Open hissssst opened 1 year ago
Here is how to run:
But this measurement is incorrect, since it completely ignores the GC and such stuff. I mean, if we start tasks slower, or induce memory limit, the runtime will be able to spawn more tasks with less memory, sacrificing time for GC
Why not use, perf stat
or massif?
to full report:
perf record --call-graph fp program
and perf report
or valgrind --tool=massif --stacks=yes program
References
Hi, thank you for the benchmark and interesting blog post!
I wanted to discuss it with my friends and tweak some code to see how it afffects the result. But I've found out that I need to know how did you measure the amount of used memory? And how did you run the benchmarks?
Plus, your blog post shows rust examples, but they're not provided here.