Hi Adam. On the whole the benchmarks look good to me. Can I ask why
those ciphers and key agreement schemes that support multiple key
lengths aren't called out in the @Param annotations? I'm thinking 192
and 256 bit for AES and maybe 1024 and 3072 and/or 4096 for DH. Do we
not need numbers across various key lengths with these microbenchmarks?
--Jamil
On 12/19/2018 10:43 AM, Adam Petcher wrote:
Webrev: http://cr.openjdk.java.net/~apetcher/8215643/webrev.00/
JBS: https://bugs.openjdk.java.net/browse/JDK-8215643
Please review this enhancement that adds two new crypto
microbenchmarks. See the JBS ticket for the motivation behind these
new benchmarks. The Cipher benchmark nearly duplicates some existing
AES benchmarks, though the measurement is done differently in order to
also address JDK-8214800[1]. I didn't want to upset anything by
removing the existing AES benchmarks, but I'm interested in advice on
how we should handle this duplication.
[1] https://bugs.openjdk.java.net/browse/JDK-8214800
