Hello Shrikanth,
On 10/7/2026 11:13 AM, Shrikanth Hegde wrote:
> Hi Prateek.
>
> I ran the series on 320 CPU power11 system today.
> hackbench shows high regression across the different number of groups and
> it is more severe at lower number of groups. I haven't run anything else yet.
>
> I haven't taken a look at the patches yet. Post your talk today will try to
> go through and see if i can spot something.
Thanks a ton for the testing. Now that I'm back, I'll take a look at
few bits again, especially stuff for optimizing traversals.
Can you please share what does your system topology look like and what
did:
sudo dmesg | grep sbm
say when you booted into your kernel?
>> +void __init sbm_set_topology(int num_instances, int
>> max_threads_per_instance)
>> +{
>> + sbm_max_threads_per_instance = max_threads_per_instance;
>> + sbm_num_instance = num_instances;
>> +
>> + pr_info("sbm topology set to %d instance(s) with a maximum of %d
>> thread(s) per instance\n",
>> + sbm_num_instance,
>> + sbm_max_threads_per_instance);
>> +}
>
> nit: Stray empty line at the end.
Ack! Thank you again for taking a look.
>
>> +
--
Thanks and Regards,
Prateek
