Bitcoin Forum

http://embedded-computing.com/atmels-cloning-secures-data

Alright, who wants to put a couple dozen of these chips onto a breadboard and wire it up as a 'getwork' device via USB for me?  Generation arms race!

It is certainly less complicated than getting my hands dirty on FPGAs, but I can't find any performance numbers. I'm inclined to believe this is more of a 'secure box' than a performance contender.

You would be hard pressed to try and scale this to a point where it was interesting, imho.

Probably, but at the claimed price point of $1 each, there could be an awful lot of these little guys on a board for much under the cost of a new GPU or FPGA

haha

Would be cool to hook this up to a solar array for a portable self powered bitcoin client. Include a slot for a 3g card for bandwidth....win!

A project that I have forming in my head is what lead me there.  I'm thinking about how one might make a purpose built product that can use a couple of digital radios (low-power WiFi in ad-hoc mode & Dash7, maybe a NFC pad as well) to run a stand-alone 'ligwithweight' bitcoin client and a wireless version of a Jabber client, and perhaps a wireless version of UUCP for email transport.  (Android to Android Wireless Copy Protocol, perhaps?)  Not even as powerful as a netbook, but something that could keep a villager in Africa (the whole village?) connected and able to transact business over a fair distance.  Dash7 is an excellent wireless protocol for low bandwidth communications over distance, and takes very little power to hit 2 klicks over a clear path.  It also does not use sessions like Wifi, and can broadcast a transaction or a text message to any Dash7 radio that can hear it.  The Wifi radio would only be for opprotunistic blockchain updates, whenever a pair of units came close enough to compare their blockchains, or one came within radioshot of a wifi hotspot.  Think android (or maybe less) qwerty smartphone without the phone part.

The hardware sha-256 acceleration wouldn't really need to be a high performance chip in this context.  I'm wondering how cheaply such a device could be built.

Such devices cat not be safe without a keyboard and display.
And such a device is to have everyone in your pocket - cellphone.

I skimmed through the datasheet for the AT88SA102S ... when I was tired...

It does report t_{EXEC_MAC} = 30ms (Delay to execute MAC command) in table 4-1.  I think this is the only SHA-256 operation where you get the output, the other is to generate some encryption key for burning in the secret.

It looks like the 512 bits it hashes are rather tightly constrained - you can include some message, but the rest is internal (secret, write-only) state.  But assume it will hash anything for you, on request.  Input is 36 bytes, output is 32 bytes; each(?) plus 3 bytes to form the IO Block.

Sending data down the serial interface (t_BIT in table 3-1) takes >37us per bit.  Receiving data takes >41us (the chip's clock spec is looser than that of the host issuing commands).

Total time per hash (MAC command) looks like
   IO: (36 + 3) byte * (37 microsec / bit) + (32 + 3) byte * (41 microsec / bit) = 23ms
   + Compute: 30ms

Per chip looks just under 19 hashes/sec.  You need 150,000 of these to match a modern (3Mhash/sec) CPU.