RFR: 8260501: [Vector API] Improve register usage for shift operations on x86

[Jie Fu]

On Wed, 27 Jan 2021 21:43:08 GMT, Vladimir Ivanov <vlivanov at openjdk.org> wrote:

> IIRC AD instructions for variable shifts were deliberately merged to reduce static footprint of the JVM on x86 (JVM library size). IMO it's preferable to leave them as they are.

Hi @iwanowww  ,

I didn't know that variable shifts were deliberately merged to reduce the size of libjvm.so.
Thanks for your clarification.

With the original patch, the size of libjvm.so on Linux/x64 will be increased by ~0.085%.

However, vshiftL_imm and vshiftI_imm can still be optimized without increasing the binary size.
And the size of libjvm.so will be further decreased by ~0.003% with the updated patch on Linux/x64.
So are you OK with it?

Thanks.
Best regards,
Jie

-------------

PR: https://git.openjdk.java.net/jdk/pull/2257