# HG changeset patch
# User drwhite
# Date 1460060334 14400
#      Thu Apr 07 16:18:54 2016 -0400
# Node ID bcdcfa84864fdcb049a03016aa6acd7a3cd25708
# Parent  5d720f8c105be3cf31263fe906fc6d0bef9d72ea
[mq]: webrev.01

diff --git a/src/share/vm/runtime/vm_version.cpp b/src/share/vm/runtime/vm_version.cpp
--- a/src/share/vm/runtime/vm_version.cpp
+++ b/src/share/vm/runtime/vm_version.cpp
@@ -288,6 +288,7 @@
                                                       unsigned int switch_pt) {
   if (FLAG_IS_DEFAULT(ParallelGCThreads)) {
     assert(ParallelGCThreads == 0, "Default ParallelGCThreads is not 0");
+    unsigned int threads;
     // For very large machines, there are diminishing returns
     // for large numbers of worker threads.  Instead of
     // hogging the whole system, use a fraction of the workers for every
@@ -295,9 +296,20 @@
     // and a chosen fraction of 5/8
     // use 8 + (72 - 8) * (5/8) == 48 worker threads.
     unsigned int ncpus = (unsigned int) os::active_processor_count();
-    return (ncpus <= switch_pt) ?
-           ncpus :
-          (switch_pt + ((ncpus - switch_pt) * num) / den);
+    threads = (ncpus <= switch_pt) ?
+             ncpus :
+             (switch_pt + ((ncpus - switch_pt) * num) / den);
+#ifndef _LP64
+    // On 32-bit binaries the virtual address space available to the JVM
+    // is usually limited to 2-3 GB (depends on the platform).
+    // Do not use up address space with too many threads (stacks and per-thread
+    // data). Note that x86 apps running on Win64 have 2 stacks per thread.
+    // GC may more generally scale down threads by max heap size (etc), but the
+    // consequences of over-provisioning threads are higher on 32-bit JVMS,
+    // so add hard limit here:
+    threads = MIN2(threads, (unsigned int)(2*switch_pt));
+#endif
+    return threads;
   } else {
     return ParallelGCThreads;
   }