The Z Garbage Collector (ZGC) is a scalable low latency garbage collector. ZGC performs all expensive work concurrently, without stopping the execution of application threads for more than a millisecond. It is suitable for applications which require low latency. Pause times are independent of the heap size that is being used. ZGC works well with heap sizes from a few hundred megabytes to 16TB.
ZGC was initially introduced as an experimental feature in JDK 11, and was declared Production Ready in JDK 15. In JDK 21 was reimplemented to support generations.
At a glance, ZGC is:
- Using colored pointers
- Using load barriers
At its core, ZGC is a concurrent garbage collector, meaning all heavy lifting work is done while Java threads continue to execute. This greatly limits the impact garbage collection will have on your application's response time.
|JDK 15 (Experimental since JDK 11)
|JDK 15 (Experimental since JDK 13)
|JDK 15 (Experimental since JDK 14)
|JDK 15 (Experimental since JDK 14)
|Requires Windows version 1803 (Windows 10 or Windows Server 2019) or later.
If you're trying out ZGC for the first time, start by using the following GC options:
For more detailed logging, use the following options:
See below for more information on these and additional options.
Configuration & Tuning
ZGC has been designed to be adaptive and to require minimal manual configuration. During the execution of the Java program, ZGC dynamically adapts to the workload by resizing generations, scaling the number of GC threads, and adjusting tenuring thresholds. The main tuning knob is to increase the maximum heap size.
ZGC comes in two versions: The new, generational version and the legacy, non-generational version. The Non-generational ZGC is the older version of ZGC, which doesn't take advantage of generations (see Generations) to optimize its runtime characteristics. It is encouraged that users transition to use the newer Generational ZGC.
The Generational ZGC is enabled with the command-line options
The Non-generational ZGC is enabled with the command-line option
The following JVM options can be used with ZGC:
|General GC Options
|ZGC Diagnostic Options (-XX:+UnlockDiagnosticVMOptions)
In addition to these the following flags are available when the generational mode is enabled with -XX:+UseZGC -XX:+ZGenerational:
|General GC Options
|ZGC Diagnostic Options (-XX:+UnlockDiagnosticVMOptions)
Setting Heap Size
The most important tuning option for ZGC is setting the maximum heap size, which you can set with the
-Xmx command-line option. Because ZGC is a concurrent collector, you must select a maximum heap size such that the heap can accommodate the live-set of your application and there is enough headroom in the heap to allow allocations to be serviced while the GC is running. How much headroom is needed very much depends on the allocation rate and the live-set size of the application. In general, the more memory you give to ZGC the better. But at the same time, wasting memory is undesirable, so it’s all about finding a balance between memory usage and how often the GC needs to run.
ZGC has another command-line option related to the heap size named
-XX:SoftMaxHeapSize. It can be used to set a soft limit on how large the Java heap can grow. ZGC will strive to not grow beyond this limit, but is still allowed to grow beyond this limit up to the maximum heap size. ZGC will only use more than the soft limit if that is needed to prevent the Java application from stalling and waiting for the GC to reclaim memory. For example, with the command-line options
-XX:SoftMaxHeapSize=4g ZGC will use 4GB as the limit for its heuristics, but if it can't keep the heap size below 4GB it is still allowed to temporarily use up to 5GB.
Setting Concurrent GC Threads
Note! This section pertain to the non-generational version of ZGC. Generational ZGC has a more adaptive implementation and you are less likely to need to tweak the GC threads.
The second tuning option one might want to look at is setting the number of concurrent GC threads (
-XX:ConcGCThreads=<number>). ZGC has heuristics to automatically select this number. This heuristic usually works well but depending on the characteristics of the application this might need to be adjusted. This option essentially dictates how much CPU-time the GC should be given. Give it too much and the GC will steal too much CPU-time from the application. Give it too little, and the application might allocate garbage faster than the GC can collect it.
NOTE!! Starting from JDK 17, ZGC dynamically scales up and down the number of concurrent GC threads. This makes it even more unlikely that you'd need to adjust the concurrent number of GC threads.
NOTE!!! In general, if low latency (i.e. low application response time) is important for you application, then never over-provision your system. Ideally, your system should never have more than 70% CPU utilization.
Returning Unused Memory to the Operating System
By default, ZGC uncommits unused memory, returning it to the operating system. This is useful for applications and environments where memory footprint is a concern, but might have a negative impact on the latency of Java threads. You can disable this feature with the command-line option
-XX:-ZUncommit. Furthermore, memory will not be uncommitted so that the heap size shrinks below the minimum heap size
(-Xms). This means this feature will be implicitly disabled if the minimum heap size
(-Xms) is configured to be equal to the maximum heap size
You can configure an uncommit delay using
-XX:ZUncommitDelay=<seconds> (default is 300 seconds). This delay specifies for how long memory should have been unused before it's eligible for uncommit.
NOTE! Allowing the GC to commit and uncommit memory while the application is running could have a negative impact on the latency of Java threads. If extremely low latency is the main reason for running with ZGC, consider running with the same value for
-Xms, and use
-XX:+AlwaysPreTouch to page in memory before the application starts.
NOTE!! On Linux, uncommitting unused memory requires
FALLOC_FL_PUNCH_HOLE support, which first appeared in kernel version 3.5 (for tmpfs) and 4.3 (for hugetlbfs).
Enabling Large Pages On Linux
On Linux x86, large pages (also known as "huge pages") have a size of 2MB.
Let's assume you want a 16GB Java heap. That means you need 16GB / 2MB = 8192 huge pages.
The heap requires at least 16GB (8192 pages) of memory to the pool of huge pages. The heap along with other parts of the JVM will use large pages for various internal data structures (such as code heap and marking bitmaps). In this example you will reserve 9216 pages (18GB) to allow for 2GB of non-Java heap allocations to use large pages.
Configure the system's huge page pool to have the required number of pages (requires root privileges):
Note that the above command is not guaranteed to be successful if the kernel cannot find enough free huge pages to satisfy the request. Also note that it might take some time for the kernel to process the request. Before proceeding, check the number of huge pages assigned to the pool to make sure the request was successful and has completed.
NOTE! If you're using a Linux kernel >= 4.14, then the next step (where you mount a hugetlbfs filesystem) can be skipped. However, if you're using an older kernel then ZGC needs to access large pages through a hugetlbfs filesystem.
Mount a hugetlbfs filesystem (requires root privileges) and make it accessible to the user running the JVM (in this example we're assuming this user has 123 as its uid).
Now start the JVM using the
If there are more than one accessible hugetlbfs filesystem available, then (and only then) do you also have to use
-XX:AllocateHeapAt to specify the path to the filesystems you want to use. For example, assume there are multiple accessible hugetlbfs filesystems mounted, but the filesystem you specifically want to use it mounted on
/hugepages, then use the following options.
NOTE! The configuration of the huge page pool and the mounting of the hugetlbfs file system is not persistent across reboots, unless adequate measures are taken.
Enabling Transparent Huge Pages On Linux
NOTE! On Linux, using ZGC with transparent huge pages enabled requires kernel >= 4.7.'
Use the following options to enable transparent huge pages in the VM:
These options tell the JVM to issue
madvise(..., MADV_HUGEPAGE) calls for memory it maps, which is useful when using transparent huge pages in
To enable transparent huge pages, you also need to configure the kernel by enabling
shmem huge pages for the heap, so the following kernel setting also needs to be configured:
It is important to check these kernel settings when comparing the performance of different GCs. Some Linux distributions forcefully enable transparent huge pages for private pages by configuring
/sys/kernel/mm/transparent_hugepage/enabled to be set to
always, while leaving
/sys/kernel/mm/transparent_hugepage/shmem_enabled at the default
never. In this case all GCs but ZGC will make use of transparent huge pages for the heap. See Transparent Hugepage Support for more information.
Enabling NUMA Support
ZGC has NUMA support, which means it will try it's best to direct Java heap allocations to NUMA-local memory. This feature is enabled by default. However, it will automatically be disabled if the JVM detects that it's bound to only use memory on a single NUMA node. In general, you don't need to worry about this setting, but if you want to explicitly override the JVM's decision you can do so by using the
When running on a NUMA machine (e.g. a multi-socket x86 machine), having NUMA support enabled will often give a noticeable performance boost.
Enabling GC Logging
GC logging is enabled using the following command-line option:
For general information/help on this option:
To enable basic logging (one line of output per GC):
To enable GC logging that is useful for tuning/performance analysis:
gc* means log all tag combinations that contain the
gc tag, and
:gc.log means write the log to a file named
- Support for generations (-XX:+ZGenerational) (JEP 439)
- Support for String Deduplication (-XX:+UseStringDeduplication)
- Linux/PowerPC support
- Various bug-fixes and optimizations
- Dynamic Number of GC threads
- Reduced mark stack memory usage
- macOS/aarch64 support
- GarbageCollectorMXBeans for both pauses and cycles
- Fast JVM termination
- Concurrent Thread Stack Scanning (JEP 376)
- Support for in-place relocation
- Performance improvements (allocation/initialization of forwarding tables, etc)
- Production ready (JEP 377)
- Improved NUMA awareness
- Improved allocation concurrency
- Support for Class Data Sharing (CDS)
- Support for placing the heap on NVRAM
- Support for compressed class pointers
- Support for incremental uncommit
- Fixed support for transparent huge pages
- Additional JFR events
- macOS support (JEP 364)
- Windows support (JEP 365)
- Support for tiny/small heaps (down to 8M)
- Support for JFR leak profiler
- Support for limited and discontiguous address space
- Parallel pre-touch (when using -XX:+AlwaysPreTouch)
- Performance improvements (clone intrinsic, etc)
- Stability improvements
- Increased max heap size from 4TB to 16TB
- Support for uncommitting unused memory (JEP 351)
- Support for -XX:SoftMaxHeapSIze
- Support for the Linux/AArch64 platform
- Reduced Time-To-Safepoint
- Support for concurrent class unloading
- Further pause time reductions
- Initial version of ZGC
- Does not support class unloading (using -XX:+ClassUnloading has no effect)
What does the "Z" in ZGC stand for?
It doesn't stand for anything, ZGC is just a name. It was originally inspired by, or a homage to, ZFS (the filesystem) which in many ways was revolutionary when it first came out. Originally, ZFS was an acronym for "Zettabyte File System", but that meaning was abandoned and it was later said to not stand for anything. It's just a name. See Jeff Bonwick's Blog for more details.
Is it pronounced "zed gee see" or "zee gee see"?
There's no preferred pronunciation, both are fine.
JVMLS 2023 - Video (33 min)
Oracle Developer Live 2022 - Slides | Video (28 min)
Jokerconf 2021 - Slides
Inside Java Podcast - ZGC - Sound (30 min)
Oracle Developer Live 2020 - Slides | Video (40 min)
Oracle Code One 2019 - Slides
PL-Meetup 2019 - Slides
Jfokus 2019 - Slides | Video (21 min)
Devoxx 2018 - Slides | Video (40 min)
Oracle Code One 2018 - Slides | Video (45 min)
Jfokus 2018 - Slides | Video (45 min)
FOSDEM 2018 - Slides
- No labels