Open Intel Matrix Storage Console
Right click on Array_0000 and Enable Hard Drives Write Cahce
Right click on each Volume and Enable Volume Write-Back Cache
Run your tests again and watch your Burst Speed
Printable View
Open Intel Matrix Storage Console
Right click on Array_0000 and Enable Hard Drives Write Cahce
Right click on each Volume and Enable Volume Write-Back Cache
Run your tests again and watch your Burst Speed
awesome tip! burst is 2051.5MB/s now!!!Quote:
Originally Posted by Raver
seqential read speed averages 160-170MB/s. peak is 230MB/s and slowest is ~100MB/s
nice!
I took the jumpers off the barracudas and enabled write back cache and here's the revised hd tune:
http://i169.photobucket.com/albums/u...dnojumpers.jpg
i enable hard drive write cache, and enable write cache back for my volume
my burst rates improved a ton, but my average reads are done. The graph is more random. Any suggestions???
I see people are still tweaking when it comes to maximum speed of the array.
But i see little people with trouble on getting a RAID0 array to work on Vista 64 Bit nowadays.
Is this correct or am i overlooking something?
I recently bought a new pc as well and will buy 2 Raptor X 150GB drives next week to put in RAID0.
I hope i didn't buy this motherboard for nothing cause i dont feel like spending another 200-300 euros on a good RAID controller.
no problems here. i have the maximus and 2 150gig raptors and its been running fine for months. the raid runs nice and fast.
74's here, but yes, no issues worth remember now. There's just mad speed and I'm ready for a few WD 640's. :D My current storage drives are woefully inadequate in their mirrored configuration.Quote:
Originally Posted by fitseries3
How's that sound? Any recommendations of array type w/
2 x 250GB Hitachi's 55MB STR, 13 RAT
3 x 640GB New WD's w/ 100MBish STR, ? RAT
2 x 74GB raptors w/ 75MB STR, 7 RAT
Look what happened after i turned Write Back Cache, On. :eek:
http://members.home.nl/djthesp/PC/As...27-04-08-2.jpg
The ICH9R really struggles with any more than 4 drives. This is all I could squeeze out of 6 drives.
http://i8.photobucket.com/albums/a11..._32kstripe.jpg
http://i8.photobucket.com/albums/a11..._32kstripe.jpg
My results with 4 drives:
http://i8.photobucket.com/albums/a11...ch_benchie.jpg
http://i8.photobucket.com/albums/a11...ecachingON.jpg
For a desktop setup thats still sick. :eek:
What do you use such I/O power for?
I am now officially in the RAID 0 Club, with my two 150GB Raptors!
I now have 'Hard Drive Write Cache Enabled' set to Yes on both of my Raptors.
Do my scores seem a tad bit low?
http://img360.imageshack.us/img360/2301/run23fx9.jpg
I am looking for a good defrag program that will run on Vista 64 Ultimate, and does a good job with RAID 0. I would like to be able to download it, try it out, then buy it if I like it. (I probably will buy it too if it's good.) ;)
They are fine. You can slice your raid smaller maybe 60-100 gb for first array to get better and constant I/O and acces time.
I use RAID 3x500gbAAKS raid 0/5. My average transrate are bout 240, the burst bout 2000 something and the acces time bout 8.6s something. It on dfi lt p35 but it the same since they use ICH9R. Now waiting for my maximus and eyeing for newer WD6400AAKS. To bad I cant screen shot since I already sold my p35.
I use PerfectDisk 2008 and it work fine. I think it vista capable but it not freeware. Maybe there try version somewhere.
Here a good matrix raid thread.
http://www.ocforums.com/showthread.php?t=467848
My setup indeed lost it snappy feeling a bit after loading the antivirus and all the windows services.
edit : that is on the first raid 0 array. But another raid 5 array got only bout 130mb/s on average transfer and acces time about 15s or something. It less than 100mb/s if the write back cache disable. Other number on the raid 5 array I already forggotten but it ok for redundancy array.
Thanks for the post! ;)
I will check out that thread.
4 X 320GB 7200rpm Seagate Sata II drives.
http://img179.imageshack.us/img179/6...chbenchfm5.png
Ya it's a Rampage Formula but really no different than my Maximus Rampage. ;)
Talonman, Maximum PC just did a comparison between Perfect Disk, Diskeeper 2008 and Auslogics Defrag. Benchmarks proved the feature free Vist degrag provided best performance. Whoa! I've always been a PD fan and paid for PD 2008 but their results really shook me up. Pick up the June issue if you can find it.
1st picture is 3x 36Raptor in Raid 0
2nd picture is 2x 500gb YS drives
I ran HD Tune on the rest of my system too, just to see how the non-RAID drives do in comparison.
(C-Drive) (2) WD 150GB 10K RPM Raptor's in RAID 0. 279GB's free after formatting.
Minimum Transfer Rate: 87.8 MB/sec
Maximum Transfer Rate: 158.8 MB/sec
Average Transfer Rate: 132.5 MB/sec
Access Time: 8.0 ms
Burst Rate: 1861.0 MB/sec
CPU Usage 2.1%
(D-Drive) Samsung 232GB's free after formatting non-RAID.
Minimum Transfer Rate: 37.9 MB/sec
Maximum Transfer Rate: 70.0 MB/sec
Average Transfer Rate: 58.7 MB/sec
Access Time: 13.8 ms
Burst Rate: 118.3 MB/sec
CPU Usage .7%
(E-Drive) Caviar 698GB's free after formatting non-RAID.
Minimum Transfer Rate: 44.9 MB/sec
Maximum Transfer Rate: 93.6 MB/sec
Average Transfer Rate: 74.5 MB/sec
Access Time: 13.8 ms
Burst Rate: 128.1 MB/sec
CPU Usage .7%
Looks like RAID smokes them pretty bad... :D
Thanks for the info WFO. I will try Auslogics Defrag if it offers a performance increase.
Update: Downloaded Auslogics Defrag, and ran fine under Vista 64bit. It is free too! Looks like a keeper to me. :up:
Now I need a backup program, that will allow me to back up my RAID array onto my Samsung 232GB drive, and that runs on Vista 64bit.
Any suggestions would be appreachiated! ;)
@Vixx
@Lestat
Weird....why you all using 1 TB of raid 0. It hurt the acces time. Cut the array to what your OS need for programs...usually below 60-100gb is enough. Then watch it fly. :D
Talonman I use Acronis. It also not free either but I got my PerfectDisk and Acronis from the torrents. :P
There are free version backups like BartPE and DriveXML...but you must do some research to use it....Acronis is user friendly.
Damn I wish my maximus and 3x6400aaks was here for screen shots.
Thanks for the info. Just my 2 Raptors are in RAID, my other 2 drives are not. :)
Rebuilding my system with the OS and Games on RAID, my first 103GB's are gone! ;)
Trivia: The backup program that comes standard with Vista does backup a RAID array to another HD. I was under the incorrect impression that it only backed up to CD or DVD. My 103GB's of data was successfully backed up onto my Samsung drive and only used 38GB's of disk space. Not too bad! There must be some good compression involved. :D
http://img211.imageshack.us/img211/5964/run25sm9.jpg
omg...already use 100gb already on your C:. Maybe I dont game that much or install all games simultaneously. Use 80Gb for years and I delete the game after I finish play it to the final boss.
Question: I was reading the RAID link posted above, and it was talking about removing your 1.5G limiting jumper off the back of the HD's for added speed. I didn't even look at the jumpers on the back of my Raptors.
Do you think I have jumpers that need to be removed on the back of my two 150GB Raptors?
I thought jumpers on the back of the drive were for determining if it was operating in Slave, or as the Main drive...
One other question. Is the reason my speed test started out around 150MB/sec, and ended around 87MB/sec is due to the size if my RAID volume, and the read/write head moving down into the slower portion of the drives?
They are 10,000RPM drives, and seem like it should beat 7,500RPM drives to me.
Never use Raptor before because it not in my price league. [~.~] Afaik MOST sata hd doesnt have jumper for master or slave except for some hd. But it just only for sata modus selection. The new Raptor should be on sata 2 therefore no jumper. Just go check.
Yep you are right bout speed test of the smaller hd portion. That is one of the main idea of matrix raid. Maybe you should slice up you raid array to two raid 0 array and put 128k for both strip. Just make it smaller for your 1 array of your OS than second one.
I think the raptor still king of the seek time. But you might check out the new WD6400AAKS or Samsung 1TB F1(high failure rate). They both has good I/O.
Thanks for the reply! ;)
Looking around the net, it appears to me that my jumpers in the factory setting is fine on Raptors. Some post said that a couple of the incorrect options would even prevent the system from booting. The other info that I read thought there was no, or very little performance difference between jumper totally removed, and the factory settings. (At least on the WD 10K RPM 150GB Raptors.)
I think I am just going to keep my jumpers in the default position for now. :up:
If anybody believes this is incorrect, please speak up.
If I ran my defrag program after the OS and all Microsoft patches were installed, then also ran it after each game I installed, dosen't the OS already have the fastest part of my array already? I am wondering if my OS is already allowing the system to boot AFAP just due to it's location on the disks, and the only thing that might look slower, is the benchmark as it covers 279GB's of shere animal storage. :p:
If I look at dividing my graph into thirds, to show how long I actually hold my 158MB/sec transfer rate, it looks like it coveres most if the first 80GB or so. I can tell by making it into a 100GB array, my minimum transfer rate would rise allot for sure, but would my actual system performance? If my data is loaded to the fastest portion of the drive with the defrag program, I might already be golden. At 50% into my array, about 140GB's worth of actual storage, I still look to be getting around 140MB/sec. I am hoping there is nothing even written onto the slowest half of the drive.
My newbie eye's are starting to appreachiate my graph more, or at least to better understand it...
The performance graph dosen't mean much if you don't know exactly the size of array your looking at.
I bet we could make a smoken 10GB 2 disk array performance graph! :D
I was simply thinking flat = better, and looking at the minimum, average, maximum, and burst transfer rates.
I was having a hard time understanding why my Raptors performance dropped off so much. Answer = I am using the entire drives in my array, and that is to be expected when doing so. I would do better to focus on how many GB's of storage, that I am holding my 'good speed' before taking any major performance drop. That is the better way to judge ones probable array speed and boot time.
I haven't read it yet, but I sure hope Auslogics Defrag program automatically selects the fastest portion of the array to write on... I gotta believe any respectable RAID defrag program would.
My lesson for a large 2 disk array is that it might be wasteful yes, but extra system performance by making it into 2 drives, say 100GB and a 197GB, probably wouldn't offer much more speed for me to be had. That's providing my OS and games were located on the fastest half of the array already. Would you agree? :)
http://img360.imageshack.us/img360/2301/run23fx9.jpg
I would say just keep it simple, less headache in the long run for a little bit of performance gain.Quote:
Originally Posted by Talonman
I have been running my 2x 74GB Raptors in Raid 0 for a long time with np. I ran the same tests that you guys have just posted and mine is pretty spot on for your results. The only problem I have with mine is it is a very dirty system that has been up and running for 8 months or so, a good clean up or reinstall would clean up the spikes.
I am unclear as to what services I would need to disable to get things faster, nor do I really want to get into all that for a small performance gain.
I have also noticed with my SSD Raid that I get different readings from each program, with Everest being what I believe to be the closest to what it should be. There also seems to be a performance hit when running Vista over XP, at least from the graph DVnation sent me on XP vs my system and Vista. I also seemed to see a bit of difference from Vista 32 vs 64, that could be my Raid card tho.
In the end I think it is hard to bench a Raid that has a system running on it, because the OS is always trying to do things while you are benching.
I will be installing another 4x Raptor's to make another Raid 0 for my ARC-1231, just to fill out all Raid slots but it will be awhile for that. I want to get the system cooled and OC'd before I do much more. I might change out Motherboards also when the DreadNaught comes out, gezz who knows Ha !