Kevin OBrien

Corsair Performance Pro SSD (256GB) Review Discussion

35 posts in this topic

The Corsair Performance Pro SSD is powered by the Marvell 9174 processor, Toshiba 34nm Toggle MLC NAND and a SATA 6Gb/s interface. The Performance Pro adds to Corsair's already excellent line of SDD products that leverage perhaps the most robust set of combinations of NAND and SSD processors available with a SATA interface. In this case the Performance Pro replaces the Performance 3, Corsair's first Marvell 9174-based SSD which was announced about a year ago. A year goes a long way in the SSD world, the Performance Pro picks up a tremendous speed boost over the Performance 3, delivering read speeds of up to 515MB/s and an impressive 440MB/sec in sequential write performance, compared to 480Mb/s reads and 320 MB/s writes in the first generation model.

Read Full Review

Share this post


Link to post
Share on other sites

nice drive but the plextor seems more compelling to me. course will have to see how the plextor fairs once you are through with it. you going to perform the same tests on this(corsair) drive?

doesnt the plextor use 2x toggle nand? nvm found the answer:)

seems like the plex & this drive would be perfect for a endurability comparicion between 34 & 2x nand.

got another question, with owc being the exception, seems like the only 5 yr warranties out there are on 2x nand.not 34. everything i have read says it should be the opposite. whats up with that? ;)

also any idea wether the 128 vers has 512 or 256 cache? couldnt find out at site.

and, what is "supercharged nand"? is it just a gimmick or is there actually some real tech involved?

anybody?

Edited by mike2h

Share this post


Link to post
Share on other sites

Aww no Enterprise IOMeter test? Corsair claim they have advanced garbage collection and these are very suitable for RAID so would love to see how that transpires in the steady-state benchamrks!

Share this post


Link to post
Share on other sites

I second h4lf. I thought SR said that you were going to be performing the steady-state write test on all SSDs in the future. I certainly want to see it for every SSD. It is a lot more interesting than your fake "real-world" tests that do not write real world data (but rather easily compressible repeating patterns).

Edited by johnw42

Share this post


Link to post
Share on other sites

We were holding off on the Performance Pro enterprise testing until the M3S issue was worked out. Plextor is finally shipping a replacement and we should have those tested soon. We have preliminary numbers, just wanted to cross all the T's and dot the I's before adding those charts.

Share this post


Link to post
Share on other sites

The power consumption figures shown in the review are interesting, compared to the Plextor PX-M3S which uses the same hardware (different firmware).

Corsair: 5.8W sequential write, 3.12W sequential read, 0.86W random 4K read, 0.33W idle, 5.85W startup

Plextor: 3.58W sequential write, 1.98W sequential read, 0.77W random 4K read, 0.35W idle, 2.81W startup

Worst case, that's over 2W difference. Were both sets of readings taken with the drive connected to a SATA-III controller?

Would you be able to give readings with the drives connected to a SATA-II controller? My laptop only has a SATA-II controller and I guess the rate-limiting when connected to that would reduce the maximum power consumption of the Corsair drive. But the Plextor power consumption would probably drop too.

Also, what sort of power consumption do these SSDs have when the drive is stopped? (That is, after the computer tells the drive to stop/spin down, e.g. using hdparm -y or hdparm -Y in Linux.)

Finally, was there a typo in the article synopsis: "... Toshiba 34nm Toggle MLC NAND...". Should that read 24nm?

Edited by Donuts

Share this post


Link to post
Share on other sites

Very interesting how Marvell seems to be gaining some major ground in the SSD market. Even Intel integrated there controller in there own 510 SSD.

If you analyses the pricing on the intel 5 year warranty models, its rather obvious your paying a large premium just for a 5 year warranty. ( No Thank You! )

Corsair and Plextor seem to have a possible winner I really like the Idea of 32nm NAND on the Corsair!

What I'm not sure of is the NAND flash, 25nm or 32nm...... Now you could argue that based on specifications 32nm has higher r/w capability, BUT is it fact are just marketing BS.

For me and I'm sure most everyone else, they would prefer reliability and high r/w capability over just pure speed.

Share this post


Link to post
Share on other sites

The power consumption figures shown in the review are interesting, compared to the Plextor PX-M3S which uses the same hardware (different firmware).

Corsair: 5.8W sequential write, 3.12W sequential read, 0.86W random 4K read, 0.33W idle, 5.85W startup

Plextor: 3.58W sequential write, 1.98W sequential read, 0.77W random 4K read, 0.35W idle, 2.81W startup

...

Finally, was there a typo in the article synopsis: "... Toshiba 34nm Toggle MLC NAND...". Should that read 24nm?

That isn't a typo. This drive uses 34nm NAND while the Plextor M3S uses 25nm. However, the Plextor M2P uses 34nm NAND just like the Corsair drive and if you look at the review, http://www.storagereview.com/plextor_pxm2p_ssd_review_256gb the power consumption is very similar. So I suppose you can chalk the power difference up to the difference in flash process size.

Share this post


Link to post
Share on other sites

Competition in the comparison graphs was all Sandforce SF-2281 and Marvell controller drives; would removing some of the redundant SF-2281 drives and replacing with Samsung's 830 SSD or OCZ's Octane (with the new Indilinx controller) improve the comparison?

Or is there a subtle point of comparison that I am missing? (I think there's two SF-2281 drives with synchornous NAND in the comparison; are they redundant or is there a NAND type I'm missing? (25nm vs. 34nm? Toggle vs. some other type?))

Just curious if I'm missing the obvious. :)

And between this drive and other Marvell drives, unless price is a significant advantage, I'm thinking I'd stick to the Crucial m4...?

Share this post


Link to post
Share on other sites

I am curious if somebody familiar with the details of the below reviews can clear up a minor (descrepancy) between what I perceive as 2 similar hardware being benchmarked.

In the following reviews an OCZ Vertex 3 240GB is benchmarked:

http://www.storagereview.com/corsair_performance_pro_ssd_256gb_review

http://www.storagereview.com/intel_ssd_320_raid_review

A majority of the benchmark results between the two reviews for the same OCZ Vertex 3 hardware are within what one can be assumed, the acceptable margin of variations.

However IOps 4k random Transfer READ/WRITE 4k:

intel_320_160gb_4k_randomtransfer_iops_4k.png

corsair_performance_pro_256gb_repeatingrandom_4k_randomtransfer_iops_4k.png

And to me, More worrying and more noticeable:

corsair_performance_pro_256gb_repeatingrandom_iops_database_4k.png

intel_320_160gb_iops_database_4k.png

the OCZ 3 240 is pulling values of 2600-2800 versus 4400 and 3100 respectively for DB

File server 35000/23000 versus 19000/21000

Those large variances are a bit of concern to me and I'm curious what was not disclosed/what did I miss about the setup that would cause that large of a discrepancy.

Share this post


Link to post
Share on other sites

Keep in mind advertising dollars are there bread and butter. So if your expecting a non bias review you should find a independent reviewer.

Share this post


Link to post
Share on other sites

useful reply..

wanderer- those are 2 different vers of the vertex 3- vanilla vs max iops. outside of that u will have to wait for somebody that actually knows something to give u an intelligent response:)

Edited by mike2h

Share this post


Link to post
Share on other sites

I am curious if somebody familiar with the details of the below reviews can clear up a minor (descrepancy) between what I perceive as 2 similar hardware being benchmarked.

In the following reviews an OCZ Vertex 3 240GB is benchmarked:

http://www.storagere...sd_256gb_review

http://www.storagere...320_raid_review

A majority of the benchmark results between the two reviews for the same OCZ Vertex 3 hardware are within what one can be assumed, the acceptable margin of variations.

However IOps 4k random Transfer READ/WRITE 4k:

And to me, More worrying and more noticeable:

the OCZ 3 240 is pulling values of 2600-2800 versus 4400 and 3100 respectively for DB

File server 35000/23000 versus 19000/21000

Those large variances are a bit of concern to me and I'm curious what was not disclosed/what did I miss about the setup that would cause that large of a discrepancy.

You are looking at values between test rigs, old models labeled 3.0Gb/s and 6.0Gb/s were tested on the old rig based around a Dell XPS 9000. Intel ICH10R and a LSI MegaRAID 9260 for SATA 6.0. With the onboard cache it had a ramping affect in some areas and the stripe size caused a smoothing affect in others. Charts without that designation are on the newer Sandy Bridge testing platform that we currently use to this day. That is the main difference. We went back and tried to update all the older review comparisons, but some might have slipped by.

Share this post


Link to post
Share on other sites

Keep in mind advertising dollars are there bread and butter. So if your expecting a non bias review you should find a independent reviewer.

I would hope that our track record with reviews would be enough proof, but we don't lean in any direction outside of what our benchmark results come back with.

On the Corsair Performance Pro front, we now have burst/steady figures. Working up a chart now but until its finished here are the values in text:

Burst: 271.09 MB/s, 69401 IOPS, 0.46ms

Steady: 29.14 MB/s, 7460 IOPS, 4.29ms

Share this post


Link to post
Share on other sites

You are looking at values between test rigs, old models labeled 3.0Gb/s and 6.0Gb/s were tested on the old rig based around a Dell XPS 9000. Intel ICH10R and a LSI MegaRAID 9260 for SATA 6.0. With the onboard cache it had a ramping affect in some areas and the stripe size caused a smoothing affect in others. Charts without that designation are on the newer Sandy Bridge testing platform that we currently use to this day. That is the main difference. We went back and tried to update all the older review comparisons, but some might have slipped by.

Thanks, that's what I was looking for. Would be nice if each review had a table listing the pertinent hardware setup that everything was tested on. Seemed to be the one thing that was missing from most of the benchmark.

And Thanks to everybody for putting in the effort that goes into making these thorough and fairly standardized reviews.

Edited by The_Wanderer

Share this post


Link to post
Share on other sites

Keep in mind advertising dollars are there bread and butter. So if your expecting a non bias review you should find a independent reviewer.

In my honest opinion and my limited 10 years of reading and comparing benchmark reviews this is about the only benchmark review site where I could consistently find multiple reviews that had a common set of hardware and the numbers were mostly inline/comparable.

That was why when I noticed that two/three discrepancies I wanted to know what was the root.

Share this post


Link to post
Share on other sites

On the Corsair Performance Pro front, we now have burst/steady figures. Working up a chart now but until its finished here are the values in text:

Burst: 271.09 MB/s, 69401 IOPS, 0.46ms

Steady: 29.14 MB/s, 7460 IOPS, 4.29ms

Thanks for the update, the numbers look good. Very similar to the Plextor M2P, I guess the firmware must be near identical like the hardware? :)

Share this post


Link to post
Share on other sites

I would hope that our track record with reviews would be enough proof, but we don't lean in any direction outside of what our benchmark results come back with.

On the Corsair Performance Pro front, we now have burst/steady figures. Working up a chart now but until its finished here are the values in text:

Burst: 271.09 MB/s, 69401 IOPS, 0.46ms

Steady: 29.14 MB/s, 7460 IOPS, 4.29ms

Not knocking your benchmarking, it would be nice though if you showed some clarity on the testing platform used and reliability issues. The review of the PX-M2P you stated that

Plextor isn't really pitching this as a client SSD; in fact they'd prefer we look at this like an enterprise drive.

Could you explan in more detail why consumers should take this view when its listed and marketed as a consumer SSD? If you believe this is a note worthy statement shouldn't the Corsair Performance Pro be listed as a Enterprise Drive also?

If anyone wants a real world samples of user experiences, Newegg reviews are a great place to start!
Edited by Ricky_005

Share this post


Link to post
Share on other sites

i actually use newegg reviews to. problem is seperating the real from the fake. and other silly shyt. it is a good place to start, dont know about great.

i usually start in places like this that deal with specific hardware, find parts that look interesting(&get get more technical info) then go to places like newegg & amazon-wich imo is way worse than newegg for company shills- and see what 'real' users are saying.

Share this post


Link to post
Share on other sites

i actually use newegg reviews to. problem is seperating the real from the fake. and other silly shyt. it is a good place to start, dont know about great.

i usually start in places like this that deal with specific hardware, find parts that look interesting(&get get more technical info) then go to places like newegg & amazon-wich imo is way worse than newegg for company shills- and see what 'real' users are saying.

If a product on Newegg has lets say for example purposes 100+ reviews. I find the sample rate adequate and amazingly accurate in most situations. What is troubling is when 5 eggs dips below 60% and 1 egg approaching 20% or greater is a sign of serious problems with the product. If market/product expectations aren't met, this is the pattern you'll find in the Newegg statistical Egg rating!

Edited by Ricky_005

Share this post


Link to post
Share on other sites

Could you explain in more detail why consumers should take this view when its listed and marketed as a consumer SSD? If you believe this is a note worthy statement shouldn't the Corsair Performance Pro be listed as a Enterprise Drive also?

It really depends on how the manufacturer positions the drive in both price and marketing. Many SSDs have similar hardware, using unique firmware to customize it per the application. Plextor wanted to push the M2P into the business/enterprise space while Corsair is pushing the Performance Pro at the consumer space.

Share this post


Link to post
Share on other sites

It really depends on how the manufacturer positions the drive in both price and marketing. Many SSDs have similar hardware, using unique firmware to customize it per the application. Plextor wanted to push the M2P into the business/enterprise space while Corsair is pushing the Performance Pro at the consumer space.

1. Do you believe Plextor has the in-house talent to develop enterprise grade firmware?

2. Do you think the M2P has enterprise class firmware now?

3. Do you believe its possible to create a reliable enterprise drive based on consumer hardware by installing enterprise class firmware?

Edited by Ricky_005

Share this post


Link to post
Share on other sites

1. We haven't worked with Plextor that much, but that remains to be seen at the business/enterprise level. I know most will offer fast turn around times for defective drives and other perks for business support. Sadly since we are outside of that channel its hard to talk about what places might offer without experiencing it. Are they to the level of Micron, LSI, Intel, etc? Not in full enterprise but for small business they might have compelling options.

2. Not counting the secure erase issue, it held up pretty darn will in steady state at least in 4K. We didn't put it through out expanded steady state benchmarks which you will see in a review upcoming shortly, but most of those offerings are SLC so these guys are competing in the heavy read market, instead of heavy 24/7 write.

3. Quick answer might be to look at Intel (SSD 320 vs SSD 710), long answer would be its hard to tell at our level outside of power regulation methods. Many drives offer similar hardware with different NAND/firmware configurations for different scenarios. Enterprise drives generally offer better power regulation internally to handle the server environment, which obviously the Plextor drive lacks. This isn't unheard-of though as Kingston pushes some drives in the business market for workstations that are setup in a similar way.

Share this post


Link to post
Share on other sites

You are looking at values between test rigs, old models labeled 3.0Gb/s and 6.0Gb/s were tested on the old rig based around a Dell XPS 9000. Intel ICH10R and a LSI MegaRAID 9260 for SATA 6.0. With the onboard cache it had a ramping affect in some areas and the stripe size caused a smoothing affect in others. Charts without that designation are on the newer Sandy Bridge testing platform that we currently use to this day. That is the main difference. We went back and tried to update all the older review comparisons, but some might have slipped by.

I'm coming to this discussion a little late, but aren't those two different SSDs? Certainly the label has one as being the Vertex 3, and the other the Vertex 3 Max IOPS, and those ARE two different SSDs. So either your label is wrong, or the poster is comparing two different models and expecting to get the same results...which is obviously a poor expectation. The Max IOPS should be faster (and it is, in SR's tests). Isn't that "the main difference"?

Another poster (mike2h) already mentioned this, but no one else seems to have noticed so I thought I'd mention it again.

Edited by johnw42

Share this post


Link to post
Share on other sites

Competition in the comparison graphs was all Sandforce SF-2281 and Marvell controller drives; would removing some of the redundant SF-2281 drives and replacing with Samsung's 830 SSD or OCZ's Octane (with the new Indilinx controller) improve the comparison?

I agree. Another problem with too many Sandforce drives is that they take up two slots (compressible and random data).

I suggest the following SSDs be included for comparison with the SSD under test in the future:

OCZ Vertex 3

Corsair Performance Pro

Crucial m4

Samsung 830

Those are all close competitors at the very top of the consumer performance spectrum, so they make good benchmarks for comparing new SSDs with. And it is a good selection of manufacturers, controllers (1 Sandforce, 2 Marvell, 1 Samsung) and flash (IMFT, Toshiba toggle, Samsung).

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!


Register a new account

Sign in

Already have an account? Sign in here.


Sign In Now