Hi Paul,
Good idea. Although the downside is the test will depend on
java.management and jdk.management modules, I think it's acceptable.
Also disable to run ParallelPrefix.java in concurrency.
Please check the new webrev:
http://cr.openjdk.java.net/~mli/8179242/webrev.01/
Thank you
-Hamlin
On 2017/6/13 0:06, Paul Sandoz wrote:
Hi Hamlin,
Do you know if it is possible to adjust the array sizes based on max memory
thresholds?
For example in the os.maxMemory is less than 2g then omit the LARGE_ARRAY_SIZE
from the test (assuming that reduces the memory pressure, which it likely
should).
Paul.
On 12 Jun 2017, at 02:38, Hamlin Li <huaming...@oracle.com> wrote:
Would you please review the below patch?
bug: https://bugs.openjdk.java.net/browse/JDK-8179242
webrev: http://cr.openjdk.java.net/~mli/8179242/webrev.00/
Thank you
-Hamlin
------------------------------------------------------------------------
the test takes about 700M on linux 64, so propose to fix the issue with below
patch:
diff -r 9200df3d3c0b test/java/util/Arrays/ParallelPrefix.java
--- a/test/java/util/Arrays/ParallelPrefix.java Sun Jun 11 18:36:23 2017
-0700
+++ b/test/java/util/Arrays/ParallelPrefix.java Mon Jun 12 02:35:20 2017
-0700
@@ -25,7 +25,8 @@
* @test 8014076 8025067
* @summary unit test for Arrays.ParallelPrefix().
* @author Tristan Yan
- * @run testng ParallelPrefix
+ * @requires os.maxMemory >= 2g
+ * @run testng/othervm -Xms512m -Xmx1024m ParallelPrefix
*/
