02-14-2012 01:26 PM
I run my single SSD in raid mode too. Mostly because I don't have AHCI but I'm just pointing that as illogical as it sounds it is a perfectly valid way to run them and Intel even added support solely for that scenario in v9.6 of their drivers. (trim for a single ssd in raid mdoe)
02-15-2012 09:07 AM
Yes you can run Single disks in RAID mode, that is how I always run my SSD's, so I can ensure the latest RAID ROM is in use as it's often a little faster that way
02-15-2012 09:31 AM
I don't see the point in RAID mode installation.That shouldn't give me better results as it's using the AHCI mode for a single drive thats not part of an array.
I don't want the latest RAID rom, 10.8 is quite good.
And that doesn't explain the poor benchmark results. If I follow your suggestion, how much performance could I win? 5%? I guess less.
But I miss 30%.
Would you please show me a comparison? AHCI vs. RAID for a single SSD. Are there any benchmarks, tests, reviews?
For now I will test with another m4 128 GB tomorrow. AHCI mode and 10.8.
If I get around 800 points with it, I will return the drive.
If I get the same results, I will return the mainboard.
02-15-2012 10:37 AM
It runs through the RAID ROM though, that's why it's sometimes faster. I saw you mentioned you wanted to stay with the current ROM your BIOS is using, that's why I didn't comment further about that.
And yes, I agree, you shouldn't have speeds that slow with 10.8 ROM, so I'm not sure what's going on here!
You can test RAID or AHCI yourself right now, without having to do a clean install. The differences are often minimal, I was just suggesting that to you because that is how I always run my disks, and I thought it might possibly help you with these issues. All you have to do is enable RAID and AHCI in the registry, then you can switch back and forth anytime you want, here's how
Enable switching between all IDE/AHCI/RAID modes by changing "Start" Values in these keys to 0 (Win 7 / Vista ONLY)
02-15-2012 01:16 PM
I just switched to RAID and as expected it didn't help.
In fact, the performance was even worse.
The Windows flower at startup managed to show up. The boot is much slower with RAID not only due to the RAID ROM, that has to load.
I benched with Prime95 @ 1 core:
02-16-2012 09:45 AM - edited 02-16-2012 09:53 AM
your numbers are not indicative of a 'bad' drive. looks like some power saving features are in place, to me, and/or a dirty drive. when you said you secure erased it, how did you go about this? encryption will slow the performance, from what i've seen (whole drive). what software/AV, etc., do you have running in the background when benching? again, your basic results look 'normal', just that your deeper q results are a bit less.
as-ssd scores heavily on the deepQ performance. you might try anvils storage utilities basic ssd bench, also an atto run.
when benching, it is also advised you run on all cores. i notice when i have turned off speedstep, and c states, and then oc'ed a little, scores go up. everyone i know that benches these drives turns off speedstep, c states, and runs under 'performance' power settings in windows.
02-16-2012 09:54 AM
Well, I get 500 points with drive A and 800 points with drive B.
Both drives are a Crucial m4 128 GB with firmware 0309.
Both drives were tested under exactly the same circumstances:
same BIOS settings (all energy saving features disabled)
same Windows x64 SP1
I secure erased with Parted Magic.
All the same except the 2 SSDs. Really! Exactly the same!
One drive gives me 500 points, the other one 800 points. Same circumstances. I don't lie.
02-16-2012 10:02 AM - edited 02-16-2012 10:06 AM
when doing the parted magic erase, do you use the basic internal erase, or the 'enhanced' option?
same port when testing?
if you feel you should rma the drive, by all means do so. i'm just curious, as your results are average, until you get to the final as-ssd score.
02-16-2012 10:27 AM - edited 02-16-2012 10:30 AM
Ya, sorry bootup will be a few seconds more for the RAID ROM to go by, I didn't think of mentioning that as it doesn't affect the SSD speeds itself
Why are you benchmarking the SSD with prim95 running, that sounds odd to me!? Just wondering?
Are your drives exactly the same? Came in the same style and shape box, same initial firmware listed on the sticker, same color grey metal casing, made in the same place (Product of xx)? I ask because I have 3, and one of them is new and came in a different box, has different firmware on the sticker of course, also has a slightly different color casing, and is also not made in the USA like my two originals.