https://community.intel.com/t5/Intel-Integrated-Performance/Timing-question-involving-the-IPPL/m-p/990623#M22257 <DIV></DIV> <DIV>This is question that involves the Intel Performance<BR />Primitives Library (IPPL), but it is more a question about<BR />timing on a Pentium 4 with hyperthreading using the IPPL.</DIV> <DIV></DIV> <DIV>I have an application that uses the IPPL on a Windows XP<BR />system running a Pentium 4 with hyperthreading turned on,<BR />i.e it has two logical processors, and one physical processor.<BR />When we time calculations using the Windows GetProcessTimes<BR />interface we get the following curious result.</DIV> <DIV></DIV> <DIV>When we do a calculation that shows 100% total CPU usage<BR />(i.e. 100% on both CPUs) we get a time that is very close to<BR />twice wallclock time. That is expected.</DIV> <DIV></DIV> <DIV>When we do a different calculation that shows 60% total CPU usage<BR />(according to the Windows performance monitor) we get agreement<BR />with wallclock time to say three figures. Why is this happening?<BR />I would have expected that the total process time would differ<BR />from wallclock by say 20%. In other words, I would have expected<BR />a gradual increase in process time as the CPU utilization<BR />goes from 50% to 100%. But at 60% my data shows no change relative<BR />to wallclock time.</DIV> <DIV></DIV> <DIV>I would appreciate any insight into this matter.</DIV> Thu, 19 May 2005 04:45:01 GMT martin3 2005-05-19T04:45:01Z https://community.intel.com/t5/Intel-Integrated-Performance/Timing-question-involving-the-IPPL/m-p/990623#M22257 <DIV></DIV> <DIV>This is question that involves the Intel Performance<BR />Primitives Library (IPPL), but it is more a question about<BR />timing on a Pentium 4 with hyperthreading using the IPPL.</DIV> <DIV></DIV> <DIV>I have an application that uses the IPPL on a Windows XP<BR />system running a Pentium 4 with hyperthreading turned on,<BR />i.e it has two logical processors, and one physical processor.<BR />When we time calculations using the Windows GetProcessTimes<BR />interface we get the following curious result.</DIV> <DIV></DIV> <DIV>When we do a calculation that shows 100% total CPU usage<BR />(i.e. 100% on both CPUs) we get a time that is very close to<BR />twice wallclock time. That is expected.</DIV> <DIV></DIV> <DIV>When we do a different calculation that shows 60% total CPU usage<BR />(according to the Windows performance monitor) we get agreement<BR />with wallclock time to say three figures. Why is this happening?<BR />I would have expected that the total process time would differ<BR />from wallclock by say 20%. In other words, I would have expected<BR />a gradual increase in process time as the CPU utilization<BR />goes from 50% to 100%. But at 60% my data shows no change relative<BR />to wallclock time.</DIV> <DIV></DIV> <DIV>I would appreciate any insight into this matter.</DIV> Thu, 19 May 2005 04:45:01 GMT https://community.intel.com/t5/Intel-Integrated-Performance/Timing-question-involving-the-IPPL/m-p/990623#M22257 martin3 2005-05-19T04:45:01Z https://community.intel.com/t5/Intel-Integrated-Performance/Timing-question-involving-the-IPPL/m-p/990624#M22258 <DIV>Hi,</DIV> <DIV></DIV> <DIV>I did not understood your question. Did you see performance problem in IPP?</DIV> <DIV></DIV> <DIV>Regards,</DIV> <DIV> Vladimir</DIV> Wed, 25 May 2005 00:49:19 GMT https://community.intel.com/t5/Intel-Integrated-Performance/Timing-question-involving-the-IPPL/m-p/990624#M22258 Vladimir_Dudnik 2005-05-25T00:49:19Z