If you think TRIM is working, think again!

[G73S]

Most users with SSDs install the IRST drivers as they believe they offer a slightly better performance BS the native MSAHCI driver, while that may be true in certain cases (not on my Samsung 840 EVO), you may or may not TRIM functionality which will hurt your SSD's performance in the long run.

What most people do is run the TRIM check command in a DOS command prompt:

[B]fsutil behavior query DisableDeleteNotify[/B]

It will give you one of two results, either a 0 or a 1. A zero indicates that TRIM is enabled correctly, a one means that it is not.

That will tell you if the TRIM command was sent successfully or not, but that's only HALF the story, we need to check if the TRIM command ACTUALLY reached the drive not only if the TRIM command was sent or not...

So how can one really verify if TRIM is working?

Please read this article: TRIMcheck: Does Your SSD Really have TRIM Working? | The SSD Review

trimcheck Download

Just copy the file to the drive you want to test, then right click on it and choose "Run As Administrator"

When it launches, press Enter on your keyboard to have it create a file and delete it. then close the Window.

Wait for about 1 minute, then run it again and press Enter to do another test and see the result!

 

[RU482]

is this a paid advertisement? sure reads like one

 

[G73S]

is this a paid advertisement? sure reads like one

where the hell did you see any advertisements?

I have Adblock Plus installed on my Pale Moon browser maybe that's why I didn't see any ads. Either way, I was just googling how to really check if TRIM is working and came across that article so I posted it. you don't even need to click it, simply download the file and follow my instructions

 

[KingFatty]

This seems to depend on a characteristic of the SSD drive itself, and not any particular installation/software right?

So the only solution would be to try to find updated BIOS that fixes your bum SSD if it's affected?

 

[BonzaiDuck]

where the hell did you see any advertisements?

I have Adblock Plus installed on my Pale Moon browser maybe that's why I didn't see any ads. Either way, I was just googling how to really check if TRIM is working and came across that article so I posted it. you don't even need to click it, simply download the file and follow my instructions

The TRIM issue has been on the table for a few years now. Supposedly, we have other forum veterans here who've tested newer chipsets, BIOS updates etc. to see if drives are being Trimmed in various RAID modes as well as AHCI.

Then there's the concurrent thread about using MSAHCI versus the Intel IASTOR with BIOS configured for AHCI mode. Of course, there are many manufactures and models of SSDs -- and I'm personally less interested in those at the moment because the only ones that counts for me are the Sammy 840 units. I can see that my Magician software will TELL me if TRIM is implemented, and adds a feature whereby it can force TRIM as a manual operation.

I might take the time to download the Panteleev TRIMCHECK "test" program which is supposed to show whether TRIM is actually being run as opposed to the command-line FSUTIL.

But it's a little bit silly, don't you think -- to imagine that somehow TRIM wouldn't be implemented by the Intel AHCI driver, when Intel had been at the forefront of SSD development? And after Intel had announced their intention a few years back to tie up loose ends as pertains to RAID and TRIM on their SATA controllers?

At this point -- MSAHCI, IASTOR or whatever -- I think it's probably certain you're getting TRIM with BIOS' AHCI-mode for a standalone SSD.

 

[G73S]

The TRIM issue has been on the table for a few years now. Supposedly, we have other forum veterans here who've tested newer chipsets, BIOS updates etc. to see if drives are being Trimmed in various RAID modes as well as AHCI.

Then there's the concurrent thread about using MSAHCI versus the Intel IASTOR with BIOS configured for AHCI mode. Of course, there are many manufactures and models of SSDs -- and I'm personally less interested in those at the moment because the only ones that counts for me are the Sammy 840 units. I can see that my Magician software will TELL me if TRIM is implemented, and adds a feature whereby it can force TRIM as a manual operation.

I might take the time to download the Panteleev TRIMCHECK "test" program which is supposed to show whether TRIM is actually being run as opposed to the command-line FSUTIL.

But it's a little bit silly, don't you think -- to imagine that somehow TRIM wouldn't be implemented by the Intel AHCI driver, when Intel had been at the forefront of SSD development? And after Intel had announced their intention a few years back to tie up loose ends as pertains to RAID and TRIM on their SATA controllers?

At this point -- MSAHCI, IASTOR or whatever -- I think it's probably certain you're getting TRIM with BIOS' AHCI-mode for a standalone SSD.

with the MSAHCI driver TRIM works no question. The verification needs to be done with the IASTOR driver. Many people think just because they send a FSUTIL command and it reports successfully that TRIM is REALLY working when that is not the case.

For what I know, TRIM only works on RAID 0 and no other RAID mode. still, since it takes a minute to check, always good to double check and ensure that the TRIM commands are actually reaching your SSD

PS: we are in the same sammy 840 club and these drives do not play well with the IASTOR when RAPID is enabled. Seems like Samsung programmed RAPID keeping in mind that the majority of users are using the default MSAHCI driver provided by Windows

 

[Coup27]

Have we gone back in time 2 years? Grief.

 

[Prime_Time]

TRIM are working just fine with IRST (IASTOR) driver.

And if my memory serve me well you are the one on OCN who repeatedly post your AS SSD score with RAPID mode many and many time right?

And that is the same guy who always said that TRIM is not working if IRST on single drive mode and after many people told you that TRIM working well with IRST you here, and again TRIM may not working on IRST ????

Dude TRIM working fine on IRST and for most user IRST give a better result compare to msahci or stoahci.

RAPID is just something stupid Windows own ram cache work even better and RAPID also make your SSD performance drop in many scenarios it just marketing gimmick that make benchmark score look like WOWWW but in the real world usage it good for nothing.

above are my TRIMcheck result and I can confirm that TRIM is work out of box with IASTOR for every rig I have.

 

[G73S]

TRIM are working just fine with IRST (IASTOR) driver.

And if my memory serve me well you are the one on OCN who repeatedly post your AS SSD score with RAPID mode many and many time right?

And that is the same guy who always said that TRIM is not working if IRST on single drive mode and after many people told you that TRIM working well with IRST you here, and again TRIM may not working on IRST ????

Dude TRIM working fine on IRST and for most user IRST give a better result compare to msahci or stoahci.

RAPID is just something stupid Windows own ram cache work even better and RAPID also make your SSD performance drop in many scenarios it just marketing gimmick that make benchmark score look like WOWWW but in the real world usage it good for nothing.

above are my TRIMcheck result and I can confirm that TRIM is work out of box with IASTOR for every rig I have.

who the hell are you? and what is OCN? you registered here just to trash talk?

No need for the rude post. Member call outs are not allowed. -Shmee

 

[BonzaiDuck]

PS: we are in the same sammy 840 club and these drives do not play well with the IASTOR when RAPID is enabled. Seems like Samsung programmed RAPID keeping in mind that the majority of users are using the default MSAHCI driver provided by Windows

I've had RAPID enabled with my 840-Pro, check up on it from time to time with Magician. I don't have any problem with it. Magician doesn't have any problem with it. What sort of "trouble" of not-playing-well do you mean?

Incidentally. Before I got the BIOS particulars of my system sorted out (from RAID to AHCI-mode), I did a lot of reading about the Sammy drives and the Magician software. I ran across an enthusiast web site in UK -- can't remember which precisely but it was a well-established review site with forums that pops up in searches. People were asking similar questions. The strongest answer I saw suggested that Samsung had focused its efforts with Magician and their hardware on the Intel controllers. There was an explanation as to how the Sammy drives play well with the Intel controllers. If that were the case, I don't think they would've missed much pertaining to an AHCI driver.

 

[Fernando 1]

@ G73S:

What is the concrete reason for starting a new thread about a very old topic?
If you would have used the Search function, you would have seen >this< discussion, which has been started by me in February 2013.

 

[imagoon]

Hey I remember that app. That is the one that doesn't actually tell you anything. TRIM can be enabled, it is up to the drive for if and when (if ever) a block gets TRIMed, assuming the OS is actually sending the TRIM command. TRIM commands are noncommittal and are not required to performed. The ATA specifications in use now don't even provide a way for the drive to send a success or fail back to the OS. Once drives implemented slack space, TRIM became even less valuable in the consumer space. Add in that the LBA to block mapping is completely up to the drive meaning even if a single 4KB file system page is deleted, the drive may never bother TRIMing. There is an entire long in depth thread in this forum about it.

 

[G73S]

I've had RAPID enabled with my 840-Pro, check up on it from time to time with Magician. I don't have any problem with it. Magician doesn't have any problem with it. What sort of "trouble" of not-playing-well do you mean?

Incidentally. Before I got the BIOS particulars of my system sorted out (from RAID to AHCI-mode), I did a lot of reading about the Sammy drives and the Magician software. I ran across an enthusiast web site in UK -- can't remember which precisely but it was a well-established review site with forums that pops up in searches. People were asking similar questions. The strongest answer I saw suggested that Samsung had focused its efforts with Magician and their hardware on the Intel controllers. There was an explanation as to how the Sammy drives play well with the Intel controllers. If that were the case, I don't think they would've missed much pertaining to an AHCI driver.

oh no trouble bro, just much higher benchmark results while RAPID is on using MSAHCI vs. RAPID with IRST driver which wsa very strange since all SSD enthusiasts usually recommend the IRST driver as it is a better controller driver

 

[G73S]

Hey I remember that app. That is the one that doesn't actually tell you anything. TRIM can be enabled, it is up to the drive for if and when (if ever) a block gets TRIMed, assuming the OS is actually sending the TRIM command. TRIM commands are noncommittal and are not required to performed. The ATA specifications in use now don't even provide a way for the drive to send a success or fail back to the OS. Once drives implemented slack space, TRIM became even less valuable in the consumer space. Add in that the LBA to block mapping is completely up to the drive meaning even if a single 4KB file system page is deleted, the drive may never bother TRIMing. There is an entire long in depth thread in this forum about it.

oh crap, now I am completely lost. so you mean that app is useless and cannot really verify TRIM as it claims?

 

[G73S]

@ G73S:

What is the concrete reason for starting a new thread about a very old topic?
If you would have used the Search function, you would have seen >this< discussion, which has been started by me in February 2013.

thanks i will read that thread seems interesting

and no I don't seach for every topic that comes in my mind before posting it. do you own the forums all of a sudden?

 

[imagoon]

oh crap, now I am completely lost. so you mean that app is useless and cannot really verify TRIM as it claims?

Pretty much. The TRIM command is just a suggestion to the drive. The drive can simply ignore it or delay it. Most drives now rarely process the TRIM immediately. They just queue it in to the garbage collection cycle. Also due to spare area techniques, the drive may not TRIM the block until a write is inbound. Some redirect the write to a spare block then go back and copy the rest of the old block, deallocate the old block and then send the old block to GC to be stuck back in to the spare pool. During these times the old data maybe read back from that block because an incoming write hasn't triggered the LBA remap.

SSD's do a lot more in firmware than magnetic and they are actively being tuned and changed. TRIM simply doesn't have a set standard on how to actually handle and verify the request in this ATA standard.

 

[Carson Dyle]

C:\>fsutil behavior query DisableDeleteNotify
DisableDeleteNotify = 0

 

[BonzaiDuck]

thanks i will read that thread seems interesting

and no I don't seach for every topic that comes in my mind before posting it. do you own the forums all of a sudden?

Sometimes we're all in a hurry and hope someone can provide a quick answer to save us from sifting through volumes of information to find some relevant set of details.

Here's for the record:

And the TrimCheck results:

I hope this puts the matter to rest. As for the benchmarks: It seems I was advised at the time of my first SSD to avoid running too many benchmarks, as the results would slowly get worse. However -- and I could post the screenie if I went through all the steps to get the screen capture -- my RAPID-enabled 840-Pro shows "sequential read" of about 1,200 MB/s, and "sequential writes" of about 1,100 MB/s. For the smaller or random tests, I'd have to run CrystalDiskMark or something. But -- not now.

I'm happy. Happy with my 840-Pro, happy with my configuration and everything is wonderful. :twisted:

 

[Prime_Time]

who the hell are you? and what is OCN? you registered here just to trash talk?

I'm nobody but is you are Maximus666 on OCN (overclock dot net)?

I'm personally doesn't post much so I forget my old Anandtech user but I can't stand your post about TRIM not working on IRST that wrong, Actually TRIM work fine on IRST like many people told you.

I don't mind lying, but I hate inaccuracy.

 

[Fernando 1]

Most users with SSDs install the IRST drivers as they believe they offer a slightly better performance BS the native MSAHCI driver, while that may be true in certain cases (not on my Samsung 840 EVO), you may or may not TRIM functionality which will hurt your SSD's performance in the long run.

Here is my comment:

1. All Intel RST drivers since v9.6.0.1014 (except v12.5.0.1066) do support TRIM, if the Intel SATA ports are running in AHCI mode. That means, that they let the TRIM command pass through the Intel SATA AHCI Controllers into the SSD.
   That is why I don't see any reason for the users of an Intel RST AHCI driver to be in doubt regarding the TRIM activity within the SSD after having verified, that the OS sends TRIM commands at all.
   Things are different for SSDs, which are running in RAID mode or which are even members of an Intel RAID array.
2. Since the TRIM activity within the SSD depends on the amount of TRIM commands, which were recently sent from the OS or from a TRIM trigger tool, it is not always easy to detect.
   The TrimCheck tool is fine and very easy to use, but according to my experience not able to detect very low TRIM activities. That is why I recommend to trigger TRIM (can be done by running the Win8 "Optimizer" or Anvil's Storage Utilities), before executing the TrimCheck tool.
   Users, who want to detect low TRIM activities, should better use the manual Hex Editor TRIM test (look >here<).

 

[BonzaiDuck]

Here is my comment:

1. All Intel RST drivers since v9.6.0.1014 (except v12.5.0.1066) do support TRIM, if the Intel SATA ports are running in AHCI mode. That means, that they let the TRIM command pass through the Intel SATA AHCI Controllers into the SSD.
   That is why I don't see any reason for the users of an Intel RST AHCI driver to be in doubt regarding the TRIM activity within the SSD after having verified, that the OS sends TRIM commands at all.
   Things are different for SSDs, which are running in RAID mode or which are even members of an Intel RAID array.
2. Since the TRIM activity within the SSD depends on the amount of TRIM commands, which were recently sent from the OS or from a TRIM trigger tool, it is not always easy to detect.
   The TrimCheck tool is fine and very easy to use, but according to my experience not able to detect very low TRIM activities. That is why I recommend to trigger TRIM (can be done by running the Win8 "Optimizer" or Anvil's Storage Utilities), before executing the TrimCheck tool.
   Users, who want to detect low TRIM activities, should better use the manual Hex Editor TRIM test (look >here<).

Nobody has mentioned this, and I'm guessing most folks with their iron in this fire are using the Intel onboard controller. My Z68 (ASUS) board has two auxiliary controllers in addition to the Intel: A Marvell 92xx SATA-III AHCI (only) controller, and a JMicron eSATA controller (driving the eSATA motherboard ports).

Obviously, one would feel fairly certain that TRIM works on such a Marvell controller: it uses its own AHCI driver, and doesn't provide RAID. But I thought there was evidence that HDD performance on the Marvell controller was a bit lackluster. Anyone have bench results for an SSD?

 

[Makaveli]

thanks i will read that thread seems interesting

and no I don't seach for every topic that comes in my mind before posting it. do you own the forums all of a sudden?

No he doesn't but he has done more for this forum and its members when its comes to SSD's and raid than you reposting old information you should show alittle respect.

And searching for a topic is a skill you should learn as the mods are very active on anandtech and will close or merge topics that are already open.

 

[Coup27]

This thread should be closed.

 

[BonzaiDuck]

This thread should be closed.

What-evah! What-evah! Truth is -- there were recent threads gone stale covering this topic per the Intel Controllers, RAID and TRIM.

 

[MongGrel]

A few people might have needed a boink on the head though lately, but I doubt they reading it anyway.