On Sun, 5 Oct 2025 13:00:44 GMT, Jonas Norlinder <[email protected]> wrote:
>> Hi all,
>>
>> This PR augments the CPU time sampling measurement capabilities that a user
>> can perform from Java code with the addition of
>> `MemoryMXBean.getGcCpuTime()`. With this patch it will be possible for a
>> user to measure process and GC CPU time during critical section or
>> iterations in benchmarks to name a few. This new method complements the
>> existing `OperatingSystemMXBean.getProcessCpuTime()` for a refined
>> understanding.
>>
>> `CollectedHeap::gc_threads_do` may operate on terminated GC threads during
>> shutdown, but thanks to JDK-8366865 by @walulyai we can piggyback on the new
>> `Universe::is_shutting_down`. I have implemented a stress-test
>> `test/jdk/java/lang/management/MemoryMXBean/GetGcCpuTime.java` that may
>> identify reading CPU time of terminated threads. Synchronizing on
>> `Universe::is_shutting_down` and `Heap_lock` resolves this problem.
>>
>> FWIW; To my understanding we don't want to add a
>> `Universe::is_shutting_down` check in gc_threads_do as this may introduce a
>> performance penalty that is unacceptable, therefore we must be careful about
>> the few places where external users call upon gc_threads_do and may race
>> with a terminating VM.
>>
>> Tested: test/jdk/java/lang/management/MemoryMXBean/GetGcCpuTime.java,
>> jdk/javax/management/mxbean hotspot/jtreg/vmTestbase/nsk/monitoring on Linux
>> x64, Linux aarch64, Windows x64, macOS x64 and macOS aarch64 with release
>> and fastdebug.
>
> Jonas Norlinder has updated the pull request incrementally with one
> additional commit since the last revision:
>
> Move from j.l.management to com.sun.management etc.
I introduced a JDK-specific MemoryMXBean in `com.sun.management` and added
`getTotalGcCpuTime` there. I believe the Java idiomatic way to retrieve GC CPU
time and process CPU time would be something like:
import java.lang.management.ManagementFactory;
import com.sun.management.MemoryMXBean;
import com.sun.management.OperatingSystemMXBean;
public class Main2 {
public static void main(String[] args) {
final MemoryMXBean mxMemoryBean =
ManagementFactory.getPlatformMXBean(MemoryMXBean.class);
final OperatingSystemMXBean mxOSBean =
ManagementFactory.getPlatformMXBean(OperatingSystemMXBean.class);
System.out.println("getTotalGcCpuTime: " +
mxMemoryBean.getTotalGcCpuTime());
System.out.println("getProcessCpuTime: " +
mxOSBean.getProcessCpuTime());
}
}
Given that `getTotalGcCpuTime` is most useful when you also sample process CPU
time I like this symmetry, thanks for the suggestion.
I updated the language in the documentation:
/**
* Returns the CPU time used by garbage collection.
*
* <p> CPU time used by all garbage collection. In
* general this includes time for all driver threads,
* workers, VM operations on the VM thread and the string
* deduplication thread (if enabled). May be non-zero even if no
* GC cycle occurred. This method returns {@code -1} if the
* platform does not support this operation or if called during
* shutdown.
*
* @return the total accumulated CPU time for garbage collection
* in nanoseconds.
*
* @since 26
*/
Since we include time om VM thread which strictly is not a garbage collection
thread I think it is better to constrain it to "garbage collection" in general.
Let me know what you think.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/27537#issuecomment-3369043846
