New Intel RST 10.0.0.1043 BETA Out.

I never saw another post here but could be wrong so here goes. A beta release of the intel RST driver has hit the web and can be found here:

USER BEWARE as there are beta drivers of course...
intel drivers pour Raid/Sata/Ata/Ahci

There are two claims with respect to its abilities, one being that your system boots faster and the second is that something is going on which just may be TRIM in RAID mode. I haven't got a RAID system but can say my average system loading time before the new drivers on the test system was 16-17 seconds and has now dropeed to 12-13 seconds consistently.

With respect to RAID and TRIM, I wouldn't dare say it but alot of posts similar to this are popping up real quick:

XtremeSystems Forums - View Single Post - Intel RST 10 Series Driver released!!

USER BEWARE as there are beta drivers of course...

 

ooooh, goody. Beta disk drivers. I'll jump right on that.

Not.

 

Thanks for the post...
But ah, which one? The Rapid Storage Tech, right? Not the Matrix manager...
And I'm guessing the very first link, the 10.0.0.1043 ones, is the one to get?
Thanks.

 

should you uninstall previous version before installing this one, or install over it?

 

the ssd guys and momentus xt guys should jump on this, if they havent yet

 

In my opinion no one should jump at this. They're beta drivers and I couldn't measure any performance improvement over the existing drivers. If anything, it was slightly slower.

And we had some people having to restore Windows because the driver didn't work for them.

 

ok phil thats good to know...

 

The Intel RST 10.0.0.1043 BETA indeed improves the performance for me.
The Intel RST 10.0.0.1026 BETA also available on this website installs well, but loops during Windows boot. Avoid it!

 

Do you mean it improves performance in synthetic benchmarks? Because I was talking about real world performance.

If you are talking about real world performance, what did you measure and how large were the improvements?

 

No I checked them with a build of project "Win7RescuePE" with WinBuilder.
The turnaround time dropped from

Project time: 5 minutes, 56 seconds and 757 miliseconds to

Project time: 5 minutes, 28 seconds and 599 miliseconds

 

maximinimaus, that's interesting. Could you run the same test for Microsoft AHCI driver?

I just checked the new driver with Boottimer. Indeed 10.0.0.1043 boots one second quicker than IRS 9.6.1014 (or something like that).

Now it's as fast as MS AHCI.

 

http://forum.notebookreview.com/har...marks-brands-news-advice-685.html#post6719444

 

I already checked Intel Rapid Storage 9.6.5.1001 against the Microsoft AHCI driver.
The Intel Rapid Storage 9.6.5.1001 improved the turnaround time by about 30 seconds.
In synthetic benchmarks like CDM the Microsoft AHCI scored better.

 

I'm not sure if this is related but here's my post regarding a potential problem: http://forum.notebookreview.com/6734412-post4239.html

 

HWMonitor doesn't show the temperatures of my
Drive Model: Hitachi HTS723225L9A360
But HWiNFO32 shows them. Other tools like SIV and SIW show them too.

 

New feats of RST 10.0:
- Optical Disk Drive (ODD) Video Read-ahead (VRA) - REMOVED later due to problems with certain BluRay protections
- Zero Power ODD – Power-off Unused Optical Disk Drives (ODD)
- Email Alert and Notification
- RAID Volume Cache Size Increased to 16 MB of System Memory
- SATA Link Power Management (LPM) Support on Desktop

 

Vostro1400user...Apologies as not having see that but thought it merited its own thread. If a Mod wants to move thread OP to yourself, i have abs no prob with that as you were first at bat.

 

can more ppl use them? We need more results.. i'm trying them now..

They improved my 4K scores a bit.. this is so troublesome... these tweaks each are casuing turbo boost sideffects... managed to make JJB's oe work.

 

Guys I got a question... you know all the intel startup processes...does disabling them make any performance differences?

 

has anyone **really** found a difference between 9.6.4.1002 (or even 9.6.0.1014) and these beta drivers obtained from station-driver

 

The newer driver got 1 second of my boot time with a Kingston v+ 64GB. Before it was 13.7, now it's 12.7.
Here's another improvement:

I've installed the driver via Device manager though. According to maximinimaus I'll get better performance through setup.

 

...that would make plenty of sense as a device manager driver update grabs the .sys files only while an install from the setup.exe (or whatever name) installer will push in the drivers as well as the IRST control application and services.

It's probably a good idea to keep the drivers and their control application sync'd up whenever possible.

 

I finally decided to give IRST 10.0.0.1043 a fair chance, since I've only been impressed with the MS AHCI driver. I ran BootTimer 3x/driver and the results are worth noting:

MS AHCI - avg over 3 boots
17.467 seconds

IRST 10.0.0.1043 - avg over 3 boots
15.615 seconds

The results were consistently within the same range and there was immediate improvement with IRST.

Benchmarks schmenchmarks. Boot time improvement shows performance improvement.

 

anseio, did you install the driver via setup.exe?

 

I downloaded the driver from here, extracted the files, open the x64 folder and ran the setup file, named PNPINST64.

 

Since the package contains only the driver files, they can also be installed via Device Manager.

 

So IAStorDataMgrSvc would have to be installed from a previous version?

 

First I installed RST 9.6.5.1001 via Setup, than the drivers of version 10.0.0.1043beta via Device Manager.
I've done further testing and the biggest advance was the step from MS AHCI to Intel RST.
The improvements by the various versions of RST or the method installing either via Setup or Device Manager are neglibible.
For the reason of possible backlevel or regression, I went back to RST version 9.6.5.1001.

 

Is this still true though?

"I experienced better scores when this Intel Rapid Storage service is installed and running."

 

make up your own mind and look at the thumbnails.
The test were done with a

Drive Model: SAMSUNG SSD PB22-JS3 FDE 2.5" 128GB

As I said, the biggest gain was the install of IRST.
The turnaround time of my "WinRescuePE" project with "WinBuilder" dropped by about 30 secs.
MSHCI:
40 scripts processed.
Project time: 5 minutes, 52 seconds and 607 miliseconds

IRST 9.6.5.1001:
40 scripts processed.
Project time: 5 minutes, 22 seconds and 780 miliseconds

 

Just want to note that I've been using these latest (beta) drivers for a while and what I notice most (no stopwatches) is that the drive performs more consistent - whether that drive is my Inferno SSD or the Momentus XT hybrid.

A point that jumped out at me (when I finally noticed it!) was the Windows 7 startup animation would pause (on the XT) with IRST 9.6. While with the 1043 beta, the pause is gone.

I can definitely say that the beta drivers are making my systems faster/smoother - no matter what the benchmarks say.

 

It could be that the larger read-ahead cache in the drivers are actually reading full tracks instead of selecting sectors. Just suck data off of the drive as fast as the platters fly underneath the head(s) and let the drive electronics, driver cache, and filesystem pick out what is really wanted/needed.

Lots of layers there, each one with a chance for errors, but if Intel has their layer (drivers) fully debugged, it should be reliable when the 10.x drivers come out of beta. NTFS (and the linux equivs like JFS, XFS, etc, etc) are already capable of managing data that come in via multiple cache layers.

'enterprise' class storage has been doing this for years.

 

No gain for me either...

My guess is they offer the same performance on a single SSD on the ICH8 chipset...

(My Queued writes were a tad slower actually, but that might be just impacted by "other software" that runs alongside the OS, no safe mode test)

 

That comment was based on my Crucial C300 .

On the Kingston V+ the newer 10.*.043 driver took 1 second of my boot consistently.

 

Interesting.
Question is, can you measure it up to 1s? My Boot times oscillate a fair amount... don't ask me why. But the final performance is what you'll notice more in the end.
Obviously, a boot time of say 2min vs. 1min is a huge different, but can you accurately measure the boot time to 1s? I'm asking because even when I didn't have a lot of software on my Vaio my boot time was less than half of what it is now (LiveSynch.. grr... it's slow) it was still oscillating by maybe +/- 5s and some outliers.

 

I took 3 runs, boot time went from average 13.7 seconds to 12.7 seconds.

My boottimes are quite consistent measured with Boottimer.exe.

 

If anyone cares I have a 7200 rpm internal and I have noticed no differance. Granted I also have a fresh install.

 

Interesting. Might give these a shot with my system and SSD when i get home.