You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Which operating system and hardware are you running on?
Linux genoa3 6.2.16 #1 SMP PREEMPT_DYNAMIC Tue Jul 11 23:25:59 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
Details of the problem
I thought the physical indexes are OS provided. But those are not consistent with the result of lscpu command. They are not even continuous. Am I missing something?
lscpu isn't reporting core IDs here. Core IDs are identical between sockets because they are hardwired in each CPUs, not recomputed depending on the number of sockets. Hence they should be 0-31 twice on your system, not 0-31 in one socket and 32-63 in the other. I don't know lscpu enough to understand what they mean here, but my lscpu manpage talks about logical and physical IDs, at least in the -y option description, so maybe you're just showing their logical core IDs.
Anyway, beware that physical core IDs have been strange for years. For instance, they've been non-contiguous on many Intel CPUs (those where you don't have the maximal number of cores enabled in your SKUs).
Run this to see the list of core IDs reported by Linux and used by hwloc by default:
What version of hwloc are you using?
lstopo 2.7.0
Which operating system and hardware are you running on?
Linux genoa3 6.2.16 #1 SMP PREEMPT_DYNAMIC Tue Jul 11 23:25:59 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
Details of the problem
I thought the physical indexes are OS provided. But those are not consistent with the result of
lscpu
command. They are not even continuous. Am I missing something?The text was updated successfully, but these errors were encountered: