On Tue, 23 Nov 2010, stephane eranian wrote: > Seems like 0x1c8 is returning something more sensible, though I > believe it overcounts by 2x. Are you seeing this as well?
yes, though it's more like 1.75x +/- 10% (it varies run to run). It's unfortunate because the HW_INT_RCV results are very useful on other machines for adjusting to get the actual retired instruction count. All of the other machine types I've tested on can get an exact value on my one big microbenchmark, but on Atom I can't get the 0x1c8 value to correspond exactly, it varies from run to run in a way a can't correlate to anything. Vince ------------------------------------------------------------------------------ Increase Visibility of Your 3D Game App & Earn a Chance To Win $500! Tap into the largest installed PC base & get more eyes on your game by optimizing for Intel(R) Graphics Technology. Get started today with the Intel(R) Software Partner Program. Five $500 cash prizes are up for grabs. http://p.sf.net/sfu/intelisp-dev2dev _______________________________________________ perfmon2-devel mailing list perfmon2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
