NVIDIA Engineer Continues Working On Proactive Memory Compaction For Linux

NVIDIA's Nitin Gupta continues working on proactive compaction for the Linux kernel's memory management code.

This proactive compaction is designed to avoid the high latency introduced right now when the Linux kernel does on-demand compaction when an application needs a lot of hugepages. With this proactive compaction, a large number of hugepages can be requested while avoiding high latencies.

The proactive compaction code is now up to its second revision and exposes a sysfs tunable for the low/high thresholds for external fragmentation for triggering the proactive compaction. Ultimately it's for addressing the performance/latency when a lot of hugepages are needed for an application on an otherwise memory fragmented system.

The proactive memory compaction code is still flying under a "request for comments" flag but the results from this NVIDIA engineer are quite promising. Hopefully the work will be received well and can make it to the mainline kernel in the not too distant future. While NVIDIA isn't known for their general code contributions to the Linux kernel, this isn't Nitin's first rodeo - while he's been at NVIDIA just a year, prior to that he was working on Linux kernel engineering for Oracle and Intel.

Those wishing to learn more about the proactive compaction work can do so via this mailing list thread.