Half write speed in SSD

[fw1374]

Hello to all,

I get Transcend SSD320 256 GB a few days ago. Naturally I run some tests. In ATTO I get abnormal write speeds. Read speed is very close to advertised speed (560 MB/s), but write speed is almost half of expected speeds:

http://img199.imageshack.us/img199/600/attossd.jpg

Before any answers things I have checked or done:

-My OS is Windows 7 (Clean Install)

-SSD is connected to Intel Sata 3.0 controller in Sata_0, not Marvel or another controller. (Gigabyte GA-Z68AP-D3)

-AHCI was always enabled

-I tried both Microsoft and Intel drivers (Msahci- IaStor)

-My firmware is latest (5.0.4), checked this from Transcend's website and SSD Scope program.

-TRIM was always enabled, also checked this with Intel's SSD Toolbox and cmd from Windows.

-Write Caching is enabled.

What could be the problem?

Any help is appreciated, thanks.

 

[Dereck47]

-Write Caching is enabled.

Disable Write Caching, re-enable Write Caching, reboot your pc, and re-test with ATTO.
See if that works.

 

[fw1374]

-Write Caching is enabled.

Disable Write Caching, re-enable Write Caching, reboot your pc, and re-test with ATTO.
See if that works.

Still similar results. I also recognised that disabling or enabling Write Caching made almost no difference in benchmark. :\

Add to my first post, I tried Paragon Alignment Tool 3.0, program says SSD is optimally aligned. Alignment doesn't seem to be the problem.

 

[Dereck47]

Connect the drive to the 2nd Intel SATA 6Gb/s port (SATA3_1) and re-test.
If you are getting the same results you might want to look at getting the drive replaced.

 

[GlennSkene]

I have seen similar behavior with a Samsung 840 Pro 256GB. Crystaldiskmark showed sequential read speed of 461 MB / sec and sequential write speed of 266 MB / sec. The problem was with the SATA cable. With another shorter cable, Crystaldiskmark showed sequential read speed of 519 MB/sec and sequential write speed of 502 MB/sec.

Does anyone know if a SATA III controller will drop to SATA II speeds if it encounters errors at SATA III speeds?

...glenn