Jump to content
Corsair Community

Performance Pro firmware update?


jim21

Recommended Posts

There do not seem to be any firmware updates posted for the Performance Pro. Are they available somewhere and I am missing them?

 

Similar models from other manufacturers have posted firmware updates during the last couple months to improve performance (low QD sequential reads, garbage collection) and compatibility (eg., LSI SAS RAID cards).

 

Can we expect a firmware update for the Performance Pro in the near future?

Link to comment
Share on other sites

  • Replies 52
  • Created
  • Last Reply
There is no firmware update for that model, and they typically don't announce firmware updates before they're out.

 

Who is "they"? Who maintains the firmware for the Performance Pro?

 

Can you at least ask them whether they are currently working on a firmware update?

Link to comment
Share on other sites

This is Corsair's forum, so "they" = Corsair.

 

Yes, I got that. The problem is that Corsair is not a "they", it is a company (an "it"). I was asking who, specifically, is in charge of the Performance Pro firmware. And can he or she or they be contacted to ask them if an updated firmware is currently being worked on.

 

I understand that it may be against policy to try to give an estimated availability date, but I am only asking whether there is any work at all going on to create an improved firmware.

Link to comment
Share on other sites

So, am I to assume from the lack of a response that no one at Corsair knows if there is any work being done to update the Performance Pro firmware?

 

The alternative is to assume that someone is working on updated firmware, but no one at Corsair will admit to this. That seems perverse, so I am inclined to go with the former assumption, in which case I think my next SSD purchase dollars would be better spent on a PX-256M3P rather than a Corsair Performance Pro. At least the PX models receive updated firmware.

Link to comment
Share on other sites

At this point, there is no need for updated firmware on our Marvell based Perf Pro SSDs. Our current firmware avoided a lot of the early bugs that plagued other drives thus avoiding the need for constant firmware updates. You can't compare our drives directly to other drives as we released much later than others for this specific reason.
Link to comment
Share on other sites

I don't have anything I can post showing read speed vs. QD and block size, but you can see the difference in AS-SSD

 

Sequential read speed (AS-SSD)

 

Corsair Performance Pro 256GB: 474MB/s

 

PX-128M2Pv1.06: 482MB/s

 

PX-128M2Pv1.08: 501MB/s

 

That difference may look small, but that is because AS-SSD is using a fairly large block size. If you measure sequential read with, say, 128KiB block size, QD=1, then the difference between the old and new firmware is larger. Most fast SSDs can achieve over 500MB/s on AS-SSD for sequential read. The Performance Pro cannot.

 

The new PX firmware makes a noticeable difference. All other things being equal, I'd rather get a Performance Pro than an M3P (the PP is cheaper), but the performance difference is enough that I am probably going to get an M3P instead.

Link to comment
Share on other sites

Actually, there is a need:

 

1) Current firmware has incompatibility with many SAS RAID cards

 

2) Current firmware has poor sequential read performance at low QDs

 

1) Define "many" and identify by specific model and controller BIOS / firmware.

 

2) The fact that we have not matched competitor's firmware release schedules does not mean that we are not/have not/will not evaluate current or future firmwares. At this time, there is no update available.

Link to comment
Share on other sites

Are you really unaware of the issue with LSI SAS based RAID cards and the Performance Pro? This issue, which at its root is a problem with the Marvell controller and the reference Marvell firmware, has been all over various Internet forums for months.

 

There is a whole range of RAID cards that the Performance Pro does not work with using the current firmware. Examples include LSI 9260 and Areca 1880. Many RAID cards are LSI based, and the current firmware is incompatible with them.

 

The m4 and the PX SSDs recently got firmware updates that addressed this issue.

 

I don't understand why the Performance Pro does not have a similar update.

 

Does Corsair need to pay someone to do the firmware updates? It seems clear there is no full-time firmware person at Corsair who works on the Performance Pro.

Link to comment
Share on other sites

In the universe of SSDs, there are only a handful of SSD controllers, and SATA III controllers are even less abundant. Many manufactures use the same controller as the Performance Pro SSD, and as you noted, they have different performance. That is common with SSD using the same controller, and is not simply due to the type of NAND chips used, such as asynchronous or synchronous.

 

Unfortunately, SSD/controller performance that is determined by it's firmware is a give and take proposition. Performance increased in one area can and does cause performance elsewhere to be decreased. That type of interaction can even be seen in Intel's IRST software, when Write-back cache is enabled, write speed is greatly increased, at the cost of some read speed performance.

 

The designers of a SSD may decide that they will trade a bit of performance in one area to gain some elsewhere. That may be the case with the Performance Pro. Given reviews of it I just checked, it's sequential read speed of large files is a bit lower than the best SSDs using the same controller, but it's sequential write, 4K read, high queue depth 4K read, and write access time test results are all significantly better than the others. It's overall score in AS SSD, for example, is also better than those others.

 

I much rather have the performance increases elsewhere, and the resulting superior real-world performance, than just seeing 500 in the sequential read speed test results. Some SSD manufactures may tune their product's FW to to attract the one number, sequential read speed spec is everything crowd. The Performance Pro is not one of those products, and I would hope that it remains that way.

Link to comment
Share on other sites

Given reviews of it I just checked, it's sequential read speed of large files is a bit lower than the best SSDs using the same controller, but it's sequential write, 4K read, high queue depth 4K read, and write access time test results are all significantly better than the others. It's overall score in AS SSD, for example, is also better than those others.

 

Wrong. The PX-128M3P, with the latest firmware update, meets or exceeds the benchmarks of the 128GB Performance Pro in all categories. Specifically, the M3P beats the PP in sequential read, and the M3P matches or exceeds the PP in 4K read, high QD 4K read, and write.

 

The Performance Pro really needs a firmware update in order to be competitive with the PX models.

 

If you disagree, show me a benchmark of the Performance Pro that beats one of these in any category:

 

Here is an example of the 256GB M3P:

http://www.xtremesystems.org/forums/showthread.php?278415-New-Plextor-M3-Pro-push-Marvell-push-D&p=5064133&viewfull=1#post5064133

 

and the 128GB M3P:

http://www.xtremesystems.org/forums/showthread.php?278415-New-Plextor-M3-Pro-push-Marvell-push-D&p=5091520&viewfull=1#post5091520

Link to comment
Share on other sites

As I have already said above and in other threads, we don't announce firmwares ahead of time. We are of course aware of what Marvell has in development and what our competitors have been doing. When/if we decide to release an update, it will be posted.
Link to comment
Share on other sites

Wrong. The PX-128M3P, with the latest firmware update, meets or exceeds the benchmarks of the 128GB Performance Pro in all categories. Specifically, the M3P beats the PP in sequential read, and the M3P matches or exceeds the PP in 4K read, high QD 4K read, and write.

 

The Performance Pro really needs a firmware update in order to be competitive with the PX models.

 

If you disagree, show me a benchmark of the Performance Pro that beats one of these in any category:

 

Here is an example of the 256GB M3P:

http://www.xtremesystems.org/forums/showthread.php?278415-New-Plextor-M3-Pro-push-Marvell-push-D&p=5064133&viewfull=1#post5064133

 

and the 128GB M3P:

http://www.xtremesystems.org/forums/showthread.php?278415-New-Plextor-M3-Pro-push-Marvell-push-D&p=5091520&viewfull=1#post5091520

 

You are correct, I had not seen reviews on that new SSD, very nice.

 

But there is a problem in a direct comparison between the two, they are not using identical NAND chips. As stated in reviews and the first forum post you linked to, the Plextor uses the latest ONFI 2.0 Toggle NAND, running at 166 MT/s, vs ONFI 1.0 Toggle NAND running a 133 MT/s in the Performance Pro. That is a ~20% difference in data transfer rate.

 

That is a significant difference between these two SSDs, a basic hardware change that cannot be ignored, and makes a comparison apples to oranges. ONFI 2.0 NAND chips are very new and almost certainly not available when the P. Pro was introduced. Can a FW update compensate for a hardware difference of this kind? That would be virtually impossible.

Link to comment
Share on other sites

Can a FW update compensate for a hardware difference of this kind?

 

Yes, in this case. The M2P uses the same flash as the Performance Pro. An M2P firmware update provides the same gain in sequential read speed as the M3P has. I just don't have any links handy to similar benchmarks for the M2P, so you will have to take my word for it (or search yourself). But I have no doubt that the Performance Pro could reach 500+ MB/s sequential read speeds if it got a similar firmware update as the M2P received.

 

EDIT:

 

By the way, I don't think it makes sense to talk about ONFI toggle flash. I think ONFI is one standard, and toggle-mode is another standard. I don't think "ONFI toggle" flash exists.

Link to comment
Share on other sites

jim21 is correct in his notification re hba/roc issues with the existing fw. it's real. and another 'manufacturer' has posted up a fw that increases performance on their 'ppro' type drive to the point it is wiping the floor with our one-off fw drives. this is the best drive corsair have at the moment. maybe it is some sort of thing with sandforce/LSI (figures) that binds corsair from offering the best FW for their pro drive owners. if so, you are biting yourselves in the important bits. i know you forum guys are not ssd folks, but if you could get word to paid corsair people...unless they too have no clue re ssds...
Link to comment
Share on other sites

maybe it is some sort of thing with sandforce/LSI (figures) that binds corsair from offering the best FW for their pro drive owners.
This drive uses a Marvell controller, so Sandforce / LSI has nothing to do with it.

 

 

but if you could get word to paid corsair people
Yellowbeard works for Corsair and has already posted in this thread.
Link to comment
Share on other sites

::pirate::

 

As I have already said above and in other threads, we don't announce firmwares ahead of time. We are of course aware of what Marvell has in development and what our competitors have been doing. When/if we decide to release an update, it will be posted.
Link to comment
Share on other sites

hey jim, doorules is beating my a$$ with his updated ssds! corsair! please provide an update to help out the hba/roc guys, and us performance minded folk. the drive is named 'performance pro' after all!

 

Good luck keeping up with DooRules, he changes his SSDs more often than his underwear. He also has many more tricks up his sleeves than FW updates.

 

Are you using your SSD(s) as the OS volume? Using empty drives makes a big difference. Are your SSD(s) partitioned and formatted? It's a trick pro reviewers use, the SSD is initialized but that's it. Only good for single drives, not for RAID 0 volumes. Is your CPU overclocked over 4.5GHz, and the BCLK tweaked up as far as it will go? Can't go above ~105MHz with Sandy Bridge boards, but that buys you some extra SATA interface speed. You might know what he does with SATA drivers and Intel RST OROMs, still more magic. Do you have a PC that is dedicated to SSD testing, and not a general purpose PC?

 

That's all just the basic stuff. If you have less than six SSDs and don't have two or three PCs with the latest hardware to play with, you (and me) are simply not in the same league as he is. Benching and testing SSDs is his hobby, and if you want to play, ya gotta pay...

Link to comment
Share on other sites

You are correct, I had not seen reviews on that new SSD, very nice.

 

But there is a problem in a direct comparison between the two, they are not using identical NAND chips. As stated in reviews and the first forum post you linked to, the Plextor uses the latest ONFI 2.0 Toggle NAND, running at 166 MT/s, vs ONFI 1.0 Toggle NAND running a 133 MT/s in the Performance Pro. That is a ~20% difference in data transfer rate.

 

That is a significant difference between these two SSDs, a basic hardware change that cannot be ignored, and makes a comparison apples to oranges. ONFI 2.0 NAND chips are very new and almost certainly not available when the P. Pro was introduced. Can a FW update compensate for a hardware difference of this kind? That would be virtually impossible.

 

 

If the NAND itself is 20 percent faster, then Corsair did a stellar job with the "slower" NAND. From what I'm seeing, the performance difference between those benchmarks and some of the better Corsair Performance Pro benchmarks is more like 8-10 percent, not 20.

 

 

I jumped in this thread because of the mention of interoperability problems with LSI raid cards. I can confirm this. I have a Megaraid 9265-8i with 6 WD RE4's in Raid5 that worked perfectly until I added a pair of 128MB Corsair Performance Pro's in Raid 0 to it. They work wonderfully as long as I limit myself to only a cold boot. It is a bit of a royal pain as I have to shut down my system to change from Windows to Linux, or when I have to do updates. I have to kill the power to it if a restart is done by accident. This is pretty poor, but from what I understand, any drive using this particular Marvell chip is affected.

 

According to the thread linked on the first page though, this issue is known and it appears that Corsair may be working on a fix for it, which is good news!

 

 

 

And why isn't anyone mentioning that the PPro 128GB drives have the full 512MB of DDRIII and the competitor's product only has 256MB of DDRIII? That's what drove my decision in the first place when I bought the PPro's. More cache is always a good thing, and I would bet it makes garbage collection more efficient. Of course, it's pretty hard to do a review about long term performance as by the time one could confirm it, the product has ceased to be relevant.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.


×
×
  • Create New...