00:54
claudiuinberlin has joined #jruby
00:55
claudiuinberlin has quit [Client Quit]
03:51
bga57 has quit [Ping timeout: 244 seconds]
06:16
_whitelogger has joined #jruby
06:50
ChrisSeatonGitte has quit [*.net *.split]
06:50
TimGitter[m] has quit [*.net *.split]
06:50
TimGitter[m]1 has quit [*.net *.split]
06:50
rdubya[m] has quit [*.net *.split]
06:50
enebo[m] has quit [*.net *.split]
06:50
lopex[m] has quit [*.net *.split]
06:50
subbu has quit [*.net *.split]
06:50
ThomasEEneboGitt has quit [*.net *.split]
06:50
MarcinMielyskiGi has quit [*.net *.split]
06:50
MattPattersonGit has quit [*.net *.split]
06:50
CharlesOliverNut has quit [*.net *.split]
06:50
Freaky has quit [*.net *.split]
07:16
MattPattersonGit has joined #jruby
07:16
ThomasEEneboGitt has joined #jruby
07:16
MarcinMielyskiGi has joined #jruby
07:16
CharlesOliverNut has joined #jruby
07:16
Freaky has joined #jruby
07:16
ChrisSeatonGitte has joined #jruby
07:16
TimGitter[m]1 has joined #jruby
07:16
subbu has joined #jruby
07:16
TimGitter[m] has joined #jruby
07:16
rdubya[m] has joined #jruby
07:16
lopex[m] has joined #jruby
07:16
enebo[m] has joined #jruby
07:22
bga57 has joined #jruby
07:33
claudiuinberlin has joined #jruby
08:26
MattPattersonGit has quit [*.net *.split]
08:26
ThomasEEneboGitt has quit [*.net *.split]
08:26
CharlesOliverNut has quit [*.net *.split]
08:26
MarcinMielyskiGi has quit [*.net *.split]
08:26
Freaky has quit [*.net *.split]
08:27
ChrisSeatonGitte has quit [*.net *.split]
08:27
TimGitter[m] has quit [*.net *.split]
08:27
TimGitter[m]1 has quit [*.net *.split]
08:27
enebo[m] has quit [*.net *.split]
08:27
rdubya[m] has quit [*.net *.split]
08:27
lopex[m] has quit [*.net *.split]
08:27
subbu has quit [*.net *.split]
08:35
subbu has joined #jruby
08:35
lopex[m] has joined #jruby
08:36
ChrisSeatonGitte has joined #jruby
08:37
Freaky has joined #jruby
08:37
enebo[m] has joined #jruby
08:37
rdubya[m] has joined #jruby
08:37
TimGitter[m] has joined #jruby
08:37
TimGitter[m]1 has joined #jruby
08:41
CharlesOliverNut has joined #jruby
08:41
ThomasEEneboGitt has joined #jruby
08:41
MarcinMielyskiGi has joined #jruby
08:42
MattPattersonGit has joined #jruby
08:59
shellac has joined #jruby
09:13
drbobbeaty has quit [Ping timeout: 276 seconds]
10:58
drbobbeaty has joined #jruby
13:09
lucasb has joined #jruby
14:02
whitingjr has joined #jruby
15:10
ebarrett has quit [Ping timeout: 245 seconds]
15:11
ebarrett has joined #jruby
15:56
claudiuinberlin has quit [Read error: Connection reset by peer]
16:03
xardion has quit [Remote host closed the connection]
16:08
xardion has joined #jruby
16:10
shellac has quit [Ping timeout: 250 seconds]
16:30
rusk has quit [Remote host closed the connection]
18:14
<
lopex >
wow, lol @G1
18:16
<
headius[m] >
yeah shocking
18:16
<
headius[m] >
I mean I'd accept 10-15% degradation but this seems like too much
18:18
<
lopex >
then it would be worth to benchmark a rails app
18:18
<
lopex >
also also wonder if -XX:NewRatio=1 still helps
18:19
<
headius[m] >
it would be worth re-testing Java 11+
18:19
<
headius[m] >
NewRatio=1 for G1 supposed to help?
18:20
<
lopex >
no, just increase new generation
18:20
<
lopex >
well, it could ?
18:20
<
headius[m] >
I don't recall seeing significant drop in Rails perf on Java 11 though...it is surprisingly less GC-intensive than pragtob's Go solver
18:20
<
headius[m] >
I'll give it a try anyway
20:04
travis-ci has joined #jruby
20:04
travis-ci has left #jruby [#jruby]
20:16
_whitelogger has joined #jruby
20:19
<
headius[m] >
lopex: NewRatio did not seem to help G1 results much
20:20
<
lopex >
headius[m]: but does it affect older gen ?
20:20
<
lopex >
or too many get promoted anyways ?
20:20
<
lopex >
is that for default heap size ?
20:23
travis-ci has joined #jruby
20:23
travis-ci has left #jruby [#jruby]
20:32
travis-ci has joined #jruby
20:32
travis-ci has left #jruby [#jruby]
20:58
lucasb has quit [Quit: Connection closed for inactivity]
21:01
<
headius[m] >
lopex: I think the problem is the write barriers that G1 and Shenandoah use
21:02
<
headius[m] >
JRuby churns through so many objects it's paying a heavier cost for those write barriers. The other strategy of using read barriers (Zing, ZGC) might have less impact, but I can't test those on MacOS
21:16
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
23:36
travis-ci has joined #jruby
23:36
travis-ci has left #jruby [#jruby]
23:45
travis-ci has joined #jruby
23:45
travis-ci has left #jruby [#jruby]