2% 23.56Ki ± 10% -65.46% (p=0.002 n=6)
RandomPfxSize/1_000 521.2Ki ± 0% 202.8Ki ± 4% -61.10% (p=0.002 n=6)
RandomPfxSize/10_000 4.400Mi ± 0% 1.943Mi ± 4% -55.85% (p=0.002 n=6)
RandomPfxSize/100_000 39.02Mi ± 0% 18.78Mi ± 5% -51.86% (p=0.002 n=6)
RandomPfxSize/1_000_000 309.1Mi ± 0% 171.9Mi ± 7% -44.39% (p=0.002 n=6)
geomean 3.652Mi 2.011Mi -44.95%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/size.bm │ cidranger/size.bm │
│ bytes │ bytes vs base │
Tier1PfxSize/100 42.82Ki ± 4% 56.03Ki ± 5% +30.85% (p=0.002 n=6)
Tier1PfxSize/1_000 288.9Ki ± 1% 525.9Ki ± 3% +82.02% (p=0.002 n=6)
Tier1PfxSize/10_000 1.822Mi ± 0% 5.091Mi ± 2% +179.41% (p=0.002 n=6)
Tier1PfxSize/100_000 7.545Mi ± 0% 50.258Mi ± 2% +566.14% (p=0.002 n=6)
Tier1PfxSize/1_000_000 34.33Mi ± 0% 477.45Mi ± 2% +1290.93% (p=0.002 n=6)
RandomPfx4Size/100 43.33Ki ± 4% 54.16Ki ± 6% +24.99% (p=0.002 n=6)
RandomPfx4Size/1_000 301.1Ki ± 1% 514.0Ki ± 3% +70.69% (p=0.002 n=6)
RandomPfx4Size/10_000 2.422Mi ± 0% 4.900Mi ± 3% +102.35% (p=0.002 n=6)
RandomPfx4Size/100_000 18.27Mi ± 0% 45.76Mi ± 3% +150.47% (p=0.002 n=6)
RandomPfx4Size/1_000_000 109.8Mi ± 0% 396.2Mi ± 3% +260.87% (p=0.002 n=6)
RandomPfx6Size/100 150.16Ki ± 1% 61.08Ki ± 3% -59.33% (p=0.002 n=6)
RandomPfx6Size/1_000 1350.5Ki ± 0% 579.3Ki ± 0% -57.11% (p=0.002 n=6)
RandomPfx6Size/10_000 12.291Mi ± 0% 5.582Mi ± 0% -54.59% (p=0.002 n=6)
RandomPfx6Size/100_000 115.32Mi ± 0% 54.75Mi ± 0% -52.52% (p=0.002 n=6)
RandomPfx6Size/1_000_000 1074.5Mi ± 0% 534.1Mi ± 0% -50.29% (p=0.002 n=6)
RandomPfxSize/100 68.22Ki ± 2% 55.07Ki ± 5% -19.27% (p=0.002 n=6)
RandomPfxSize/1_000 521.2Ki ± 0% 524.9Ki ± 2% +0.70% (p=0.026 n=6)
RandomPfxSize/10_000 4.400Mi ± 0% 5.024Mi ± 2% +14.18% (p=0.002 n=6)
RandomPfxSize/100_000 39.02Mi ± 0% 47.53Mi ± 2% +21.82% (p=0.002 n=6)
RandomPfxSize/1_000_000 309.1Mi ± 0% 424.9Mi ± 2% +37.44% (p=0.002 n=6)
geomean 3.652Mi 5.068Mi +38.75%
In the lookup, art is the champion, closely followed by bart.
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/lookup.bm │ art/lookup.bm │
│ sec/op │ sec/op vs base │
LpmTier1Pfxs/RandomMatchIP4 59.84n ± 31% 48.52n ± 16% ~ (p=0.075 n=10)
LpmTier1Pfxs/RandomMatchIP6 72.45n ± 38% 47.73n ± 4% -34.12% (p=0.022 n=10)
LpmTier1Pfxs/RandomMissIP4 97.48n ± 23% 31.57n ± 0% -67.61% (p=0.000 n=10)
LpmTier1Pfxs/RandomMissIP6 21.00n ± 4% 31.20n ± 0% +48.54% (p=0.001 n=10)
LpmRandomPfxs100_000/RandomMatchIP4 77.81n ± 28% 56.62n ± 15% -27.24% (p=0.019 n=10)
LpmRandomPfxs100_000/RandomMatchIP6 63.97n ± 26% 48.90n ± 17% ~ (p=0.052 n=10)
LpmRandomPfxs100_000/RandomMissIP4 92.05n ± 20% 31.51n ± 1% -65.77% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMissIP6 77.34n ± 37% 31.16n ± 1% -59.71% (p=0.000 n=10)
geomean 64.89n 39.72n -38.79%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/lookup.bm │ cidrtree/lookup.bm │
│ sec/op │ sec/op vs base │
LpmTier1Pfxs/RandomMatchIP4 59.84n ± 31% 1144.50n ± 35% +1812.60% (p=0.000 n=10)
LpmTier1Pfxs/RandomMatchIP6 72.45n ± 38% 1443.50n ± 15% +1892.55% (p=0.000 n=10)
LpmTier1Pfxs/RandomMissIP4 97.48n ± 23% 1183.50n ± 44% +1114.10% (p=0.002 n=10)
LpmTier1Pfxs/RandomMissIP6 21.00n ± 4% 70.51n ± 2012% +235.71% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMatchIP4 77.81n ± 28% 1236.50n ± 34% +1489.03% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMatchIP6 63.97n ± 26% 1432.00n ± 32% +2138.72% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMissIP4 92.05n ± 20% 1286.50n ± 23% +1297.61% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMissIP6 77.34n ± 37% 1431.50n ± 14% +1750.92% (p=0.000 n=10)
geomean 64.89n 904.9n +1294.42%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/lookup.bm │ critbitgo/lookup.bm │
│ sec/op │ sec/op vs base │
LpmTier1Pfxs/RandomMatchIP4 59.84n ± 31% 373.90n ± 16% +524.83% (p=0.000 n=10)
LpmTier1Pfxs/RandomMatchIP6 72.45n ± 38% 699.95n ± 56% +866.18% (p=0.000 n=10)
LpmTier1Pfxs/RandomMissIP4 97.48n ± 23% 772.05n ± 49% +692.01% (p=0.000 n=10)
LpmTier1Pfxs/RandomMissIP6 21.00n ± 4% 757.05n ± 21% +3504.14% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMatchIP4 77.81n ± 28% 337.50n ± 53% +333.72% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMatchIP6 63.97n ± 26% 697.65n ± 12% +990.67% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMissIP4 92.05n ± 20% 560.25n ± 40% +508.64% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMissIP6 77.34n ± 37% 631.00n ± 24% +715.88% (p=0.000 n=10)
geomean 64.89n 579.6n +793.11%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/lookup.bm │ lpmtrie/lookup.bm │
│ sec/op │ sec/op vs base │
LpmTier1Pfxs/RandomMatchIP4 59.84n ± 31% 245.50n ± 19% +310.26% (p=0.000 n=10)
LpmTier1Pfxs/RandomMatchIP6 72.45n ± 38% 250.45n ± 63% +245.71% (p=0.000 n=10)
LpmTier1Pfxs/RandomMissIP4 97.48n ± 23% 233.20n ± 67% +139.23% (p=0.029 n=10)
LpmTier1Pfxs/RandomMissIP6 21.00n ± 4% 88.63n ± 21% +321.92% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMatchIP4 77.81n ± 28% 259.75n ± 9% +233.80% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMatchIP6 63.97n ± 26% 199.95n ± 7% +212.59% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMissIP4 92.05n ± 20% 278.10n ± 5% +202.12% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMissIP6 77.34n ± 37% 235.70n ± 4% +204.76% (p=0.000 n=10)
geomean 64.89n 213.6n +229.08%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/lookup.bm │ cidranger/lookup.bm │
│ sec/op │ sec/op vs base │
LpmTier1Pfxs/RandomMatchIP4 59.84n ± 31% 351.55n ± 44% +487.48% (p=0.000 n=10)
LpmTier1Pfxs/RandomMatchIP6 72.45n ± 38% 292.70n ± 42% +304.03% (p=0.000 n=10)
LpmTier1Pfxs/RandomMissIP4 97.48n ± 23% 206.50n ± 56% +111.84% (p=0.000 n=10)
LpmTier1Pfxs/RandomMissIP6 21.00n ± 4% 160.50n ± 24% +664.10% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMatchIP4 77.81n ± 28% 1352.50n ± 61% +1638.10% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMatchIP6 63.97n ± 26% 481.05n ± 43% +652.05% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMissIP4 92.05n ± 20% 452.45n ± 24% +391.53% (p=0.000 n=10)
LpmRandomPfxs100_000/RandomMissIP6 77.34n ± 37% 400.55n ± 21% +417.91% (p=0.000 n=10)
geomean 64.89n 376.3n +479.88%
bart is by far the fastest algorithm for updates under all competitors.
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/update.bm │ art/update.bm │
│ sec/op │ sec/op vs base │
Insert/Insert_into_100 54.36n ± 0% 105.85n ± 1% +94.72% (p=0.000 n=10)
Insert/Insert_into_1_000 54.35n ± 0% 125.45n ± 1% +130.82% (p=0.000 n=10)
Insert/Insert_into_10_000 54.36n ± 0% 129.05n ± 1% +137.40% (p=0.000 n=10)
Insert/Insert_into_100_000 54.37n ± 0% 120.55n ± 2% +121.74% (p=0.000 n=10)
Insert/Insert_into_1_000_000 64.71n ± 0% 121.00n ± 0% +86.99% (p=0.000 n=10)
Delete/Delete_from_100 18.53n ± 0% 34.04n ± 3% +83.70% (p=0.000 n=10)
Delete/Delete_from_1_000 43.07n ± 0% 59.06n ± 0% +37.13% (p=0.000 n=10)
Delete/Delete_from_10_000 43.09n ± 0% 58.93n ± 0% +36.75% (p=0.000 n=10)
Delete/Delete_from_100_000 45.93n ± 0% 58.71n ± 0% +27.82% (p=0.000 n=10)
Delete/Delete_from_1_000_000 45.68n ± 0% 58.50n ± 0% +28.08% (p=0.000 n=10)
geomean 45.82n 79.57n +73.66%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/update.bm │ cidrtree/update.bm │
│ sec/op │ sec/op vs base │
Insert/Insert_into_100 54.36n ± 0% 945.20n ± 0% +1638.78% (p=0.000 n=10)
Insert/Insert_into_1_000 54.35n ± 0% 1452.00n ± 0% +2571.57% (p=0.000 n=10)
Insert/Insert_into_10_000 54.36n ± 0% 2125.00n ± 0% +3809.12% (p=0.000 n=10)
Insert/Insert_into_100_000 54.37n ± 0% 1595.00n ± 0% +2833.87% (p=0.000 n=10)
Insert/Insert_into_1_000_000 64.71n ± 0% 2627.50n ± 0% +3960.42% (p=0.000 n=10)
Delete/Delete_from_100 18.53n ± 0% 1502.00n ± 0% +8005.77% (p=0.000 n=10)
Delete/Delete_from_1_000 43.07n ± 0% 1765.00n ± 0% +3997.98% (p=0.000 n=10)
Delete/Delete_from_10_000 43.09n ± 0% 3444.00n ± 0% +7892.57% (p=0.000 n=10)
Delete/Delete_from_100_000 45.93n ± 0% 3493.50n ± 0% +7506.14% (p=0.000 n=10)
Delete/Delete_from_1_000_000 45.68n ± 0% 4819.00n ± 0% +10449.47% (p=0.000 n=10)
geomean 45.82n 2.125µ +4538.26%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/update.bm │ critbitgo/update.bm │
│ sec/op │ sec/op vs base │
Insert/Insert_into_100 54.36n ± 0% 109.95n ± 1% +102.26% (p=0.000 n=10)
Insert/Insert_into_1_000 54.35n ± 0% 121.75n ± 1% +124.01% (p=0.000 n=10)
Insert/Insert_into_10_000 54.36n ± 0% 125.50n ± 1% +130.87% (p=0.000 n=10)
Insert/Insert_into_100_000 54.37n ± 0% 147.50n ± 1% +171.31% (p=0.000 n=10)
Insert/Insert_into_1_000_000 64.71n ± 0% 152.20n ± 2% +135.20% (p=0.000 n=10)
Delete/Delete_from_100 18.53n ± 0% 99.93n ± 1% +439.29% (p=0.000 n=10)
Delete/Delete_from_1_000 43.07n ± 0% 108.40n ± 1% +151.68% (p=0.000 n=10)
Delete/Delete_from_10_000 43.09n ± 0% 112.55n ± 1% +161.20% (p=0.000 n=10)
Delete/Delete_from_100_000 45.93n ± 0% 131.40n ± 2% +186.09% (p=0.000 n=10)
Delete/Delete_from_1_000_000 45.68n ± 0% 127.40n ± 2% +178.90% (p=0.000 n=10)
geomean 45.82n 122.6n +167.67%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/update.bm │ lpmtrie/update.bm │
│ sec/op │ sec/op vs base │
Insert/Insert_into_100 54.36n ± 0% 314.55n ± 0% +478.64% (p=0.000 n=10)
Insert/Insert_into_1_000 54.35n ± 0% 352.45n ± 2% +548.48% (p=0.000 n=10)
Insert/Insert_into_10_000 54.36n ± 0% 362.60n ± 1% +567.03% (p=0.000 n=10)
Insert/Insert_into_100_000 54.37n ± 0% 507.25n ± 1% +833.05% (p=0.000 n=10)
Insert/Insert_into_1_000_000 64.71n ± 0% 656.80n ± 1% +914.99% (p=0.000 n=10)
Delete/Delete_from_100 18.53n ± 0% 77.03n ± 3% +315.73% (p=0.000 n=10)
Delete/Delete_from_1_000 43.07n ± 0% 123.30n ± 0% +186.28% (p=0.000 n=10)
Delete/Delete_from_10_000 43.09n ± 0% 149.50n ± 1% +246.95% (p=0.000 n=10)
Delete/Delete_from_100_000 45.93n ± 0% 266.55n ± 0% +480.34% (p=0.000 n=10)
Delete/Delete_from_1_000_000 45.68n ± 0% 309.70n ± 3% +577.98% (p=0.000 n=10)
geomean 45.82n 262.8n +473.53%
goos: linux
goarch: amd64
cpu: Intel(R) Core(TM) i5-8250U CPU @ 1.60GHz
│ bart/update.bm │ cidranger/update.bm │
│ sec/op │ sec/op vs base │
Insert/Insert_into_100 54.36n ± 0% 2442.50n ± 1% +4393.19% (p=0.000 n=10)
Insert/Insert_into_1_000 54.35n ± 0% 4130.00n ± 2% +7498.90% (p=0.000 n=10)
Insert/Insert_into_10_000 54.36n ± 0% 4849.00n ± 1% +8820.16% (p=0.000 n=10)
Insert/Insert_into_100_000 54.37n ± 0% 8110.50n ± 1% +14818.61% (p=0.000 n=10)
Insert/Insert_into_1_000_000 64.71n ± 0% 7579.00n ± 1% +11612.25% (p=0.000 n=10)
Delete/Delete_from_100 18.53n ± 0% 385.05n ± 3% +1977.98% (p=0.000 n=10)
Delete/Delete_from_1_000 43.07n ± 0% 421.45n ± 0% +878.52% (p=0.000 n=10)
Delete/Delete_from_10_000 43.09n ± 0% 442.30n ± 1% +926.46% (p=0.000 n=10)
Delete/Delete_from_100_000 45.93n ± 0% 534.85n ± 8% +1064.49% (p=0.000 n=10)
Delete/Delete_from_1_000_000 45.68n ± 0% 549.85n ± 0% +1103.70% (p=0.000 n=10)
geomean 45.82n 1.514µ +3205.41%