Hi....here are my results....HD TACH and HD TUNE should be a lot higher. ATTO from .5 to 16 are too low...
To see what the result should be here is a good review with the results. http://hothardware.com/Articles/Fusionio-vs-Intel-X25M-SSD-RAID-Grudge-Match/?page=4 http://hothardware.com/Articles/Fusionio-vs-Intel-X25M-SSD-RAID-Grudge-Match/?page=4
Using the same controller as me ICH10R SB Sata controller.
My OS is Windows 64bits. Core I7 920 and Aus P6T deluxe mb.
Any suggestions ???
I just went from a single non-RAID0 160GB G2 drive to a pair of 160's RAID0 as well (tonight!) and Iam getting fairly similar measures. Using Crystal I get about 504 MB/sec sequential read and 210 MB/sec sequential write -which is twice what I was getting with only the single SSD, and I can feel the improved performance at boot up and other operations. I then teusing HD Tune and like similar measures - 278 MB/sec seq read. I don't know why HDD Tune shows only 278 MB/sec for seq read. When I was testing just a single SSD using HD Tune ir was slightly lower than Crystal Disk, but I expect with two SSDs in RAID0 it would be much closer to Crystal Disk's readings (maybe a little lower). I also then installed the Intel ICH10R raid driver on my Win 7 64 and P6T board, retested and get pretty much the same results. I am not sure we have a problem - could be just that HD Tune can't correctly measure the RAID SSDs properly?
I think the reviewer is using a controller card since his tests were done last year. I emailed him directly, i will post his answer. Or either some benchmarks don't recognize software raid (ICH10R)
Could be; I just installed and ran HD Tach and got about the same results as you did. I am also running an i920 @ 3.8GHz on a P6T MOB. I tested the SSD under both Win7 64 and Vista 64 - same results.
Message was edited by: Firstlight&# 13; &# 13; I re-readthe test in the link and it does say &# 13; &# 13; "...note that we performed all of our SSD RAID testing with the Intel X25-M drives on an Intel X58 chipset-based motherboard via its ICH10R Southbridge SATA controller. This controller offered peak RAID 0 performance versus even the hardware-based RAID controllers we had in the lab for testing.?"&# 13; &# 13; So I don't think the controller/driver is the issue.&# 13; &# 13;
I think I solved the mystery. When using HD Tune its default block size out of the box is 64K; I believe the defult for Crystal Disk is 128K block size (I formatted my X25-M 160's using 128KB block size - so I'd rather test using 128K with each test tool).
In HD Tune you have to go to Preferences and change the block size. When I changed it to 128 I got seq read of 430 MB/sec, which is what I expected (I've noticed HD Tune seems to be consistently more conservative than Crystal). When I upped the block size to 512 and then to 1024, the seq read went up to 460 MB/sec and just shy of 500 MB/sec repectively.
Now for HDTach I don't know how to set the block size - It shows only 2 choices, a short 8K bench and a long 32K bench (I assume these might be the block sizes, and if so, explains why the seq reads are maxing at 280 MB/sec.
Thank you !! I will left you with an appetizer ... I want 2 of them :-)
a read throughput speed of up to 355MB/s and a write throughput speed of up to 215MB/s
Just imagine in RAID 0 !!
Dag! SATA 6 - wonder how much those babies will cost - impressive.
For now I am a happy camper. I do a lot of Photoshop work with thousands of 12MB RAW files. So here is my real works (if unscientific) test on my i920 3.8 GHz rig:
Photoshop CS4 64-bit opened fresh in about 2 seconds. I then opened over 200 12MP full size RAW files (D300) which opened in the ACR dialog in about 2.5 seconds. Now the real test - 16-bit RAW conversion: When opening all 100 RAW NEFs (about 16MB each) from ACR to CS4 in 16-bit into CS4 edit mode (the actual RAW conversion), all 100 files were opened in 2 min 30 sec. It used 6.84GB of RAM and I could see all 8 threads evenly busy across all 4 CPU core.
Another test, those 100 RAW files amount to 1.5GB total; a straight folder copy took between 1.2 and 1.5 seconds.
HDTune and HDTach will not measure Sequential Read correctly unless Volume WriteBack Cache is turned on, and since this setting is best left off, disregard those apps when measuring sequential speeds. Real world Seq. speeds are very close to ATTO and CDM, so then you know what to use...