View Single Post
Old 02-15-2012, 09:29 AM   #26
lothar
Diamond Member
 
Join Date: Jan 2000
Posts: 6,430
Default

Quote:
Originally Posted by DLeRium View Post
Uhm. Just because that Sandisk card is damn good and its only rated Class 4 doesn't mean it is an automatic win and any other "Class 10" cards suck.

What solves it all is benchmarks. If you have benchmarks of that specific Sandisk card slapping the hell out of the Class 10 Patriot card, then by all means, show the data. Otherwise there's no need for this.
Here you go(Both are duplicates):
http://www.ccereviews.com/reviews/pa...memory-card/4/
http://www.overclockersonline.net/reviews/50003341/



Here's another one for the 16GB version of that same exact card:
http://www.tomshardware.com/reviews/...e,3011-12.html




Proof of engineering concept for those who may be confused as to why I'm also linking the 16GB version of the card:
The second set of charts is for the 16GB version. In other for the 32GB version of that same card to show the same exact performance displayed in this charts by the 16GB version, it needs to have double the NAND channels/chips on board. Most manufacturers don't double the NAND channels/chips when they double RAM capacity, and certainly absolutely NONE of the manufacturers on the market more than double the NAND channels/chips when they increase capacity 2x.

Class 10 cards are the best on the market if all you're talking about is playing MP3's or Video on it because those are sequential. That Patriot card would slam SanDisk in those results.
However, if you're talking about OS functions on the card such as app2sd, custom ROMS, and others; then random write performance is more important than any sequential test.

I'm still sticking to the opinion that poofyhairguy's random write performance results were an anomaly.
Saying that they are an anomaly doesn't necessarily mean it's wrong. It just means it's an outlier and not likely to happen in the real world.
There are lots of outliers in statistics. That doesn't mean the statistical data collected for that particular outlier was wrong.


User data on various cards from this website if you want to know how well SanDisk performs on random write performance: http://forums.whirlpool.net.au/forum....cfm?t=1582172
Quote:
Random 4k R/W (QD=1) results, sorted by best 4k write speed Card Size Class ReadMB/s WriteMB/s Purchased/Bus/User
Sandisk 32GB 2 1.130 2.863 Q410/USB/OP
Sandisk 32GB 4 3.241 1.752 Q211/USB/OP
Sandisk 32GB 4 2.888 1.663 Q211/USB/Hansel
Team 32GB 4 2.560 1.613 Q211/USB/OptPess
Sandisk 32GB 4 2.747 1.471 Q211/USB/Zadax
Sandisk 16GB 4 3.154 1.464 Q311/USB/dzx
Sandisk 16GB 2 0.761 1.295 Q310/USB/warakawa
Sandisk 4GB 4 2.473 1.035 Q???/USB/bamboo
Sandisk 16GB 2 1.263 0.307 Q310/USB/CzarKhan
Kingston 16GB 2 3.452 0.056 Q410/USB/MrMaestro
Sandisk 8GB 4 3.457 0.039 Q408/USB/OP
Sandisk 8GB 2 3.744 0.018 old/USB/piers888
Sandisk 8GB 2 3.321 0.018 Q408/USB/OP
Nokia 8GB 4 3.381 0.017 Q109/USB/pspiso
Kingston 16GB 4 3.473 0.016 Q410/PCI/OP
Toshiba 4GB 4 3.599 0.016 Q110/USB/MrInsane
Lexar 16GB 4 3.750 0.010 Q111/USB/flannigan
Lexar 32GB 10 5.468 0.008 Q111/USB/bchliu
Lexar 32GB 10 5.233 0.008 Q111/USB/nick Fritz
Lexar 16GB 4 4.275 0.008 Q111/USB/Siggi
Patriot 8GB 10 2.949 0.007 Q311/USB/eqx
Team 16GB 10 3.487 0.006 Q111/USB/CzarKhan
Team 16GB 10 3.248 0.006 Q211/USB/OptPess
Team 16GB 6 2.841 0.006 Q111/PCI/effdubya
I'm pretty sure there's at least 3 SanDisk 32GB Class 4 results thrown in there somewhere, so feel free to compare them to all the Class 10 cards at the bottom and the different manufacturers.

Last edited by lothar; 02-15-2012 at 11:25 AM.
lothar is offline   Reply With Quote