-
Notifications
You must be signed in to change notification settings - Fork 115
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[GC] Add IO related size policy for Parallel GC #847
Conversation
Summary: add new parallel gc option UseIOPrioritySizePolicy to early shrink heap when io wait is high. Testing: jtreg Reviewers: maoliang.ml, yude.lyd Issue: dragonwell-project#846 CR: dragonwell-project#847
7fd0e9d
to
93dc2e5
Compare
93dc2e5
to
55132e3
Compare
Summary: add new parallel gc option UseIOPrioritySizePolicy to early shrink heap when io wait is high. Testing: jtreg Reviewers: maoliang.ml, yude.lyd Issue: dragonwell-project#846 CR: dragonwell-project#847
55132e3
to
52efa83
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Summary: add new parallel gc option UseIOPrioritySizePolicy to early shrink heap when io wait is high. Testing: jtreg Reviewers: maoliang.ml, yude.lyd Issue: dragonwell-project/dragonwell11#846 CR: dragonwell-project/dragonwell11#847
Summary: add new parallel gc option UseIOPrioritySizePolicy to early shrink heap when io wait is high. Testing: jtreg Reviewers: maoliang.ml, yude.lyd Issue: dragonwell-project/dragonwell11#846 CR: dragonwell-project/dragonwell11#847
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
When heap is automatically reduced, how and when will it grow back?
|
||
static bool should_update_eden_stats(GCCause::Cause cause) { | ||
return AdaptiveSizePolicy::should_update_eden_stats(cause) || | ||
(UseIOPrioritySizePolicy && GCCause::_gc_locker == cause); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
comment explaining why (cause == _gc_locker) is checked here?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
When running spark on TPC-DS, we observed IO occurs simultaneously with jni calls. So we intend to update eden size upon GC locker caused young gc.
After IO finish, we will restore _throughput_goal to its original value (like 0.99), and typically this goal won't be satisfied. As a result, Adaptive Size Policy will try to increase the heap by 20% each time. |
Summary: add new parallel gc option UseIOPrioritySizePolicy to early shrink heap when io wait is high. Testing: jtreg Reviewers: maoliang.ml, yude.lyd Issue: dragonwell-project#846 CR: dragonwell-project#847
52efa83
to
3b0ee54
Compare
Summary: add new parallel gc option UseIOPrioritySizePolicy to early shrink heap when io wait is high. Testing: jtreg Reviewers: maoliang.ml, yude.lyd Issue: dragonwell-project#846 CR: dragonwell-project#847
3b0ee54
to
70a4662
Compare
Summary: add new parallel gc option UseIOPrioritySizePolicy to early shrink heap when io wait is high.
Testing: jtreg
Reviewers: maoliang.ml, yude.lyd
Issue: #846
CR: #847