Jump to content
Corsair Community

Corsair F60 on AMD 780G-SB700 - MSahci vs AMDahci


PierreBr

Recommended Posts

I have made use of a an Asrock 780G/SB700 motherboard (which supports AHCI) to install my Corsair F60 drive.

 

I tested both with MS and AMD Ahci drivers...both seem to support TRIM (even though an AMD Support communication was quoted to me, in which it was said that TRIM is supported only for newer 8** AMD Chipsets)

 

The performance is good but not great, and both drivers have stronger and weaker areas of performance.

 

Here are the bench results with MS AHCI

attachment.php?attachmentid=7781&stc=1&d=1293212363

attachment.php?attachmentid=7782&stc=1&d=1293212363

attachment.php?attachmentid=7783&stc=1&d=1293212363

 

 

And here are the results with AMD AHCI

attachment.php?attachmentid=7784&stc=1&d=1293212468

attachment.php?attachmentid=7785&stc=1&d=1293212468

^^ This result is weird...because the first time I run it after install AMD ahci drivers, the write performance in the sequential test was a little short of 100mb/sec and the 4k 64Thrd numbers were much better also...after I went back to MS ahci and then switched back to AMD ahci, these tests were negatively affected. The overall score the first time round was 346...

attachment.php?attachmentid=7786&stc=1&d=1293212548

This is the result I got running the bench again, after having cycled between the drivers

attachment.php?attachmentid=7787&stc=1&d=1293214077

 

So, what's your take on these numbers? The AMD ahci drivers does give better performance in 4k write, but may suffer elsewhere...

Also, could the severe degradation in performance while running with the AMD AHCI drivers suggest that they in fact do not support TRIM even though crystaldiskinfo and that command-prompt query show it as enabled?

Link to comment
Share on other sites

I have made use of a an Asrock 780G/SB700 motherboard (which supports AHCI) to install my Corsair F60 drive.

 

I tested both with MS and AMD Ahci drivers...both seem to support TRIM (even though an AMD Support communication was quoted to me, in which it was said that TRIM is supported only for newer 8** AMD Chipsets)

 

The performance is good but not great, and both drivers have stronger and weaker areas of performance.

 

Here are the bench results with MS AHCI

attachment.php?attachmentid=7781&stc=1&d=1293212363

attachment.php?attachmentid=7782&stc=1&d=1293212363

attachment.php?attachmentid=7783&stc=1&d=1293212363

 

 

And here are the results with AMD AHCI

attachment.php?attachmentid=7784&stc=1&d=1293212468

attachment.php?attachmentid=7785&stc=1&d=1293212468

^^ This result is weird...because the first time I run it after install AMD ahci drivers, the write performance in the sequential test was a little short of 100mb/sec and the 4k 64Thrd numbers were much better also...after I went back to MS ahci and then switched back to AMD ahci, these tests were negatively affected. The overall score the first time round was 346...

attachment.php?attachmentid=7786&stc=1&d=1293212548

This is the result I got running the bench again, after having cycled between the drivers

attachment.php?attachmentid=7787&stc=1&d=1293214077

 

So, what's your take on these numbers? The AMD ahci drivers does give better performance in 4k write, but may suffer elsewhere...

Also, could the severe degradation in performance while running with the AMD AHCI drivers suggest that they in fact do not support TRIM even though crystaldiskinfo and that command-prompt query show it as enabled?

 

Command prompt or CrystalDisk are no indicator that TRIM is working. CrystalDisk shows that the drive supports it. The command prompt trick shows that it is enabled. Neither of these confirm that the driver is passing it to the drive.

 

The latest AMD driver (10.12) looks like it has removed support for TRIM. People keep posting articles from November which predate the new driver where support looks to have been removed.

Link to comment
Share on other sites

Well, having read the latest posts in the thread in the forum, it seems that this might not be the case about the drivers, i.e. AMD having removed TRIM support...rather that support for TRIM is not a new feature for the latest drivers...

...a member also quoted an AMD personnel response that tRIM is supported for 8** chipsets version and upwards...

This might in fact be the case...at least the performance degradation in my case seems to support it...

 

EDIT:

Well, actually, my example only provides support for the case that with SB700 and AMD ahci driver TRIM is not functioning...whether this is because of drivers' generally unsupported feature or of unsupported feature for the combination of drivers and chipset, it is not clear...

Link to comment
Share on other sites

* To the administrator:

Some of the images I had embedded in the first post have vanished. I don't know why.

 

To the point:

The initial performance with the MS-AHCI VS AMD-AHCI drivers was slightly in favor of the second...

With AMD-AHCI driver I got a bit slower performance in sequential & QD324K read tests, but I got some more significant performance increases in 4K Write tests (additional 20mb/s + in write)

 

With MS-AHCI drivers the write performance did seem to suffer in 4k write tests, never exceeding 48mb/s...

 

That's why I had decided to keep the AMD-AHCI drivers, until I realized that they did not allow me to update firmware and that probably TRIM does not work with them, since I got a significant performace degradation within a couple of days.

 

Since the initial benches posted above, I have updated to firmware 2.0 and secure erased the drive.

 

However, I am seeing significant performance degradation with MS-AHCI drivers also...at least with AsSSD bencmark and CrystaldiskMark in default configuration.

 

 

- After secure erasing-formatting the drive I got these results with CrystalDiskMark in default configuration

http://img574.imageshack.us/img574/9623/crystaldiskmark25122010.jpg

 

- Two and half days after that, I rerun the test in default setting and got the following results...as you can see, the write performance has deteriorated significantly:

http://img820.imageshack.us/img820/6016/crystaldiskmark28122010.jpg

http://img338.imageshack.us/img338/8996/asssdbench28122010.jpg

 

 

With ATTO and CrystalDiskMark run in 0x0 0Fill configuration, the decrease was not obvious:

http://img337.imageshack.us/img337/6016/crystaldiskmark28122010.jpg

http://img522.imageshack.us/img522/8963/atto28122010a.jpg

 

 

So, I am again puzzled...the performance deterioration with the AMD drivers was not worse than that with the MS drivers, and in a very short time and in a really, really far from intensive use.

 

So, what's going on? I can't seem to draw a conclusion from this...

Link to comment
Share on other sites

Is there a way to definitely determine whether TRIM is actually enabled and used?

 

Based on the results I'm having with MS ahci drivers, there does not seem to be any difference with AMD ahci, the latter presumably not supporting TRIM for this chipset or generally...

...which can mean that either TRIM is not enabled in both cases or that it is enabled but it is not effective...

 

Concerning system use, I can say that I have many tabs open in Firefox and the system stays open 24/7 in P2P...other than that, it's just doing Word processing, video playing and periodically video recording/encoding (NOT on the OS drive)

Link to comment
Share on other sites

TRIM Frequently Asked Questions

 

 

How to check if TRIM is active

 

To check if the TRIM command is active on your PC, start a Command Prompt window (type “CMD” in the Search bar from the Windows Start Menu) and enter the following command:

 

fsutil behavior query DisableDeleteNotify

 

If the result is “0” then the TRIM command is enabled, and if the result is “1” then the TRIM command is disabled.

 

Use the following command to enable TRIM:

 

fsutil behavior query|set DisableDeleteNotify = 0

 

The DisableDeleteNotify command only indicates that the operating system is passing the TRIM command on to the storage drivers. It does not indicate whether or not the storage drivers are passing the command on to the storage controller IC in the SSD, or whether the storage controller IC supports TRIM. As such, as result of “0” is not a guarantee that TRIM is functioning correctly.

 

To determine whether TRIM is functioning correctly, you can periodically measure the performance of your SSD using tools such as ATTO and CrystalDiskMark. If the performance of your drive is generally at the level specified on the Corsair website for your specific drive then you can be confident that TRIM is functioning.

 

Please note that the performance levels of all SSDs will fluctuate from the maximum due to normal usage, so you should not expect the drive to produce completely consistent benchmark numbers, or attain the maximum theoretical performance. Only a clean (Secure-Erased) SSD will deliver the absolute maximum theoretical performance figures.

Link to comment
Share on other sites

Thanks.

 

I had read that...TRIM seems to be enabled with regard to registry settings, but the benchmarks indicate a noticeable difference within a few days...

...so I was looking for a way to determine whether TRIM is actually functioning in all respects or not and if it is, whether something I'm doing is producing these mediocre results...

 

I have no experience with SSDs and I don't know if this sizeable performance drop is "normal" or indicative of non-functioning TRIM...this is the main question...I notice here again that this performance drop is measurable and obvious only with CrystalDiskMark in default settings and As SSD Benchmark.

 

If I had to guess, I'd say that TRIM is probably not functioning...or maybe I hope it isn't and that it can be fixed, otherwise, if performance deteriorates at such pace, maybe I don't need an SSD...

Link to comment
Share on other sites

There will always be performance degradation, especially if a) you are filling the drive b) you make constant writes and deletes eg. a lot of benchmarking.

 

IMO, the crystalmark benchmark is somewhat unreliable and inconsistent. I would stick with ATTO.

 

Another thing that Wired didn't point out about TRIM is that the only way you can tell if it's WORKING is if you put your computer to sleep after a lot of writes/deletes and then come back and benchmark to see the difference. If you're not letting your computer idle or put it to sleep in between benchmarks..you won't see the advantage of TRIM. TRIM will only kick in when the drive is not doing anything.

Link to comment
Share on other sites

There will always be performance degradation, especially if a) you are filling the drive b) you make constant writes and deletes eg. a lot of benchmarking.

From what I gather this is a general principle...stated as such, however, it does not answer the issue at hand: is such performance drop justified on the basis of 1) time elapsed between the secure erase/install and the point where the bench tools were rerun 2) the use of the computer during that time.

 

As far as I'm concerned, the facts concerning the above two questions are quite discouraging...two days passed, two series of benchmarks (6 benches total) run, 10-12 programs installed (with Office being the only major one) plus all the windows updates, casual internet browsing (this for me includes lots of tabs, however)...and the result is 1/3 decrease in the performance in some write tests...

 

IMO, the crystalmark benchmark is somewhat unreliable and inconsistent. I would stick with ATTO.

I can't know that...however, if it is unreliable, it would be unreliable the first time I ran it also...but there has been a great decrease in bench results between executions of this same program, I am not only comparing its results with those of another one...the results are also confirmed by As SSD benchmark.

 

Another thing that Wired didn't point out about TRIM is that the only way you can tell if it's WORKING is if you put your computer to sleep after a lot of writes/deletes and then come back and benchmark to see the difference. If you're not letting your computer idle or put it to sleep in between benchmarks..you won't see the advantage of TRIM. TRIM will only kick in when the drive is not doing anything.

I read somewhere here in the forums that this is not a condition anymore for modern SSD drives and that TRIM works continuously, not only in idle or sleep periods...

 

Maybe someone else can confirm that...I will edit this message if I find the thread where this was mentioned.

 

In case what you say is true, do you know if logging off and putting the pc to sleep mode works the same for TRIM?

If I log off my user account, will programs like utorrent continue to run actively or will they stop interacting?

 

Thanks a lot for your support and suggestions

Link to comment
Share on other sites

Well, I did log off for about 8 hours and performance keeps degrading...which means that if TRIM is not working (which does not need the user to log off), then GC is not working either...

...this is really frustrating...it suggests that I might have to run secure erase every few days to prevent the performance from dropping 1/3 in some write tests within a matter of days and with very casual system use.

attachment.php?attachmentid=7805&stc=1&d=1293611558

The results with As SSD Benchmark are even more staggering in performance drop measured.

attachment.php?attachmentid=7806&stc=1&d=1293613395

 

But why isn't TRIM and maybe even GC working? When TRIM is enabled, the performance restoration is quite surprising, if one looks at relevant tests

 

To keep the puzzlement alive, the bench results of CrystalDiskMark set at 0x00 0Fill setting hardly seem changed...

attachment.php?attachmentid=7807&stc=1&d=1293613483

crystaldiskmark-29122010.jpg.adfaa45477ec62c9221927821bd5bf1e.jpg

asssdbench-29122010.jpg.cb7468e435e9d873312baec73e295a10.jpg

crystaldiskmark-29122010-0x0-0fill.jpg.720dba896c4b7586938380a268f26f7b.jpg

Link to comment
Share on other sites

Things are getting frustrating a bit...

 

I did a real-world bench to test the performance of the my Corsair F60 by copying two video files of 1,96GB and 880mb file size from my HDD to my SSD and it did not measure above 36-7mb/sec...(according to HDD sentinel)

That's at the same speed as the one measured by As SSD Benchmark...

 

I'm really starting to believe that TRIM is not actually functioning even with the MS ahci drivers...

 

After installing OS + drivers the performance was measured at high level...so it seems that the installation of the programs, the use of the page file on the SSD and FF browsing has impacted the performance in my case.

What seems weird, though, is that apart from MS Word, which is a large program, all the other software that was installed, were both small in number and small in size....antivirus software, K-lite codec pack, CCleaner, FF, Opera, keyboard/mouse software, an Uninstallation program, a registry fix program, Notepad++, Filezilla, HDD sentinel (yesterday in order to test the transfer speed) and two very small dictionaries...that's all I have installed...

I can't really believe that the installation of those programs - after the OS + drivers were installed - and the relaxed use of Firefox for even less than 2 days has impacted the performance by such a degree...concerning sequential write speed, that's a 2/3 deterioration in performance.

That's not easy to accept, unless maybe TRIM is not working

 

Another reason why I believe TRIM is not working, is that I got the same results with the AMD ahci drivers within the first two days of installation/use...and people here at the Corsair forums have quoted support questions/answers to AMD in which it was said that TRIM was supported only for the newer 800 series chipsets...

- So, if MS ahci drivers are giving me results that show the same level and rate of degradation (or worse) as the AMD ahci drivers, which by all accounts don't seem to support TRIM for my chipset, then doesn't that mean that TRIM is probably not functioning in my case?

 

Assuming that FF has played a part in performance deterioration, last night I deleted all history, cache etc closed the program + I deleted the video files I had copied to the SSD and left the pc idling overnight...I rerun the AS SSD bench this morning and did not get any improvement measured (2-3mb/s can be a difference between different runs of the bench test)...shouldn't that give me some practical indication of functioning TRIM?

I understand this is not some large volume of data, but then again this very small volume of data seemed to cause the performance deterioration...

Link to comment
Share on other sites

I am experiencing a similar slowdown.

 

I have the lates Intel RST 9.6.0.1014 drivers.

RAID (AHCI implied in Dell bios) is ON and drives in non-RAID mode and the driver supports TRIM for this configuration.

I checked and TRIM is enabled. Is it working?

 

It is the sequential random reads that go from 100 MB/s to around 60 MB/s:

1)First time. I was testing the new drive with just a lot of data and this drooped after a day even after erasing data and formatting drive.

I did a secure erase and this restored it to 100 MB/s read.

2) Second time. I then installed W7 Pro and it is now doing the same thing after 2 days.

 

Is this normal?

How do I verify that TRIM is indeed working? It is enabled and the driver supports it.

Link to comment
Share on other sites

Ok, I did a little bit more research today and it seems the issue could be complicated...

 

The following are based on information drawn from sources cited below

 

- Sandforce SSDs do not play well even with the default MS ahci driver...this may be the fault of both of them, MS's working for a universal driver and SF not doing optimizations...testing for TRIM support with Sandforce SSDs seems to yield confusing results.

- Secondly, Sandforce TRIM/GC are very mild, and allow much degradation to happen before kicking in, and they may be slow at it also. They could have opted for such configuration in order to benefit longevity.

- Support for TRIM seems to depend on many factors: OS > Os_driver > Chipset_Controller > SSD_Firmware > SSD_Controller...

- There is no evidence or proof (or real indication) that 780G/SB700 chipsets don't support TRIM, but on the AMD side driver-support is the main issue...a TRIM supporting driver was released on September (?), but it was withdrawn...even on that case it is unclear if SB700 sb was supported by the driver to implement TRIM

- Corsair might be pushing for Intel controllers/shpisets for its Sandforce SSD drives (e.g. I had asked about getting an addon PCIe controller and was advised to get a specific chipset Intel-based one), because the intel released driver overcomes the shortcomings of the Sandforce-MS match...

 

 

AMD and TRIM

780G + Sandforce

AMD AHCI 10.9 and Trim?

Trim Test

AMD AHCI Driver SSD TRIM Support

 

 

Could we get more detailed info and commenting on the TRIM workings for Sandforce SSDs?

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...