PDA

View Full Version : Interesting article on Air flash degradation...




TheRealDamager
Feb 7, 2011, 11:57 AM
Final conclusion is that use and reuse, as well as low available capacity, doesn't affect performance:

http://www.macworld.com/article/156227/2010/12/macbookair_torturetests.html



Hellhammer
Feb 7, 2011, 12:02 PM
Thanks for the heads up. Looks like Apple has really worked on the firmware to make it that good. All this crying about the lack of TRIM seem to be history now.

jamesryanbell
Feb 7, 2011, 12:11 PM
"Those 'in the know' understand that TRIM doesn't matter anymore for anything new running OSX".

KPOM
Feb 7, 2011, 12:15 PM
It would be interesting to see how the MacBook Airs with the new Samsung drives compare to the ones with the Toshiba drives.

nowstime
Feb 7, 2011, 01:18 PM
Yep...this is the article made me jump for the Macbook Air recently. Thanks for posting...

altecXP
Feb 7, 2011, 01:33 PM
Hows that different from the OWC Sandforce based benchmarks?

altecXP
Feb 7, 2011, 01:35 PM
"Those 'in the know' understand that TRIM doesn't matter anymore for anything new running OSX".

The technology they are using is in the firmware, so that means it applies to any OS, not just OS X. TRIM is ment to be a level above that to help even more.

ecib
Feb 7, 2011, 01:50 PM
From the forums of the article cited:

"Don't the manufacturers of Flash storage cite very high numbers of writes to a given block, before the cells start getting flaky? Any reason to think that these tests came anywhere close to that level? Since there are no "faster tracks" on a SSD, the OS (or better, drive firmware) simply has to always write to the least-recently-used block and map it to wherever it supposedly is, and you now have very few hot spots that would have crossed the over-use threshold of the devices.

If modern SSD controllers are half as smart as they should be, your tests came nowhere close to being stress tests of the storage itself."


This poster pretty much summed up what I was wondering. As I understood it, it was not the fact that *all* cells could be written X times without degradation, but that *individual* cells could only be written X times without degradation. I am by absolutely no means an expert, but as I understood the critique of flash memory's degradation, this test failed to address what was cited as the weakness.

That being said, I have a new maxed out MBA, and I am in no way worried about memory degradation, -I'm certain I will have a new computer long before this ever becomes an issue, but it seems to me at a glance that this stress test wasn't designed to test the alleged shortcomings of flash memory as they were defined.

PeckhamBog
Feb 7, 2011, 01:57 PM
But what happened to the 128 GB Air? Was the fault ever discovered?