Re: RFR: 8287785: Reduce runtime of java.lang.invoke microbenchmarks [v2]
On Fri, 3 Jun 2022 12:53:38 GMT, Claes Redestad wrote: >> - Add explicit run configurations to java.lang.invoke micros, aiming to >> reduce runtime while maintaining a decently high confidence that there's >> enough warmup to produce good enough data. >> >> - Remove several trivial baseline micros, mainly those that only return a >> static object: It's reasonable to have baseline microbenchmarks when the >> baseline op is complex and you're mostly interested in checking the overhead >> of doing the same thing via some MH API, but blackhole operations are now >> shortcutting very quickly and timings doesn't differ from one type of object >> to another, so we don't need a multitude of such baseline tests. >> >> Estimated runtime of `make test TEST=micro:java.lang.micro` (excluding >> build) drops from just above 28 to just above 3 hours. > > Claes Redestad has updated the pull request incrementally with one additional > commit since the last revision: > > Remove unused cached field in LookupAcquire Marked as reviewed by mchung (Reviewer). - PR: https://git.openjdk.java.net/jdk/pull/9012
Re: RFR: 8287785: Reduce runtime of java.lang.invoke microbenchmarks [v2]
On Fri, 3 Jun 2022 12:37:16 GMT, liach wrote:
>> Claes Redestad has updated the pull request incrementally with one
>> additional commit since the last revision:
>>
>> Remove unused cached field in LookupAcquire
>
> test/micro/org/openjdk/bench/java/lang/invoke/LookupAcquire.java line 59:
>
>> 57: @Setup
>> 58: public void setup() {
>> 59: cached = MethodHandles.lookup();
>
> Since the benchmark using the cached object is removed, should this field be
> removed as well?
Fixed!
-
PR: https://git.openjdk.java.net/jdk/pull/9012
Re: RFR: 8287785: Reduce runtime of java.lang.invoke microbenchmarks [v2]
> - Add explicit run configurations to java.lang.invoke micros, aiming to > reduce runtime while maintaining a decently high confidence that there's > enough warmup to produce good enough data. > > - Remove several trivial baseline micros, mainly those that only return a > static object: It's reasonable to have baseline microbenchmarks when the > baseline op is complex and you're mostly interested in checking the overhead > of doing the same thing via some MH API, but blackhole operations are now > shortcutting very quickly and timings doesn't differ from one type of object > to another, so we don't need a multitude of such baseline tests. > > Estimated runtime of `make test TEST=micro:java.lang.micro` (excluding build) > drops from just above 28 to just above 3 hours. Claes Redestad has updated the pull request incrementally with one additional commit since the last revision: Remove unused cached field in LookupAcquire - Changes: - all: https://git.openjdk.java.net/jdk/pull/9012/files - new: https://git.openjdk.java.net/jdk/pull/9012/files/f351ac70..0a1f7ec7 Webrevs: - full: https://webrevs.openjdk.java.net/?repo=jdk&pr=9012&range=01 - incr: https://webrevs.openjdk.java.net/?repo=jdk&pr=9012&range=00-01 Stats: 11 lines in 1 file changed: 0 ins; 11 del; 0 mod Patch: https://git.openjdk.java.net/jdk/pull/9012.diff Fetch: git fetch https://git.openjdk.java.net/jdk pull/9012/head:pull/9012 PR: https://git.openjdk.java.net/jdk/pull/9012
