Hi, I am planning to benchmark the performance of my AMD GEODE LX platform with CONFIG_PREEMPT_PATCH. I am planning to do this with LPPTEST utility that comes the patch. My question is: For how long should I run the testlpp script in order to get the maximum latency? Right now I ran the testlpp for 15 mins. from my source machine and the max latency that it recorded for the my target system was around 677 usecs with 100% CPU load. Below is a part of the output: new max latency: 13.36 usecs (6681 cycles) new max latency: 33.11 usecs (16555 cycles) new max latency: 44.66 usecs (22330 cycles) new max latency: 156.87 usecs (78430 cycles) new max latency: 221.22 usecs (110605 cycles) new max latency: 386.23 usecs (193105 cycles) new max latency: 427.48 usecs (213730 cycles) new max latency: 450.58 usecs (225280 cycles) new max latency: 618.89 usecs (309430 cycles) new max latency: 640.34 usecs (320155 cycles) new max latency: 677.19 usecs (338575 cycles) [ interrupted - exiting ] total number of responses: 390755 average reponse latency: 12.70 usecs minimum latency: 1.71 usecs maximum latency: 677.19 usecs LPP latency histogram: 1 usecs: 22 8 usecs: 2690 9 usecs: 1325 ............................. ............................. ............................. I wonder if I run it for an entire day my max. latency would increase? Thanks Kushal Koolwal _________________________________________________________________ Get Windows Live and get whatever you need, wherever you are. Start here. http://www.windowslive.com/default.html?ocid=TXT_TAGLM_WL_Home_082008-- To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
