View Single Post
Old 03-12-2013, 11:08 PM   #14
Shai-Hulud
Silicon Book Worm
Shai-Hulud has much to be proud ofShai-Hulud has much to be proud ofShai-Hulud has much to be proud ofShai-Hulud has much to be proud ofShai-Hulud has much to be proud ofShai-Hulud has much to be proud ofShai-Hulud has much to be proud ofShai-Hulud has much to be proud ofShai-Hulud has much to be proud ofShai-Hulud has much to be proud ofShai-Hulud has much to be proud of
 
Shai-Hulud's Avatar
 
Posts: 129
Karma: 27430
Join Date: Jul 2012
Location: England
Device: Kobo Touch
Quote:
Originally Posted by TechniSol View Post
Actually, what I was saying is that for short random writes class 2 cards can be 50-100 times faster than a class 4 or 10 card.
There is too much variability to say either way. It's down to the individual card I'm afraid and testing is the only reliable deciding factor unless you happen to see the exact card already benchmarked on a trusted site e.g. Tom's Hardware. I was merely pointing out that for a slow device like the KT you bash into the law of diminishing returns much sooner.

Quote:
Originally Posted by TechniSol View Post
I think this, perhaps more than cost, was why the lower class cards are more popular in Android/Linux devices when used for internal storage/boot disk applications.
Not strictly true as inevitably a trade off between sequential and random read/writes is necessary depending on typical usage requirements e.g. Of the half dozen cards I tested on my Raspberry Pi, an Ultra class 10 Sandisk works best for all round performance. A slower device like the KT is going to be less demanding and will have a different typical usage profile such that a different card may be better suited (but not necessarily a lower class).

Notwithstanding, I would be willing to wager that the selection criteria was simply an acceptable performance profile for the lowest possible bulk price and these have a tendency to be the lower class cards anyhow.
Shai-Hulud is offline   Reply With Quote