gather/fastgather benchmarks - feb 15, 2024

[ctb]

on SRR606249 at k=31, scaled=1000

Summary

hackmd version

fastgather with zip file

v0.9.0-pre, as of #197/#211 merge (using sourmash core v0.12.1)

• 2m 5s
• 14 GB RAM

        Command being timed: "sourmash scripts fastgather SRR606249.trim.k31.sig.zip /group/ctbrowngrp/sourmash-db/gtdb-rs214/gtdb-rs214-k31.zip -o SRR606249.x.gtdb-rs214.fastgather.csv"
        User time (seconds): 6023.69
        System time (seconds): 31.83
        Percent of CPU this job got: 4837%
        Elapsed (wall clock) time (h:mm:ss or m:ss): 2:05.17
        Average shared text size (kbytes): 0
        Average unshared data size (kbytes): 0
        Average stack size (kbytes): 0
        Average total size (kbytes): 0
        Maximum resident set size (kbytes): 14118520
        Average resident set size (kbytes): 0
        Major (requiring I/O) page faults: 974
        Minor (reclaiming a frame) page faults: 2421007
        Voluntary context switches: 45047
        Involuntary context switches: 151667
        Swaps: 0
        File system inputs: 199312
        File system outputs: 56
        Socket messages sent: 0
        Socket messages received: 0
        Signals delivered: 0
        Page size (bytes): 4096
        Exit status: 0

[ctb]

sourmash gather

as of v4.8.7

42m 26s
14.5 GB of RAM

        Command being timed: "sourmash gather -k 31 SRR606249.trim.k31.sig.gz /group/ctbrowngrp/sourmash-db/gtdb-rs214/gtdb-rs214-k31.zip -o SRR606249.x.gtdb-rs214.csv"
        User time (seconds): 2494.26
        System time (seconds): 58.98
        Percent of CPU this job got: 100%
        Elapsed (wall clock) time (h:mm:ss or m:ss): 42:26.62
        Average shared text size (kbytes): 0
        Average unshared data size (kbytes): 0
        Average stack size (kbytes): 0
        Average total size (kbytes): 0
        Maximum resident set size (kbytes): 14473280
        Average resident set size (kbytes): 0
        Major (requiring I/O) page faults: 95793
        Minor (reclaiming a frame) page faults: 10369270
        Voluntary context switches: 3938
        Involuntary context switches: 210962
        Swaps: 0
        File system inputs: 24523896
        File system outputs: 336
        Socket messages sent: 0
        Socket messages received: 0
        Signals delivered: 0
        Page size (bytes): 4096

[ctb]

fastgather with pathlist

36 min, 1.9 GB!?

        Command being timed: "sourmash scripts fastgather SRR606249.trim.k31.sig.zip gtdb-rs214-k31.d.file.list -o SRR606249.x.gtdb-rs214.fastgather.csv"
        User time (seconds): 7566.15
        System time (seconds): 176.87
        Percent of CPU this job got: 357%
        Elapsed (wall clock) time (h:mm:ss or m:ss): 36:04.07
        Average shared text size (kbytes): 0
        Average unshared data size (kbytes): 0
        Average stack size (kbytes): 0
        Average total size (kbytes): 0
        Maximum resident set size (kbytes): 1916736
        Average resident set size (kbytes): 0
        Major (requiring I/O) page faults: 0
        Minor (reclaiming a frame) page faults: 788119
        Voluntary context switches: 2386015
        Involuntary context switches: 1690833
        Swaps: 0
        File system inputs: 25911264
        File system outputs: 56
        Socket messages sent: 0
        Socket messages received: 0
        Signals delivered: 0
        Page size (bytes): 4096
        Exit status: 0

[ctb]

rocksdb indexing

from zip file, 5 hours, 14 GB RAM.

        Command being timed: "sourmash scripts index -o gtdb-rs214-k31.rocksdb /group/ctbrowngrp/sourmash-db/gtdb-rs214/gtdb-rs214-k31.zip"
        User time (seconds): 47382.74
        System time (seconds): 44254.80
        Percent of CPU this job got: 506%
        Elapsed (wall clock) time (h:mm:ss or m:ss): 5:01:47
        Average shared text size (kbytes): 0
        Average unshared data size (kbytes): 0
        Average stack size (kbytes): 0
        Average total size (kbytes): 0
        Maximum resident set size (kbytes): 14553672
        Average resident set size (kbytes): 0
        Major (requiring I/O) page faults: 41
        Minor (reclaiming a frame) page faults: 46714832
        Voluntary context switches: 1483801815
        Involuntary context switches: 1785402
        Swaps: 0
        File system inputs: 142267440
        File system outputs: 243309352
        Socket messages sent: 0
        Socket messages received: 0
        Signals delivered: 0
        Page size (bytes): 4096
        Exit status: 0

[ctb]

I wonder if the zip memory stuff is related to sourmash-bio/sourmash#2340?

[ctb]

fastmultigather with a rocksdb index

2m 8s, 600 MB RAM!

        Command being timed: "sourmash scripts fastmultigather SRR606249.trim.k31.sig.zip gtdb-rs214-k31.rocksdb"
        User time (seconds): 12.93
        System time (seconds): 21.23
        Percent of CPU this job got: 26%
        Elapsed (wall clock) time (h:mm:ss or m:ss): 2:08.05
        Average shared text size (kbytes): 0
        Average unshared data size (kbytes): 0
        Average stack size (kbytes): 0
        Average total size (kbytes): 0
        Maximum resident set size (kbytes): 600260
        Average resident set size (kbytes): 0
        Major (requiring I/O) page faults: 5
        Minor (reclaiming a frame) page faults: 377146
        Voluntary context switches: 371919
        Involuntary context switches: 223163
        Swaps: 0
        File system inputs: 4682456
        File system outputs: 8
        Socket messages sent: 0
        Socket messages received: 0
        Signals delivered: 0
        Page size (bytes): 4096
        Exit status: 0

[ctb]

old fastgather (v0.8.6) with a zip file

28m 34s!! but only 1.7 GB of RAM.

        Command being timed: "sourmash scripts fastgather SRR606249.trim.k31.sig.gz /group/ctbrowngrp/sourmash-db/gtdb-rs214/gtdb-rs214-k31.zip -o SRR606249.x.gtdb-rs214.fastgather.csv"
        User time (seconds): 6148.02
        System time (seconds): 231.95
        Percent of CPU this job got: 372%
        Elapsed (wall clock) time (h:mm:ss or m:ss): 28:34.14
        Average shared text size (kbytes): 0
        Average unshared data size (kbytes): 0
        Average stack size (kbytes): 0
        Average total size (kbytes): 0
        Maximum resident set size (kbytes): 1679044
        Average resident set size (kbytes): 0
        Major (requiring I/O) page faults: 17
        Minor (reclaiming a frame) page faults: 1791154
        Voluntary context switches: 36714
        Involuntary context switches: 102717
        Swaps: 0
        File system inputs: 24407184
        File system outputs: 25830288
        Socket messages sent: 0
        Socket messages received: 0
        Signals delivered: 0
        Page size (bytes): 4096
        Exit status: 0

[ctb]

old fastgather (v0.8.6) with a pathlist

2m 24s, 1.6 GB of RAM

        Command being timed: "sourmash scripts fastgather SRR606249.trim.k31.sig.gz gtdb-rs214-k31.d.file.list -o SRR606249.x.gtdb-rs214.fastgather.csv"
        User time (seconds): 4591.34
        System time (seconds): 131.89
        Percent of CPU this job got: 3276%
        Elapsed (wall clock) time (h:mm:ss or m:ss): 2:24.16
        Average shared text size (kbytes): 0
        Average unshared data size (kbytes): 0
        Average stack size (kbytes): 0
        Average total size (kbytes): 0
        Maximum resident set size (kbytes): 1621440
        Average resident set size (kbytes): 0
        Major (requiring I/O) page faults: 14
        Minor (reclaiming a frame) page faults: 2114295
        Voluntary context switches: 1582072
        Involuntary context switches: 217324
        Swaps: 0
        File system inputs: 25911264
        File system outputs: 40
        Socket messages sent: 0
        Socket messages received: 0
        Signals delivered: 0
        Page size (bytes): 4096
        Exit status: 0

[ctb]

with #221 -

pathlists with parallel loading

2m 26s, 1.8 GB of RAM.

        Command being timed: "sourmash scripts fastgather SRR606249.trim.k31.sig.zip gtdb-rs214-k31.d.file.list -o SRR606249.x.gtdb-rs214.fastgather.csv"
        User time (seconds): 6771.21
        System time (seconds): 165.43
        Percent of CPU this job got: 4744%
        Elapsed (wall clock) time (h:mm:ss or m:ss): 2:26.21
        Average shared text size (kbytes): 0
        Average unshared data size (kbytes): 0
        Average stack size (kbytes): 0
        Average total size (kbytes): 0
        Maximum resident set size (kbytes): 1838548
        Average resident set size (kbytes): 0
        Major (requiring I/O) page faults: 38
        Minor (reclaiming a frame) page faults: 3288773
        Voluntary context switches: 1655356
        Involuntary context switches: 1765532
        Swaps: 0
        File system inputs: 81016
        File system outputs: 56
        Socket messages sent: 0
        Socket messages received: 0
        Signals delivered: 0
        Page size (bytes): 4096
        Exit status: 0

[ctb]

scripts in /home/ctbrown/scratch3/SRR606249-feb13-2024

github repo for scripts: https://github.com/sourmash-bio/2024-benchmark-branchwater-plugin

srun command I used:

srun -p bmh --time=24:00:00 --nodes=1 --cpus-per-task=64 --mem=50GB --pty /bin/bash

[ctb]

new benchmarks, for v0.9.5: sourmash-bio/sourmash#3232