Thanks Rick for the message.
My machine is a HP server rx4640 (
http://h18000.www1.hp.com/products/quickspecs/11847_div/11847_div.HTML)
with 4 montecito processors. I understand this is a single-bus system.
I run up to 8 threads (of a maximum 16) so as not to 'invoke'
hyperthreading. Then measure BUS_DATA_CYCLE_ANY to calculate bus
utilization. My intention is to see if the benchmark is saturating the
memory bus with off-chip accesses.
Vineeth
On 06/28/2010 03:52 PM, Rick Jones wrote:
Vineeth Mekkat wrote:
Hi All,
I am trying to measure bus utilization of some benchmarks on Itanium
Montecito machine. I followed the procedure in
http://www.ice.gelato.org/oct06/pres_pdf/gelato_ICE06oct_perfmon2_eranian_hp.pdf
and measured the bus utilization. Now, I am not sure at what
percentage of value is the bus saturated? Is it 100%?
What sort of system are you measuring? A single-bus system, or a
"cell-based" system?
Say, if my bus utilization is 75%, can I say the bus is saturated?
I will be grateful if someone can shed some light on this.
I'm not sure how much light I will shed, but when it comes to memory
busses, I've never been comfortable much past 80%.
------------------------------------------------------------------------------
This SF.net email is sponsored by Sprint
What will you do first with EVO, the first 4G phone?
Visit sprint.com/first -- http://p.sf.net/sfu/sprint-com-first
_______________________________________________
perfmon2-devel mailing list
perfmon2-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
