Table 2.
Identity-by-state (Hamming distance) and complete linkage clustering times (sec)
| Calculation | Dataset | Machine | PLINK 1.07 | PLINK 1.90 | Ratio |
|---|---|---|---|---|---|
| IBS matrix only | synth1p | Mac-2 | 2233.6 | 1.9 | 1.2 k |
| Mac-12 | 1320.4 | 1.2 | 1.1 k | ||
| Linux32-8 | 1937.2 | 2.8 | 690 | ||
| Linux64-512 | 1492 | 3.7 | 400 | ||
| Win32-2 | 3219.0 | 7.2 | 450 | ||
| Win64-2 | 2674.4 | 1.5 | 1.8 k | ||
| synth2p | Mac-2 | ∼190 k | 118.8 | 1.6 k | |
| Mac-12 | ∼99 k | 23.5 | 4.2 k | ||
| Linux32-8 | 152.5 k | 214.3 | 710 | ||
| Linux64-512 | ∼98 k | 25.3 | 3.9 k | ||
| Win32-2 | ∼270 k | 654.5 | 410 | ||
| Win64-2 | ∼200 k | 104.6 | 1.9 k | ||
| chr1snp | Mac-2 | ∼26 k | 17.5 | 1.5 k | |
| Mac-12 | 13.4 k | 12.6 | 1.06 k | ||
| Linux32-8 | 18.4 k | 30.9 | 600 | ||
| Linux64-512 | ∼14 k | 43.1 | 320 | ||
| Win32-2 | 32.7 k | 95.9 | 341 | ||
| Win64-2 | ∼26 k | 15.3 | 1.7 k | ||
| Basic clustering | synth1p | Mac-2 | 2315.7 | 2.7 | 860 |
| Mac-12 | 1317.9 | 2.0 | 660 | ||
| Linux32-8 | 1898.7 | 4.1 | 460 | ||
| Linux64-512 | 1496 | 4.5 | 330 | ||
| Win32-2 | 3301.7 | 9.1 | 360 | ||
| Win64-2 | 2724.5 | 1.9 | 1.4 k | ||
| synth2p | Mac-2 | ∼230 k | 245.6 | 940 | |
| Mac-12 | ∼140 k | 123.9 | 1.1 k | ||
| Linux32-8 | 197.1 k | 395.6 | 498 | ||
| Linux64-512 | ∼125 k | 143.3 | 872 | ||
| Win32-2 | ∼440 k | 976.7 | 450 | ||
| Win64-2 | ∼270 k | 127.9 | 2.1 k | ||
| chr1snp | Mac-2 | ∼26 k | 18.4 | 1.4 k | |
| Mac-12 | 13.6 k | 13.5 | 1.01 k | ||
| Linux32-8 | 18.5 k | 33.4 | 554 | ||
| Linux64-512 | ∼14 k | 44.2 | 320 | ||
| Win32-2 | 33.2 k | 95.0 | 349 | ||
| Win64-2 | ∼26 k | 15.8 | 1.6 k | ||
| IBD report | synth1p | Mac-2 | 2230.1 | 12.4 | 180 |
| Mac-12 | 1346.2 | 2.4 | 560 | ||
| Linux32-8 | 2019.9 | 12.4 | 163 | ||
| Linux64-512 | 1494 | 5.0 | 300 | ||
| Win32-2 | 3446.3 | 42.2 | 81.7 | ||
| Win64-2 | 2669.8 | 15.1 | 177 | ||
| synth2p | Mac-2 | ∼190 k | 447.1 | 420 | |
| Mac-12 | ∼99 k | 50.3 | 2.0 k | ||
| Linux32-8 | 161.4 k | 618.7 | 261 | ||
| Linux64-512 | ∼98 k | 57.4 | 1.7 k | ||
| Win32-2 | ∼270 k | 1801.1 | 150 | ||
| Win64-2 | ∼200 k | 541.0 | 370 | ||
| IBD report | chr1snp | Mac-2 | ∼26 k | 24.8 | 1.0 k |
| Mac-12 | 13.4 k | 14.6 | 918 | ||
| Linux32-8 | 18.5 k | 53.5 | 346 | ||
| Linux64-512 | ∼14 k | 46.5 | 300 | ||
| Win32-2 | 33.1 k | 199.2 | 166 | ||
| Win64-2 | ∼26 k | 25.1 | 1.0 k |
Computation of the basic distance matrix is expensive, but has an “embarrassingly parallel” structure. Clustering requires an additional serial step, while the identity-by-descent report includes a pairwise population concordance test which does not benefit from bit-level parallelism, but speedups for both remain greater than 100x on 64-bit systems.