Hi all,
I have a question about Umask value in Core 2 Duo processors. As you know,
there is an umask value for each event number. I have considered two events
from "Non-Architectural Performance Events in Core Duo Processors":
(03H: LD_Blocks, Umask 00H) and
(05H: Misalign_Mem_Ref, Umask: 00H).
A multi-threaded application has been executed as following with different
events:
sudo perf stat -r 5 -e r0300 -a my/a
21,242,257,626
r0300 ( +- 1.41% )
sudo perf stat -r 5 -e r03 -a my/a
0 r03
As you can see, output of 0300H is not the same as value of 03H
On the other hand, I have done this experiment for event 05H, and results
are shown at bellow:
sudo perf stat -r 5 -e r0500 -a my/a
0 r0500
sudo perf stat -r 5 -e r05 -a my/a
288,018 r05
( +- 27.30% )
According to the obtained results, It can be found that event 03 does not
work properly without determining of Umask value (00H). In contrast, event
05H are not able to count correctly whit Umask 00H. Now, Could you tell me
why this happens?
------------------------------------------------------------------------------
Want excitement?
Manually upgrade your production database.
When you want reliability, choose Perforce
Perforce version control. Predictably reliable.
http://pubads.g.doubleclick.net/gampad/clk?id=157508191&iu=/4140/ostg.clktrk
_______________________________________________
perfmon2-devel mailing list
perfmon2-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
