SSD roundup: Vertex 3 vs M4 vs C300 vs 510 vs 320 vs x25-M vs F120 vs Falcon II

[Likvid]

I have tested both Crucial M4 128GB and Corsair Force GT 120GB on my Lenovo Thinkpad W510, the laptop only have SATA2 but strangely enough the Crucial got better results in AS SSD, look below.

First with MSAHCI driver



Second with IASTOR driver

[Anvil]

What do you find strange with these results?

The m4 doesn't care if the data is incompressible or not, the Force GT does. (the SF controller uses compression internally)

I'll see if I can do a few runs on my W510 but the results does look OK, the somewhat low 4K speeds are typical for the chipset. (power saving)

[Likvid]

What do you find strange with these results?

(power saving)

The Force GT should perform better, not this bad, compressed or not.

[Computurd]

The Force GT should perform better, not this bad, compressed or not

do you have ACHI enabled? Problem with SF performance is the smoke and mirrors.

[Anvil]

@CT
He's using the iaStor driver and that would mean AHCI.

The Force GT should perform better, not this bad, compressed or not.

I'm asking again, what exactly should be better, sequential's, 4K QD1 or QD64, read or write?

[deathman20]

Ah the grand thing about sandforce controllers. The company never spells out what the drive should do with compressed files both read and write.

[Likvid]

@CT

I'm asking again, what exactly should be better, sequential's, 4K QD1 or QD64, read or write?

Not specific, overall performance should be better with Corsair, you read the tests first on the net?

I am returning the Force GT anyway as i experience stutters and holdups.

[Anvil]

I've got those SSD's, that's why I'm asking.

I think you are expecting to much, you are running the test on a laptop with well known performance issues.

Performance wise the GT = Vertex 3

If you are looking for better performance on the 120GB you have to look at the Vertex 3 MAX IO or the Patriot Wildfire, both are using Toggle mode NAND.

Anyways, I'll try my GT on my W510 later tonight but I don't expect to improve on your results, I'll throw in the AS SSD result for a standard PC as well.

[Anvil]

OK, this is the GT 120GB in the UltraBay of the W510.

So, it's almost spot on what you are getting and this is using the Maximum Performance Profile.


Asus P8P67 Deluxe to the right

These are the first runs on each computer.

as-ssd-bench Corsair Force GT 7.14.2011 10-57-43 PM.png as-ssd-bench Corsair Force GT 7.14.2011 9-35-47 PM.png

ATTO_Force_GT_W510.PNG atto_p8p67.PNG

[Eggcake]

I have a small question regarding the Intel 320 series, I just put it here. I'm not sure whether someone is able to give me an explanation:

The Intel 320 is available in the following capacities available to the user [GB]:
40, 80, 120, 160, 300, 600

The following NAND capacities are available for the drive [GiB]:
40, 80, 120, 176, 320, 640

This results in the following spare areas:
40-120GB: ~6.8%
160GB: 15.3%
300-600GB: ~12.7%


Main source is Anandtech (last SSD round up and test of the 300GB 320). Now, I'm not sure if the above is correct, but if it's true, can someone explain to me why the 160GB 320 has so much more spare area then the others? I mean..it's about nearly 2.5x as much as the smaller drives which is pretty impressive...

Edit: Sorry, I can't link directly to the articles at the moment as I can't reach anandtech.com

[johnw]

I don't think your figures for on-board flash memory are all correct. But the answer to your question lies in the fact that the Intel 320 series use a RAID-4 like XOR parity, and they have extra flash devoted to the parity data.

[Eggcake]

Okay, that answers part of my question (although I don't understand why they need more with higher capacities (ratio wise)) and why the 160GB has so much more spare area than the rest. And I'm pretty sure my data is correct.

Link to the Anandtech article:

http://www.anandtech.com/show/4421/t...ge-ssd-roundup
176GiB raw NAND capacity for the 160GB Intel 320 (15.3% spare area)

http://www.anandtech.com/show/4244/i...d-320-review/2
320GiB raw NAND capacity for the 320GB Intel 320 (12.7% spare area)

And I don't have a source but I'm 99% sure that the 80GB has 10x8GiB = 80GiB, the 40GB has 5x8GiB=40GiB and the 120GB has 5x8GiB+5x16GiB=120GiB, resulting in 6.8% spare area.

Edit: Okay, I may be wrong. Actually the 80GB seems to have 9x8GiB and 1x16GiB:
http://en.expreview.com/2011/03/23/w...b/15569.html/2

I can see 9x"105217" and 1x"110417". The second one is 16GiB according to Anand.
Okay, that's news to me

[johnw]

Good luck with that 99% pretty sure. Looks like it is the 1 in 100 time to me.

Would you be willing to back that 99% sure with a bet? I'm talking about the 40, 80, 120 models that you said you were 99% sure about. How about you give me only 50 to 1 odds on the bet, as if you were only 98% sure? I'll pay you $10 if your are right, and in the 1% chance that you are wrong, you pay me $500?

Or are you not really 99% sure? Maybe not even 50% sure? Maybe just wrong?

[Da_maniaC]

Okay so i just did some AS SSD benchmarks on my Intel 510 120GB RAID0 setup.
I think i've seen ppl with better results and i'm a bit dazzled as to what might be the cause (my reads wont go near 1 GB/sec for example).
Im using:
Intel Core i7 2600K
Gigabyte Z68X-UD7-B3
G.Skill Ripjaws-X 2 x 4 GB PC12800 CL7
Seasonic X-850.

Here are the screens:
Controller Cache disabled:


Controller Cache enabled:


Any thoughts?
(I thought i'd seen ppl get 920-960 MB/sec on this kind of setup).

[johnw]

Do you have an AS-SSD screenshot for a single Intel 510 on your system? Or two, if by some odd chance they were different?

[Anvil]

Okay so i just did some AS SSD benchmarks on my Intel 510 120GB RAID0 setup.
I think i've seen ppl with better results and i'm a bit dazzled as to what might be the cause (my reads wont go near 1 GB/sec for example).

A single 510 does about 415-425MB/s sequentially so you can't expect miracles, you won't get more than 2x the performance of a single drive.

~850MB/s is pretty good for two of those drives imho.

[Computurd]

my reads wont go near 1 GB/sec for examp

they may be using controllers with cache?
those are excellent results!

[Da_maniaC]

A single 510 does about 415-425MB/s sequentially so you can't expect miracles, you won't get more than 2x the performance of a single drive.

~850MB/s is pretty good for two of those drives imho.

Thanks for all the replies guys.
It does indeed look like you are able to hit 910 MB/sec but only on random occasions (probably a spike in the benchmark).
I hit it last night while doing the benchmark on an empty partition of the drive.

I will bench a single drive sometime this week (because i want those results also, for a review i am making.

[koc]

Hi guys,
I bought crucial m4 256gb , and i installed win7 and other programs smoothly , and these are some benchmarks

1.JPG

But i have freezing issue , i followed this guide , but still the same problem ? any solutions plz.

http://forum.crucial.com/t5/Solid-St...-me/td-p/38766

Thank you

[Timur]

Update the firmware to 0002. Fixes LPM issues and thus for many people freezing, but not for everyone. Fixed freezing in OS X for me.

[koc]

Update the firmware to 0002. Fixes LPM issues and thus for many people freezing, but not for everyone. Fixed freezing in OS X for me.

I have a bad experience with SSD firmware update , i sent 2x vertex and one G.Skill SSDs to RMA because of updating the firmware , and i don't want this happen again.
Is it easy to update the crucial FW ?
Is there any guide how to update ?
Thank you

[koc]

Well, I downloaded the 0002 firmware and then flashed the m4 with the new firmware ,
every think works fine , and no issue until now .
i hope its work good now

[RealTelstar]

Koc, pls keep us updated cause i plan to get the same drive (but 128gb).

[deathman20]

I have a 2x 128Gig M4's. Had to update the firmware on a single drive to Rev 0002 as one already came with it. They've been running in Raid0 now for hmm 2 weeks about love it. Works flawlessly.

[Da_maniaC]

The latest benchmarks on my new RAID0 array (2 x Intel 510 SSD 120GB). Im happy.