Thank you for trying that Martin. If you are not understandably fed up by now I would like to know which SSD's you are using.

Hi CPL0! Thank you for helping me. The SSD's are 2x 64GB Samsung MMCRE64G, also known as PB22-J.
I bought this SSDs in 2008. At this time Trim was not supported. But I think I made a firmwareupdate to Ver. VBM19C1Q which supports RAID a few years ago, but I'm not sure. Is there any possibility to check the Firmware Version within the RAID array ? In the RAID BIOS I only see the serial numbers, and in the device manager -> drives -> Hardware id's I only see IDE\DiskBOOT1.0.00__ ... so nothing interesting there...

@Martin_Q

Use latest BIOS 0802. In BIOS -> Advanced -> Onboard Devices Configuration disable LAN Boot Rom (NOT Gigabit LAN) and also IDE/eSATA Boot Rom from Marvell. If you don't have anything connected to Marvell ports, disable all together Marvell IDE/eSATA. Try again with 10.1.0.1008. You had too many devices with boot capabilities in limited shadowing space.

Zitat von Martin_Q im Beitrag #17

---

Is there any possibility to check the Firmware Version within the RAID array ?

---

I already use the 0802 Bios. Boot from LAN is disabled, also the 2nd LAN port. The Marvell Storage Controller must be enabled, because I use all 8 SATA ports from this Mainboard (6x Intel, 2x Marvell) , and the PATA port is connected to an optical drive. All PCI/PCIe Slots except of the one PCIe x4 is in use... I know, there is no shadowing space left... but I can't do without one of the expansion cards. Before I do this I can easier live without Trim...

@CPL0: thank you, i will look for the firmwareversionnumber when I am at home.

@Martin_Q
I do admit about being zealously short in words, in an attempt to not confuse or bore you with too much details. Here is the long story. Note that I am not an hardware engineer and I might post things not entirely true.

The system scans for devices with boot capabilities (Boot Rom) to add to the boot sequence, looking especially for PCI/PCIe devices. Your BIOS is designed to fill a certain buffer (my guess in memory, shadowing) with Boot Roms and when it overflows, it dumps one Boot Rom. Use Aida64 or HWiNFO and look in PCI Devices for a device with Bus - Device - Function of 0 / 0 / 4 or 4 /0 /0. In order to bypass this error you either receive a BIOS upgrade with an increase buffer for Boot Roms, or disable certain Boot Rom. Please not that disabling a Boot Rom does not necessarily disables that controller. For instance, disabling "IDE/eSATA Boot Rom" under Marvell is NOT the same with disabling Marvell IDE/eSATA. The first one is the Autoload, which adds the controller to the boot sequence, thus enabling Windows to boot from a HDD/SSD attached to Marvell, the second one is the entire controller.

My understanding is that you are booting from a HDD/SSD attached to Intel controller, so you need to shorten the boot sequence. Disable any "LAN Boot Rom" but leave "Gigabit LAN" - the Lan will work but yo will no longer be able to boot from LAN. Disable "IDE/eSATA Boot Rom" under Marvel but leave "Marvell IDE/eSATA" to AHCI/Legacy/RAID. The Marvell controller will still work after boot, but you will not be able to boot from a HDD/SSD attached to Marvell. If you will post the other expansion cards connected to the system, maybe we will find a way to disable the Boot Rom from them too. They will still work, but will not count for boot devices.

You can then try 10.1.0.1008 again. If you still have doubts about my words. please read this.

@CPL0 the installed Firmware is the newest one which supports trim (see attachment)

@lordkag thank you for the lesson :-) when i configure the bios disabling all not-needed bootroms i will get no errors at bootup even at Option Rom Version 10.1

@all but trim is still not working :-(

Zitat von Martin_Q im Beitrag #23

---

but trim is still not working :-(

---

@Fernando I don't know if it definitly support trim, but I hope that the manufacturer's information is right.
Additionally there are a few Blogs, like the one from TechNet :
https://blogs.technet.com/b/sieben/archi...-windows-7.aspx
And: Crystaldiskinfo also tells that this drive supports trim

Following your advice, I tried it with a Hex Editor opening some file - After deleting the file (waiting 2 minutes) the offset stays the same; no "FF FF FF..."

@CPL0 DisableDeleteNotify = 0

Thanks Marrtin.

Would you try the following program placed somewhere on your RAID0 drive. It will write a file "TrimTest01.txt" the size of one cluster (usually 4kB) and trim it without deletion. The file contains the word "Test" repeated.

Note that this is just a rough test to try and see what is going on, possibly buggy, and not intended as a fully fledged trim tester.

when i run this exe file i get a textfile with 1024x "Test" (=4kB), and as result I get (beside the funny smiley) "The operation has completed successfully" (see attachment).

What does that mean ? Does trim work in the way it should ?

Zitat von Martin_Q im Beitrag #28

---

"The operation has completed successfully". What does that mean ?

---

TRIM WORKS !!!!

Now I installed the old 8.0 Option ROM again, because since I made BIOS Update with the 10.1 Option ROM i got many Windows Errors (Program XY doesn't work and have to be closed.. This happened with Winamp, WinWord, Windows Explorer.....). Since I installed the old 8.0 Option ROM I didn't get any errors.

The Intel RST 11.2 driver makes problems with my data-disks (4 disks - 2x RAID1). They can't get managed with the RST console, because the RAID1 arrays are "incompatible" (but still accessable) I get asked to delete this array and create it new... but I don't have enough space on external disks to back up all data from this arrays. So I installed the newest RST (12.9) I don't have any errors, all arrays (2x RAID1 HDD and the RAID0-SSD-array) are ok.

So the combination old ROM and newest RST is the key :-) (but I don't know why it didn't work earlier, because I think I already tested old ROM-new RST)...

Now I will check with hex Editor...