Hi Rajaram, > I want to practically see the benefits of CKRM CPU resource controller > on my PC. > > For example, I would like to see something like : > > how a process in a class C1 with a high share of CPU runs faster than > another process in a class C2 with a low share of CPU. > > Could anyone please suggest me some effective ways of trying this ?
Please notice the following two things: 1) Choose CPU bound processes to run CPU controller cannot regulate I/O bound and mostly sleeping process's speed. 2) Total number of CPU bound processes must be more than the number of CPU It is because surplus CPU can be consumed by any processes regardless of share. The share is not limit, but minimum guarantee. For example, if only two CPU hog process run on 2CPU machine, each of them can consume one whole CPU regardless of share. Have you downloaded the pdf file, cpurc-v0.3-2615-kernbench.pdf on the project web site? It shows the results of kernbench on several conditions. > I remember seeing details about kernbench in this mailing list. Would > kernbench be useful for my purpose ? One flaw of kernbench for this purpose is that measuring kernbench is time consuming, but it is useful for demonstrating CPU controller. Thanks, MAEDA Naoaki ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642 _______________________________________________ ckrm-tech mailing list https://lists.sourceforge.net/lists/listinfo/ckrm-tech
