SATA 6Gb/s / SATA III - PCIe Card - Works in LION !

[Uglycasanova]

Getting the same speed I was when using my on board SATA 2 ports...But my ASMedia1061 card is recognized in my system info.
Anyone have the same problems?
Im using a GA-H61N-USB3 mobo and ocz agility 3 ssd, getting 180mb read and 70mb write...

 

[fishead]

Corsair swapped out my SSD and it now benches over 450Mbs write/520Mbs read in Windows 7. When I use Black Magic Speed Test with Lion I get 140/200. When I use AJA System Test I get 450/460. Doesn't Matter if I'm on the Marvel, ASM1601, or the Native Mobo Sata III Controller. Starting to think Black Magic speed test is ****. I know I heard the compressible data argument but when I ran on a cheaper Gigybyte 1155 board w/i5 using Snow Leapord I was benching over 450read/over 500write. Is there something Lion/Black Magic Test related? I remember I had to install SL using the Sata II port and after could move it over to the Sata III otherwise the install crawled. Funny thing. I wish I still had that board. I might just install SL using the old method just for kicks. The machine is running great besides a shitty Black Magic test. Maybe I should just waste time on other senseless projects.

 

[ectoplasmosis]

The machine is running great besides a shitty Black Magic test.

There's nothing wrong with your drive or the Blackmagic test. It's the way Sandforce-based drives are by design: viewtopic.php?p=366605#p366605

Avoid Sandforce if you need good incompressible performance/Blackmagic benchmark scores.

 

[149113]

Have it in a 16x PCIe 2.x slot on a EX58-UD5 (v1.0) and getting the following in AS SSD score:

That reviews say this SSD should get 500 MB/s Read & 500 MB/s Write. So what gives? I have a GT9800 in the other 16x slot. Do I need to move it to another slot?


*** EDIT ***
I have now moved it to all available PCIe 2.x slots and the scores do not improve from what was posted below. I see that other folks with the same MB (EX58-UD5 (v1.0) are getting much higher scores. So where is the difference?

 

[149113]

Here is my Blackmagic score...

Seems way slow for a drive rated as high as this one. Has to be the controller.

 

[thelostswede]

Here is my Blackmagic score...

Seems way slow for a drive rated as high as this one. Has to be the controller.

Ugh... Blackmagic has changed their benchmark, so if you have an SSD with a SandForce controller, then it WILL PERFORM SLOWER than the rated speeds due to the fact that they're using INCOMPRESSIBLE data. SandForce controllers reach those kind of speeds with help of data compression, so if you use a test like the Blackmagic one or AS SSD which tests using incompressible data, then the scores will be MUCH slower than the claimed speeds for the drives. Try a different benchmark like CrystalDiskMark and you'll see a massive performance increase.

On top of that, any third party controller will NOT perform as well as the Intel controllers, although this might not apply to the latest generation of controllers from Marvell IF connected to two PCI Express lanes, but this would either require that the board maker has done this or that the controller is fitted to a x4 PCI Express card.

Please do your research before you start blaming your components, as the past two posts are clearly a case of users not knowing what they're doing.

 

[149113]

Here is my Blackmagic score...

Seems way slow for a drive rated as high as this one. Has to be the controller.

Ugh... Blackmagic has changed their benchmark, so if you have an SSD with a SandForce controller, then it WILL PERFORM SLOWER than the rated speeds due to the fact that they're using INCOMPRESSIBLE data. SandForce controllers reach those kind of speeds with help of data compression, so if you use a test like the Blackmagic one or AS SSD which tests using incompressible data, then the scores will be MUCH slower than the claimed speeds for the drives. Try a different benchmark like CrystalDiskMark and you'll see a massive performance increase.

On top of that, any third party controller will NOT perform as well as the Intel controllers, although this might not apply to the latest generation of controllers from Marvell IF connected to two PCI Express lanes, but this would either require that the board maker has done this or that the controller is fitted to a x4 PCI Express card.

Please do your research before you start blaming your components, as the past two posts are clearly a case of users not knowing what they're doing.

Understood on the SandForce controller issue but there are folks in this thread that posted benchmarks with the same controller, same SandForce and same MB (Gigabyte EX58) that are getting much higher scores. Just browse the thread to see what I am referring to. Also I have tried Crystal DiskMark and here are those scores... again these are not great by any means and do not compare to the numbers posted in online reviews done by third parties.

I can understand a slight difference between a native Intel SATA III and an add on card such as this. But The difference here is almost 30-40% less performance and is closer to SATA II.

 

[thelostswede]

You're clearly not reading what I said.
Blackmagic CHANGED their benchmark, so post earlier in this thread is using the OLD benchmark with used compressible data, the NEW benchmark uses incompressible data, hence the difference in performance.

The CrystalDiskMark numbers aren't very impressive, have you tried moving the card to another slot? Some of the x1 PCI Express slots share bandwidth with other devices and as such you might not be getting the full bandwidth needed for the SATA controller. Sadly it looks like Gigabyte doesn't provide this information in their manuals any more, so I couldn't say which ones are shared and which ones aren't, with the exception of the ones connected directly to the CPU.

 

[149113]

You're clearly not reading what I said.
Blackmagic CHANGED their benchmark, so post earlier in this thread is using the OLD benchmark with used compressible data, the NEW benchmark uses incompressible data, hence the difference in performance.

The CrystalDiskMark numbers aren't very impressive, have you tried moving the card to another slot? Some of the x1 PCI Express slots share bandwidth with other devices and as such you might not be getting the full bandwidth needed for the SATA controller. Sadly it looks like Gigabyte doesn't provide this information in their manuals any more, so I couldn't say which ones are shared and which ones aren't, with the exception of the ones connected directly to the CPU.

Clearly you are not either reading or understanding what I am saying. I understand the Blackmagic scores. I know they are lower on a Sandforce controller because of the compression/uncompressed issue that impacts that specific controller. Those scores were only included as an additional data point to underscore the overall point I am making... The AS106x Asmedia SATA III controller is not performing at SATA III speeds in my EX58-UD5 per the AS SSD and CrystalMark scores I have provided.

And yes, if you re-read my first post you will see that I have placed the AS106x Asmedia SATA III controller in EVERY PCIe slot (1x-16x) that this motherboard allows. It makes no difference in the scores.

So again... the question is why is this card not getting SATA III speeds in my EX58-UD5 (v 1.0)?

 

[thelostswede]

So you're expecting me to read every single damn post in this forum?
I missed the earlier post and I'm sorry I didn't see it, but considering just about everyone else has managed to get the card working... and on top of that you keep going on about benchmark scores and seemingly ignoring any replies to your posts as to why the scores are lower than you expected.

Two things you can try, 1. upgrade the motherboard BIOS/UEFI and 2. upgrade the firmware on the card, there should be information earlier in this thread about how to do that.
I very much doubt there are any other solutions that will solve your problem.

Besides, we're not tech support, this is a forum and we try to help the best we can. If you run into a problem that no-one has experienced or have managed to solve, how can you expect a solution to it? **** happens, it's hardware and some stuff just doesn't work, no matter what, such is life when it comes to computers.