RST Final 10.1.0.1008 - 64 & 32 bit plus RST now available

[Ao1]

http://downloadcenter.intel.com/Sear...0.1.0.1008+%22

[zalbard]

Both are here (apparently): http://www.station-drivers.com/page/intel%20raid.htm

[Computurd]

nice find A01

[Ao1]

RST/ F6 32/ 64 but drivers now posted on Intel's site.

http://downloadcenter.intel.com/Sear...0.1.0.1008+%22

[Anvil]

Downloaded
About to reinstall my bench computer today with these drivers.

(just need to salvage my hIOmon results)

[flamenko]

Have we confirmed... Does TRIM work or is it still a very aggressive ITGC (garbage collection)???

[GargaZol]

Changed from MS own driver to this new one

My pc now freeze every ~30 min and can only come alive again by hard reboot.

ICH10 running raptors in raid 0 in windows 7 X64

Rolled back the driver and all is running fine again

[Ao1]

Changed from MS own driver to this new one

My pc now freeze every ~30 min and can only come alive again by hard reboot.

ICH10 running raptors in raid 0 in windows 7 X64

Rolled back the driver and all is running fine again

That sucks. I've had no problems so far and I had no problems with the beta versions. Not applicable to you but it seems that RST only supports AHCI & RAID. From the read me file:

"No other modes are supported by Intel(R) Rapid Storage Technology and you should continue with step 7 below. 7. Your system does not appear to be running in RAID or AHCI mode. If you feel that your system is running in RAID or AHCI mode and you do not see any of the controllers listed above, you may choose to contact your system manufacturer or place of purchase for assistance".

Flamenko, If you mean TRIM for raid then no. Intel keep claiming it is a hardware constraint not a software limitation.

[flamenko]

I just went through this elsewhere....software is the limitation of Intel and not hardware at all....tx.

[Ao1]

I know the likes of Creative have claimed hardware limitations, when in fact they had deliberately disabled hardware features via the software to make you buy new hardware when a new OS came out. I’m slightly suspicious of Intel’s claim due to the fact that that chose not to give G1 drives TRIM, when it should have been easy to do, but it is what they claim. Specifically they are saying it is a hardware limitation of the controllers. In contrast they have also said there are trying to develop TRIM support for raid, but maybe they were referring to support for the next gen controllers.

[flamenko]

http://forum.notebookreview.com/hard...ml#post6974953

[Mitchb]

http://communities.intel.com/thread/12004?tstart=120

Re: RST V9.6.0.1014. Does it pass the trim command through a raid array?

Hello, I would like to let you know that the RAID limitation regarding the trim command is related directly with the hardware manufacture and the RAID controller not the Solid State Drive is what limits the trim command to work within the operating system.

[flamenko]

Ya that a reliable source. Hi I am Les@Intel eheh. If you check his previous threads he gives the same quote time after time in bad English.

http://communities.intel.com/message/103669#103669

[alfaunits]

Well, he is right.
On one hand if you have a stripe that is less than the page size, TRIM cannot work, because a TRIM command would span half a page on separate SSDs. Larger TRIM commands would work in some area and not in others (because of how the data is split around).
On the other hand, if stripe size is a multiple of page size, then the controller can split the TRIM command to separate SSDs accordingly. E.g. if locations 512K-524K should be TRIMmed and stripe size of 4K, it would separate it as two TRIM commands - one at 256K of size 8K on one SSD, and the other of size 4K, same location on the other SSD.

Hardware RAID would need a bit more calculation to do this, especially for non R0/R1, but Intel RST could split the command in a few lines of code.

[Ao1]

Ya that a reliable source. Hi I am Les@Intel eheh. If you check his previous threads he gives the same quote time after time in bad English.

http://communities.intel.com/message/103669#103669

Mauriceb@Intel is an Intel employee. I know this for a fact, but you can also check out his profile:

Member Since: Sep 23, 2010
Last Logged In: Oct 28, 2010 11:17 AM

Occupation: Intel technical support

Job Title: IT support

[flamenko]

Well, he is right.
On one hand if you have a stripe that is less than the page size, TRIM cannot work, because a TRIM command would span half a page on separate SSDs. Larger TRIM commands would work in some area and not in others (because of how the data is split around).
On the other hand, if stripe size is a multiple of page size, then the controller can split the TRIM command to separate SSDs accordingly. E.g. if locations 512K-524K should be TRIMmed and stripe size of 4K, it would separate it as two TRIM commands - one at 256K of size 8K on one SSD, and the other of size 4K, same location on the other SSD.

Hardware RAID would need a bit more calculation to do this, especially for non R0/R1, but Intel RST could split the command in a few lines of code.

Intel recommends stripe size of 16k for SSDs right in the RST package,

[alfaunits]

Yes, higher stripes are usually better for most users, but at least RAID stripe size = SSD page size is good enough for the controller to be able to pass TRIM down. Higher stripe size only changes the calculation result, but not the outcome.

[AnXioZ]

Intel recommends stripe size of 16k for SSDs right in the RST package,

Isn't a single page 4K?

[alfaunits]

It is, the two are unrelated, since RST does not support TRIM in RAID0.
Namely, 16K stripe should provide best performance in most situations. (so it says, I did not run RAID on onboard RAID for years)