You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We have benchmarks that do each and that do one and then the other, but not one where some workers are reading objects and other workers are writing the same objects. This is a test of invalidation processing and whether reads and writes block each other.
The text was updated successfully, but these errors were encountered:
I'm unsure how to write that as a benchmark. The benchmark functions get to return one timing number, but there are two entirely separate kinds of tasks running whose times aren't related. I could spawn threads inside the benchmark function and take an average, or just return the time of the reader.
Maybe the bench_read_after_write is a good enough test of this.
We have benchmarks that do each and that do one and then the other, but not one where some workers are reading objects and other workers are writing the same objects. This is a test of invalidation processing and whether reads and writes block each other.
The text was updated successfully, but these errors were encountered: