RAID0 TRIM Another Look Editorial   Page 1

Reviewer: Jon Coulter Article Publish Date: 8/27/2012

 

 

 

 

 

Being the first to publish an article on something new can be somewhat treacherous territory. RWL was as far as we know the first site to release conclusive proof that RAID 0 TRIM is now functioning. Our first article (click here) was a result of searching for proof of RAID 0 TRIM based upon the efforts of individuals searching for answers in this now famous thread (click here) at Xtreme Systems Forums. Their hard work in combination with our article and a call to Intel by Chris Ramseyer forced this brief press release from Intel: "Trim on RAID 0 for SSDs is supported in the Intel RST driver versions 11.0 and newer. Currently available for the general public on Intel’s downloads site is RST driver version 11.2 which offers TRIM support on RAID 0 compatible with MS Windows 7 OS on Intel 7 series chipsets (earlier chipsets NOT supported). Intel is also working on a future release providing support for TRIM on RAID 0 on Microsoft Windows 8 OS for Intel 7 series chipsets." Were it not for proof being provided in the thread at Xtreme we would not have pursued our own proof for RWL’s first RAID 0 TRIM article and, were it not for RWL’s first RAID 0 TRIM article, Intel would have not made the press release when they did.

 

We have been further testing RAID 0 TRIM since publishing the first article which brings me back to the opening statement of this article. Treacherous territory. We got a few things wrong in our first article that need correcting so, we are giving you an update on the facts of RAID 0 TRIM as we have come to know them.

 

Previously we thought you had to have an 11.5 series OROM as well as an 11.5 series RST driver to enable RAID 0 TRIM. This of course has proven to be incorrect. Any 11 series RST driver will pass TRIM through a RAID 0 array. We have also come to find through our testing that any 11 series OROM will also work. Let’s talk for a moment about OROM’s because this is where Intel’s statement and other articles written after ours have missed the boat. Neither Intel or another famous review site bothered to mention you must be running a BIOS that has an 11 series OROM built into it AND running an 11 series RST driver. An OROM is a built-in part of your motherboards BIOS not something that’s part of a driver. Think of the OROM as firmware for your RAID controller. Bottom line: If you want RAID 0 TRIM the first step is a BIOS with an 11 series OROM built into it.

 

We also speculated that a Z77 board may not be a requirement. We made this assumption because there is nothing hardware related that would prevent RAID 0 TRIM from working on 6 series hardware. We were wrong because while there is nothing hardware related that would prevent functioning RAID 0 TRIM on a 6 series chipset there is another factor…..INTEL. Intel has chosen to not make RAID 0 TRIM function on 6 series chipsets through restrictions built into the RST driver. So if you want RAID 0 TRIM and don’t have a Z77 board you will need to upgrade. That’s just the way it is, at least for now you must be running a 7 series chipset and no…..X79 is not a 7 series chipset. Hopefully Intel will succumb to pressure and extend RAID 0 TRIM functionality to 6 series chipsets.

 

As a way to further show we had functioning TRIM we cited CDI now displaying individual drives in an array as re-enforcing TRIM functionality. Wrong again. I had never seen CDI display drives that were part of an array before so I assumed this was a result of the RST driver passing TRIM. I’ve since come to find CDI was already doing this with version 4.6 on arrays without TRIM.

 

The same thing happened when we looked at Anvil’s SSD benchmark. Anvils offered to TRIM the array which I had not seen before but, after thinking about it I’ve decided that Anvil’s may not be providing proof of TRIM just passage of the command not execution of it. Bottom line: CDI and possibly Anvil’s are no proof of functioning RAID 0 TRIM.

 

Vantage fill testing is where we got it RIGHT on the money. Our exclusive Vantage fill testing procedure is unmatched as a real world method of validating TRIM functionality. In-fact Vantage fill testing is THE reason 5.03 firmware that restores TRIM to SandForce drives is now starting to be released. It was Vantage fill testing done by Chris Ramseyer that got SandForce to recognize there was no TRIM. Bottom line: Vantage Fill testing is the best method of real world TRIM validation and that’s the reason we do it.