[Cryptech Tech] FPGA core and use case estimations
Joachim Strömbergson
joachim at secworks.se
Wed Mar 11 12:40:33 UTC 2015
Aloha!
Ok, I've spent some time starting to do real estimates on FPGA resources
we need and what is available in different FPGAs. I've built all of our
major use case functionality cores. I've also added (really bad)
estimates for RSA, EC and other cores that we don't yet have. I've tried
to err on the conservative side (quite heavily) for these cores.
The result seems to show that the Artix-7 T200 is a bit overkill even if
we add everything. The T100 should suffice for our needs. Altera Cyclone
A7 should also provide good margin. Performance wise, Artix looks really
promising. Running our cores at 66 MHz should be no problem.
If we then consider to allow having use case adapted configurations, the
margin becomes even better. I've added a first estimate for a DNSSEC
specific configuration.
Yes there are a whole bunch of problems with the estimates. I've not for
example added numbers for max resources for Cyclone IV GX or IV. But
this is at least a start and we can do some reasoning about amount of
resources we need.
The estimates are in the included open document spreadsheet file.
--
Med vänlig hälsning, Yours
Joachim Strömbergson - Alltid i harmonisk svängning.
========================================================================
Joachim Strömbergson Secworks AB joachim at secworks.se
========================================================================
-------------- next part --------------
A non-text attachment was scrubbed...
Name: fpga_estimates.ods
Type: application/vnd.oasis.opendocument.spreadsheet
Size: 22086 bytes
Desc: not available
URL: <https://lists.cryptech.is/archives/tech/attachments/20150311/ad6ef214/attachment-0001.ods>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 931 bytes
Desc: OpenPGP digital signature
URL: <https://lists.cryptech.is/archives/tech/attachments/20150311/ad6ef214/attachment-0001.sig>
More information about the Tech
mailing list