Add digest.Compute benchmark

[bduffany]

BenchmarkDigestCompute/SHA256/1-8                2624952               455.7 ns/op
BenchmarkDigestCompute/BLAKE3/1-8                 713340              1601 ns/op
BenchmarkDigestCompute/SHA256/10-8               2636890               463.4 ns/op
BenchmarkDigestCompute/BLAKE3/10-8                812553              1558 ns/op
BenchmarkDigestCompute/SHA256/100-8              2019846               587.7 ns/op
BenchmarkDigestCompute/BLAKE3/100-8               904671              2409 ns/op
BenchmarkDigestCompute/SHA256/1000-8              515154              2291 ns/op
BenchmarkDigestCompute/BLAKE3/1000-8              546352              3174 ns/op
BenchmarkDigestCompute/SHA256/10000-8              55933             19288 ns/op
BenchmarkDigestCompute/BLAKE3/10000-8             238008              8266 ns/op
BenchmarkDigestCompute/SHA256/100000-8              6258            188327 ns/op
BenchmarkDigestCompute/BLAKE3/100000-8             49518             23969 ns/op

Related issues: N/A

[sluongng]

LGTM.

nit: it would be nice to have a separate BUILD target (perhaps manual tagged) to run this with -test.bench=. in args.

[bduffany]

LGTM.

nit: it would be nice to have a separate BUILD target (perhaps manual tagged) to run this with -test.bench=. in args.

I have done this in the past - defining a separate {package}_benchmark_test target - one downside I noticed is that gazelle doesn't seem to work on the benchmark target, but haven't looked into whether there is a good workaround