Linux 4.15 Will Have A Scheduler Change To Benefit AMD EPYC

[phoronix]

Phoronix: Linux 4.15 Will Have A Scheduler Change To Benefit AMD EPYC

Linux 4.15 will be exciting for AMD Zen systems not only for working temperature reporting (finally) being in place for Ryzen/EPYC, but AMD EPYC CPUs should also benefit from a scheduler topology improvement...

http://www.phoronix.com/scan.php?page=news_item&px=AMD-EPYC-Sched-Linux-4.15

[chuckula]

I'd be a little more impressed if this was a generic improvement to all NUMA schedulers instead of an AMD-specific workaround.

For example, an 8 socket Xeon Platinum system also uses 8 NUMA nodes and it would be more beneficial if improvements to NUMA would benefit all systems with that topology.

[torsionbar28]

I'd be a little more impressed if this was a generic improvement to all NUMA schedulers instead of an AMD-specific workaround.

For example, an 8 socket Xeon Platinum system also uses 8 NUMA nodes and it would be more beneficial if improvements to NUMA would benefit all systems with that topology.

Core count may be the same, but the topology sure isn't. It makes sense that each has their own unique scheduler optimizations.

[agd5f]

The patch does not appear to be AMD specific.

[schmidtbag]

Interesting - Epyc already had impressive performance to begin with, and this should only help further improve it.

[pegasus]

Hope this gets backported to RHEL.

[waxhead]

A decade of wasted cores comes to mind. https://www.phoronix.com/scan.php?pa...-Scheduler-Bad

How about a follow up story? Really miss something there...

[torsionbar28]

Hope this gets backported to RHEL.

If the improvement is that significant, I'm sure 7 will get it. No luck for 6 though.

[pavlerson]

A decade of wasted cores comes to mind. https://www.phoronix.com/scan.php?pa...-Scheduler-Bad

How about a follow up story? Really miss something there...

x86 servers are typically 1-2 sockets. The 8-16 socket server domain are exclusively Unix and RISC running Unix (Solaris/AIX). With that said, if you have a crappy scheduler on 1-socket machine, it is no big difference to a good scheduler. However, if you have a bad scheduler on 8 sockets then performance will suffer a lot. Thus, as all Linux servers are 1-2 sockets, a bad scheduler is not a big problem. Unless you see Linux on 8 socket servers, you will not need a good scheduler. There is no need for follow up story until 4-8 sockets are common.

[sleeplessclassics]

Bit of noob question, I understand why the kernel will have optimizations for different platforms, so that all platforms can use the same kernel.

But from an end user perspective, isn't it a waste and inefficient to have code that won't be used.
i.e Having an AMD system but kernel will still have intel stuff??