Solid-State Drives (SSD)

[johnyeah2]

So here's what going on with my SSD.

I have updated the firmware to 1819 the first day it came out. I have also ran the wiper tool and checked for proper disk alignment. However, my sequential write performances have been terrible, even after the wiper tool (seems like the same problem has been plaguing many other 1819 firmware users). It was originally at 100 mb/s. I used crystaldiskmark to test this.

I have noticed though, I haven't performed wiper for a week, and my write performances (sequential write) went up to 115mb/s. And now I haven't used the wiper tool for two weeks, and my performances have been up to 138 mb/s. Of course, it's still far from the advertised 190 mb/s. But at least the benchmarking performances have been increasing. I was wondering if anyone had an explanation to this weird phenomenon.

[johnyeah2]

That's the current benchmark I got with my M225 128GB.

[targetbsp]

I can get 10's of megabytes of difference from 1 benchmark to another in a space of a couple of minutes.  1819 seems to invariably produce lower - but not consistent benchmarks.  I've also seen a recovery of about 20mb/sec simply from rebooting.  A lot of the theory into the problem on OCZ's forum (and sfaure of this forum has noted similar in his own experiments) is that the controller spending so much time trying to sort out the trim in conjunction with Windows 7 that it has less time to handle our own requests to read and write.  So your benchmark performance can vary based on how busy the controller is.

[sfaure]

Question:

Did you update to 1819 w/ enabled trim on an existing install?...  

My thinking is that you suffered degredation until the trim/free lists of the new firmware actually "matched" the drives use/non-use of blocks. 

This would jive up with some who have increased performance (myself included) by running the OCZ SecureErase tool, which completely cleans the drive.

[johnyeah2]

I updated 1819 first. Then I proceeded with a clean install of Win 7 64-bit. I ran wiper a couple times before and I've consistently got only 100mb/s writes after. And now, after not running wiper for a while, I've consistently gotten 130 mb/s+ on my CrystalDiskMarks.

[Tekno]

Hrm, that is interesting. Since your last wiper run, have you been using the drive heavily? Meaning, have you filled the drive to near capacity and then deleted files to free up space? You say you used it for a long while with running wiper, but I wasn't sure what you mean by use. Can you clarify?

[johnyeah2]

No, I never filled my SSD up (other than when wiper is executed). I only use my SSD to run boot Windows, run programs, games and browsers (which store temproary cache files).  There's still 80gb's of free space left. I don't do any "heavy writing" work, I just use it like a boot drive.

I'll try running wiper in a bit and check my benchies afterwards...(hopefully, it won't drop back to the 100 mb/s region again, lol).

[johnyeah2]

So my results.

Before running wiper: 205 mb/s sequential read and 128 mb/s sequential write.

After running wiper: 202 mb/s sequential read and 103 mb/s sequential write.

I did the benchmarking with CrystalDisk right before I ran wiper and right after I ran wiper (and I also restarted my comp and benchmarked again, results were the same).

[targetbsp]

There's all sorts of stuff you can play with to improve the performance: free space defrag in something like perfectdisk, AS freespacecleaner, wiper but they only seem to improve it temporarily, take a load of time and add wear to the drive - particularly the benchmarking.  So I'd just wait for the new firmware if I was you. As for the changing benchmarks - I can run 2 consecutive benchmarks and get results differing by the amount your seeing.  It just doesn't produce consistent performance for some reason.

[sonicsonic]

the best method i found is to leave the ssd alone and not to wiper it or anythink else until normal firmware comeout maybe with GC support