Hi, I'm currently doing some benchmarking on a Maramba T5240. With cpustat, I have seen that TLB misses are quite an issue. Now I want to know whether these misses mainly occur in 8k or 4m pages. However, trapstat -T is not working properly: It only reports some tsb misses, but the columns for tlb misses are all 0 (instruction and data).
I have found bug ids 2127245 (fixed in s10u1_12) and 6277304 (fixed in snv_17) which might be my problem (the description is not very extensive...). I am currently running S10u4 (5.10 Generic_127111-07) and have no chance to install a Nevada build. Does anyone know whether this is one of those two bugs? Is there a fix available for S10u4? If there is no fix for trapstat, is there any way to get the page size where the misses occur (with DTrace or something like that)? Thanks a lot, Nick. This message posted from opensolaris.org _______________________________________________ perf-discuss mailing list perf-discuss@opensolaris.org
