OCZ Technology Vertex 3 SSD 120GB and 240GB Review @ [H]

Michaelius said:
Why don't you test them in actual real world application?

I don't care about how nice transfers those drives can get in canned benchmark (where Sand Force rocks due to data compression)

Show us how much time we can save when loading games/applications (os is done in article fortunatly) by paying price premium for Sata 3 drives over sata 2 cheaper ssd.
Click to expand...

Most of those tests would be like splitting hairs with the difference being measured in milliseconds. I'm working on a large round up now that uses some homebrew benchmarks and shows not only the current crop of drives but also a couple of generations back.

The best SATA II and the new SATA III drives aren't that much different when just opening programs.

And I shouldn't have said that about the RevoD, the forum might get me into a bit of trouble sometimes:)
 
Well the things I'd like to know from storage review


One article about:
single drive vs 2 SSD raid 0 vs 3 SSD raid 0 vs big ramdisk scenarios
to know how two smaller drives work against one bigger
and raid of 2 drives vs raid of 3 drives to see at what point built in mobo raid controller becomes bottleneck
- this doesn't have to be big sample size - one representative of Sata 3 ssds (intel 510/Crucial M3, Vertex 3 or other sandforce 22xx drive) and one from Sata 2 (Intel 320 or one of first gen sandforce) would be enough.

As for software to test in normal reviews:
- os load (already done)
- go to hibernation/wake up from hibernation (i do this few times a day on my gen 2 intel)

- graphics - I don't work on those files but people who do should be able to come with some kind of demanding scenario (they have been using performance raid for work for years already)

- gaming - beetween [H] users we should be able to come up with a group of worst offenders where it comes to loading levels and or save games
On the top of my head - Total War games loading big campaign map and loading save game
Starcraft II
Witcher 1 (or 2 but first one have longer and more common loadings)

PS. That's also kinda part of my agenda to show that real world diffrences beetween 500 Mb/s and 270 Mb/s are similar to ram timings - something that looks cool on benches but it's hard to notice it IRL or to have my theory proven wrong by a pair of C4 in raid 0 :)
PS2. For me Vantage tst is just a bunch of fancy numbers which are hard to translate into real world gains
 
ChrisRam said:
Most of those tests would be like splitting hairs with the difference being measured in milliseconds. I'm working on a large round up now that uses some homebrew benchmarks and shows not only the current crop of drives but also a couple of generations back.

The best SATA II and the new SATA III drives aren't that much different when just opening programs.

And I shouldn't have said that about the RevoD, the forum might get me into a bit of trouble sometimes:)
Click to expand...

Hey Chris - on page 5, you mentioned that the Corsair P3 didn't lose much performance when filled. By "not much," how much are we talking? Is it still slower in Vantage than a filled Agility 3? And by chance do you have the same numbers on the Intel 510?
 
OCZ have removed the latest 2.09 firmware, and you can only get 2.08 now. No real explanation, but their method of updating the firmware may have had issues. They've also closed the thread for the moment, which had become the main source of information since they decided to 'reorganize' by deleteing all the other threads on the issue.
 
Last edited:
Here is the breakdown of some other drives.

Corsair Performance 3 256GB

Light Use / Empty 39154
25% 38199
50% 37015
75% 36875
Dirty / Empty (TRIM Check) 37380



Corsair Performance 3 128GB

Light Use / Empty 38598
25% 37325
50% 37896
75% 37325
Dirty / Empty (TRIM Check) 38590



Crucial m4 128GB

Light Use / Empty 71637
25% 48919
50% 47073
75% 43688
Dirty / Empty (TRIM Check) 71637



I have data on the Crucial m4 256GB, Intel 510 120GB and Intel 510 250GB but am not sitting with it in front of me. The first Crucial m4 256GB I took possession of was more of an ES drive. I was told they were the same as retail which they pretty much were except the drive had been tested and I didn't know it. When I say tested I pretty much mean the drive had been through hell and back. This made it so I couldn't do an Apples to Apples run with all of the drives tested under identical conditions. ie exact conditioning prior to testing.

As far as the Intel numbers, I'll have them for the round up article coming soon.


I'm looking at a chart now with all of the numbers in it. The Agility 3 240GB dips to 33661 at 50%. The m4 128GB dips to 47073 and the Performance 3 240GB doesn't really dip and stays around 37896.

To put this into perspective the SF-1200 Series of drives achieve around 40 - 42K in zero fill / light use and around 33K at 50% fill. I am talking 3Xnm SF-1200 drives and not the 25nm knock offs that came later. If you are still looking for a source of [ame="http://www.amazon.com/Silicon-Power-Velox-2-5-Inch-SP120GBSSDV20S25/dp/B0053ORUGU/ref=sr_1_2?ie=UTF8&qid=1309317147&sr=8-2"]Amazon.com: Silicon Power 120 GB Velox Series V20 SATA II 2.5-Inch Solid State Drive (SSD) SP120GBSSDV20S25: Electronics@@AMEPARAM@@http://ecx.images-amazon.com/images/I/41sWmfRKl1L.@@AMEPARAM@@41sWmfRKl1L[/ame]already or any day now. I hear the 120GB model will cost around 200 USD.

I know in this article I also talked a little bit about the Corsair Force 3 and said that the F3 is not suppose to do what my Agility 3 did. I just surfaced from the underground lab and can tell you it does dip to around 40K at 50% fill. It's not as bad as the Agility 3's 33K but it's still not what I'd consider a proper next gen SF drive. It is cheap, around 200 USD for the 120GB model but I sure as hell wouldn't buy a new motherboard and proc to run it.
 
Last edited:
ChrisRam said:
Here is the breakdown of some other drives.

Corsair Performance 3 256GB

Light Use / Empty 39154
25% 38199
50% 37015
75% 36875
Dirty / Empty (TRIM Check) 37380



Corsair Performance 3 128GB

Light Use / Empty 38598
25% 37325
50% 37896
75% 37325
Dirty / Empty (TRIM Check) 38590



Crucial m4 128GB

Light Use / Empty 71637
25% 48919
50% 47073
75% 43688
Dirty / Empty (TRIM Check) 71637



I have data on the Crucial m4 256GB, Intel 510 120GB and Intel 510 250GB but am not sitting with it in front of me. The first Crucial m4 256GB I took possession of was more of an ES drive. I was told they were the same as retail which they pretty much were except the drive had been tested and I didn't know it. When I say tested I pretty much mean the drive had been through hell and back. This made it so I couldn't do an Apples to Apples run with all of the drives tested under identical conditions. ie exact conditioning prior to testing.

As far as the Intel numbers, I'll have them for the round up article coming soon.


I'm looking at a chart now with all of the numbers in it. The Agility 3 240GB dips to 33661 at 50%. The m4 128GB dips to 47073 and the Performance 3 240GB doesn't really dip and stays around 37896.

To put this into perspective the SF-1200 Series of drives achieve around 40 - 42K in zero fill / light use and around 33K at 50% fill. I am talking 3Xnm SF-1200 drives and not the 25nm knock offs that came later. If you are still looking for a source of Amazon.com: Silicon Power 120 GB Velox Series V20 SATA II 2.5-Inch Solid State Drive (SSD) SP120GBSSDV20S25: Electronicsalready or any day now. I hear the 120GB model will cost around 200 USD.

I know in this article I also talked a little bit about the Corsair Force 3 and said that the F3 is not suppose to do what my Agility 3 did. I just surfaced from the underground lab and can tell you it does dip to around 40K at 50% fill. It's not as bad as the Agility 3's 33K but it's still not what I'd consider a proper next gen SF drive. It is cheap, around 200 USD for the 120GB model but I sure as hell wouldn't buy a new motherboard and proc to run it.
Click to expand...

Awesome, awesome info. Thanks so much for the reply.
 
As an Amazon Associate, HardForum may earn from qualifying purchases.
Well I`ve bought an agility 240gb for $335 and remember 50% of it is 100% of any 128gb version. I can live with the 33k dips for the price I`ve paid, actually as a OS drive I can keep it always under 50%... And still it can push burst sequential reads really high... well I`ll see when it gets here =)
 
this part is not clear where it says

"The pairing of SandForce SF-2281 and async flash just isn’t the right combination for real world performance. You can get some nice benchmark results but as soon as you fill the drive to 50% capacity the performance isn’t there."

Is it a limitation of these SSDs? Will this limitation be overcome by firmware upgrade?
 
For 335 I'd buy a 240GB Agility 3. I haven't seen that deal yet.

As far as the firmware goes, these drives will get faster. I was looking at my early benchmarks of last year's SF-1200 drives and the performance gain over 12 months was really good. The RAID numbers went up quite a bit as well.
 
ChrisRam said:
For 335 I'd buy a 240GB Agility 3. I haven't seen that deal yet.

As far as the firmware goes, these drives will get faster. I was looking at my early benchmarks of last year's SF-1200 drives and the performance gain over 12 months was really good. The RAID numbers went up quite a bit as well.
Click to expand...

I'm (understandably) leery of the SF-2281 based drives right now. But the potential performance gap is just too large to ignore. I have an Intel 510 120gb in my desktop and it's been great. Looking at something to drop in my laptop (it does have SATA III), same size/generation. Three questions:

1. I see that the Corsair P3 is practically the same drive as the Intel 510. It's also $65 cheaper. Is Corsair quality/support enough worse to justify the price gap? Always had good luck with their stuff, though it sounds like they've gone through some teething problems with SSDs (not sure how much that's a SF thing v. a Corsair thing).

2. In real life usage, is a "slow" current gen drive like the P3 really that much slower than the fastest ones (e.g. Force GT, Vertex 3 Max IOPS)?

3. Given that I had sporadic BSOD problems with the Agility 3, would it be safe to say I'd have the same problems with other SF-2281 drives like the new Force GT?

Thanks in advance!
 
Last edited:
Corsair has some of the best quality and support around. I just wrote an article last night over the Force 3 and talked about the company ethics and how they handled the recall. Corsair is the customer service benchmark in which all others should follow.

As far as speed goes, it all depends on what you do. For 95% of people the current SATA III generation SSDs are all the same. Actually, let me change that to, for 99% of people, the current range of SATA III SSDs are the same 90% of the time. The main benifit is access times and then you have that file transfer stuff that you done every once in a while. When benchmarking drives I personally don't see a one to two second different on massive tasks to be a big deal, even we were talking about a task that takes either 8 or 10 seconds. Witht that said, I personally don't think a few milliseconds is even something to question. If it takes you 5 seconds to open Crysis or 7 seconds, do you personally think this is a game changer? What if the price difference was 100 USD for the same capacity point, is that 2 seconds worth 100 USD?

If you are having BSOD issues with an A3 then you will most likely have the same issue with a Force GT right now. SF and Team SandForce partners are working to figure out what the heck is going on with the handful of systems that are having issues.

As far as the issues go, I'm stumped. I have systems here with these drives in use and not a single one has BSODed. I even installed two drives in my boys systems. It's Summer time and they are home on break, playing on the PC like 12+ hours a day when I don't kick them to the back yard. Last night they were quiet and I didn't know they were still playing and it was 2AM. Neither system BSODed yet and my boys are really hard on their systems, shutting them down with the button, leaving games running paused all night, you name it they do it.

My personal systems are using SF-2281 drives and they are fine too.This last week I've been on a Lenovo W701ds for 18 hours a day trying to get caught up on reviews and it hasn't had an issue.
 
I have been coming to this site since it was just a wee little web page. I have to admire how it has tried to be true all this time. This review reminded me of that. I love the tone Kyle and Brent use in their reviews. They provide a read that makes it seem as if you are sitting down have a discussions with them about the product. It is great. Keep up the good work!

I try to hit your links when ever I buy something from your adverts.
 
SRT8 said:
I have been coming to this site since it was just a wee little web page. I have to admire how it has tried to be true all this time. This review reminded me of that. I love the tone Kyle and Brent use in their reviews. They provide a read that makes it seem as if you are sitting down have a discussions with them about the product. It is great. Keep up the good work!

I try to hit your links when ever I buy something from your adverts.
Click to expand...

Thanks!
 
ChrisRam said:
Corsair has some of the best quality and support around. I just wrote an article last night over the Force 3 and talked about the company ethics and how they handled the recall. Corsair is the customer service benchmark in which all others should follow.

As far as speed goes, it all depends on what you do. For 95% of people the current SATA III generation SSDs are all the same. Actually, let me change that to, for 99% of people, the current range of SATA III SSDs are the same 90% of the time. The main benifit is access times and then you have that file transfer stuff that you done every once in a while. When benchmarking drives I personally don't see a one to two second different on massive tasks to be a big deal, even we were talking about a task that takes either 8 or 10 seconds. Witht that said, I personally don't think a few milliseconds is even something to question. If it takes you 5 seconds to open Crysis or 7 seconds, do you personally think this is a game changer? What if the price difference was 100 USD for the same capacity point, is that 2 seconds worth 100 USD?

If you are having BSOD issues with an A3 then you will most likely have the same issue with a Force GT right now. SF and Team SandForce partners are working to figure out what the heck is going on with the handful of systems that are having issues.

As far as the issues go, I'm stumped. I have systems here with these drives in use and not a single one has BSODed. I even installed two drives in my boys systems. It's Summer time and they are home on break, playing on the PC like 12+ hours a day when I don't kick them to the back yard. Last night they were quiet and I didn't know they were still playing and it was 2AM. Neither system BSODed yet and my boys are really hard on their systems, shutting them down with the button, leaving games running paused all night, you name it they do it.

My personal systems are using SF-2281 drives and they are fine too.This last week I've been on a Lenovo W701ds for 18 hours a day trying to get caught up on reviews and it hasn't had an issue.
Click to expand...

Once again, thanks a bunch. Picked up a P3 locally for $169 AR, slapped it in my laptop. Can't tell any subjective difference from the 510 in my desktop - you were right. It benchmarks nearly identically to the 510, too, which is to say pretty poorly compared to the A3 I used to have.

But what they say is true - benchmarks don't tell the whole story. My former Agility 3 was never consistent - sometimes it felt blazingly fast, other times it felt sluggish. I thought it was in my head so I whipped out a stopwatch to measure cold boot times. With the Agility 3, over 5 runs, I had a variability of 7 seconds (from 15s to 22s). Sometimes the Win7 boot animation wouldn't finish before bringing up the desktop; other times it would. Just tried the same totally unscientific test with the 510 - boot times only varied by ~2 seconds (14s to 16s). The P3 in my laptop (though it's not apples to apples) feels similarly consistent and very snappy.

Chris, it's worth noting that my new P3 came with the 1.1 firmware. Every other source I've seen has benchmarked the P3 with 1.0 firmware. My CDM and AS_SSD results with my 1.1 drive are consistently about 10-15% higher in random writes and reads than results I've seen on the web. (No change to sequential numbers.) Corsair does not have 1.1 available to the public, though - I wonder why?
 
I'm sure it's on the way. Yellowbeard is on vacation next week but I'll shoot an email over to the behind the scenes guys and see about a change log and inquire about a public release. I'd like to check it out at well!!!!

Yesterday I started exploring something that is going on with SandForce drives. Well most of them. I caught something in a Mushkin press release for their new SF-2281 drives. Off the top of my head it said something about they weren't going to throttle their drives.

What the throttle does is slows down the rate in which you write to the drive. Drives are rated to last for three to five years per the warranty terms. Well it seems that SandForce has allowed the manufactures this throttling feature so you don't reach the breaking point before your warranty expires. Different data types act differently, compressible and incompressible.

What you may have been hitting is this imaginary wall that slowed your SF drive down for a short period of time. It's all a bit complicated but I'll be calling SF on Tuesday to get my hands on some colorful charts and graphs that will help explain just how much BS is actually going.

The way I see it, you paid for 525MB/s read and 500MB/s write, not 500 some of the time and 100 the rest of the time. I'm already getting out my boots because what's down there is going to be thick and stink.
 
Im suprised so few people realise than Sand Force is crap ;)
Jmicron/toshiba/intel make much better controlers even if they can't put such a nice artificially increased numbers on the box.
 
I wouldn't go that far. JM and Toshiba have really poor consumer offerings. The write latency is way too high. They try to mask it with a large buffer but once the buffer fills up your in the slow lane.
 
Well it seems that SandForce has allowed the manufactures this throttling feature so you don't reach the breaking point before your warranty expires.
Click to expand...
The way I see it, you paid for 525MB/s read and 500MB/s write, not 500 some of the time and 100 the rest of the time. I'm already getting out my boots because what's down there is going to be thick and stink.
Click to expand...


well this has been known about for a long time. their lifetime throttling is old news. its been public since the first gen SF was released :) we've known it stinks for awhile.
over at OCZ and many, many other storage forums, this is very old news.


I caught something in a Mushkin press release for their new SF-2281 drives. Off the top of my head it said something about they weren't going to throttle their drives.
Click to expand...

this applies to a read throttle with teh new SF. not the writes. all SF drives are throttled. this has been confirmed by Mushkin.
 
Last edited:
"DuraClass management functionality is still active. “Unthrottled” in this context refers to write IOPS bursting up to 90,000+ but being governed down to 20,000 after a few seconds which is typical behavior with standard firmware with SF-2281. The firmware we have on the Chronos and Chronos Deluxe drives will not have that governor activated."
Click to expand...

this is the official Mushkin response to the already-asked question :)
 
Question...is the Intel sata III controller on the msi p67a-gd65 b3 mobo the same as the one used in the review (msi z68) ?
 
You must log in or register to reply here.
Back
Top