Hi Zhengyu,
On 2020-02-17 15:51, Zhengyu Gu wrote:
Hi Stefan,
Thanks for the review and suggestions, updated accordingly:
http://cr.openjdk.java.net/~zgu/JDK-8238633/webrev.01/
Thanks for moving the code. I think this looks good.
If you're up for it, I have a couple of style change suggestions:
1) ObjectMarker uses two verbs to describe the same thing: "mark" and
"visit". I propose that we only use "mark" in ObjectMarker and leave the
usage of "visited" to the Jvmti code.
2) Some updates to odd whitespaces
3) Using forward declarations in Shenandoah code.
I've bundled those changes into webrevs:
https://cr.openjdk.java.net/~stefank/8238633/webrev.01.delta
https://cr.openjdk.java.net/~stefank/8238633/webrev.01
Regarding performance testing, the HeapWalkTests you used seems to use a
very small heap. I think it would be good to redo the measurements on a
larger heap. Could you take the HeapWalkTest and add a few GBs of small,
linked objects?
Thank,
StefanK
---
Previously, the calls to 'mark' and 'visited' were inlineable, but
now every GC has to take a virtual call when marking the objects. My
guess is that this code is slow anyway, and that it doesn't matter
too much, but did you measure the effect of that change with, for
example, G1?
I did rough measurement, timing
vmTestbase/nsk/jvmti/unit/heap/HeapWalkTests/TestDescription.java test.
If you know any tests/benchmarks I should measure, please let me know.
Thanks,
-Zhengyu
Thanks,
StefanK
Test:
hotspot_gc
vmTestbase_nsk_jdi
vmTestbase_nsk_jvmti
Thanks,
-Zhengyu
