• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

OCZ Vertex 2 with 25nm NAND flash reported slower

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
i posted sanforce versus x25-m benchmarks and they are really close (64gb sandforce to 80gb x25-m G2). probably so close you'd never tell the difference in real life.

for $99 the G2 64gb microcenter special is not a bad drive - it has the best firmware there is. much better than the other sandforce drives which tend to sleep-death more often.
 
jwilliams4200 said:
You missed the context. Some people believe Sandforce's absurd claims that for the data most people are writing to the drives, the compression will be much higher than that achieved with random data like AS-SSD writes. People who believe such things will not be convinced by writing random data. They would need to write actual data that they store on their drive.
Click to expand...
Fair enough, but you can do the same thing with a memory mapped file too - that limits the sample size to a few gb but that should be good enough (and you can loop it if you need a larger filesize for exact measurement)
 
frostedflakes said:
To be fair, it's not like you've provided anything factual to refute Sandforce's claims. You're basically asking me to believe your hot air instead of theirs. :hmm:
Click to expand...

I did not provide anything factual to refute the existence of the tooth fairy either. No doubt you believe in it, too.
 
Voo said:
Fair enough, but you can do the same thing with a memory mapped file too - that limits the sample size to a few gb but that should be good enough (and you can loop it if you need a larger filesize for exact measurement)
Click to expand...

No, you are still missing the point. The only way to get a good measurement of the compression achieved for the files on a users drive is to use the actual files on a user's drive. If you only choose a few GB out of 40GB on the drive, then anyone could argue that those GB are not typical of all the data on the drive.

Anyway, most people who do that experiment on a Sandforce drive will only achieve about 90% compression factor. Sandforce's compression is just not very good.
 
I'd be nice if you guys just had the drives set-up side by side for a comparision.

I have and the OCZ 2s couldn't match the overall desktop experience of the Intel G2s or the Crucial C300 drives.

I'm a believer in specs but also value first hand experience.
 
czesiu said:
and atto shows 270?
could you also benchmark file transfer writes in windows?
Click to expand...

I tried copying a 4 GB video file from my 7200 rpm drive to my Vertex 2. It started off fast (120 MB\s) but was constantly losing speed up until it got to 38 MB\s. It maintained that speed for the duration of the file transfer.
 
Nebor said:
I tried copying a 4 GB video file from my 7200 rpm drive to my Vertex 2. It started off fast (120 MB\s) but was constantly losing speed up until it got to 38 MB\s. It maintained that speed for the duration of the file transfer.
Click to expand...

thank you
(please also add info about drive model - 60gb?)

and more questions... why some people get 67MB/s or 47MB/s...
http://www.ocztechnologyforum.com/f...rk-results-and-load-times&p=602217#post602217
http://www.ocztechnologyforum.com/f...GB-Performance&p=596662&viewfull=1#post596662
 
Nebor said:
Those ATTO benchmarks seem to favor OCZs drives. Yes I have the 60gb Vertex 2 E model. When I ran the ATTO benchmark I was told on the OCZ forums that my scores were fine.
Click to expand...
I'm just looking at as-ssd for the moment, as atto always shows 270... 🙂
 
frostedflakes said:
Sandforce claims a typical write amplification of 0.5 for their controllers, which means the controllers are actually pretty good at compressing data. Keep in mind that write amplification <1.0 isn't possible without compression, any amount under that is achieved by compressing what you're writing to the drive. So the typical WA of 0.5 suggests that on average what you write to the drive can be compressed by at least 50%.

I mean if you think about it, most people aren't writing video/music/other compressed stuff to an SSD. It contains the OS and program files, a lot of which is probably pretty compressible.


This is the kind of stuff you generally put on an SSD and it benefits a ton from DuraClass.
Click to expand...

By write amplification, I think they mean write amplification factor. 1+write amplification factor=actual write amplification.
 
czesiu said:
I'm just looking at as-ssd for the moment, as atto always shows 270... 🙂
Click to expand...

Ah. I don't see that this even merits much more discussion. OCZ screwed everyone. Why anyone would buy one of their inferior drives at this point is beyond me.

If mine weren't buried under a pile of cables and screwed to the bottom of the case, I'd trash it and buy an Intel drive.
 
OCZ built up a good reputation with their SSDs. They're also the company that took a risk on Sandforce drives and got them off the ground. What remains to be seen is how OCZ handles this situation. For starters they need to start providing information up front on NAND & controllers. I'd like to see this from all companies.

OCZ has also worked to devlop high speed drives using PCI Express and their new IBIS drives. Because SATA 6GBps is already getting maxed out.
 
frostedflakes said:
There's nothing yet to suggest the drives have lower lifetime. 25nm MLC can handle fewer P/E cycles, which is why the Sandforce controller requires more over-provisioning. The extra over-provisioning is supposed to compensate for the lower P/E cycles, higher error rate, etc. of the 25nm flash.

That's the trade-off, Sandforce has said many times that this is the direction they think SSDs are heading, more over-provisioning will become necessary. But in the end you still should come out way ahead. For example the shrink from 34nm to 25nm provides twice the storage for the same cost, but at the expense of 4% more over-provisioning. So in the end, for the same price you're still getting a lot more space with 25nm (this is assuming that IMFT and SSD manufacturers decide to pass the savings onto consumers, which hasn't happened yet, but hopefully will start to soon).
Click to expand...

This is an irony since shinking will likely lead to lower prices, but also lesser space. So is SSD going to be much cheaper at the end of the day? :hmm:
 
Dark Shroud said:
OCZ built up a good reputation with their SSDs. They're also the company that took a risk on Sandforce drives and got them off the ground. What remains to be seen is how OCZ handles this situation. For starters they need to start providing information up front on NAND & controllers. I'd like to see this from all companies.

OCZ has also worked to devlop high speed drives using PCI Express and their new IBIS drives. Because SATA 6GBps is already getting maxed out.
Click to expand...

I agree. They innovate, and most others follow. I like the way they do things, though they could have done better in this case. Anyway, my Vertex 2 is also crawling at the moment. Seq read at approx 172, write at 52... Even startup seem to take much longer than my Intel 160GB now... :\
 
my microcenter G2 (adata s599) is solid. sadly my other sandforce 120gb takes about 3-5 tries on a cold boot to catch. with the latest firmware. lame.

never had issues with intel controllers/samsung/toshiba. they might not be the fastest but they have reliability you can count on
 
Dark Shroud said:
OCZ built up a good reputation with their SSDs. They're also the company that took a risk on Sandforce drives and got them off the ground. What remains to be seen is how OCZ handles this situation. For starters they need to start providing information up front on NAND & controllers. I'd like to see this from all companies.
Click to expand...
yeah surprised by this move as in past they have differentiated the slower NAND i.e. Agility series. But not now ? At least a new model number/revision and packaging sticker is needed.

43% less write speed on 25nm 64Gbit NAND version http://www.ocztechnologyforum.com/f...data-and-0-1-fill-partially-compressible-data 😱
 
Based on the new drives performance in real world, it takes a 100 to 120GB drive to match the write speed of an Intel G2 80GB. :|
 
sandforce is a liability. nobody has perfected timing with them. for business i skip on sandforce always. i'd rather have indilinx or samsung
 
Emulex said:
sandforce is a liability. nobody has perfected timing with them. for business i skip on sandforce always. i'd rather have indilinx or samsung
Click to expand...
Well the old Vertex drives were perfectly fine, I don't see how that's a SF issue
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top