RFR: 8316592: RISC-V: implement poly1305 intrinsic [v3]

[Hamlin Li]

On Thu, 2 Nov 2023 13:14:46 GMT, Andrew Haley <aph at openjdk.org> wrote:

> > Just a general question about the test data, from data size of 64 to 16384, the performance gain is more and more? But, for data size 1048576, the performance gain is less than data size of 16384, can some give any apprerant reason for performance data change?
> 
> Maybe the 1M test maybe has not had sufficient warmup. Just look at the size of those error bars.

Thanks for explanation, it make sense.

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

PR Comment: https://git.openjdk.org/jdk/pull/16417#issuecomment-1791236883