On Fri, 3 May 2013, Dan Terpstra wrote: > Since you've already hacked libpfm4 to make IVB58 look like SNB42, could > you check to see if the IVB deprecated floating point events actually > work? That's FP_COMP_OPS_EXE and SIMD_FP_256, which were removed from > the tables, but (rumor has it) not from the hardware. If they work as > well (badly) as on SNB, which we're beginning to understand, it might be > useful to reinstate them on IVB... - d
My "hack" was only for the uncore, the regular Ivybridge events have been supported properly for some time. You can look at the results in Table VI of my recent ISPASS paper http://web.eece.maine.edu/~vweaver/projects/deterministic/ispass2013_deterministic.pdf and see that x87 fp results using the undocumented events on IVB match pretty well with SNB. The SSE results do not though. And as for SIMD_FP_256 I don't think I have any AVX test code handy. It's not easy to re-add those events to PAPI though because libpfm4 doesn't support them. Vince ------------------------------------------------------------------------------ Get 100% visibility into Java/.NET code with AppDynamics Lite It's a free troubleshooting tool designed for production Get down to code-level detail for bottlenecks, with <2% overhead. Download for free and get started troubleshooting in minutes. http://p.sf.net/sfu/appdyn_d2d_ap2 _______________________________________________ perfmon2-devel mailing list perfmon2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/perfmon2-devel
