[Cryptech Tech] core size in cryptech.h

Joachim Strömbergson joachim at secworks.se
Mon May 4 06:55:26 UTC 2015


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Aloha!

Paul Selkirk wrote:
> Ah, okay. I saw what appeared to be an extra 0: #define
> MODEXP_CORE_SIZE          0x1000
> 
> The actual effect is nil, since it doesn't change MODEXP_ADDR_BASE, 
> and there isn't another math core at the moment, but feel free to 
> revert it.
> 
> Looking ahead, will modexp actually need as many as 4096 addresses?
> I see prefixes from 0x000 to 0x400.

No not 4096 addresses, that is a bit on the high side.

The modulus, message, exponent and result max sizes are all 8192 bits,
or 256 32-bit addresses each. If we want to be able to extract the
residue (in order to speed up key operations), that is another 256
addresses. Finally we need to have a general address area for control,
status, config etc.

In total we should need 6 x 256 addresses or 1536 addresses. So a size
of 0x600 (or even 0x800) is a better fit. I did revert your fix I think,
but drop the size to 0x800.


> Coming at it from another angle, will the math segment need more
> than 2 cores of this size?

Pavel probably has better answers re how much address space the ecdsa
core will occupy.

And the question is then also if/what other math cores we might want to
include in the future. I'm unable to come up with one right now though.
But some oher bignum, dsp functions perhaps?

So, probably not more than two.

- -- 
Med vänlig hälsning, Yours

Joachim Strömbergson - Alltid i harmonisk svängning.
========================================================================
 Joachim Strömbergson          Secworks AB          joachim at secworks.se
========================================================================
-----BEGIN PGP SIGNATURE-----
Version: GnuPG/MacGPG2 v2
Comment: GPGTools - http://gpgtools.org
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/
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=NkfA
-----END PGP SIGNATURE-----


More information about the Tech mailing list