Hello,
I have run some tests to troubleshoot application performance issues and i found something weird
I used passmark and cpu z to test the cpu performance, i'll put the cpuz results here
So on a vm on windows 2012R2 on an esxi 6.x host with 2 xeon e5 2643 @3.4ghz with 2 sockets with 2 cores/socket configured so 4 cores at 3.2ghz
with cpu Z i got this
Single-Thread : 113
Multi-Thread (4T) : 385
If i vmotion the vm to an esxi with 2 xeons e5 2670 @2.3ghz same socket core config
with cpu Z i got this
Single-Thread : 122
Multi-Thread (4T) : 449
First of all i'm surprised that the second xeon with much less ghz performs better....?? i was expecting the contrary, what do you think?
Then i run the same test on our old vmware env, an esxi with 2 xeons e5649 @2.53ghz
with cpu Z i got this
Single-Thread : 219
Multi-Thread (4T) : 870
Wel well, our 10 years old infra works better than the 2 years old one....?????
Any idea about what i can do for this?
Thank you
Are you sure the VM you tested, it's the same in your old environment than in the new one?
Anyway, do you have the same settings from your host or the Datastore that is running this VM?
well i have 200+ vm
on each vm i test in the cisco infra i have +-120 on single thread cpu result
on each vm i test in the ibm infra (the old one) i have +-240 on single thread
i did the test on more than 10 vm, on both 2012r2, 2016, win10
Also if i run the same test on some physical machine with old xeon proc i have something like +-240 single thread, who make me think i have an issue with vmware or cisco
I did create a ticket on both vmware and cisco sides...
Also to be clear, our hosts are absolutely not under cpu stress, never more than 20% usage ....
I just did a cpu z on my surface 4 pro i7 i have 288 in single thread....
after that i tested 2 of our DC on win2008R2, same spec, one on cisco one on ibm, same results