Hi, LTP has a fair program (http://ltp.cvs.sourceforge.net/ltp/ltp/tools/genload/stress.c?revision=1.3&view=markup) which is capable of generating: 1) CPU stress, 2) Memory Stress, 3) Hard Disk Stress, 4) IO Bus Stress, 5) Network Stress (through netpipe)
in the Background, before your testcases (LTP) are run. The stress they generate are wholesome and cannot be discrete. We need to device a mechanism, where we can generate stress in matters of percentage (%), say: 1) X% of CPU should be stressed during run, 2) Y% of Memory Stress, 3) Z% of Hard Disk Stress, 4) P% of IO Bus Stress, 5) Q% Network Stress (through netpipe) And, all these should hold good on both Single CPU vs SMP machines as well as to NUMA architecture. Can somebody provide some Idea on how to design such a system ? We may also face a situation like: 1) On an SMP machine with 8 CPUs, only 2 should be subjected to 60% stress and others left normal, and lots of other combinations as well 2) On a NUMA machine, only particular node(s) should be stressed and other left normal. --Regards-- Subrata-- - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
