The Mrs. and I get our new sd cards today. We went with Sandisk 32GB Class 4, part number SDSDQM-032G-B35N, $30 on Newegg. This choice was largely based on the small block random write ability of this card. I couldn't find a faster one that's 32GB. This thread is where I first learned of why one cares about small block random write: SD Strange-results - or - How I learned to love CM7 on SD - Dec-11 results compiled! - xda-developers. This thread struck me as very noteworthy, as many assume class is the best, even only, thing to pay attention to for apps to sd applications.
That thread has inspired me to play with some benchmarking. Not just standard i/o, but to measure performance of running some app or rom from the sd card itself in a qualitative way, not just "the app/rom runs better on this this/that card", and compare this against the Random Write 4KB (QD=32) that the folks in the above thread claim is the key to smooth running.
The short term plan is to run some benchmark app, so far planning on AnTuTu v2.7.3, from:
1. app on phone
2. app on the stock 8GB/class2
3. app on the new 32GB/class4
and to do this from both my 3vo and wife's evo.
This will all be compiled along with the results of the Crystal Disk Mark 3.0.1 (x64)'s analyses of the sd cards, taken first just through via the phone/usb. I don't know if I should be using a card reader (and not the phone) to test the sd cards like this. I'll prolly try both and see the difference.
When running the benchmark app from the cards, I'd like to move everything over (i.e. entire app + data + app cache). TB only moves the app data over. I'm looking for something like the Darktremor script that the evo folks get to use ([Script][A2SD]Darktremor 2.7.5.3 Beta 04 [Evo] (01/29/2011) - xda-developers), which right now our phone isn't on the compatibility list.
Anyhoo...
If anyone has any comments concerning:
- is this a good plan, do you think running the app from the phone will test what I want it, which ultimately is to determine which card will be the best card to run Ubuntu from (which may eventually offer additional benchmarking options)?
- any suggestions of a method of moving the app over, ala Darktremor?
- have you seen this done somewhere already?
they are greatly appreciated!
That thread has inspired me to play with some benchmarking. Not just standard i/o, but to measure performance of running some app or rom from the sd card itself in a qualitative way, not just "the app/rom runs better on this this/that card", and compare this against the Random Write 4KB (QD=32) that the folks in the above thread claim is the key to smooth running.
The short term plan is to run some benchmark app, so far planning on AnTuTu v2.7.3, from:
1. app on phone
2. app on the stock 8GB/class2
3. app on the new 32GB/class4
and to do this from both my 3vo and wife's evo.
This will all be compiled along with the results of the Crystal Disk Mark 3.0.1 (x64)'s analyses of the sd cards, taken first just through via the phone/usb. I don't know if I should be using a card reader (and not the phone) to test the sd cards like this. I'll prolly try both and see the difference.
When running the benchmark app from the cards, I'd like to move everything over (i.e. entire app + data + app cache). TB only moves the app data over. I'm looking for something like the Darktremor script that the evo folks get to use ([Script][A2SD]Darktremor 2.7.5.3 Beta 04 [Evo] (01/29/2011) - xda-developers), which right now our phone isn't on the compatibility list.
Anyhoo...
If anyone has any comments concerning:
- is this a good plan, do you think running the app from the phone will test what I want it, which ultimately is to determine which card will be the best card to run Ubuntu from (which may eventually offer additional benchmarking options)?
- any suggestions of a method of moving the app over, ala Darktremor?
- have you seen this done somewhere already?
they are greatly appreciated!