Comment #2 on issue 2814 by [email protected]: Increase CPU profiler
sampling rate
http://code.google.com/p/v8/issues/detail?id=2814
I did some measurements of the current(old) profiler performance and the
new one with a higher sampling rate on different platforms. The tests and
the results are attached.
The test creates 100*62 different functions so that there are 100 different
call stacks of depth 62: f_x_0() -calls-> f_x_1() ->...->f_x_61() so we
have 100 different JS stacks of depth 63 (profiler captures at most 64 top
frames). Those functions are invoked in a loop for 2 seconds 5 times first
without the profiler and then with the profiler enabled. The performance
metric is the number of iterations the main thread does in 2 seconds. The
table shows how the CPU profiler affects that performance.
Attachments:
Sampling rate performance - Sheet1 (5).pdf 44.4 KB
perf-test-old-profiler.patch 4.6 KB
perf-test-new-profiler.patch 4.6 KB
--
You received this message because this project is configured to send all
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings
--
--
v8-dev mailing list
[email protected]
http://groups.google.com/group/v8-dev
---
You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/groups/opt_out.
