solana/core/benches/sigverify.rs

#![feature(test)]

extern crate test;

use solana::packet::to_packets;
use solana::sigverify;
use solana::test_tx::test_tx;
use test::Bencher;

#[bench]
fn bench_sigverify(bencher: &mut Bencher) {
    let tx = test_tx();

    // generate packet vector
    let batches = to_packets(&vec![tx; 128]);

    // verify packets
    bencher.iter(|| {
        let _ans = sigverify::ed25519_verify(&batches);
    })
}
Boot criterion (#1032) * Revert benchmarks back to libtest Criterion has too many dependencies, it's execution as slower, and we didn't see the kind of precision we had hoped for to use it to block CI builds. * Ignore benchmarks that take more than a few milliseconds per iteration * Revert "Ignore benchmarks that take more than a few milliseconds per iteration" This reverts commit b87cdf6ef4a2549ec09e282014391de695473699. * Don't run benchmarks in CI They are already built in the nightly build. Executing them in CI doesn't add much value until the results are precise enough to act on. 2018-08-22 08:57:07 -06:00			`#![feature(test)]`
cargo +nightly fix --features=bpf_c,cuda,erasure,chacha --edition-idioms 2018-12-08 22:44:20 -07:00
Boot criterion (#1032) * Revert benchmarks back to libtest Criterion has too many dependencies, it's execution as slower, and we didn't see the kind of precision we had hoped for to use it to block CI builds. * Ignore benchmarks that take more than a few milliseconds per iteration * Revert "Ignore benchmarks that take more than a few milliseconds per iteration" This reverts commit b87cdf6ef4a2549ec09e282014391de695473699. * Don't run benchmarks in CI They are already built in the nightly build. Executing them in CI doesn't add much value until the results are precise enough to act on. 2018-08-22 08:57:07 -06:00			`extern crate test;`
Only send sigverify to GPU if batch size is >64 Seems to be a decent crossover point for Xeon E5-2620 v4 8c,16t vs. nvidia 1080ti 2018-08-01 14:10:39 -07:00
Enable bench and fix upload-perf 2018-09-24 10:20:58 -07:00			`use solana::packet::to_packets;`
Only send sigverify to GPU if batch size is >64 Seems to be a decent crossover point for Xeon E5-2620 v4 8c,16t vs. nvidia 1080ti 2018-08-01 14:10:39 -07:00			`use solana::sigverify;`
Move system_transaction out of src/ 2018-12-04 15:37:11 -08:00			`use solana::test_tx::test_tx;`
Boot criterion (#1032) * Revert benchmarks back to libtest Criterion has too many dependencies, it's execution as slower, and we didn't see the kind of precision we had hoped for to use it to block CI builds. * Ignore benchmarks that take more than a few milliseconds per iteration * Revert "Ignore benchmarks that take more than a few milliseconds per iteration" This reverts commit b87cdf6ef4a2549ec09e282014391de695473699. * Don't run benchmarks in CI They are already built in the nightly build. Executing them in CI doesn't add much value until the results are precise enough to act on. 2018-08-22 08:57:07 -06:00			`use test::Bencher;`
Only send sigverify to GPU if batch size is >64 Seems to be a decent crossover point for Xeon E5-2620 v4 8c,16t vs. nvidia 1080ti 2018-08-01 14:10:39 -07:00
Boot criterion (#1032) * Revert benchmarks back to libtest Criterion has too many dependencies, it's execution as slower, and we didn't see the kind of precision we had hoped for to use it to block CI builds. * Ignore benchmarks that take more than a few milliseconds per iteration * Revert "Ignore benchmarks that take more than a few milliseconds per iteration" This reverts commit b87cdf6ef4a2549ec09e282014391de695473699. * Don't run benchmarks in CI They are already built in the nightly build. Executing them in CI doesn't add much value until the results are precise enough to act on. 2018-08-22 08:57:07 -06:00			`#[bench]`
Rename sig variables to signature We'll avoid introducing three-letter terms to free up the namespace for three-letter acronyms. But recognize the term "sigverify", a verb, to verify a digital signature. 2018-08-09 09:26:21 -06:00			`fn bench_sigverify(bencher: &mut Bencher) {`
Only send sigverify to GPU if batch size is >64 Seems to be a decent crossover point for Xeon E5-2620 v4 8c,16t vs. nvidia 1080ti 2018-08-01 14:10:39 -07:00			`let tx = test_tx();`

			`// generate packet vector`
Enable bench and fix upload-perf 2018-09-24 10:20:58 -07:00			`let batches = to_packets(&vec![tx; 128]);`
Only send sigverify to GPU if batch size is >64 Seems to be a decent crossover point for Xeon E5-2620 v4 8c,16t vs. nvidia 1080ti 2018-08-01 14:10:39 -07:00
			`// verify packets`
			`bencher.iter(\|\| {`
			`let _ans = sigverify::ed25519_verify(&batches);`
			`})`
			`}`