SSD performance degredation.

[.Alan.]

Does an SSD slow down after a period of use?

I've only just starting looking at SSDs but noticed a few articles on this subject.
Anyone with an SSD have some real experience of this?

[coiler]

didn't they fix this with the latest generation of SSD's and it was down to an inferior JMicron controller used in the early SSd's

[RJackson]

IIRC the performance issue was caused by data pointers on the disk being erased, rather than erasing the actual data. Which meant when it was time to write over that data, it had to be erased first.

Most SSD manufacturers had released firmware updates to solve that problem.

[Gazanimal]

SSD performance does degrade over time.

For a good lot of info on it, go to the OCZ forums & there's a plethora of discussions about it.

Trim & Garbage Collection features that are now built into the firmware should alleviate the problem of slowdown but not totally eradicate it.

The controller always looks for clean nand to write too so once a drive starts getting fuller & there's less clean nand available the controller will clean the nand before it can write to it, thus causing slow down.

There's manual ways to help with this, doing `free space consolidation` & then wiping the free space help immensely which is basically what Trim command should do.

As I said, the guys over at OCZ really know their stuff & are on the ball with regards SSD's & their usage so it's probably the best place I can recommend.

[JoeJones]

Gazanimal has almost conveyed correctly the concept of SSD degridation....perhaps this article will help clear up the finer points.

http://www.bit-tech.net/hardware/storage/2010/02/04/windows-7-ssd-performance-and-trim/1

[.Alan.]

Thanks for that Gaz.:thumb:
I'd skimmed a few articles generally, but I'll check out OCZ as I'm seriously considering an SSD as my next upgrade.

[Gazanimal]

Once you go SSD you'll never go back :D

[mikerr]

I REALLY think this issue is GREATLY overblown.

An SSD that has "degraded" is perhaps 70% of its original speed, still 10x the speed of a conventional spinning HDD in real world usage.

[marsey99]

I REALLY think this issue is GREATLY overblown.

An SSD that has "degraded" is perhaps 70% of its original speed, still 10x the speed of a conventional spinning HDD in real world usage.

I REALLY think your figures are GREATLY overblown.

[JoeJones]

*YAWN*...try reading the link I posted instead of trying to look clever and posting guestimates.


In all seriousness though it is an excellent article explaining why SSD's degrade, what TRIM does and shows the performance hit of with/without TRIM.


For those the still think they know everything about SSD's and TRIM...

RAID arrays are supported, but only via the manual TRIM command in Intel's SSD toolbox for Intel X25-M G2, Intel X18-MG2, Kingston SSDNow M Series (need to check if the Intel X25-V is supported as well since I believe there is very little difference between that and the M version).

[/)zh]

wasnt it proven that ccleaner free space wiper was just as effective as trim?

there were some results knocking around

[Gazanimal]

wasnt it proven that ccleaner free space wiper was just as effective as trim?

there were some results knocking around

Sometimes Trim just wasn't kicking in so using the `consolidate free space` & `free space wiper` method was actually doing what Trim should have done.

One of the tech guys at OCZ called Tony actually devised the manual Trim method & called it Tony Trim. But since OCZ updated the latest firmware on Vertex drives to 1.5 the manual option isn't needed as much as Trim & GC seems to be working fine.

[nessundorma]

Sounds the performances are not consistent so far, better wait for the technolgy to become mature and stick with the normal platters...

[Gazanimal]

They are very consistent nessun. Even like spinners that need defragmenting, there has to be some kind of maintenance regime even for SSD's.

Logging off or idling is enough to kick in a Trim, so there's no need to start a program to do a defrag etc.