June 24, 2013 at 4:08 am
I had format the disk to 64KB and perform a test case using CrystalDiskMark. The result is under expectation. Why i cant get the 20% performance improvement if compare to the default disk format 4096 bytes?
June 24, 2013 at 4:25 am
Why were you expecting a 20% performance improvement?
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
June 24, 2013 at 5:22 am
I was expecting at least 10-20% improvement. However the result is worst than the 4096bytes. It is degraded the performance. please see my test result using SQLIOSIm. What's wrong for my disk format 64kb? Do i need to set anything to improve SQL server performance?
FileNameTargetIODuration_msRunningAverageIODuration_ms (Low values are good)NumberOfTimesIOThrottled (Low numbers are better)"IORequestBlocks
(High numbers are good)""Reads
(High numbers are good)""ScatterReads
(High numbers are good)""Writes
(High numbers are good)""GatherWrites
(High numbers are good)"TotalIOTime_ms
D:\sqliosim.mdx 10036749331014905202099842141218676013
F:\sqliosim.mdx 1004753916911969181036932064245129383
F:\sqliosim.ldx 1003629800128320320531
D:\sqliosim.ldx 100320800129520300425
June 24, 2013 at 5:30 am
SQLIOSim is not a performance testing tool. It's a stress tester designed to hammer an IO subsystem and see if it throws errors. Perhaps try an IO performance test tool?
Gail Shaw
Microsoft Certified Master: SQL Server, MVP, M.Sc (Comp Sci)
SQL In The Wild: Discussions on DB performance with occasional diversions into recoverability
June 24, 2013 at 5:34 am
I also using SQLIO to test it, below is my result. Any other IO performance tool that i can use to test the disk 64KB?
D:\testfile.dat (4096 bytes)F:\testfile.dat (64 kilobytes)
File size : 20GBFile size : 20GB
Test will last 120 secondsTest will last 120 seconds
Random Writes TestRandom Writes Test
IOs/sec: 184.48IOs/sec: 179.49
MBs/sec: 1.44MBs/sec: 1.40
latency metrics:latency metrics:
Min_Latency(ms): 1Min_Latency(ms): 1
Avg_Latency(ms): 346Avg_Latency(ms): 355
Max_Latency(ms): 542Max_Latency(ms): 528
Random Reads TestRandom Reads Test
IOs/sec: 193.39IOs/sec: 175.43
MBs/sec: 1.51MBs/sec: 1.37
latency metrics:latency metrics:
Min_Latency(ms): 6Min_Latency(ms): 25
Avg_Latency(ms): 329Avg_Latency(ms): 363
Max_Latency(ms): 2274Max_Latency(ms): 2614
Sequential Writes TestSequential Writes Test
IOs/sec: 1409.14IOs/sec: 1024.48
MBs/sec: 88.07MBs/sec: 64.03
latency metrics:latency metrics:
Min_Latency(ms): 0Min_Latency(ms): 1
Avg_Latency(ms): 45Avg_Latency(ms): 61
Max_Latency(ms): 181Max_Latency(ms): 221
Sequential Reads TestSequential Reads Test
IOs/sec: 1445.88IOs/sec: 1073.58
MBs/sec: 90.36MBs/sec: 67.09
latency metrics:latency metrics:
Min_Latency(ms): 6Min_Latency(ms): 13
Avg_Latency(ms): 43Avg_Latency(ms): 59
Max_Latency(ms): 3198Max_Latency(ms): 3100
Viewing 5 posts - 1 through 4 (of 4 total)
You must be logged in to reply to this topic. Login to reply