G73JH-A1, OCZ Vertex 2, Terrible Write Speeds

-----------------------------------------------------------------------
CrystalDiskMark 3.0.1 x64 (C) 2007-2010 hiyohiyo
Crystal Dew World : Crystal Dew World
-----------------------------------------------------------------------
* MB/s = 1,000,000 byte/s [SATA/300 = 300,000,000 byte/s]

Sequential Read : 203.970 MB/s
Sequential Write : 41.076 MB/s
Random Read 512KB : 196.728 MB/s
Random Write 512KB : 45.250 MB/s
Random Read 4KB (QD=1) : 19.288 MB/s [ 4709.0 IOPS]
Random Write 4KB (QD=1) : 41.404 MB/s [ 10108.4 IOPS]
Random Read 4KB (QD=32) : 100.065 MB/s [ 24429.8 IOPS]
Random Write 4KB (QD=32) : 38.723 MB/s [ 9453.9 IOPS]

Test : 50 MB [C: 34.1% (19.0/55.8 GB)] (x3)
Date : 2011/02/14 22:39:44
OS : Windows 7 Home Premium Edition [6.1 Build 7600] (x64)

I'm a bit stumped on the terrible write speeds I've been getting, even the read is considerably lower than what I've seen other people getting. Not to mention I'm getting HDD speeds from my brand new OCZ SSD.

I've followed two guides on tweaking registry values, bios settings and changing device manager settings, really to no avail.

I just recently installed the Intel RST drivers, latest version on the website. The above numbers from CDM reflect no gain or relative loss in doing so. I did a clean install of Win7 when I got the SSD and made it the main drive, while slaving a 500GB HDD as a secondary drive; a USB2.0 External enclosure with another 500GB HDD in it.

I am under the belief that I have updated all drivers to the most recent. Is there something simple I probably am missing here?

Please and Thank you!

More information may help.

I have made sure I am in AHCI mode, both in BIOS and Win7. Is there a command line that gives me a direct confirmation, though? I am positive the hex value is "0" in the msahci portion of the registry.

I have unparked the cores and followed the following guides:

http://forum.notebookreview.com/har...-intel-series-4-5-965-chipsets-jjb-tweak.html

The SSD Optimization Guide | The SSD Review

Intel RST version 10 Drivers, INF Drivers updated as per Chastity's Drive and Application post on the ASUS ROG Forums.

Not that I think it matters, but also soft modded for the X-Fi MB sound drivers.

The only thing I can really say I haven't done yet is disabled the Pagefile, which from what I understand really only opens up data space on the SSD; not necessarily a performance boost? - Going to do that right now as I may as well anyways with 8GB of RAM.

Any thoughts or ideas would be appreciated. I will post back with new CDM results after a little while of playing around..

 

Well, on my G53 I only used the JJB Tweaks and you can see from the screenshots posted in that thread that it worked.
I have to mention that i also used the commands that makes the 6-7 hidden tweaks to appear in the W7 Power Plans and made the modifications in the P4G max performance plan not directly in the registry.


I'm using the latest Intel RST drivers and i made sure that Trim is enabled.
I also made sure that my partition is properly aligned (not entirely convinced that it matters in terms of performance, but it cannot make things worse... hopefuly )

PS. I didnt disabled the page file nor i did any of the ssd "tweaks" except for checking that defrag is disabled for the drive C:
Most of the SSD tweaks are not geared towards performance per-se, but more to minimize the space taken by the OS and also to minimize the number of writes to the drive... Most of them i find to be counterproductive.

just seen it:
Could it be this? The new OCZ with 25nm 64GB chips that have almost half the write performance?

 

There's a nice writeup on the OCZ forums, explaining how crystal mark isn't the right benchmark tool to use...

Though I get close to the same ATTO results, my crystal mark results are crazy low like yours (lower, actually).

And I've done all the tweaks and such. Maybe that's why I'm not nearly as impressed with my SSD purchase as others have been.

I guess I should have forked over the extra $30 and got an intel drive >_< Also, the read speeds are still much faster than the hard drive across the board, and everything but sequential is also much faster. In theory.

 

There is setting on crystal that needs to be changed. Instead of "random" data select 1 or 2. Then rerun the test. Let us know what you get. DCX is correct, crustal is bad unless you change the value, ATTO is what my manufacturer uses. When I change the test values, my results are almost identical to published specs with Crystal though.

Hope that helps.

 

Yeah, let the Sandforce compress 0-data. Then you will get the pronounced speed. But you don't have such compressable data in real life.

...and only 1000 pe-cycles, and 5gb less space (60 + 120gb drives).

 

Yes this is the reason, the 25nm... Guide New update on the 25nm OCZ SSD drives

 

Set it to 1, and the speeds jumped by quite a bit - all the way up to manufacturers spec. It's one more reason why synthetic benchmarks aren't great ... though I stand by furmark as a good stress test