Repository navigation
Commit 1bc4970
committed
test(index): add the A/B script the builder benchmark was quoted from
CONTRIBUTING asks for benchmark numbers reproducible from the branch with
a command anyone can run. The three-genome table in the PR was produced
by hand, so nobody else could rerun it.
The script drives the two arms through --limitGenomeGenerateRAM, which is
the flag that actually selects the builder, so the comparison exercises
the dispatch rather than bypassing it. It reports wall clock and peak RSS
and then compares SA and SAindex with cmp, failing if they differ.
It refuses to run above a load average of 2. A genomeGenerate run is
minutes long and uses every core, so an unrelated job does not add noise
evenly: it lands on whichever arm overlaps it, and nothing in the
resulting numbers shows that it happened. I made exactly that mistake
benchmarking the BAM backends on this machine today.1 parent aaa7b69 commit 1bc4970
1 file changed
Lines changed: 53 additions & 0 deletions
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
| 1 | + | |
| 2 | + | |
| 3 | + | |
| 4 | + | |
| 5 | + | |
| 6 | + | |
| 7 | + | |
| 8 | + | |
| 9 | + | |
| 10 | + | |
| 11 | + | |
| 12 | + | |
| 13 | + | |
| 14 | + | |
| 15 | + | |
| 16 | + | |
| 17 | + | |
| 18 | + | |
| 19 | + | |
| 20 | + | |
| 21 | + | |
| 22 | + | |
| 23 | + | |
| 24 | + | |
| 25 | + | |
| 26 | + | |
| 27 | + | |
| 28 | + | |
| 29 | + | |
| 30 | + | |
| 31 | + | |
| 32 | + | |
| 33 | + | |
| 34 | + | |
| 35 | + | |
| 36 | + | |
| 37 | + | |
| 38 | + | |
| 39 | + | |
| 40 | + | |
| 41 | + | |
| 42 | + | |
| 43 | + | |
| 44 | + | |
| 45 | + | |
| 46 | + | |
| 47 | + | |
| 48 | + | |
| 49 | + | |
| 50 | + | |
| 51 | + | |
| 52 | + | |
| 53 | + | |
0 commit comments