On 2016/09/06 17:05, Paul Goyette wrote:
Kewl!  How is your performance?  I was talking with Matt Thomas the other day, 
and he suggested that having so much memory might introduce significant 
overhead in page management.

Hmm. This machine used to have 32GB memory. When it was
increased, I felt no significant regression in performance
both with scientific calculation (diagonalization of huge
matrix) and compiling something.

I notice your kernel is a couple weeks old, while mine was running a kernel 
from yesterday's sources.  I don't think anything has changed recently, but you 
never know.

I updated the kernel and userland to today's current.
But nothing special happened.

Thanks for the feedback.

My pleasure ;-)

Rin

Reply via email to