From cdirik at umd.edu Sun Jun 8 17:46:28 2008 From: cdirik at umd.edu (Cagdas Dirik) Date: Sun, 8 Jun 2008 17:46:28 -0400 (EDT) Subject: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected Message-ID: <20080608174628.DFT51156@po1.mail.umd.edu> Hi, Like Lei, when I runvalid DiskSim 4.0, I am getting mismatching results. I have tried this on a OS 10.4 with gcc version 4.0.1. Also tried it on a linux box and got the same mismatching outputs. Any ideas on what the problem might be? Thanks, Cagdas Hi, I compiled the source code of disksim 4.0 successfully on my virtual machine (Fedora Core 4), and ran runvalid and memsvalid files under the valid directory. The results are as follows, and it seems that, 1) for runvalid, some simulation results are not the same as expected, please see the results marked in red. 2) for memsvalid, it looks like there is no mems.g3.parv file under the valid directory. Thanks, Lei [root at smile valid]# ./runvalid These results represent actual drive validation experiments QUANTUM_QM39100TD-SW (rms should be about 0.378) disksim_run_simulation(): simulated 1090389 events rms = 0.377952 SEAGATE_ST32171W (rms should be about 0.349) disksim_run_simulation(): simulated 1585710 events rms = 0.347570 SEAGATE_ST34501N (rms should be about 0.318) disksim_run_simulation(): simulated 1591859 events rms = 0.317972 SEAGATE_ST39102LW (rms should be about 0.107) disksim_run_simulation(): simulated 883217 events rms = 0.106906 IBM_DNES-309170W (rms should be about 0.135) disksim_run_simulation(): simulated 2690642 events rms = 0.135884 QUANTUM_TORNADO (rms should be about 0.159) disksim_run_simulation(): simulated 1094910 events rms = 0.267721 HP_C2247_validate (rms should be about 0.090) disksim_run_simulation(): simulated 829410 events rms = 0.089931 HP_C3323_validate (rms should be about 0.306) disksim_run_simulation(): simulated 1088847 events rms = 0.305653 HP_C2490_validate (rms should be about 0.254) disksim_run_simulation(): simulated 1068808 events rms = 0.253762 DEC_RZ26_validate (rms should be about 0.438) *** warning: ignoring hpl parameters for non-hpl seek type. disksim_run_simulation(): simulated 707184 events rms = 0.340367 The remaining tests exercise other DiskSim components Open synthetic workload (avg. resp. should be about 47.3ms) disksim_run_simulation(): simulated 1044524 events IOdriver Response time average: 10.937386 Closed synthetic workload (avg. resp. should be about 87.6ms) disksim_run_simulation(): simulated 571678 events IOdriver Response time average: 87.819135 Mixed synthetic workload (avg. resp. should be about 24.4ms) disksim_run_simulation(): simulated 1313881 events IOdriver Response time average: 22.086628 RAID 5 at device driver (avg. resp. should be about 22.8ms) disksim_run_simulation(): simulated 1974909 events IOdriver Response time average: 22.861326 Set of disk arrays at device driver (avg. resp. should be about 33.3ms) disksim_run_simulation(): simulated 1895981 events Overall I/O System Response time average: 34.272035 Memory cache at controller (avg. resp. should be about 23.0ms) disksim_run_simulation(): simulated 1396705 events IOdriver Response time average: 24.651367 Cache device managed at controller (avg. resp. should be about 27.4ms) disksim_run_simulation(): simulated 2123707 events IOdriver Response time average: 28.939379 Simpledisk instead of original model (avg. resp. should be about 13.5ms) disksim_run_simulation(): simulated 377469 events IOdriver Response time average: 13.711596 3 different disks (avg. resps. should be about 10.951) disksim_run_simulation(): simulated 1044524 events IOdriver Response time average: 10.937386 3 disks on separate controllers (avg. resps. should be about 10.951) disksim_run_simulation(): simulated 1044524 events IOdriver Response time average: 10.937386 Note: this validates HP trace input, *not* the corresponding traced disk HP srt trace input (avg. resp should be about 48.8ms) disksim_run_simulation(): simulated 3422024 events IOdriver Response time average: 48.786646 ASCII input (avg. resp should be about 13.8ms) disksim_run_simulation(): simulated 10754770 events IOdriver Response time average: 13.766948 syssim: externally-controlled DiskSim (avg. resp should be about 8.9ms) response time: n=1000 average=8.894719 std. deviation=2.116511 IOdriver Response time average: 8.894719 [root at smile valid]# ./memsvalid G1 MEMS Device - synthetic workload (avg. resp. should be about 2.00 ms) disksim_run_simulation(): simulated 891324 events IOdriver Response time average: 2.004587 G2 MEMS Device - synthetic workload (avg. resp. should be about 1.18 ms) disksim_run_simulation(): simulated 800016 events IOdriver Response time average: 1.178837 G3 MEMS Device - synthetic workload (avg. resp. should be about 0.87 ms) *** assertion failed: in disksim_loadparams() (disksim_loadparams.c:86): disksim->parfile != NULL: mems.g3.parv ./memsvalid: line 15: 11021 gave up ../src/disksim mems.g3.parv mems.g3.nospring.outv ascii 0 1 mems0 "Spring constant factor" 0.00 G1 MEMS Device with 75% springs - synthetic workload (avg. resp. should be about 2.01 ms) disksim_run_simulation(): simulated 891263 events IOdriver Response time average: 2.013940 G2 MEMS Device with 75% springs - synthetic workload (avg. resp. should be about 1.16 ms) disksim_run_simulation(): simulated 799997 events IOdriver Response time average: 1.158043 G3 MEMS Device with 75% springs - synthetic workload (avg. resp. should be about 0.85 ms) *** assertion failed: in disksim_loadparams() (disksim_loadparams.c:86): disksim->parfile != NULL: mems.g3.parv ./memsvalid: line 30: 11027 gave up ../src/disksim mems.g3.parv mems.g3.outv ascii 0 1 mems0 "Spring constant factor" 0.75 From wobber at microsoft.com Mon Jun 9 13:52:52 2008 From: wobber at microsoft.com (Ted Wobber) Date: Mon, 9 Jun 2008 10:52:52 -0700 Subject: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected In-Reply-To: <20080608174628.DFT51156@po1.mail.umd.edu> References: <20080608174628.DFT51156@po1.mail.umd.edu> Message-ID: <338B06DAAFBCFD4894EFCAC38E7B811A1FCB46D695@NA-EXMSG-C113.redmond.corp.microsoft.com> We observe virtually identical results on our Win32 port of DiskSim. Ted Wobber MSR Silicon Valley -----Original Message----- From: disksim-users-bounces at ece.cmu.edu [mailto:disksim-users-bounces at ece.cmu.edu] On Behalf Of Cagdas Dirik Sent: Sunday, June 08, 2008 2:46 PM To: disksim-users at ece.cmu.edu Subject: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected Hi, Like Lei, when I runvalid DiskSim 4.0, I am getting mismatching results. I have tried this on a OS 10.4 with gcc version 4.0.1. Also tried it on a linux box and got the same mismatching outputs. Any ideas on what the problem might be? Thanks, Cagdas Hi, I compiled the source code of disksim 4.0 successfully on my virtual machine (Fedora Core 4), and ran runvalid and memsvalid files under the valid directory. The results are as follows, and it seems that, 1) for runvalid, some simulation results are not the same as expected, please see the results marked in red. 2) for memsvalid, it looks like there is no mems.g3.parv file under the valid directory. Thanks, Lei [root at smile valid]# ./runvalid These results represent actual drive validation experiments QUANTUM_QM39100TD-SW (rms should be about 0.378) disksim_run_simulation(): simulated 1090389 events rms = 0.377952 SEAGATE_ST32171W (rms should be about 0.349) disksim_run_simulation(): simulated 1585710 events rms = 0.347570 SEAGATE_ST34501N (rms should be about 0.318) disksim_run_simulation(): simulated 1591859 events rms = 0.317972 SEAGATE_ST39102LW (rms should be about 0.107) disksim_run_simulation(): simulated 883217 events rms = 0.106906 IBM_DNES-309170W (rms should be about 0.135) disksim_run_simulation(): simulated 2690642 events rms = 0.135884 QUANTUM_TORNADO (rms should be about 0.159) disksim_run_simulation(): simulated 1094910 events rms = 0.267721 HP_C2247_validate (rms should be about 0.090) disksim_run_simulation(): simulated 829410 events rms = 0.089931 HP_C3323_validate (rms should be about 0.306) disksim_run_simulation(): simulated 1088847 events rms = 0.305653 HP_C2490_validate (rms should be about 0.254) disksim_run_simulation(): simulated 1068808 events rms = 0.253762 DEC_RZ26_validate (rms should be about 0.438) *** warning: ignoring hpl parameters for non-hpl seek type. disksim_run_simulation(): simulated 707184 events rms = 0.340367 The remaining tests exercise other DiskSim components Open synthetic workload (avg. resp. should be about 47.3ms) disksim_run_simulation(): simulated 1044524 events IOdriver Response time average: 10.937386 Closed synthetic workload (avg. resp. should be about 87.6ms) disksim_run_simulation(): simulated 571678 events IOdriver Response time average: 87.819135 Mixed synthetic workload (avg. resp. should be about 24.4ms) disksim_run_simulation(): simulated 1313881 events IOdriver Response time average: 22.086628 RAID 5 at device driver (avg. resp. should be about 22.8ms) disksim_run_simulation(): simulated 1974909 events IOdriver Response time average: 22.861326 Set of disk arrays at device driver (avg. resp. should be about 33.3ms) disksim_run_simulation(): simulated 1895981 events Overall I/O System Response time average: 34.272035 Memory cache at controller (avg. resp. should be about 23.0ms) disksim_run_simulation(): simulated 1396705 events IOdriver Response time average: 24.651367 Cache device managed at controller (avg. resp. should be about 27.4ms) disksim_run_simulation(): simulated 2123707 events IOdriver Response time average: 28.939379 Simpledisk instead of original model (avg. resp. should be about 13.5ms) disksim_run_simulation(): simulated 377469 events IOdriver Response time average: 13.711596 3 different disks (avg. resps. should be about 10.951) disksim_run_simulation(): simulated 1044524 events IOdriver Response time average: 10.937386 3 disks on separate controllers (avg. resps. should be about 10.951) disksim_run_simulation(): simulated 1044524 events IOdriver Response time average: 10.937386 Note: this validates HP trace input, *not* the corresponding traced disk HP srt trace input (avg. resp should be about 48.8ms) disksim_run_simulation(): simulated 3422024 events IOdriver Response time average: 48.786646 ASCII input (avg. resp should be about 13.8ms) disksim_run_simulation(): simulated 10754770 events IOdriver Response time average: 13.766948 syssim: externally-controlled DiskSim (avg. resp should be about 8.9ms) response time: n=1000 average=8.894719 std. deviation=2.116511 IOdriver Response time average: 8.894719 [root at smile valid]# ./memsvalid G1 MEMS Device - synthetic workload (avg. resp. should be about 2.00 ms) disksim_run_simulation(): simulated 891324 events IOdriver Response time average: 2.004587 G2 MEMS Device - synthetic workload (avg. resp. should be about 1.18 ms) disksim_run_simulation(): simulated 800016 events IOdriver Response time average: 1.178837 G3 MEMS Device - synthetic workload (avg. resp. should be about 0.87 ms) *** assertion failed: in disksim_loadparams() (disksim_loadparams.c:86): disksim->parfile != NULL: mems.g3.parv ./memsvalid: line 15: 11021 gave up ../src/disksim mems.g3.parv mems.g3.nospring.outv ascii 0 1 mems0 "Spring constant factor" 0.00 G1 MEMS Device with 75% springs - synthetic workload (avg. resp. should be about 2.01 ms) disksim_run_simulation(): simulated 891263 events IOdriver Response time average: 2.013940 G2 MEMS Device with 75% springs - synthetic workload (avg. resp. should be about 1.16 ms) disksim_run_simulation(): simulated 799997 events IOdriver Response time average: 1.158043 G3 MEMS Device with 75% springs - synthetic workload (avg. resp. should be about 0.85 ms) *** assertion failed: in disksim_loadparams() (disksim_loadparams.c:86): disksim->parfile != NULL: mems.g3.parv ./memsvalid: line 30: 11027 gave up ../src/disksim mems.g3.parv mems.g3.outv ascii 0 1 mems0 "Spring constant factor" 0.75 _______________________________________________ Disksim-users mailing list Disksim-users at ece.cmu.edu https://sos.ece.cmu.edu/mailman/listinfo/disksim-users From cdirik at umd.edu Mon Jun 9 15:19:09 2008 From: cdirik at umd.edu (Cagdas Dirik) Date: Mon, 9 Jun 2008 15:19:09 -0400 (EDT) Subject: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected Message-ID: <20080609151909.DFU15290@po1.mail.umd.edu> Hi Ted, You mean identical results in the sense that they are matching expected results? Or idential results to what Lei and I observe (mismatching to expected results)? Thanks, Cagdas ---- Original message ---- >Date: Mon, 9 Jun 2008 10:52:52 -0700 >From: Ted Wobber >Subject: RE: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected >To: "cdirik at umd.edu" , "disksim-users at ece.cmu.edu" >Cc: Ted Wobber > >We observe virtually identical results on our Win32 port of DiskSim. > >Ted Wobber >MSR Silicon Valley > > >-----Original Message----- >From: disksim-users-bounces at ece.cmu.edu [mailto:disksim-users-bounces at ece.cmu.edu] On Behalf Of Cagdas Dirik >Sent: Sunday, June 08, 2008 2:46 PM >To: disksim-users at ece.cmu.edu >Subject: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected > >Hi, > >Like Lei, when I runvalid DiskSim 4.0, I am getting mismatching results. I have >tried this on a OS 10.4 with gcc version 4.0.1. Also tried it on a linux box and >got the same mismatching outputs. > >Any ideas on what the problem might be? > >Thanks, > >Cagdas > >Hi, > >I compiled the source code of disksim 4.0 successfully on my virtual machine >(Fedora Core 4), and ran runvalid and memsvalid files under the valid directory. > >The results are as follows, and it seems that, >1) for runvalid, some simulation results are not the same as expected, please >see the results marked in red. >2) for memsvalid, it looks like there is no mems.g3.parv file under the valid >directory. > >Thanks, > >Lei > > >[root at smile valid]# ./runvalid > >These results represent actual drive validation experiments > >QUANTUM_QM39100TD-SW (rms should be about 0.378) >disksim_run_simulation(): simulated 1090389 events >rms = 0.377952 > >SEAGATE_ST32171W (rms should be about 0.349) >disksim_run_simulation(): simulated 1585710 events >rms = 0.347570 > >SEAGATE_ST34501N (rms should be about 0.318) >disksim_run_simulation(): simulated 1591859 events >rms = 0.317972 > >SEAGATE_ST39102LW (rms should be about 0.107) >disksim_run_simulation(): simulated 883217 events >rms = 0.106906 > >IBM_DNES-309170W (rms should be about 0.135) >disksim_run_simulation(): simulated 2690642 events >rms = 0.135884 > >QUANTUM_TORNADO (rms should be about 0.159) >disksim_run_simulation(): simulated 1094910 events >rms = 0.267721 > >HP_C2247_validate (rms should be about 0.090) >disksim_run_simulation(): simulated 829410 events >rms = 0.089931 > >HP_C3323_validate (rms should be about 0.306) >disksim_run_simulation(): simulated 1088847 events >rms = 0.305653 > >HP_C2490_validate (rms should be about 0.254) >disksim_run_simulation(): simulated 1068808 events >rms = 0.253762 > >DEC_RZ26_validate (rms should be about 0.438) >*** warning: ignoring hpl parameters for non-hpl seek type. >disksim_run_simulation(): simulated 707184 events >rms = 0.340367 > >The remaining tests exercise other DiskSim components > >Open synthetic workload (avg. resp. should be about 47.3ms) >disksim_run_simulation(): simulated 1044524 events >IOdriver Response time average: 10.937386 > >Closed synthetic workload (avg. resp. should be about 87.6ms) >disksim_run_simulation(): simulated 571678 events >IOdriver Response time average: 87.819135 > >Mixed synthetic workload (avg. resp. should be about 24.4ms) >disksim_run_simulation(): simulated 1313881 events >IOdriver Response time average: 22.086628 > >RAID 5 at device driver (avg. resp. should be about 22.8ms) >disksim_run_simulation(): simulated 1974909 events >IOdriver Response time average: 22.861326 > >Set of disk arrays at device driver (avg. resp. should be about 33.3ms) >disksim_run_simulation(): simulated 1895981 events >Overall I/O System Response time average: 34.272035 > >Memory cache at controller (avg. resp. should be about 23.0ms) >disksim_run_simulation(): simulated 1396705 events >IOdriver Response time average: 24.651367 > >Cache device managed at controller (avg. resp. should be about 27.4ms) >disksim_run_simulation(): simulated 2123707 events >IOdriver Response time average: 28.939379 > >Simpledisk instead of original model (avg. resp. should be about 13.5ms) >disksim_run_simulation(): simulated 377469 events >IOdriver Response time average: 13.711596 >3 different disks >(avg. resps. should be about 10.951) >disksim_run_simulation(): simulated 1044524 events >IOdriver Response time average: 10.937386 >3 disks on separate controllers >(avg. resps. should be about 10.951) >disksim_run_simulation(): simulated 1044524 events >IOdriver Response time average: 10.937386 > >Note: this validates HP trace input, *not* the corresponding traced disk >HP srt trace input (avg. resp should be about 48.8ms) >disksim_run_simulation(): simulated 3422024 events >IOdriver Response time average: 48.786646 > >ASCII input (avg. resp should be about 13.8ms) >disksim_run_simulation(): simulated 10754770 events >IOdriver Response time average: 13.766948 > >syssim: externally-controlled DiskSim (avg. resp should be about 8.9ms) >response time: n=1000 average=8.894719 std. deviation=2.116511 >IOdriver Response time average: 8.894719 > > > >[root at smile valid]# ./memsvalid > >G1 MEMS Device - synthetic workload (avg. resp. should be about 2.00 ms) >disksim_run_simulation(): simulated 891324 events >IOdriver Response time average: 2.004587 > >G2 MEMS Device - synthetic workload (avg. resp. should be about 1.18 ms) >disksim_run_simulation(): simulated 800016 events >IOdriver Response time average: 1.178837 > >G3 MEMS Device - synthetic workload (avg. resp. should be about 0.87 ms) >*** assertion failed: in disksim_loadparams() (disksim_loadparams.c:86): >disksim->parfile != NULL: mems.g3.parv >./memsvalid: line 15: 11021 gave up ../src/disksim mems.g3.parv >mems.g3.nospring.outv ascii 0 1 mems0 "Spring constant factor" 0.00 > >G1 MEMS Device with 75% springs - synthetic workload (avg. resp. should be >about 2.01 ms) >disksim_run_simulation(): simulated 891263 events >IOdriver Response time average: 2.013940 > >G2 MEMS Device with 75% springs - synthetic workload (avg. resp. should be >about 1.16 ms) >disksim_run_simulation(): simulated 799997 events >IOdriver Response time average: 1.158043 > >G3 MEMS Device with 75% springs - synthetic workload (avg. resp. should be >about 0.85 ms) >*** assertion failed: in disksim_loadparams() (disksim_loadparams.c:86): >disksim->parfile != NULL: mems.g3.parv >./memsvalid: line 30: 11027 gave up ../src/disksim mems.g3.parv >mems.g3.outv ascii 0 1 mems0 "Spring constant factor" 0.75 > >_______________________________________________ >Disksim-users mailing list >Disksim-users at ece.cmu.edu >https://sos.ece.cmu.edu/mailman/listinfo/disksim-users From wobber at microsoft.com Mon Jun 9 15:53:40 2008 From: wobber at microsoft.com (Ted Wobber) Date: Mon, 9 Jun 2008 12:53:40 -0700 Subject: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected In-Reply-To: <20080609151909.DFU15290@po1.mail.umd.edu> References: <20080609151909.DFU15290@po1.mail.umd.edu> Message-ID: <338B06DAAFBCFD4894EFCAC38E7B811A1FCB46D7F7@NA-EXMSG-C113.redmond.corp.microsoft.com> Identical to what you and Lei observe. /Ted -----Original Message----- From: Cagdas Dirik [mailto:cdirik at umd.edu] Sent: Monday, June 09, 2008 12:19 PM To: Ted Wobber; disksim-users at ece.cmu.edu Subject: RE: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected Hi Ted, You mean identical results in the sense that they are matching expected results? Or idential results to what Lei and I observe (mismatching to expected results)? Thanks, Cagdas ---- Original message ---- >Date: Mon, 9 Jun 2008 10:52:52 -0700 >From: Ted Wobber >Subject: RE: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected >To: "cdirik at umd.edu" , "disksim-users at ece.cmu.edu" >Cc: Ted Wobber > >We observe virtually identical results on our Win32 port of DiskSim. > >Ted Wobber >MSR Silicon Valley > > >-----Original Message----- >From: disksim-users-bounces at ece.cmu.edu [mailto:disksim-users-bounces at ece.cmu.edu] On Behalf Of Cagdas Dirik >Sent: Sunday, June 08, 2008 2:46 PM >To: disksim-users at ece.cmu.edu >Subject: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected > >Hi, > >Like Lei, when I runvalid DiskSim 4.0, I am getting mismatching results. I have >tried this on a OS 10.4 with gcc version 4.0.1. Also tried it on a linux box and >got the same mismatching outputs. > >Any ideas on what the problem might be? > >Thanks, > >Cagdas > >Hi, > >I compiled the source code of disksim 4.0 successfully on my virtual machine >(Fedora Core 4), and ran runvalid and memsvalid files under the valid directory. > >The results are as follows, and it seems that, >1) for runvalid, some simulation results are not the same as expected, please >see the results marked in red. >2) for memsvalid, it looks like there is no mems.g3.parv file under the valid >directory. > >Thanks, > >Lei > > >[root at smile valid]# ./runvalid > >These results represent actual drive validation experiments > >QUANTUM_QM39100TD-SW (rms should be about 0.378) >disksim_run_simulation(): simulated 1090389 events >rms = 0.377952 > >SEAGATE_ST32171W (rms should be about 0.349) >disksim_run_simulation(): simulated 1585710 events >rms = 0.347570 > >SEAGATE_ST34501N (rms should be about 0.318) >disksim_run_simulation(): simulated 1591859 events >rms = 0.317972 > >SEAGATE_ST39102LW (rms should be about 0.107) >disksim_run_simulation(): simulated 883217 events >rms = 0.106906 > >IBM_DNES-309170W (rms should be about 0.135) >disksim_run_simulation(): simulated 2690642 events >rms = 0.135884 > >QUANTUM_TORNADO (rms should be about 0.159) >disksim_run_simulation(): simulated 1094910 events >rms = 0.267721 > >HP_C2247_validate (rms should be about 0.090) >disksim_run_simulation(): simulated 829410 events >rms = 0.089931 > >HP_C3323_validate (rms should be about 0.306) >disksim_run_simulation(): simulated 1088847 events >rms = 0.305653 > >HP_C2490_validate (rms should be about 0.254) >disksim_run_simulation(): simulated 1068808 events >rms = 0.253762 > >DEC_RZ26_validate (rms should be about 0.438) >*** warning: ignoring hpl parameters for non-hpl seek type. >disksim_run_simulation(): simulated 707184 events >rms = 0.340367 > >The remaining tests exercise other DiskSim components > >Open synthetic workload (avg. resp. should be about 47.3ms) >disksim_run_simulation(): simulated 1044524 events >IOdriver Response time average: 10.937386 > >Closed synthetic workload (avg. resp. should be about 87.6ms) >disksim_run_simulation(): simulated 571678 events >IOdriver Response time average: 87.819135 > >Mixed synthetic workload (avg. resp. should be about 24.4ms) >disksim_run_simulation(): simulated 1313881 events >IOdriver Response time average: 22.086628 > >RAID 5 at device driver (avg. resp. should be about 22.8ms) >disksim_run_simulation(): simulated 1974909 events >IOdriver Response time average: 22.861326 > >Set of disk arrays at device driver (avg. resp. should be about 33.3ms) >disksim_run_simulation(): simulated 1895981 events >Overall I/O System Response time average: 34.272035 > >Memory cache at controller (avg. resp. should be about 23.0ms) >disksim_run_simulation(): simulated 1396705 events >IOdriver Response time average: 24.651367 > >Cache device managed at controller (avg. resp. should be about 27.4ms) >disksim_run_simulation(): simulated 2123707 events >IOdriver Response time average: 28.939379 > >Simpledisk instead of original model (avg. resp. should be about 13.5ms) >disksim_run_simulation(): simulated 377469 events >IOdriver Response time average: 13.711596 >3 different disks >(avg. resps. should be about 10.951) >disksim_run_simulation(): simulated 1044524 events >IOdriver Response time average: 10.937386 >3 disks on separate controllers >(avg. resps. should be about 10.951) >disksim_run_simulation(): simulated 1044524 events >IOdriver Response time average: 10.937386 > >Note: this validates HP trace input, *not* the corresponding traced disk >HP srt trace input (avg. resp should be about 48.8ms) >disksim_run_simulation(): simulated 3422024 events >IOdriver Response time average: 48.786646 > >ASCII input (avg. resp should be about 13.8ms) >disksim_run_simulation(): simulated 10754770 events >IOdriver Response time average: 13.766948 > >syssim: externally-controlled DiskSim (avg. resp should be about 8.9ms) >response time: n=1000 average=8.894719 std. deviation=2.116511 >IOdriver Response time average: 8.894719 > > > >[root at smile valid]# ./memsvalid > >G1 MEMS Device - synthetic workload (avg. resp. should be about 2.00 ms) >disksim_run_simulation(): simulated 891324 events >IOdriver Response time average: 2.004587 > >G2 MEMS Device - synthetic workload (avg. resp. should be about 1.18 ms) >disksim_run_simulation(): simulated 800016 events >IOdriver Response time average: 1.178837 > >G3 MEMS Device - synthetic workload (avg. resp. should be about 0.87 ms) >*** assertion failed: in disksim_loadparams() (disksim_loadparams.c:86): >disksim->parfile != NULL: mems.g3.parv >./memsvalid: line 15: 11021 gave up ../src/disksim mems.g3.parv >mems.g3.nospring.outv ascii 0 1 mems0 "Spring constant factor" 0.00 > >G1 MEMS Device with 75% springs - synthetic workload (avg. resp. should be >about 2.01 ms) >disksim_run_simulation(): simulated 891263 events >IOdriver Response time average: 2.013940 > >G2 MEMS Device with 75% springs - synthetic workload (avg. resp. should be >about 1.16 ms) >disksim_run_simulation(): simulated 799997 events >IOdriver Response time average: 1.158043 > >G3 MEMS Device with 75% springs - synthetic workload (avg. resp. should be >about 0.85 ms) >*** assertion failed: in disksim_loadparams() (disksim_loadparams.c:86): >disksim->parfile != NULL: mems.g3.parv >./memsvalid: line 30: 11027 gave up ../src/disksim mems.g3.parv >mems.g3.outv ascii 0 1 mems0 "Spring constant factor" 0.75 > >_______________________________________________ >Disksim-users mailing list >Disksim-users at ece.cmu.edu >https://sos.ece.cmu.edu/mailman/listinfo/disksim-users From raysmile at gmail.com Sat Jun 14 19:44:45 2008 From: raysmile at gmail.com (ray smile) Date: Sat, 14 Jun 2008 18:44:45 -0500 Subject: [Disksim-users] Fwd: I can't find mems.g3.parv in disksim-4.0/valid In-Reply-To: References: Message-ID: ---------- Forwarded message ---------- From: Yinliang Yue Date: Fri, Jun 13, 2008 at 11:42 PM Subject: I can't find mems.g3.parv in disksim-4.0/valid To: raysmile at gmail.com Hi, When I run *memsvalid*, I can't find *mems.g3.parv* in disksim-4.0/valid, Could anyone tell me where *mems.g3.parv* is ? Thanks! Yinliang Yue -- ----------------------------------- -------------- next part -------------- An HTML attachment was scrubbed... URL: From vchs62003 at yahoo.com.cn Tue Jun 17 06:50:11 2008 From: vchs62003 at yahoo.com.cn (xm yang) Date: Tue, 17 Jun 2008 18:50:11 +0800 (CST) Subject: [Disksim-users] why is the number of handled request to be short always 1 less than issued request? Message-ID: <207944.23378.qm@web15204.mail.cnb.yahoo.com> hi, when the valaue of Probability of time-critical request is set to 1.0, disksim report that the number of handled request is same to the configured value of the parametre, Number of I/O requests to generate, in parv file. however, when changed it's value to 0.0, disksim always report that the number of handled request is to be short 1 less than issued request. why? following are disksim reported: OVERALL I/O SYSTEM STATISTICS ----------------------------- Overall I/O System Total Requests handled: 49151 ...... Overall I/O System Non-Critical Writes: 49151 0.999980 Overall I/O System Number of reads: 0 0.000000 Overall I/O System Number of writes: 49152 1.000000 Overall I/O System Sequential reads: 0 0.000000 0.000000 Overall I/O System Sequential writes: 24679 0.502096 0.502096 ...... SYSTEM-LEVEL LOGORG STATISTICS ------------------------------ System Logical Organization #0 System logorg #0 Number of requests: 49152 System logorg #0 Number of read requests: 0 0.000000 System logorg #0 Number of accesses: 49152 System logorg #0 Number of read accesses: 0 0.000000 ...... System logorg #0 disk Number of buffer accesses: 49151 ...... IODRIVER STATISTICS ------------------- IOdriver Total Requests handled: 49151 ...... IOdriver Number of writes: 49152 1.000000 ...... DISK STATISTICS ...... Disk Number of buffer accesses: 49151 ...... --------------------------------- ???????????? -------------- next part -------------- An HTML attachment was scrubbed... URL: From cdirik at umd.edu Tue Jun 17 13:23:17 2008 From: cdirik at umd.edu (Cagdas Dirik) Date: Tue, 17 Jun 2008 13:23:17 -0400 (EDT) Subject: [Disksim-users] why is the number of handled request to be Message-ID: <20080617132317.DGA91659@po1.mail.umd.edu> I have not tested with synthetically generated traces much, but from what i have tested with real traces, DiskSim stops simulating when it reads the last request from the file BEFORE it processes the request. This is probably due to internal timing, that, if the last request is issued at time X, then simulation should have continued for X + last_request_service_time, which is an unknown. Therefore DiskSim will stop at time X, which is a known value, after it received the last request before it handled the request. This is my guess. Thanks, Cagdas ---- Original message ---- >Date: Tue, 17 Jun 2008 12:00:07 -0400 (EDT) >From: disksim-users-request at ece.cmu.edu >Subject: Disksim-users Digest, Vol 33, Issue 4 >To: disksim-users at ece.cmu.edu > >Send Disksim-users mailing list submissions to > disksim-users at ece.cmu.edu > >To subscribe or unsubscribe via the World Wide Web, visit > https://sos.ece.cmu.edu/mailman/listinfo/disksim-users >or, via email, send a message with subject or body 'help' to > disksim-users-request at ece.cmu.edu > >You can reach the person managing the list at > disksim-users-owner at ece.cmu.edu > >When replying, please edit your Subject line so it is more specific >than "Re: Contents of Disksim-users digest..." > > >Today's Topics: > > 1. why is the number of handled request to be short always 1 > less than issued request? (xm yang) > > >---------------------------------------------------------------------- > >Message: 1 >Date: Tue, 17 Jun 2008 18:50:11 +0800 (CST) >From: xm yang >Subject: [Disksim-users] why is the number of handled request to be > short always 1 less than issued request? >To: disksim-users at ece.cmu.edu >Message-ID: <207944.23378.qm at web15204.mail.cnb.yahoo.com> >Content-Type: text/plain; charset="gb2312" > >hi, >when the valaue of Probability of time-critical request is set to 1.0, disksim report that the number of handled request is same to the configured value of the parametre, Number of I/O requests to generate, in parv file. however, when changed it's value to 0.0, disksim always report that the number of handled request is to be short 1 less than issued request. why? > >following are disksim reported: > >OVERALL I/O SYSTEM STATISTICS >----------------------------- > >Overall I/O System Total Requests handled: 49151 >...... >Overall I/O System Non-Critical Writes: 49151 0.999980 >Overall I/O System Number of reads: 0 0.000000 >Overall I/O System Number of writes: 49152 1.000000 >Overall I/O System Sequential reads: 0 0.000000 0.000000 >Overall I/O System Sequential writes: 24679 0.502096 0.502096 >...... >SYSTEM-LEVEL LOGORG STATISTICS >------------------------------ > >System Logical Organization #0 >System logorg #0 Number of requests: 49152 >System logorg #0 Number of read requests: 0 0.000000 >System logorg #0 Number of accesses: 49152 >System logorg #0 Number of read accesses: 0 0.000000 >...... >System logorg #0 disk Number of buffer accesses: 49151 >...... >IODRIVER STATISTICS >------------------- > >IOdriver Total Requests handled: 49151 >...... >IOdriver Number of writes: 49152 1.000000 >...... >DISK STATISTICS >...... >Disk Number of buffer accesses: 49151 >...... > > > >--------------------------------- > ???????????????????????? >-------------- next part -------------- >An HTML attachment was scrubbed... >URL: http://sos.ece.cmu.edu/pipermail/disksim-users/attachments/20080617/9c54cf05/attachment-0001.html > >------------------------------ > >_______________________________________________ >Disksim-users mailing list >Disksim-users at ece.cmu.edu >https://sos.ece.cmu.edu/mailman/listinfo/disksim-users > > >End of Disksim-users Digest, Vol 33, Issue 4 >******************************************** From schindjr at ece.cmu.edu Tue Jun 24 13:06:36 2008 From: schindjr at ece.cmu.edu (Jiri Schindler) Date: Tue, 24 Jun 2008 13:06:36 -0400 Subject: [Disksim-users] Fwd: I can't find mems.g3.parv in disksim-4.0/valid In-Reply-To: References: Message-ID: <8BA66E1D-718B-45EB-BADE-E10D729B16FB@ece.cmu.edu> Hello, the updated version of the DiskSim 4.0 tar balls, posted yesterday on the website, now includes the missing mems.g3.parv file. Jiri Schindler On Jun 14, 2008, at 7:44 PM, ray smile wrote: > > > ---------- Forwarded message ---------- > From: Yinliang Yue > Date: Fri, Jun 13, 2008 at 11:42 PM > Subject: I can't find mems.g3.parv in disksim-4.0/valid > To: raysmile at gmail.com > > > Hi, > > When I run memsvalid, I can't find mems.g3.parv in disksim-4.0/ > valid, Could anyone tell me where mems.g3.parv is ? > > Thanks! > > Yinliang Yue > > > > -- > ----------------------------------- > _______________________________________________ > Disksim-users mailing list > Disksim-users at ece.cmu.edu > https://sos.ece.cmu.edu/mailman/listinfo/disksim-users -------------- next part -------------- An HTML attachment was scrubbed... URL: From schindjr at ece.cmu.edu Tue Jun 24 13:07:56 2008 From: schindjr at ece.cmu.edu (Jiri Schindler) Date: Tue, 24 Jun 2008 13:07:56 -0400 Subject: [Disksim-users] DiskSim 4.0 - runvalid - fail - results do NOT match expected In-Reply-To: <20080608174628.DFT51156@po1.mail.umd.edu> References: <20080608174628.DFT51156@po1.mail.umd.edu> Message-ID: Hello, we have fixed the runvalid mismatches and updated the DiskSim 4.0 tar balls on the web site. By mistake, I included the runvalid values for previous disksim releases rather than values calibrated for DiskSim 4.0. The discrepancy for some of the runs is caused by the order in which modules such as diskmodel are loaded when disksim is started and the resulting change to the random generator seed. Jiri Schindler On Jun 8, 2008, at 5:46 PM, Cagdas Dirik wrote: > Hi, > > Like Lei, when I runvalid DiskSim 4.0, I am getting mismatching > results. I have > tried this on a OS 10.4 with gcc version 4.0.1. Also tried it on a > linux box and > got the same mismatching outputs. > > Any ideas on what the problem might be? > > Thanks, > > Cagdas > > Hi, > > I compiled the source code of disksim 4.0 successfully on my virtual > machine > (Fedora Core 4), and ran runvalid and memsvalid files under the > valid directory. > > The results are as follows, and it seems that, > 1) for runvalid, some simulation results are not the same as > expected, please > see the results marked in red. > 2) for memsvalid, it looks like there is no mems.g3.parv file under > the valid > directory. > > Thanks, > > Lei > > > [root at smile valid]# ./runvalid > > These results represent actual drive validation experiments > > QUANTUM_QM39100TD-SW (rms should be about 0.378) > disksim_run_simulation(): simulated 1090389 events > rms = 0.377952 > > SEAGATE_ST32171W (rms should be about 0.349) > disksim_run_simulation(): simulated 1585710 events > rms = 0.347570 > > SEAGATE_ST34501N (rms should be about 0.318) > disksim_run_simulation(): simulated 1591859 events > rms = 0.317972 > > SEAGATE_ST39102LW (rms should be about 0.107) > disksim_run_simulation(): simulated 883217 events > rms = 0.106906 > > IBM_DNES-309170W (rms should be about 0.135) > disksim_run_simulation(): simulated 2690642 events > rms = 0.135884 > > QUANTUM_TORNADO (rms should be about 0.159) > disksim_run_simulation(): simulated 1094910 events > rms = 0.267721 > > HP_C2247_validate (rms should be about 0.090) > disksim_run_simulation(): simulated 829410 events > rms = 0.089931 > > HP_C3323_validate (rms should be about 0.306) > disksim_run_simulation(): simulated 1088847 events > rms = 0.305653 > > HP_C2490_validate (rms should be about 0.254) > disksim_run_simulation(): simulated 1068808 events > rms = 0.253762 > > DEC_RZ26_validate (rms should be about 0.438) > *** warning: ignoring hpl parameters for non-hpl seek type. > disksim_run_simulation(): simulated 707184 events > rms = 0.340367 > > The remaining tests exercise other DiskSim components > > Open synthetic workload (avg. resp. should be about 47.3ms) > disksim_run_simulation(): simulated 1044524 events > IOdriver Response time average: 10.937386 > > Closed synthetic workload (avg. resp. should be about 87.6ms) > disksim_run_simulation(): simulated 571678 events > IOdriver Response time average: 87.819135 > > Mixed synthetic workload (avg. resp. should be about 24.4ms) > disksim_run_simulation(): simulated 1313881 events > IOdriver Response time average: 22.086628 > > RAID 5 at device driver (avg. resp. should be about 22.8ms) > disksim_run_simulation(): simulated 1974909 events > IOdriver Response time average: 22.861326 > > Set of disk arrays at device driver (avg. resp. should be about > 33.3ms) > disksim_run_simulation(): simulated 1895981 events > Overall I/O System Response time average: 34.272035 > > Memory cache at controller (avg. resp. should be about 23.0ms) > disksim_run_simulation(): simulated 1396705 events > IOdriver Response time average: 24.651367 > > Cache device managed at controller (avg. resp. should be about 27.4ms) > disksim_run_simulation(): simulated 2123707 events > IOdriver Response time average: 28.939379 > > Simpledisk instead of original model (avg. resp. should be about > 13.5ms) > disksim_run_simulation(): simulated 377469 events > IOdriver Response time average: 13.711596 > 3 different disks > (avg. resps. should be about 10.951) > disksim_run_simulation(): simulated 1044524 events > IOdriver Response time average: 10.937386 > 3 disks on separate controllers > (avg. resps. should be about 10.951) > disksim_run_simulation(): simulated 1044524 events > IOdriver Response time average: 10.937386 > > Note: this validates HP trace input, *not* the corresponding traced > disk > HP srt trace input (avg. resp should be about 48.8ms) > disksim_run_simulation(): simulated 3422024 events > IOdriver Response time average: 48.786646 > > ASCII input (avg. resp should be about 13.8ms) > disksim_run_simulation(): simulated 10754770 events > IOdriver Response time average: 13.766948 > > syssim: externally-controlled DiskSim (avg. resp should be about > 8.9ms) > response time: n=1000 average=8.894719 std. deviation=2.116511 > IOdriver Response time average: 8.894719 > > > > [root at smile valid]# ./memsvalid > > G1 MEMS Device - synthetic workload (avg. resp. should be about 2.00 > ms) > disksim_run_simulation(): simulated 891324 events > IOdriver Response time average: 2.004587 > > G2 MEMS Device - synthetic workload (avg. resp. should be about 1.18 > ms) > disksim_run_simulation(): simulated 800016 events > IOdriver Response time average: 1.178837 > > G3 MEMS Device - synthetic workload (avg. resp. should be about 0.87 > ms) > *** assertion failed: in disksim_loadparams() (disksim_loadparams.c: > 86): > disksim->parfile != NULL: mems.g3.parv > ./memsvalid: line 15: 11021 gave up ../src/disksim > mems.g3.parv > mems.g3.nospring.outv ascii 0 1 mems0 "Spring constant factor" 0.00 > > G1 MEMS Device with 75% springs - synthetic workload (avg. resp. > should be > about 2.01 ms) > disksim_run_simulation(): simulated 891263 events > IOdriver Response time average: 2.013940 > > G2 MEMS Device with 75% springs - synthetic workload (avg. resp. > should be > about 1.16 ms) > disksim_run_simulation(): simulated 799997 events > IOdriver Response time average: 1.158043 > > G3 MEMS Device with 75% springs - synthetic workload (avg. resp. > should be > about 0.85 ms) > *** assertion failed: in disksim_loadparams() (disksim_loadparams.c: > 86): > disksim->parfile != NULL: mems.g3.parv > ./memsvalid: line 30: 11027 gave up ../src/disksim > mems.g3.parv > mems.g3.outv ascii 0 1 mems0 "Spring constant factor" 0.75 > > _______________________________________________ > Disksim-users mailing list > Disksim-users at ece.cmu.edu > https://sos.ece.cmu.edu/mailman/listinfo/disksim-users