Linux Parallel CPU Bring-Up Shows Great Impact Bringing Up 1,920 Sapphire Rapids Cores

[jabl]

Close, but not quite.

The Altix UV line is actually pretty much dead at this point. The big HPE boxes now are actually replacements for Superdome/Superdome2/SuperdomeX for commercial computing (not technical computing like the Altix), but with SGI interconnect IP integrated in a vaguely Frankenstein-esque way. That's what the article refers to. The term to look up is "Superdome Flex" (not Superdome Flex 280, which is a mostly-unrelated glueless 8-socket system AFAIK.)

Oh, coming from a HPC background I had forgotten about Superdome. But, sounds reasonable. Sad to hear about the demise of the SGI heritage supercomputers, though.

[kozman]

From 71 > 14 seconds (80.28% reduction)? Full boot from 223 > 166 seconds (25.56% reduction)? Damn! With just one version rev? AMD be sweating yo. LOL! Datacenter admins will probably be happy about this coming down the pipe for a future kernel. How likely are they to backport it to 6.1 if it gets approval / merged down the road?

[onlyLinuxLuvUBack]

From 71 > 14 seconds (80.28% reduction)? Full boot from 223 > 166 seconds (25.56% reduction)? Damn! With just one version rev? AMD be sweating yo. LOL! Datacenter admins will probably be happy about this coming down the pipe for a future kernel. How likely are they to backport it to 6.1 if it gets approval / merged down the road?

add "systemd_" to the front

[dark-star]

it sounds impressive, but for big servers in the datacenter it doesn't really matter much:
- servers are rarely rebooted, many have uptimes in months. 5 minutes in a month is not much time saved
- servers spend many minutes in their POST, enumerating PCI devices, SAN storage, etc.
- often, the pre-boot IPMI takes some time as well, at least on a cold boot
- the actual services they're starting after boot (databases, ...) take minutes to start up too

[willmore]

zamadatix jabl Thank you both for that clarification!