Intel X-25 G2 160GB SSD: extraordinary sequential write, albeit mediocre 512k

Hi all,

This is what I get right after consolidating free space with Perfectdisk 10 build 129. Can't even get this number after running the Intel SSD toolbox.

512k, however, had reached 107MB/s before and it's now only 83.3MB/s

 

try using atto disk benchmark so you can better see what's going on.

http://www.attotech.com/software/files/drivers/bench32.zip

 

Tried. This is the result (Is this benchmark optimized for 64-bit OS?)

 

You know that defragging a SSD is NOT recommended?
Wait a couple of days so it can recover.

 

Yep. I've read a lot about that. But there is another camp saying two defrag programs (PerfectDisk/Diskeeper) do not actually defrag the disk per se. They rearrange and consolidate the free space. I am not sure how the operations compare to TRIM, but the performance boost is pretty obvious to me, noticeable even in regular usage rather than benchmark alone.

What would you think of this?

 

there's no need.

there's no need to defrag an SSD.   consolidating unused space is unnecessary since data isn't read with a magnetically-driven read/write arm like on an HDD.   SSDs work like a small batch of RAID volumes.   the controller chip can read from each storage chip simultaneously if necessary before combining the data for final output.

 

Well, that is defragging. Per Se or not
Again, I would wait until TRIM and Garbage-Collection did their job and rearranged "clusters" with their own algorithm.

Did you reboot and just let it sit?

 

They technically cannot rearrange files as you might commonly here from a hard drive and if they did it would be concerning. SSDs have something called wear levelling technology which ensures all cells wear evenly. You have no choice where the info is stored and it is always stored in the cells less used to ensure they all wear at an equal rate. Imagine if info was always put to the front cells....

Next, TRIM simply physically erases the index and info info rather than erasing just the index as is done with other non-TRIM ssds. There is presently alot of controversy regarding the effectiveness of trim because, in theory, it is great, yet in reality....the differences are not that astounding to the typical user. (ie if my 256Gb Sammy slows from 145mb/s write to 120mb/s write it still surpasses the typical Intel speed.

Oh and back to wear I came from....defrag should not be done and is not necessary simply because wear levelling instructions take care of that for ya...

 

@Erik: I am aware of the working principle of a SSD, and that it shouldn't require defrag at all, nor should a defrag run improve performance in any way. The intriguing point is, however, I do experience enhanced performance and much snappier regular operations (Outlook with a 4 GB PST, browsers, app launch/run and boot etc.)

@TinyRK: do you mean after the 'defrag'? The benchy was run right after the 'defrag', but I did turn it off (complete off, not hibernate) at home and back on after I arrived at my office. I usually power it off completely at night. Otherwise, it's almost always powered on when I am awake.

@Les: yeap. I understand. There shouldn't be a need to do a defrag at all. But it surprises me, as an experiment, I did get an increase in speed both with my feeling and on the benchmark. If TRIM does take care of everything, why was there an increase in sequential write and a significantly snappier environment after defrag?
Plus, I am not even sure if TRIM is running properly. I have the latest INTEL firmware. Windows 7 (Lenovo recovery disc) was reinstalled fresh after formatting the drive after the upgrade. I manually picked the MS 'Standard AHCI IDE/SATA controller' over the auto-suggested Intel one. But there's still no way to be sure if TRIM is running or not.

 

Wow, I don't know that X25-M seq write speed can be that quick. Write speed has always been the achilles' heel of the X25-M, and Intel only stated a 75MB/s write speed for the X25-M (which was increased to 100MB/s after firmware update). Although mine does have seq write speed over 100MB/s (around 120 as far as I can recall), but I have never seen seq write speed over 200MB/s for the X25-M.

 

If Perfectdisk 10 has processes running in the background(and it does) it will cause the drive to benchmark poorly. I would uninstall PD10 and just use the command line defrag c: /x from an elevated command prompt if you feel the need to consolidate free space.

 

This is only one instance of benchmark. My usual bency for the X25-M is ~108 MB/s, averaging to around 102 MB/s usually. This is on par with the spec after the firmware update. I am not sure why this suddenly shows >200MB/s in that one benchmark, so I'm asking around in here, particularly about the effect of a supposedly unrecommended defrag run.

 

Good point! Rather than uninstalling it, I have disabled the PerfectDisk services (setting it to manual). So it shouldn't run unless I manually run the program. I don't see any other PD related entry in Task Manager. Wonder if this works?