Corsair Force GT 240 GB SSD Review @ [H]

FrgMstr

FrgMstr

Just Plain Mean
Staff member
Joined
May 18, 1997
Messages
55,719
Corsair Force GT 240 GB SSD Review - We are reviewing the Corsair Force GT 240GB SSD. Corsair provides enthusiasts with both sides of the SSD controller coin by offering Marvell and SandForce controlled SSDs in its product lines. Today we will take a look at the SandForce option with its SandForce SF-2282 controller paired with high-performance IMFT synchronous NAND.
 
I have the same drive, why is the 4k write speed on your benchmark THAT high compared to other reviews and even mine?

Also the drive loses a lot of write speed ( half ) if the drive is about 33% filled.
 
To boost the speed of your SSD it is important to configure your power settings correctly. Turning the C-states on your motherboard off is important, and also configuring your power settings within the OS to "maximum performance" will give you the best results :)
 
i have a few questions

1) what are the sizes of the other drives tested? ssd performance varies with the size of the drive.

2) i am assuming that all drives were connected to the intel sata3 controller on the mb. was the c300 also connected to the same controller? or was it on the sata2 controller on the mb?
 
they are all 240 for the SF, and 256 for the Marvell. all the same capacity point.
all drives were connected to the same native sata3 port. we tested the M4, not the C300.
 
I see the reviews on HArdOCP for ssd and am looking forward to the read.
My question is why the Samsung 830 drives are not in mix for the benchmark test.
Many sites show those are the drive to get for its speed and reliablility.
 
p05ta1 said:
I see the reviews on HArdOCP for ssd and am looking forward to the read.
My question is why the Samsung 830 drives are not in mix for the benchmark test.
Many sites show those are the drive to get for its speed and reliablility.
Click to expand...

We currently do not have Samsung SSD drives on hand. Hugh is a new editor and we simply do not have the resources to build out a comparison basis from the get-go, but we will be adding as we go.
 
The bottom line should read "Man, that HyperX 3K is a hell of a deal...."
 
tricky0502 said:
i know that, but the system drive was the c300. what was that connected to?
Click to expand...

It doesn't matter if a system drive is on one 6gbps port and the test drive is on the other Intel 6gbps port. They're not sharing BW between them, meaning if you're running both at 100% (which isn't the case with an OS drive anyway) you're not going to get lower results for one or both drives. That is, if I understand what you're getting at.
 
I picked up 2 of the 240GB Corsair Force GTs in mid July when Newegg had them on a ridiculous unlisted sale price of $179 each. I consider that the deal of the year and I'm very pleased with them. Currently running them in RAID 0. I had been looking around for something to replace my 3 old 60GB Vertex drives (the originals), also in RAID 0, for a while and wasn't in any kind of crunch, but when I saw that price I had to jump on it.

I think that's the lesson I learned when shopping for SSDs. The price fluxuates a lot and there's commonly sales on them so it's smart to wait and have a couple different drives in mind that you would like so if any of them reach your target price you can jump on it.

For me, the price premium demanded by the new Marvel controlled drives in the 256GB size class wasn't worth it. And since I'm running RAID, the TRIM support doesn't come into play. Why do I run RAID and not go individual so TRIM can kick in? Well, in my experience TRIM isn't quite there yet and running RAID can improve the steady-state performance of the drives significantly.
 
Since SF doesn't have TRIM at the moment -- though it should have it again soon -- sequential writes are the best way to keep a SF in shape. SFs GC works at write-time to keep everything in shape, so if you bombard the drive with random writes and want to get performance back, sequential writes are the only way to do that -- except for an SE.
 
SSD prices seem to be rebounding a little bit, or at least are no longer in free-fall, but in general it is a competitive market. It does seem like you pay for brand name a lot of the time...although often you can get both.

Main two things to look at in a SSD are its controller and its NAND type (size, type, and unit allocation). When those match up, you often peek at other factors, like upgrade kits and warranties. That's relevant here as the HyperX 3K is only rated for 3000 P/E cycles, while many drives are 5000.

Other than that, these SF drives are pretty much fungible, except with regard to firmware. Not to much custom work going on like you have with the Vertex 4, however, but it still encourages people to lean towards certain vendors over others.
 
Last edited:
Hugh_Briggs said:
To boost the speed of your SSD it is important to configure your power settings correctly. Turning the C-states on your motherboard off is important, and also configuring your power settings within the OS to "maximum performance" will give you the best results :)
Click to expand...

This is true, a popular tool to assist with this is ThrottleStop.
 
Glad to see more ssd benchmarks , makes me want one as i still got slow sata1 drives.
 
ckryan said:
It doesn't matter if a system drive is on one 6gbps port and the test drive is on the other Intel 6gbps port. They're not sharing BW between them, meaning if you're running both at 100% (which isn't the case with an OS drive anyway) you're not going to get lower results for one or both drives. That is, if I understand what you're getting at.
Click to expand...

you are right, my info was a bit off. it is the sata2 that share their bandwidth. raid0 any 4 ssds and you will be capped at 750MB/s.
 
Maxx said:
That's relevant here as the HyperX 3K is only rated for 3000 P/E cycles, while most normal drives are 5000.
Click to expand...

The majority of SSDs are 3K rated. That is actually the base specification of 25nm NAND. 5K is from the binning process, and only Intel does this. It is Intels 'private' NAND in a sense, the previous HyperX had the 5K NAND, which is why it is special.
 
tricky0502 said:
you are right, my info was a bit off. it is the sata2 that share their bandwidth. raid0 any 4 ssds and you will be capped at 750MB/s.
Click to expand...

I don't know if SB and onward even do that anymore. I thought that was an issue with ICH10R and before.
 
man everyone wants to see a Samsung. They are a major player, so i might just send them an email and see what i can work out :)
 
You guys need to shop around a bit before posting prices. Newegg has that Performance Pro WAY cheaper than Amazon...

http://www.newegg.com/Product/Product.aspx?Item=N82E16820233227

I have 2 Performance Pro SSDs in RAID 0 on a SATA III chipset. I benchmark a little lower than Astral's in AS SSD. But I monitor my SSD/HDD usage on my 2nd monitor as I am running games, watching videos, listening to music. I NEVER see more than about 148MB/s. I fail to find a real world application that gets used everyday that will actually utilize these speeds. Even booting my PC, Windows doesn't load that much faster than it does with a single SSD installed.
 
Hexen525 said:
I NEVER see more than about 148MB/s.
Click to expand...

This is true, there are very few instances that you will use the bandwidth that is provided outside of large file transfers. there is even a cap imposed with large file transfers by the OS and the process that it uses to transfer files (has to hit DRAM first, be checked for validity once written, etc.)

This is why we focus on low QD testing in our targeted testing. This shows the true performance that can be expected under realistic use patterns. very low QD!

I do not focus on boot times, simply because any SSD will boot fast. There are limitations of the OS in the manner that it requests information from the storage solution that hold up the process. there is a definite point where it simply isn't possible to go faster, regardless of solution.
All in all, the amount of time required to boot is so finite compared to the amount of time that the user is actually working with the computer. a few seconds to boot...then the computer is used for hours. I am more focused on what happens during those hours of use personally, and that is where RAID shines.
It is also where some solutions will show their ugly side too. Steady state can become a real factor in overall OS snappiness over time, especially with heavy multitasking. RAID does a good bit of mitigating that performance drop.
 
Last edited:
While RAID may drop some performance issues, it rasies others. Namely the lack of TRIM support or a garbage collector. Thus, my SSD RAID setup slows down faster over time than a single SSD would (or so I've been told, never ran a single SSD as my OS drive yet).

And I got to looking at my own AS SSD benchies and mine aren't really all that close to Astral's. If this article paints my drives as faster, then why do his smoke mine?

as-ssd-benchAMD20StripeR210201212-48-50AM.png
 
Last edited:
Hexen525 said:
While RAID may drop some performance issues, it rasies others. Namely the lack of TRIM support or a garbage collector. Thus, my SSD RAID setup slows down faster over time than a single SSD would (or so I've been told, never ran a single SSD as my OS drive yet).

And I got to looking at my own AS SSD benchies and mine aren't really all that close to Astral's. If this article paints my drives as faster, then why do his smoke mine?
Click to expand...

I have the write-back cache enabled on the Intel Rapid Storage Technology driver. That is boosting my numbers. I always have it enabled, as that's how I run my drives, and it's part of the driver, so that's how I benched it, but here's a redo with it off so you can see the difference it makes on write scores.


 
Yer still killing me!!! Hmm, gonna go dig through my BIOS and see if I missed something. Unless my AMD controller is just utter shit.
 
Well, changing even 1 setting in my BIOS pertaining to my SATA ports and the whole system doesn't see my RAID array. However, a BIOS update did yield some slight bumps in performance. Still not on the same level as you Astral.
 
<face palm> Feel like a bit of an idiot now. I have the same options on my RAID...only I am required to DL AMD's RAIDXert app to change them. New numbers are MUCH nicer than before. And a lot closer to Astral's.

as-ssd-benchAMD20StripeR86201212-29-16AM.png
 
Hexen525 said:
<face palm> Feel like a bit of an idiot now. I have the same options on my RAID...only I am required to DL AMD's RAIDXert app to change them. New numbers are MUCH nicer than before. And a lot closer to Astral's.
Click to expand...

Very nice improvement.

Don't feel bad, the Intel driver is the same way. I was only able to change that setting after installing the RST manager from Intel and it isn't required for the array to function.
 
i got to ask as i have no clue on this, why are 4k' scores so low vs everything else?
 
Just picked up one of these on sale for $180. Never had an SSD before, I'm busy copying my OS and programs over to it and using my HDD for media storage and backups of the OS.

Excited to see what kind of difference I'll see in load times, etc. My backup drive died and I had to get something, and I thought I'd try this instead of a boring old external HDD.
 
Aluisious said:
Just picked up one of these on sale for $180. Never had an SSD before, I'm busy copying my OS and programs over to it and using my HDD for media storage and backups of the OS.

Excited to see what kind of difference I'll see in load times, etc. My backup drive died and I had to get something, and I thought I'd try this instead of a boring old external HDD.
Click to expand...

If you've never used an SSD before your jaw may hit the floor. It's that drastic.
 
You must log in or register to reply here.
Back
Top