How fast (really) should a 3.0Gb/s Hard Drive be?

[golgii]

The screenshot below shows the test I ran on my hard drives. I removed the jumper so I should be capable of 300MB/s. 20.8MB/s seems awfully slow. Any suggestions, comments?
Here is my setup:

Hard Drives
2 x Seagate Barracuda 7200.11, 1TB, 7200 RPM, SATA 3.0Gb/s Internal Hard Drive, Average Latency 4.16ms.
- Both setup in Raid 1 configuration (mirroring)

Processor
Intel Core 2 Duo E8500 3.16GHz

RAM
CORSAIR Dominator 4GB (4 x 1GB) SDRAM DDR2 800 Dual Channel

Motherboard
ASUS P5Q Pro w/ Vista Ultimate 64-bit

 

[warrenr333]

Somthing not right there i have a WD 320GB SATA 3.0Gb/s HDD

I got

Random Read : 63.51 MB/s

Buffered read : 180.71 MB/s

 

[MrLinux]

From the trace, it looks like you have something interfering with drive access, do you have an anti-virus on-access scanner running?

Using motherboard RAID and an on-access scanner will seriously hit your disk access speed due to both trying to use the CPU at the same time.

 

[Paperdoc]

Yeah, the benchmarks I've seen show SATA 3.0 Gb/s averaging about 40 to 60 MB/s random read / write. Your score is low.

 

[cjl]

40-60MB/s random is a pipedream, and not happening anytime soon. Sequentials should be somewhere between 50 and 120 MB/s on any modern drive though, so there's definitely something wrong with that trace. Access time should be between 7 and 18 ms as well. Check to make sure that there isn't any background activity that is accessing the drive, like a virus scan or a backup.

A 7200.11 test should look more like this:

 

[Paperdoc]

"40-60MB/s random is a pipedream, and not happening anytime soon. "

Check these two 2-year-old reviews from Tom's:

http://www.tomshardware.com/reviews/icydocks-mb559-happily-marries-esata-usb2-0,1347.html

http://www.tomshardware.com/reviews/seagate-500-gb-external-hard-drive-esata,1307.html

I must admit, after reviewing them, the "average data transfer rates" of 40 to 60 MB/s on eSATA do NOT specify random access. But they do show a good range of performance at that time, and the superior speed of eSATA over USB2. Note that Firewire 400 is almost as fast, and Firewire 800 is the fastest if you can get it. Or at least, that's how it was 2 years ago.

 

[cjl]

It definitely isn't random if it's that speed.

Oh, and eSATA should be the best, followed by FireWire 800.

 

[golgii]

I believe the hard drives were being sync'd via some RAID software when I ran the first test. I reran the tests and the new results are below. From what everyone says, these readings are normally what to expect.

 

[cjl]

That still isn't quite right, though it is certainly much better. It should look more like the one I posted, without the huge dips.

 

[br3nd064]

Yes, it still looks like something is accessing the hard drives when you're running the test. If you go into the vista "Resource Monitor", is the disk graph showing any major activity before you run the test?

 

[golgii]

I disabled startup services and got a worse result. So I again disabled all non-core services, and restarted in Safe Mode. This is probably the best I can get. How come the results tend to decline towards the end?

 

[cjl]

The decline is to be expected, although yet again, that isn't as fast as it really should be. The one I posted is what an ideal trace should look like, with the continual decline being a result of the progression from the outer diameter to the inner diameter of the disk.