cancel
Showing results for 
Search instead for 
Did you mean: 

trim on raid 0 with x79?

idata
Esteemed Contributor III

Various tech sites such as Anandtech have an article up about the fact that with the newest RST software that Raid 0 and trim is supported.

Is this available for X79? The X79 of course uses RSTe which has a different version number to the normal RST and the articles do not mention anything about X79 at all

74 REPLIES 74

idata
Esteemed Contributor III

If they didn't use the 9182 on most X58 boards, they sure won't on Z77, they have way fewer PCIe lanes, none to spare! Only one lane per 912x chip, just like Marvell designed them.

Speaking of Marvell, a court in the US awarded a judgement against them for Intellectual Property infringement or illegal usage, for technology they supposedly used in chips that are used in millions of HDDs. The judgement is over one Billion dollars!

Nope, do not have a Vector, I'm waiting on something from the host of this forum, that have just released their own SATA 6Gb/s SSD controller, used in the new enterprise oriented DC S3700 SSD. I've been seeing sales on 520's all over lately, and I have a feeling that is not just holiday overstock.

idata
Esteemed Contributor III

Thanks, didn't see anything about Marvell until you said it. Several similar cases like this have ended up being overturned, and I hope so because if they lose that much, we won't see a driver update till 2020, ROLOL

I understand the lane, but what strikes me odd is the same was said about 1366, but it was wrong because the 9182 proved it wrong. The Z77 can have 8 LSI Sata 3 ports, 2 native and 2 Sata 2.

I just think if it had only two native and one 9182, it would be fine, considering it can have 12 sata ports and four pci-e.

I'm surprised your not going for a Samsung, they test out hot, looks like above the Intel. Personally I will never buy one, with what I have seen with there phones, appliances, TV's, laptops, from my experience they are absolute garbage, as a company. I actually can't believe they are around the top rated for speed.

If you have a vertex 4, and get a Intel, let me know if she passes the 4, I would get one.

idata
Esteemed Contributor III

parsec,

You are close enough to how it is patched. There are no side-effects. Asrock and Gigabyte are changing the Device ID for their X79 implementation directly editing the PCI Configuration space. Once it is changed they load directly the RST 11.6 OROM.

If you check the UEFI specification you will understand how Option ROMs work. It is a driver that gets loaded for the devices it is supposed to get loaded to. My method of patching evolves around changing the device identifier in the OROM itself so that it loads on the device id 2826. Otheriwise it will only load for 2822. There is no other difference in the patch. The difference between Asrock/Gigabyte implementations and mine is just that. I load the OROM for the 2826, they change 2826 to 2822 and load it as is. So there are no side-effects. If there are any bugs, they would clearly be shared on the stock OROM and driver as well. With that in mind Raja@ASUS is reporting that the OROM/driver pair is not passing ASUS quality check and that is why they don't implement it themselves. They have noticed issues with array creation and SRT on z77. I patched it because everyone was asking for it, after careful research and testing. Hence the big fat disclaimer at my thread. My patch works identically to the other existing X79 implementations.

Regarding patching Intel boards. It is possible if you backup the SPI with Intel's FPT tool patch the OROM on the backup and flash the patched backup via FPT again. (Because I don't know how to unpack .BIO and have no personal interest in researching that:)).

idata
Esteemed Contributor III

Hey Nod, can you post a link to your thread, or where I can read how to change the ID and if it can be done with HEX on a ASRock BIOS.

You comment about the X79, are you saying you have seen a mod that allows TRIM to pass on the X79 when in RAID 0, which is the same as how you have modded the ID ?

idata
Esteemed Contributor III

Thre was a link to it above but this is it: http://rog.asus.com/forum/showthread.php?26501-RIVE-3301-Patched-for-TRIM-in-RAID! http://rog.asus.com/forum/showthread.php?26501-RIVE-3301-Patched-for-TRIM-in-RAID! RIVE 3301 Patched for TRIM in RAID!

Asrock and Gigabyte have released UEFI versions that allow you to select what OROM you want loaded via the Setup menu. Their method of doing so I explained above and I also explained mine. In both cases the OROM loaded is identical as it comes to the controller code. My patched OROM just changes the device id on the OROM itself so that it gets loaded for the proper controller. That's all there is to it. The code is the same. It's a just a different way of loading it. The way I did it is much simpler than coding an EFI module to change the PCI Configuration Space registers (or inject code in another module) and attach an option to the Setup. It's a hack but the functionality is the same nonetheless and there was no reason to code something like the Asrock/Gigabyte implementation because it will eventually get done by ASUS. The only difference is that you can't change back to the 3.x OROMS without flashing the stock UEFI image. Other than that all functionality and code is the same.

You can directly rip the OROM with APTIO MMTOOL from one of my offered images in the thread. It will work for any X79/c600 board.

EDIT: Patched 11.6.0.1702 image added to the original thread as a standalone for those of you who want to mod other boards. The OROM works for all X79 boards.