TM2 iavl BenchmarkLarge panicking #2714
Labels
🐞 bug
Something isn't working
⭐️ could have ⭐️
Optional work needed to complete a project
🤝 open for contributors 🤝
Work available to external contributors
Comments
zivkovicmilos
added
🐞 bug
|
This seems to be a timeout judging by the final lines. In practice, I think it's just that especially for the bad usage we make of it, using diff --git a/tm2/pkg/iavl/benchmarks/bench_test.go b/tm2/pkg/iavl/benchmarks/bench_test.go
index 88de3634..7bada156 100644
--- a/tm2/pkg/iavl/benchmarks/bench_test.go
+++ b/tm2/pkg/iavl/benchmarks/bench_test.go
@@ -173,6 +173,9 @@ func BenchmarkLarge(b *testing.B) {
ls := db.BackendList()
bs := make([]benchmark, 0, len(ls))
for _, backend := range ls {
+ if backend == db.BoltDBBackend {
+ continue
+ }
bs = append(bs, benchmark{backend, 1_000_000, 100, 16, 40})
}
runBenchmarks(b, bs)Maybe add this in #2716, with an explanatory comment linking to this issue. goleveldb seems to run fine; took less than a minute on my laptop. |
sw360cab
added a commit
that referenced
this issue
Sep 26, 2024
## Things changes * using new action plugin: [benchmark-action/github-action-benchmark](https://github.com/benchmark-action/github-action-benchmark) * previous library was [bobheadxi/gobenchdata](https://github.com/bobheadxi/gobenchdata) * creating alerts when benchmarks results are worse ## Things unchanged * Pushing benchmarks results onto a local branch, then `benchmark` repository will fetch from this into his own `gh-pages` * direct push is possible but requires PAT token to be used ## Things TODO (before merging) * restore benchmarks actions which were manually disabled directly in the `Action` menu in Github * drop and recreate branch `gh-repository` * fix benchmark tests in Gno repo (see #2711 and #2714) * Evaluate alternative configuration possibilities [here](https://github.com/benchmark-action/github-action-benchmark?tab=readme-ov-file#action-inputs) ## Things to consider (performance) * Should solve #2432 * are resources available in the Github basic runners enough for benchmarks? * are multiple runs spawned in the same hw conditions? * consider using [Larger Runner](https://docs.github.com/en/actions/using-github-hosted-runners/using-larger-runners/managing-larger-runners#adding-a-larger-runner-to-an-organization) or restoring a [Self-Hosted Runner](https://docs.github.com/en/actions/hosting-your-own-runners/managing-self-hosted-runners/adding-self-hosted-runners#adding-a-self-hosted-runner-to-an-organization)
|
Fixed at #2854 |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
BenchmarkLarge faling
A benchmark case for the package
github.com/gnolang/gno/tm2/pkg/iavl/benchmarksis panicking.
It is not clear if this is due to a bug or to a hardware / memory leak.
Your environment
8bd07c971e7ed9ee94af03d7145f247224683694Steps to reproduce
go test -benchmem -run=^$ -bench ^BenchmarkLarge$ github.com/gnolang/gno/tm2/pkg/iavl/benchmarks -tags=wireinjectLogs
The text was updated successfully, but these errors were encountered: