SSD performance below average?

[Edgar R. Zakarian]

Hi guys.

 

So I took the test of userbenchmark, and it says my SSD seems to be underperforming. Can someone help clarify the reason?

 

 

 

Please see the full test link here: http://www.userbenchmark.com/UserRun/2897618

 

I have the SSD connected to a sata 6 port, as far as I know, this motherboard only HAS the fast type of sata ports.

I've just updated the SSD to the lastest firmare, MU04, not sure what else I could do, to make it perform as expected?

 

Also, is there anything else I could improve on my PC, if you look at the test results? (Misconfigured stuff etc.)

Memory is running dual channel, as far as I know.  Not a Whiz at bios settings

[Phentos]

On your motherboard, is the SATA bus only controlled by the AMD chipset on the motherboard? Or is there a second controller?

 

Some motherboards feature two controllers, and the AMD/Intel chipset controller is usually always superior.

[Mira Yurizaki]

http://www.anandtech.com/show/9258/crucial-mx200-250gb-500gb-1tb-ssd-review/7

 

Actually that looks on par with what Anandtech got, that is, sequential reads (usually the fastest test) are floating around 450 MB/sec. 4K Random is always slow and not really a benchmark you need to worry about

 

Though I would argue raw bandwidth is nothing to be concerned about considering few applications peak 300MB/sec. It's the latency that matters.

[Edgar R. Zakarian]

15 hours ago, Phentos said:

On your motherboard, is the SATA bus only controlled by the AMD chipset on the motherboard? Or is there a second controller?

 

Some motherboards feature two controllers, and the AMD/Intel chipset controller is usually always superior.

 

I have NO idea??

How do I figure that out?

 

It's an Asus Sabertooth 990 FX rev 1.0

[Edgar R. Zakarian]

15 hours ago, M.Yurizaki said:

http://www.anandtech.com/show/9258/crucial-mx200-250gb-500gb-1tb-ssd-review/7

 

Actually that looks on par with what Anandtech got, that is, sequential reads (usually the fastest test) are floating around 450 MB/sec. 4K Random is always slow and not really a benchmark you need to worry about

 

Though I would argue raw bandwidth is nothing to be concerned about considering few applications peak 300MB/sec. It's the latency that matters.

 

I can see that one is green / fine as well. And you are right.

On MX200

They get something like 412 MB/s

I get 429 MB/s

 

But what about the "Deep Queue 4k" performance?

I'm guessing that is the same as this? http://www.anandtech.com/show/9258/crucial-mx200-250gb-500gb-1tb-ssd-review/8

 

They get 71.3 MB/s

I get 59.5 MB/s

 

Which is a bit slow. Just wondering, is there anything I could do, or is it just an unlucky specimen?

[Mira Yurizaki]

1 hour ago, Edgar R. Zakarian said:

I can see that one is green / fine as well. And you are right.

On MX200

They get something like 412 MB/s

I get 429 MB/s

 

But what about the "Deep Queue 4k" performance?

I'm guessing that is the same as this? http://www.anandtech.com/show/9258/crucial-mx200-250gb-500gb-1tb-ssd-review/8

 

They get 71.3 MB/s

I get 59.5 MB/s

 

Which is a bit slow. Just wondering, is there anything I could do, or is it just an unlucky specimen?

As I said before 4K performance is always slow and isn't really a benchmark you have to worry about.  4K performance basically means "how fast can I randomly get the smallest amount of data?"

[Phentos]

9 hours ago, Edgar R. Zakarian said:

I have NO idea??

How do I figure that out?

 

It's an Asus Sabertooth 990 FX rev 1.0

When you boot your computer, does it initialize something related to Marvell?

 

For my motherboard, which is a Sabertooth P67, there is a Marvell SATA controller for two of the SATA ports. It initializes on boot and it shows it on the screen when it does. The other four SATA ports are directly controlled by the Intel chipset.

 

And like the above poster said, 4k speeds are not the most important factor. It's all about latency, queue depth consistency, and how the SSD controller deals with garbage collection and write amplification.