Angry_Games Posted May 30, 2006 Posted May 30, 2006 I'll do ATTO later when I get home. The drives are Hitachi, 160Gb. yes maybe you can re-read my previous statement though and provide THOROUGH detail saying it is a 160GB hitachi drive tells us ZERO what stripe size? cluster? Share this post Link to post Share on other sites More sharing options...
Mdebrito Posted May 31, 2006 Posted May 31, 2006 This is a complicated question... in my opinion, the stripe/cluster ratio is a PART of it, but not necessarily all of the difference. To help better compare your results to mine, you'll need more info than what I've provided. My result above was obtained by building the stripe set at 16k stripe size and then partition/format in XP to 4k clusters. The other detail not included is that I'm not booting from the RAID array (YET), because I'm not finished testing and fully satisfied that I'm not going to have corruption hell. My data is my work and livelihood and I cannot take any chances with it. After I move my stuff to the array, I'll be making auto-backups to other drives and network storage with Ghost Enterprise. So questions I cannot answer are whether or not you are booting your RAID0 and if there are any significant differences in access time and/or firmware configuration between my ST3160812AS spindles and your 7200.9 ST-250s.... and whether or not booting from the RAID set you're testing increases, decreases or does nothing to a bench of the drive with these benchmarks. The Burst speed itself, if that is the only part of your numbers you feel is 'slow', can be effected by the configuration of your mainboard, sometimes your BIOS (what version NVRAIDBIOS), the Driver version, version of Windows, and finally by the drive firmware again. The 7200.9 drives have algorithms in the firmware that affect write configuration and sometimes the read speed to an extent, but I just don't know about how Burst is effected by firmware exactly... I'm not that 'in the know' on that side of these disk setups. Hi thanks for your help My raid is the boot drive and the strip is at 64K, to format the drive I´ve used the normal windows configuration About the firmware where can I find it ? do I need do format the drive again ? Share this post Link to post Share on other sites More sharing options...
wevsspot Posted May 31, 2006 Posted May 31, 2006 You'll have to break your array and rebuild with 16kbs stripe sizes if that's what you're after. The default Windows NTFS is already 4kbs. So there you'll get your 4/1 perfect ratio (but remember that rule of thumb applies to 2 drive RAID arrays only). However, if you're thinking man that's gonna be a pain in the butt, I can guarantee you that in real world normal day to day use you'll never know that you're at 64/4. I've done it both ways and the only way I can tell the difference is in benchmarking. So if you're after the best benchies with a 2 drive array, then 16/4 is the way to go. Share this post Link to post Share on other sites More sharing options...
Mdebrito Posted May 31, 2006 Posted May 31, 2006 I have tryed 16K once but it gave me some errors of missinfiles, so I went back to 64K Can be a hard drive problem ? I have tested on seagate tools and both passed the test can be something else ? Share this post Link to post Share on other sites More sharing options...
Mdebrito Posted June 1, 2006 Posted June 1, 2006 Here are some tests ! Just after booting second test and next tests ATTO Can anyone explain Why i have such a huge diference from ther first boot test and all others ? Share this post Link to post Share on other sites More sharing options...
BadWillHunting Posted June 1, 2006 Posted June 1, 2006 Here are some tests !------------------------------ Can anyone explain Why i have such a huge diference from ther first boot test and all others ? I note that your CPU utilization from the second test is in at 3%, half of what your initial test shows (6%)... and the burst speed is also around half.... might be related. What else do you have running in the background and how does it test if you kill all unnecessary processes in Task Manager? Kill the virus checker first, if one is running, and test again. Share this post Link to post Share on other sites More sharing options...
Mdebrito Posted June 1, 2006 Posted June 1, 2006 No virus checker running, only the drivers, creative, ATI and nvidia, and windows I´ve beem thinking can be the dual core fix path of windows ? it is no installed Share this post Link to post Share on other sites More sharing options...
wevsspot Posted June 1, 2006 Posted June 1, 2006 The two most important performance indicators pretty much stayed the same. They are your Random Access Test and Average Read. Don't even concern yourself with the Burst Test reading in HDTach. Your performance is on par for your RAID setup and the drives you are using. If you aren't getting any file corruption and your system feels snappy I'd leave everything just the way it is. Just IMHO Share this post Link to post Share on other sites More sharing options...
soundx98 Posted June 2, 2006 Posted June 2, 2006 Zoom, Zoom, Zoom 3x 250G WD2500KS 16M Cache, RAID 0 Stripe 32 Cluster 4, Read Caching DISABLED, running Win XP Pro x64 Edition on 1st 50G. Share this post Link to post Share on other sites More sharing options...
Angry_Games Posted June 2, 2006 Posted June 2, 2006 hey guys....make sure you run ATTO as per the original post's settings... Share this post Link to post Share on other sites More sharing options...
wevsspot Posted June 3, 2006 Posted June 3, 2006 SoundX and AG, I've noticed the same thing on my main rig, with read cache disabled my benchmark read speeds are higher. I'm not sure I understand why that happens, can either of you shed any light on that? Also, even though benchies are higher, for an online game server what do you think offers the best performance? Share this post Link to post Share on other sites More sharing options...
soundx98 Posted June 3, 2006 Posted June 3, 2006 I'm not the brightest bulb on the Christmas tree but I thought I read somewhere (and this may just be a figment of my very warped mind) that Read Caching Enabled and NCQ/TCQ are really designed for multi-user (server) applications where they enhance drive performance. I can tell you that both my WD Raptors and these WD2500KS drives when in RAID test much faster with them DISABLED. May want to search www.StorageReview.com for info. Both their tests of the WD740GD and their 250G Drive roundup address this point. Sometimes this stuff gets a little over my haid so I just do the OL' trial and error and compare. Does sorta make sense since Read Caching Enabled and Command Queing points towards multiple users. Maybe "Happy" or ExRoadie can offer a "smarter" answer the n00bster Share this post Link to post Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now