Perfasm attributes 39% of cycles to JMH infrastructure in my benchmark

[Aleksey Shipilev]

On 04/16/2015 03:36 PM, Aleksey Shipilev wrote:
> Implementation-wise, that is done to make benchmark threads always hot
> during the iteration, otherwise we may just block and wait, consuming no
> cycles. I wonder, however, if we can demote the CAS-ed busy wait into
> more light-weight test:
>   http://cr.openjdk.java.net/~shade/jmh/busy-wait-test-1.patch

Okay, this seems to be a good improvement anyway, so I pushed:
  https://bugs.openjdk.java.net/browse/CODETOOLS-7901393

But, for your case, exponential backoff for @TearDown(Trial) should help
immensely:
  https://bugs.openjdk.java.net/browse/CODETOOLS-7901394

> Please try to build it with 1.9-SNAPSHOT, and try on your config. Does
> it help your case?

Please see if current tip works for you. We'd like to start the release
preparations soon.

Thanks,
-Aleksey