64Gb Sdxc Cards

[neko]

SDHC devices will only support the SDXC cards which use UHS104 speeds; SDHC devices will not recognize the SDXC cards which use the faster (SD 4.0), final specification of SDXC.

Just to be clear, does this mean that SD4.0 SDXC cards will not work on Pandora?

Possibly, but no such cards are on the market right now.

The Pandora seems to be able to do 16MB/sec, so anything up to class 16 cards should be okay.

 

[WizardStan]

The Pandora seems to be able to do 16MB/sec, so anything up to class 16 cards should be okay.

That is not how classes work. The class is the minimum sustained write speed the card is guaranteed to go at. They should (and actually must) have the ability to write slower, in devices that are only capable of reaching that lower speed.
The reason SDHC readers might not be able to read these faster cards is not virtue of the fact that they are faster, but because they've changed things about the card themselves in the process of giving them the ability to go faster.
I am highly suspicious of the claim that SDHC devices will absolutely not support cards which use the finalized spec. Seems to me it is just a software thing, but I've been wrong (many times) before. We shall have to wait for such a card to actually come out to try and see.

 

[Pontaloon]

In Australia, the price of a 32GB SDHC card is roughly the same as a 750GB Passport HD - around $120. It looks like from the comments above that the USB2.0 port is faster than the card readers. I'd figure that the cards are much more portable and not as breakable (if you're out exercising, etc.) but are there any other reasons why the SD cards are better than the external harddrives? I'd probably get one of each anyway.

 

[slaeshjag]

SD cards won't drain your battery as fast.

EDIT: And won't risk destroying your USB-port.

 

[Pleng]

Yea you can say goodbye to at least half your battery life by using an external HDD

 

[Yoyobuae]

From what I understand the main limiting factor for SD card speeds is the controller in the SD card. This controller must do wear leveling for the flash chips inside. All the while, the controller needs provide the block device abstraction to whatever the SD card is inserted into.

The controller is a processor, with a program that executes the above functions. So you can imagine that some controller chips might be faster or the software they run might be better optimized. This is why SD card speeds varies.

If I'm not mistaken, within the same SD spec, the electrical interface (including clock speed) is the same. But in a new SD spec that clock speed can change. Devices not designed to handle those speeds will then be incompatible.

I'm guessing that so far the clock speed has remained the same and SD card controllers have been playing catch up. But now (with SD 4.0) they are increasing the clock speed, so that the controllers can continue improving.

 

[davidgro]

...
The controller is a processor, with a program that executes the above functions. So you can imagine that some controller chips might be faster or the software they run might be better optimized. This is why SD card speeds varies.

I never thought about it that way, it really does explain a lot! (such as why SSDs are so much more expensive and faster compared to the same capacity in SD cards) - and why some cards use more power, even when sitting idle (bugs in the code + inefficient processor)

Thanks for the insight.

 

[joseluisjazz]

I want to say that I was wrong about ext2 being more convenient than ext3 for SD cards. I had my root and another heavily used partitions formatted as ext2, and had some corruption problems. So I created a journal and converted them to ext3. Since then I had 0 problems, and I don't notice any drawback in speed. And that's my experience now.