-
Notifications
You must be signed in to change notification settings - Fork 916
Speedup filter_bytes
~-20-40%, filter_native
low selectivity (~-37%)
#7463
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Conversation
filter_native
low selectivity (~-37%)
filter_native
low selectivity (~-37%)filter_bytes
~20-40%, filter_native
low selectivity (~-37%)
filter_bytes
~20-40%, filter_native
low selectivity (~-37%)filter_bytes
~-20-40%, filter_native
low selectivity (~-37%)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @Dandandan -- I have queued up some benchmark runs on this branch too. The code makes sense to me, though I have a suggestion to make the code easier to understand.
🤖 |
🤖: Benchmark completed Details
|
🤖 |
These are my results:
|
Somehow these results are all over the place, even for untouched kernels 🤔 |
🤖: Benchmark completed Details
|
🤯 |
I can try to reproduce the results manually -- it could be:
What are the specs of your machine @Dandandan ? |
I run it on: It's weird to me some of the benchmarks like ("filter context fsb") are reporting 1.67x slowdown, because nothing has changed in the code path (I expect only filtering bytes and everything using |
I wonder what gcp vm are you using? No machine that uses bursting by any chance? |
It is a c2-standard-16 (16 vCPUs, 64 GB Memory) I don't think it has bursting |
Which issue does this PR close?
Closes #7465
Vec
api forfilter_native
to generate some faster code.Rationale for this change
What changes are included in this PR?
Are there any user-facing changes?