For those interested in the horse's mouth:
http://www.marvell.com/products/storage/storage_system_solutions/sata_controllers_pc_consumer/6_gbs_sata_raid_controller_88se91xx_product_brief.pdf
 
Marvell's pdf on this controller states AHCI support. Yet EVGA says no? Um, gotchya.

EVGA_JacobF
Hi,

We already looked into this a while ago with Marvell and determined that the limitation is a PCIE bandwidth interface limitation with using high speed SSD's in RAID0. This is not limited to EVGA. Also unfortunately there is no workaround.

Hi Jakob and thanks for your reply.
 
However, i find your reply a little vague to say the least.
 
Obviously, these limitations are not only in RAID-setups as my single drive is underperforming using the Marvell-controller. Also, it only works in IDE-mode and EVGA-Support claims AHCI is not supported by Marvell and a post above says that AHCI is supported, so what is going on???
 
What is EVGA's message to those who purchased one of your Mobo's with the Marvell-controller expecting full SATA 3/6 support?

perptt

EVGA_JacobF
Hi,

We already looked into this a while ago with Marvell and determined that the limitation is a PCIE bandwidth interface limitation with using high speed SSD's in RAID0. This is not limited to EVGA. Also unfortunately there is no workaround.

Hi Jakob and thanks for your reply.

However, i find your reply a little vague to say the least.

Obviously, these limitations are not only in RAID-setups as my single drive is underperforming using the Marvell-controller. Also, it only works in IDE-mode and EVGA-Support claims AHCI is not supported by Marvell and a post above says that AHCI is supported, so what is going on???

What is EVGA's message to those who purchased one of your Mobo's with the Marvell-controller expecting full SATA 3/6 support?

 
For this I would need some more details, I am not aware of any limitations on single drive setups, can you send me a PM with the drive you are using and the performance you are seeing.
 
Secondly, we are in the process of testing a new firmware version that may improve a bit of the performance, or hopefully also resolve the IDE bug.
 
Lastly I would like to mention that just like onboard graphics, onboard sound, etc. Sometimes unfortunately there are limitations when using onboard controllers, a dedicated card provides the full performance.

ebednar11

If this is all true, I will be requesting a refund today on my Classified 3. There is no denotation on EVGA's product info that states the Marvell controller works only in IDE. This is complete BS. Done with this company. Can't update RAID ROM, 580 multi monitor is borked, Marvell controller is defective. Just done.
Instead of addressing their issues, they release new hardware then market it's new features to sweep things under the rug. No thanks...Bye EVGA

Anyone interested in class action, PM me

 
I agree with you totally BUT I think our first port of call would be to submit a support ticket requesting a refund as the product is not fit for the purposes as advertised and sold to us.
 
This will be my course of action forthwith. If i get nowhere with this route I will indeed be talking to my lawyer to start proceedings with eVGA (I have done this before with other companies for the same reasons and been successful).

EVGA_JacobF

Topboy007

So maybe we could ask eVGA directly right now.

TO ALL EVGA STAFF READING THIS PLEASE REPLY - 

Question 1 - Are eVGA or Marvell aware of slow performance issues with the Marvell sata III controller when used in multiple SSD disk RAID solutions?

Question 2 - Many users are reporting performance issues as stated in Question 1. Can eVGA please post some technical details about what if anything is being done to resolve these issues?

Question 3 - Could eVGA confirm or deny that the Marvell sata III controller in question has a design flaw that is limiting performance as stated in question 1?


I do not think we can spell this out in english any easier WE WANT ANSWERS PLEASE.

I will send this post link to eVGA staff via private message stright away and lets see what if any reply we get. Going by the replies we have had in the bios 78 thread it will be "we are looking into this"

Hi,

We already looked into this a while ago with Marvell and determined that the limitation is a PCIE bandwidth interface limitation with using high speed SSD's in RAID0. This is not limited to EVGA. Also unfortunately there is no workaround, it is a chipset limitation.

In that case i will be requesting a full refund for my E770 motherboard as nowhere in the documentation, website, or advertised media was there any notices about limitations with this controller and motherboard.
 
If these are known limitations then it should have been advertised for me to see as a potential buyer before i purchased this board. As far as I am aware to this date there are no notices on the products website or in the products documentation about "limitations under certain conditions when using the Marvell SATA III controller".
 
If these design flaws have only just come to light thru investigation then eVGA should be doing what Intel are doing with Sandy Bridge and either offer a full refund OR a replacement when the design flaw has been rectified.

ebednar11

Goodluck topboy. I already called EVGA tech support. No refund. Submitted support ticket. No response. They told me to return to newegg. It's outside 30 days but called newegg anyhow. They told me to deal with EVGA. Not worth the attorney fees, my board is already out, photographed, and boxed. It's going up on eBay. I will never do business with this company again. Nor will my company, customers, or friends. Jacob only hinted at a FW to defuse this momentarily as they have a new x58 board up their sleeves. This thread is 5 months old. What a joke.

 
Have raised a support ticket about my motherboard which is already an RMA board and am on the phone to US Customer service right now.
 
If EVGA refund me i will use them again.
 
If they do not then i will be taking the same action as you ebednar11
 
EDIT.....
 
Well that was a very very nice experiance.
 
Matt in EVGA customer service has agreed to try and resolve my issue and will be talking to the UK offices on Monday to either get a refund OR if a NEW model board is available with the NEW revision Marvell controller.
 
Either way its good news and EVGA has just re-installed my confidence with them 

wish my experience went like that...funny how different they were...

if they do refund or replace your board that would be what's called a precedent, and evga will be obligated to do so for every other customer interested in a refund/replacement...

I am amazed and dissapointed by this whole experience.
 
Clearly there are problems and limitations to the Marvell SATA 3/6-controller spec'd on Evga-Mobo's, that's a fact and admitted by EVGA-staff in this post. There is no mentioning in the marketing, product spec or user manual that full SATA3/6 performance will not be achieved, neither in general or in RAID-setup using "high-performance SSD's". What a surprise, the SATA3/6 opens up for performance improvements over SATA2, again i am amazed that EVGA/Marvell did not expect that.
 
EVGA obviously is blaming all this on Marvell. i can understand that but again, i bought a EVGA-mobo and it is covered by EVGA-warranty.
 
i am quite confident that if Toyota (only as an expample as they have had their problems ^^) specified, marketed and sold a car with 150 HP but was only able to deliver 100 HP, they would have either made a recall or offered the consumer another product. Same thing if you buy a TV that was supposed to deliver full HD but only reached HD resolutio
 
But not EVGA or Marvell, no they tell you "we found limitations, sorry" and "This is not our mistake" and basically that's that, live with it.
 
I will give this a couple of more days and see what happens, next week at the most. If we still have no resolution i will file a complaint about this the consumer agency in my country and the ECC, European Consumer Centre. I have printed this forum-thread, the Spec-PDF of my board and the user manual to follow the complaint.