Raid0 and TRIM Support on Intel 11.5 OROM and Latest v11.5.0.1184 RST Drivers ?

[B Gates]

I'm certain that unless these people had not made many tests to check and stirred the subject up on this and other forums, we wouldn't have had an Intel statement about this (not clear at all, but even though an official word).

You are 100% correct. As someone with inside knowledge I can ASSURE you that Intel's statement is a DIRECT result of this thread and would not have been made yet if this thread didn't exist.

[Fernando]

After having "downgraded" the Intel RAID ROM module of my P8Z77-V mainboard BIOS to v11.2.0.1527 I have repeated the Win Hex test to verify, if Trim is still working within the RAID0 without using any v11.5 driver and OROM.
Here is the result with my Z77 RAID system:
Trim works on Win7 even with the combination Intel RST driver v11.2.0.1006 with Intel RAID ROM v11.2.0.1527.
That means, that the installation of the RSTe drivers v11.5 containing the additional SCSI filter driver iaStorF and the related RAID OROM v11.5 is not needed to get Trim support.
I think, that this is good news for users with an Intel Z77 system, because many of them reported about problems with the Intel RST(e) RAID drivers of the v11.5 series.

As next step I will repeat the test while running Windows 8.

[Zaxx]

Originally Posted by Dazog
So we need to OROM hack or driver hack to have Z68 support? WOW, pretty bad of intel to do this.

gives you smart guys something to have a go at

One of the popular tech/review sites needs to offer up a bounty (free ssd maybe) to whomever can pull it off first...

[Highendtoys]

One of the popular tech/review sites needs to offer up a bounty (free ssd maybe) to whomever can pull it off first...

Yes and watch their Intel samples stop showing up.

[Zaxx]

Ouch...that would suck. I imagine there will be lots of appropriately skilled folks looking into it at least. With any luck it won't be a physical controller limitation of some sort.

[Kain665]

It's not a physical limitation.

[Fernando]

Test result with Windows 8 RTM running the Intel RST driver v11.2.0.1006 and Intel RAID ROM v11.2.0.1527: Trim is active within the RAID0!
With this combination I got the best benchmark results I have ever seen under Win8. Interested visitors can see the related benchmark results within >this< post.

[bluestang]

WinHex test result on P67-UD4-B3 running on Win7 with OROM 11.5.0.1111 and RST driver 11.5.2.1001:

TRIM NOT ACTIVE within RAID0

I'll try OROM 11.5.0.1582 next.

[andressergio]

Test result with Windows 8 RTM running the Intel RST driver v11.2.0.1006 and Intel RAID ROM v11.2.0.1527: Trim is active within the RAID0!
With this combination I got the best benchmark results I have ever seen under Win8. It is a pity, that I cannot show you some screenshots.

msdn released a free trial rtm for 90 days r u using that fernando ? cauz the release preview sucked bad full of locks and freezes

[Fernando]

msdn released a free trial rtm for 90 days r u using that fernando ?

Yes, I have downloaded the Evaluation version of Windows 8 x64 Enterprise from Microsoft and done a fresh install.

[Stelios]

I confirm that TRIM works on Z77 with 11.2 RST drivers and raid rom 11.0.0.1339 . Win7 is the OS .
I did the winhex test and after deleting the file i saw the magic zeros !

Board is Asus maximus V formula with latest 804 bios

[Zaxx]

It's not a physical limitation.

Good deal...there's still hope then.

[m.oreilly]

was just looking at asus z77 boards...tempting...but i think i'll wait till there is some sort of ivy 'e' board out, unless a 'magic' orom shows up, or intel pities us non 7 series consumers...lol, yeah, right!

[B Gates]

Just installed a new MSI Z77A-GD65 I plan on using for reviews It has 11.0.0.1339 OROM on it so It should TRIM Im going to fill test it and see

[Fernando]

Very good job done by Fernando, Andres Sergio, Btester, Chispy, Medeirosdez, B. Gates and others.
I'm certain that unless these people had not made many tests to check and stirred the subject up on this and other forums, we wouldn't have had an Intel statement about this (not clear at all, but even though an official word).

You are 100% correct. As someone with inside knowledge I can ASSURE you that Intel's statement is a DIRECT result of this thread and would not have been made yet if this thread didn't exist.

If you are right, I suspect the reason why Intel would have prefered to remain silent:
They are afraid to get more criticism than acclamation for this new Intel RAID driver/OROM feature, because only users with a Z77 chipset mainboard get benefit of it. That is very bad news especially for users, who recently bought a very expensive X79 system for professional RAID usage and expected the best possible RAID support by Intel.
I believe, that the development of the X79 chipset and maybe even of the "Enterprise Edition" RST drivers was a big mistake. Windows 8 doesn't need and doesn't even use the SCSI filter driver iaStorF.sys.

[JR.]

Just installed a new MSI Z77A-GD65 I plan on using for reviews It has 11.0.0.1339 OROM on it so It should TRIM Im going to fill test it and see

Hi Bill,

I've been observing what seems to me to be 'trim' enabled behaviour with 11.0.0.1337 and an 11.x driver for some time. I'll be very interested in what your testing shows

Regds, JR

[B Gates]

Just ran a quick fill test
TRIM confirmed working with OROM 11.0.0.1339 and 11.2.0.1006 driver
specs:


clean:


25% full:


50% full:


75% full:


deleted data and run within 30 seconds of doing so:


I'm not sure I waited long enough for the drives to recover from being TRIMMED but performance is clearly restored enough to show TRIM is functioning. I short stroked the volume so I wouldn't have to fill it so much so it wouldn't take so long.

[JR.]

Bill,

What type of controller in the SSDs you are using above?

Cheers, JR

p.s. Bill if you disable trim what result do you get?

[B Gates]

SandForce

[B Gates]

here is a balls out Vantage run with (2) 120GB 520's

NON-TRIM setup

[JR.]

Hi Bill,

Ummm...

.... I think you should try the same test cycle with trim disabled. I feel that we need to exclude the possibility that the recovery you are seeing is a direct function of the sandforce controllers and not a result of trim activity.

Cheers, JR

[B Gates]

Hi Bill,

Ummm...

.... I think you should try the same test cycle with trim disabled. I feel that we need to exclude the possibility that the recovery you are seeing is a direct function of the sandforce controllers and not a result of trim activity.

Cheers, JR

already have

this is non trim enabled same drives different firmware (faster non-trim enabled firmware) and different stripes

clean:


25% full


50% full


75% full


data deleted


6 hours of garbage collection


compare data deleted to clean. recovery is very minimal because there is no trim. if there were trim it would have recovered to at least 150,000 and not be stuck in the 90,000's. then look at garbage collection incase you think sf drives have garbage collection they do not

[JR.]

Hi Bill,

Forgive me - I'm not up to speed with the Intel SF fw versions. What happens if you use the trim enabled firmware with trim disabled in Windows?

Regds, JR

[B Gates]

Stasio modded a BIOS for my MSI Z77A-GD65 that has Intel RAID ROM v11.2.0.1527 in it going to compare performance to 11.0.0.1339 tonight when I re-do my rig again.

[B Gates]

Hi Bill,

Forgive me - I'm not up to speed with the Intel SF fw versions. What happens if you use the trim enabled firmware with trim disabled in Windows?

Regds, JR

Without testing I am pretty sure TRIM will not function.