Jie Fu (傅杰)

The performance improvement is impressive on x86 with avx2. Can you clarify what tests did you do? Thanks.

Thanks for the update. May I ask can we do it like this? ``` diff --git a/src/hotspot/share/opto/vectorIntrinsics.cpp b/src/hotspot/share/opto/vectorIntrinsics.cpp index 66bacf0..c4e807a 100644 --- a/src/hotspot/share/opto/vectorIntrinsics.cpp +++ b/src/hotspot/share/opto/vectorIntrinsics.cpp @@ -2494,8 +2494,9 @@ bool...

> So you're trying to make sure we have a minimum COMPILATION_TIMEOUT value of 5 seconds, but I'm not sure why you want that. With the current patch, `COMPILATION_TIMEOUT` may...

> Definitely not. Traditional default value for no specified timeoutFactor value is 1.0. Okay, you're right. This is because the timeoutFactor may be used by other tests too.

> > LGTM > > Thank you! Trivial, or? Yes, I think it's trivial. Thanks.

I came across the same kind of bugs today. Finally, I got the failing reason and fixed it just by dumping the error msg: https://github.com/vllm-project/vllm/pull/5282 .

Thanks @hmellor for the review and help.

Hi @mgoin , are you fine with the current change in `cpu-installation.rst`? Thanks.

> Hi @DamonFool, yes this is fine! It would be nice to make a new section for avx2, but I'll leave that up to you. The patch aims to tell...

Hi @WoosukKwon , are you OK with this change? Thanks.