RFR: 8318562: Computational test more than 2x slower when AVX instructions are used
Sandhya Viswanathan
sviswanathan at openjdk.org
Fri Nov 17 20:10:34 UTC 2023
On Fri, 17 Nov 2023 04:53:57 GMT, Jatin Bhateja <jbhateja at openjdk.org> wrote:
>> src/hotspot/cpu/x86/x86_64.ad line 11092:
>>
>>> 11090: %{
>>> 11091: match(Set dst (ConvD2F src));
>>> 11092: effect(TEMP dst);
>>
>> You don't need `TEMP dst`, if `dst` is an alias of `src` then a destructive `xor` is not emitted.
>
> Without TEMP annotation dst and src may be aliased if src live range does not survives beyond this instruction.
I had checked before submitting the PR, the cvt xmm0, xmm0, xmm0 form was slower than xorps xmm1, xmm1 followed by cvt xmm1, xmm1, xmm0.
-------------
PR Review Comment: https://git.openjdk.org/jdk/pull/16701#discussion_r1397833825
More information about the hotspot-compiler-dev
mailing list