• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Which would U go for MSAHCI vs IRST (840 EVO)

G73S

Senior member
On my 1 TB Samsung 840 Evo (NO RAPID mode)

CrystalDiskMark with Intel Chipset Drivers v9.4.0.1023 (W7)

15y9abn.jpg


CrystalDiskMark with IRST 12.8.0.1016 (W7)

oqgard.jpg


Based on these scores, which would you prefer? to go with the Intel Chipset Drivers (which is the default msahci) or with the IRST Driver?
 
Methinks U R jus showing off a little.:sneaky: 😉

Maybe it would be interesting to see the comparison in RAPID mode. Dunno. Would depend on his BIOS SATA configuration, which has to be AHCI for RAPID.

He says that one of the compared options is the "default" MSAHCI driver, but that it is an "Intel" driver. As I understand it, M$ provides a "Native" driver of MSAHCI. If Intel provides their own version of it, then perhaps he'd want to compare those.


But at this point, I don't see any difference. Maybe I don't fully understand the MSAHCI options, but that's the reason I chose to put a certain hardware controller in my system: No driver installation required, and it uses the Native driver.
 
He says that one of the compared options is the "default" MSAHCI driver, but that it is an "Intel" driver. As I understand i

As I understand it now too! As I just learned when the same OP attacked MY AHCI driver and that launched me into my own delving mode.

But at this point, I don't see any difference

As per my initial comment, methinks either does he. See a difference. Was not the real purpose of the thread. Who knows? If I had a 1TB SSD....I think he has TWO....perhaps I would be moved to show off too in this community.

But were that the case, I would not obfuscate that celebration under the guise of whatever. I would make a thread blatantly, joyfully titled... I HAVE DIED AND GONE TO HEAVEN!

Course, even if I could afford one, I would NEVER IN THIS LIFE own/carry an Hermes handbag. And those cost 20 times what a 1T SSD does.....some 100X more!

If you can't impress in little low risers and a simple cotton top & sandals.....well, that says a lot.
 
Last edited:
Methinks U R jus showing off a little.:sneaky: 😉
no I don't need to show off with an SSD. There are thousands of people with Samsung SSDs with similar scores. Everyone knows they're the best and most reliable. I am simply trying to see which is better IRST vs. MSAHCI
 
Maybe it would be interesting to see the comparison in RAPID mode. Dunno. Would depend on his BIOS SATA configuration, which has to be AHCI for RAPID.

He says that one of the compared options is the "default" MSAHCI driver, but that it is an "Intel" driver. As I understand it, M$ provides a "Native" driver of MSAHCI. If Intel provides their own version of it, then perhaps he'd want to compare those.


But at this point, I don't see any difference. Maybe I don't fully understand the MSAHCI options, but that's the reason I chose to put a certain hardware controller in my system: No driver installation required, and it uses the Native driver.

Score with RAPID:

AS SSD Benchmark with Intel Chipset Drivers 9.4.0.1023 (W7):

ny9z7q.png



AS SSD Benchmark with IRST 12.8.0.1016 (W7)

mb5ysl.png


But then, RAPID just lies to the benchmark telling it to benchmark my RAM not the actual SSD so I don't think that's a fair comparision that would translate into real world usage scenarios. That`s why I posted the intial scores without RAPID.

all I wanted to know from this thread, is there is a slight difference between the msahci driver vs. the IRST one, and I wanted to know which in your opinion is better.


Now regarding the MSAHCI driver, when I install the Intel Chipset Drivers and run AS SSD Benchmark, it still says msahci, but when I install IRST driver, AS SSD benchmark says iastor

So the Intel Chipset Driver is still using the default msahci driver for the ATAPI/IDE controller right?
 
. . . all I wanted to know from this thread, is there is a slight difference between the msahci driver vs. the IRST one, and I wanted to know which in your opinion is better.


Now regarding the MSAHCI driver, when I install the Intel Chipset Drivers and run AS SSD Benchmark, it still says msahci, but when I install IRST driver, AS SSD benchmark says iastor

So the Intel Chipset Driver is still using the default msahci driver for the ATAPI/IDE controller right?

That's the way it looks to me. I wouldn't know the particulars, but the second driver name you mention is familiar. I only recall it from situations where the Intel drivers had to be ready for a Wiindows installation requiring the F6 key. Apparently, with AHCI established as a standard, MS provided MSAHCI as a "native" standard.

I don't know for sure what Intel is doing. For some motherboards, drivers are segregated between SATA and RAID. In Device Manager, AHCI drivers and devices show up under "IDE/ATA/ATAPI controllers." For devices under that node, I see that iastor is the driver for AHCI mode with Intel drivers installed. Without them, I must deduce from another controller in another machine I have that the MSAHCI driver also appears under the same node when it is used instead.

Usually, RAID drivers are exposed under devices shown in "Storage Controllers." But I have a Marvell SATA controller which only works as AHCI on the first machine, with the Marvell drivers showing up under this "Storage" node. The second controller I mentioned in the previous paragraph is also Marvell, but uses the native MSAHCI driver, and shows up under the IDE/ATA/ATAPI node.

So they must be different driver options (Intel versus MS "native"), both meeting the AHCI spec.
 
That's the way it looks to me. I wouldn't know the particulars, but the second driver name you mention is familiar. I only recall it from situations where the Intel drivers had to be ready for a Wiindows installation requiring the F6 key. Apparently, with AHCI established as a standard, MS provided MSAHCI as a "native" standard.

I don't know for sure what Intel is doing. For some motherboards, drivers are segregated between SATA and RAID. In Device Manager, AHCI drivers and devices show up under "IDE/ATA/ATAPI controllers." For devices under that node, I see that iastor is the driver for AHCI mode with Intel drivers installed. Without them, I must deduce from another controller in another machine I have that the MSAHCI driver also appears under the same node when it is used instead.

Usually, RAID drivers are exposed under devices shown in "Storage Controllers." But I have a Marvell SATA controller which only works as AHCI on the first machine, with the Marvell drivers showing up under this "Storage" node. The second controller I mentioned in the previous paragraph is also Marvell, but uses the native MSAHCI driver, and shows up under the IDE/ATA/ATAPI node.

So they must be different driver options (Intel versus MS "native"), both meeting the AHCI spec.
if I only install the Intel Chispet Drivers, then it still shows as MSAHCI in AS SSD. only when I install the IRST driver it changes to iastor
 
if I only install the Intel Chispet Drivers, then it still shows as MSAHCI in AS SSD. only when I install the IRST driver it changes to iastor

I suppose that makes sense. I have IRST installed, and the controller property tab shows iastor. But your tests confirm for me that my indifference to the matter is probably spot-on: it doesn't seem to make much difference.

Don't take this as condescending, but I wouldn't make judgments in comparisons over single bench results unless they vary in dozens of MB/s results. The differences you show are insignificant. Consider that there is going to be some random variation around a mean to generate a bench result as a "single observation."

Best way to look at it -- some may criticize or find fault: A single benchmark result is a single "observation," and you could choose to generate a "sample" of several such observations -- to compute an average and standard error.*[see my note] My engineering "QC" statistics text refers to "fleas on the backs of fleas, and fleas on the backs of those . . . " as a quote from some 17th century author. Each benchmark test is itself a set of sample results, or so it would seem.

* The problem with that: supposedly taking more and more benchmarks from the same SSD would show successively worse results -- even in increments. So the complication there: each benchmark result would successively bias the ones that follow. An average of so many benchmarks would have a downward bias -- and you wouldn't want a statistic with such a bias built-in to the method for generating it.
 
Last edited:
MSAHCI. IRST means you're operating in RAID mode. That means, as you noticed, no RAPID (Benchmarking the RAM in RAPID mode is valid, since you're accessing the RAM as a 1 GB cache for the SSD). And my understanding is that Intel has only implemented TRIM passthrough for RAID0 arrays under IRST, so you wouldn't get it, and it'll hurt performance some.
 
MSAHCI. IRST means you're operating in RAID mode. That means, as you noticed, no RAPID (Benchmarking the RAM in RAPID mode is valid, since you're accessing the RAM as a 1 GB cache for the SSD). And my understanding is that Intel has only implemented TRIM passthrough for RAID0 arrays under IRST, so you wouldn't get it, and it'll hurt performance some.

I see, better go back to MSAHCI then.

thanks man
 
MSAHCI. IRST means you're operating in RAID mode. That means, as you noticed, no RAPID (Benchmarking the RAM in RAPID mode is valid, since you're accessing the RAM as a 1 GB cache for the SSD). And my understanding is that Intel has only implemented TRIM passthrough for RAID0 arrays under IRST, so you wouldn't get it, and it'll hurt performance some.

I think you're in error there, and thought I'd better post to call your attention -- and the OP's -- to it.

First. Acronyms can be confusing. ISRT refers to the SSD-caching technology which requires a RAID-mode configuration in BIOS so the SSD can be in a peculiar "RAID0" with itself. In these circumstances, TRIM is enabled on the SSD.

IRST (R and S reversed) is the software for configuring or monitoring drives attached to an Intel controller, regardless of whether the BIOS mode is RAID or AHCI. My system, for example, is currently configured with BIOS AHCI-mode for the Intel onboard controller. The controller appears in Device Manager under "IDE/ATA/ATAPI" controllers, and the driver for it is "iastor.sys." If I had not installed the IRST software, Windows would have configured its native MSAHCI driver.
 
I think you're in error there, and thought I'd better post to call your attention -- and the OP's -- to it.

First. Acronyms can be confusing. ISRT refers to the SSD-caching technology which requires a RAID-mode configuration in BIOS so the SSD can be in a peculiar "RAID0" with itself. In these circumstances, TRIM is enabled on the SSD.

IRST (R and S reversed) is the software for configuring or monitoring drives attached to an Intel controller, regardless of whether the BIOS mode is RAID or AHCI. My system, for example, is currently configured with BIOS AHCI-mode for the Intel onboard controller. The controller appears in Device Manager under "IDE/ATA/ATAPI" controllers, and the driver for it is "iastor.sys." If I had not installed the IRST software, Windows would have configured its native MSAHCI driver.

Interesting explanation...except no one's talking about ISRT except you. 😉 All of us where all talking about IRST the entire time. As you can see in the AS-SSD screenshots, when the OP is in IRST mode using the iastor driver, it is showing up as a SCSI disk device, not an ATA one, which means it's not operating in AHCI mode, but a pseudo-SCSI mode, which is a very different animal (to say the least). Again, with IRST in RAID mode, everything I've read changelog-wise says that it only passes along TRIM commands in RAID0 mode, not in single-drive mode or RAID1 mode. I will be exceedingly happy if you can show me something that says otherwise, a changelog or a report or something, I'll be exceedingly happy to see it.

Now, it's entirely possible that IRST can operate in ATA mode...but in this particular case it isn't.
 
everything I've read changelog-wise says that it only passes along TRIM commands in RAID0 mode, not in single-drive mode or RAID1 mode. I will be exceedingly happy if you can show me something that says otherwise, a changelog or a report or something, I'll be exceedingly happy to see it.

Now, it's entirely possible that IRST can operate in ATA mode...but in this particular case it isn't.

this post made me learn something really good that I haven't ever read about

thank you so much Dr. SSD
 
Interesting explanation...except no one's talking about ISRT except you. 😉 All of us where all talking about IRST the entire time. As you can see in the AS-SSD screenshots, when the OP is in IRST mode using the iastor driver, it is showing up as a SCSI disk device, not an ATA one, which means it's not operating in AHCI mode, but a pseudo-SCSI mode, which is a very different animal (to say the least). Again, with IRST in RAID mode, everything I've read changelog-wise says that it only passes along TRIM commands in RAID0 mode, not in single-drive mode or RAID1 mode. I will be exceedingly happy if you can show me something that says otherwise, a changelog or a report or something, I'll be exceedingly happy to see it.

Now, it's entirely possible that IRST can operate in ATA mode...but in this particular case it isn't.

I'm not trying to "pull a fast one," but over the last few months here, some of our colleagues had reported on the "TRIM in RAID" resolutions with newer BIOS and IRST versions. You can look for the threads; I can't remember the titles or I'd post them. They ran the tests to prove TRIM was operating.

If the OP had his BIOS configured for "RAID-mode," then what you say may be true. I've had the recent experience of mis-installing the wrong drivers for a controller, which would then show up in Device Manager as a "SCSI" storage device -- a different node than "IDE/ATA . . . " All I can say is this: my BIOS is in AHCI mode; I have IRST installed; the driver shown under "IDE/ATA/ATAPI" specifies "AHCI 1.0" compliance and it is iastor.sys.

I'm pretty sure without installing the IRST software that it would configure MSAHCI as the driver -- still AHCI compliant, and still under the "IDE/ATA/ATAPI" node in device manager.
 
I'm not trying to "pull a fast one," but over the last few months here, some of our colleagues had reported on the "TRIM in RAID" resolutions with newer BIOS and IRST versions. You can look for the threads; I can't remember the titles or I'd post them. They ran the tests to prove TRIM was operating.

If the OP had his BIOS configured for "RAID-mode," then what you say may be true. I've had the recent experience of mis-installing the wrong drivers for a controller, which would then show up in Device Manager as a "SCSI" storage device -- a different node than "IDE/ATA . . . " All I can say is this: my BIOS is in AHCI mode; I have IRST installed; the driver shown under "IDE/ATA/ATAPI" specifies "AHCI 1.0" compliance and it is iastor.sys.

I'm pretty sure without installing the IRST software that it would configure MSAHCI as the driver -- still AHCI compliant, and still under the "IDE/ATA/ATAPI" node in device manager.
I can't remember where I read but if you are in RAID, even if TRIM seems to be working, internally, the TRIM commands are NOT reaching the drive.

Best for these SSDs is to never touch the IRST driver if you're not in RAID 0 mode.

I am using RAPID mode so I stuck to the MSAHCI driver which gave me better scores anwyay
 
I can't remember where I read but if you are in RAID, even if TRIM seems to be working, internally, the TRIM commands are NOT reaching the drive.

Best for these SSDs is to never touch the IRST driver if you're not in RAID 0 mode.

I am using RAPID mode so I stuck to the MSAHCI driver which gave me better scores anwyay

Yes -- that was also part of the discussion: the command-line query that "proves" TRIM "enabled" doesn't assure that it's actually "TRIMming." But some Russian enthusiast produced a utility that addresses that issue. This had been a longstanding problem -- not just a question -- since SSD usage gathered momentum in 2011 and later. First, the question as to whether two SSDs in RAID0 had TRIM (they didn't), and then whether a single standalone drive connected to a controller port with BIOS "RAID-mode" implementation was getting "Trimmed" -- such as with the peculiar "RAID0" of a caching SSD, or just a standalone drive (it was, but without the testing, you had to wonder). Then Intel forums and techs said it had been resolved. Finally, our colleagues here in earlier threads since beginning of this year.

You can stick to the MSAHCI driver, and there shouldn't be any "down side." The Intel IRST software had more general application than just managing "matrix" storage and RAID : I just chose to install it for AHCI mode. I use the MSAHCI driver for some PCI-E controllers using the Marvell 9230 chip on another system.

But since the Intel and MS drivers are both AHCI compliant, I doubt it would make any difference, and indeed the OP's tests tend to prove it.
 
I can't remember where I read but if you are in RAID, even if TRIM seems to be working, internally, the TRIM commands are NOT reaching the drive.

Best for these SSDs is to never touch the IRST driver if you're not in RAID 0 mode.

I am using RAPID mode so I stuck to the MSAHCI driver which gave me better scores anwyay

Here's a twist I wasn't counting on.

I have another system for "strickly bidnis" usage. I've been doing maintenance on it; discovered, out of two RAM kits, one is bad. Got that solved.

The system has an onboard Intel ICH10 controller, and the controller shows up as "Intel [with detail]" in the IDE/ATA/ATAPI device manager node, just as does the controller on this machine (referenced earlier). But it doesn't have the IRST software installed, and the driver shows up as MSAHCI -- among other parts of it.

I guess the need for clarification, if desperate enough, would suggest "Ask Intel." Otherwise, neither system appears to be "misconfigured."
 
MSAHCI. IRST means you're operating in RAID mode.
No, it doesn't. Non-RAID boards can have and use IRST, including the Iastor driver, as can RAID-supporting boards not in RAID mode. Now, aside from basic disk health checking, the IRST application does little more than use up a systray icon, without RAID, but it does work fine, with the controller set to AHCI.

<- Lots of B75 and B85 PCs, all with IRST pre-installed, and Iastor configured for the driver.

As long as it's not set to IDE (QD=1, ew!), I leave it alone.
 
Last edited:
Based on these scores, which would you prefer? to go with the Intel Chipset Drivers (which is the default msahci) or with the IRST Driver?
The generic Win7 AHCI driver named MSAHCI.SYS is a universally usable AHCI driver, which supports the SATA AHCI Controllers of various chipset manufacturers (Intel, AMD, Marvell, ASMedia etc.), whereas the Intel RST driver has been especially developed and designed for being used with Intel SATA AHCI Controllers.
That is the simple reason why I recommend to use the "specialist" and not the "allrounder" under the available AHCI drivers, especially for Intel chipset systems.
Recently I have done intensive benchmark tests comparing the generic MS AHCI driver with the latest Intel RST AHCI driver versions. You can find my test results >here<.
 
The generic Win7 AHCI driver named MSAHCI.SYS is a universally usable AHCI driver, which supports the SATA AHCI Controllers of various chipset manufacturers (Intel, AMD, Marvell, ASMedia etc.), whereas the Intel RST driver has been especially developed and designed for being used with Intel SATA AHCI Controllers.
That is the simple reason why I recommend to use the "specialist" and not the "allrounder" under the available AHCI drivers, especially for Intel chipset systems.
Recently I have done intensive benchmark tests comparing the generic MS AHCI driver with the latest Intel RST AHCI driver versions. You can find my test results >here<.

Interesting. I have two controllers in my server box -- identical -- using the 9230 Marvell chip and offering RAID0, 1, 1+0. Serious tech reviews of the product promote the controller's ability to configure as AHCI with the native MSAHCI driver. Nothing need be done with the controller BIOS, unless one choose to configure any of the RAID modes. The BIOS doesn't include a JBOD option.

I have an onboard Marvell controller on my workstation motherboard, currently enabled and hosting a single SATA drive. It only provides for AHCI mode. The driver is the "92xx . . . " storage driver, and the device appears under a separate "Storage Controllers" node where you'd usually expect to see RAID or SCSI controllers.

Now it would seem to me that I could also set up my server controllers with that 92xx driver. But since everything is working tip-top with MSAHCI, I don' wanna mess with it.
 
Now it would seem to me that I could also set up my server controllers with that 92xx driver. But since everything is working tip-top with MSAHCI, I don' wanna mess with it.
If the generic MS AHCI driver is running fine, I wouldn't replace it by the Marvell AHCI driver.
My previous statement is only valid for Intel SATA AHCI Controllers, because Intel's AHCI and RAID drivers usually are better than the "special" AHCI/RAID drivers from other chipset manufacturers.
 
If the generic MS AHCI driver is running fine, I wouldn't replace it by the Marvell AHCI driver.
My previous statement is only valid for Intel SATA AHCI Controllers, because Intel's AHCI and RAID drivers usually are better than the "special" AHCI/RAID drivers from other chipset manufacturers.

Well, Fernando, I'm only a little bit confused at this point. As one of the posters mentioned, if he only installs the Intel chipset drivers, his SATA AHCI driver is MSAHCI. If he installs the IRST software, it shows up under the "IDE/ATA/ATAPI" device manager node as "Intel(R) Desktop/Workstation/Server Express Chipset SATA AHCI Controller" and the driver shown in "details" of the Properties window is IASTOR.sys, together with something called ACHEMAR.sys.

My Samsung Magician software shows my SSD 840-Pro as "AHCI" -- everything is fine -- able to implement "RAPID" with no problem. Didn't have a problem (first) with the MSAHCI driver installed; didn't have anything appear wrong with the IRST software installation; don't have anything showing any problem with this "IASTOR" driver.

I could always uninstall the IRST software and I'm betting the driver would revert to MSAHCI. Do I need the IRST software? I don't think so, but then I don't know. Maybe it would prove useful later even if there are no RAID arrays to administer and the system is not in RAID mode.

UPDATE: Here's some more on "MSAHCI or IRST/IASTOR." A forum dialog in another forum (SPCR):

http://www.silentpcreview.com/forums/viewtopic.php?t=66916p=580772

"Six of one, half dozen of the other" seems to be the conclusion of one poster who seems knowledgeable about the Intel drivers.
 
Last edited:
@ BonzaiDuck:
1. The Intel chipset "drivers" are only text files (named *.INF) and do nothing more with the Intel SATA Controllers than to rename them.
2. The RST Software, which let the Intel SATA AHCI/RAID Controller use the Intel RST driver, is just the installer, which is part of the complete "Intel RST Drivers & Software Set".
3. For AHCI users I generally recommend to install the Intel RST driver manually from within the Device Manager. The RST Software (RST Console and RST Services) is not needed at all for the AHCI mode and may have a negative impact on the AHCI performance, when the RST Software is running in the background.
 
@ BonzaiDuck:
1. The Intel chipset "drivers" are only text files (named *.INF) and do nothing more with the Intel SATA Controllers than to rename them.
2. The RST Software, which let the Intel SATA AHCI/RAID Controller use the Intel RST driver, is just the installer, which is part of the complete "Intel RST Drivers & Software Set".
3. For AHCI users I generally recommend to install the Intel RST driver manually from within the Device Manager. The RST Software (RST Console and RST Services) is not needed at all for the AHCI mode and may have a negative impact on the AHCI performance, when the RST Software is running in the background.

Thanks for the insight. there is so much bloatware that printer software, IS/AV and other installations put in the system tray. Even so, for "negative performance," it doesn't seem to have much of an impact on my 840-Pro -- with or without RAPID.
 
Back
Top