`BenchmarkValueEqual` does not run the target code with `b.N`, as
https://pkg.go.dev/testing#hdr-Benchmarks states. Thus, it cannot
produce a stable benchmark result. And, it would fail the benchmark:
https://github.com/open-telemetry/opentelemetry-go/actions/runs/10412186663.
This PR fixes this issue.
Here is the result of `benchstat`
```
goos: darwin
goarch: arm64
pkg: go.opentelemetry.io/otel/log
│ old.txt │ new.txt │
│ sec/op │ sec/op vs base │
ValueEqual/2_with_1-10 0.0000002000n ± 100% 3.1450000000n ± 1% +1572499900.00% (p=0.000 n=10)
│ old.txt │ new.txt │
│ B/op │ B/op vs base │
ValueEqual/2_with_1-10 0.000 ± 0% 0.000 ± 0% ~ (p=1.000 n=10) ¹
¹ all samples are equal
│ old.txt │ new.txt │
│ allocs/op │ allocs/op vs base │
ValueEqual/2_with_1-10 0.000 ± 0% 0.000 ± 0% ~ (p=1.000 n=10) ¹
¹ all samples are equal
```
