| 00:45.31 | *** join/#openjtag ByronT (i=byron-po@nslu2-linux/ByronT) |
| 03:10.05 | *** join/#openjtag ka6sox (n=ka6sox@nslu2-linux/ka6sox) |
| 08:52.18 | *** join/#openjtag vmaster_ (i=vmaster@p549B6157.dip.t-dialin.net) |
| 10:53.57 | *** join/#openjtag bullet (n=bullet@123.210.203.62.cust.bluewin.ch) |
| 13:11.27 | *** join/#openjtag prpplague (n=billybob@69.73.211.46) |
| 15:33.07 | *** join/#openjtag kay2 (n=ashdown@sd-420.dedibox.fr) |
| 15:33.08 | kay2 | hey |
| 15:33.21 | prpplague | howdy |
| 15:33.48 | kay2 | someone knows with a XC2VP30, what's the max frequency ? |
| 15:34.04 | kay2 | after the PLL |
| 15:51.45 | vmaster | according to a usenet post the dcm can output up to 450 mhz |
| 15:51.59 | vmaster | the datasheet should have more reliable information |
| 16:02.55 | kay2 | ok |
| 16:02.58 | kay2 | so basically |
| 16:03.40 | kay2 | If I use 20 DES core from opencores in it (since there is enough room for that), I could crypt up to 450M*20 password a second |
| 16:04.12 | kay2 | which is 9 Billion pwd a second |
| 16:05.08 | kay2 | and if I use 50 of those it would be 450 Billion pwd a second |
| 16:05.13 | kay2 | that's a lot ! |
| 16:06.00 | vmaster | the maximum dcm frequency doesn't mean that every cricuit you put into the device can run at that speed |
| 16:06.20 | kay2 | vmaster: if it's logic gates |
| 16:06.21 | kay2 | why not ? |
| 16:09.43 | kay2 | vmaster: if someone could crack 450 Billion pwd a second, it means that in 3 days, he could get the key of a bank used for generating Credit card's pin |
| 16:10.00 | vmaster | the design tools have to honor setup- and hold-times for all gates in the path |
| 16:10.43 | kay2 | vmaster: yeah but if it could run at that speed |
| 16:10.51 | kay2 | it means what I say |
| 16:10.58 | vmaster | i believe there are several papers out that discuss the efficiency you can expect from an fpga implementation |
| 16:11.06 | vmaster | dj bernstein wrote one, iirc |
| 16:12.15 | kay2 | $299 for a board, which means 50 boards would cost $15000 |
| 16:12.19 | kay2 | which is basically nothing |
| 16:12.38 | vmaster | yeah, but you can't do that many calculations per second |
| 16:13.02 | kay2 | well in opencores.org it says that the core could do 1 DES crypt a second |
| 16:13.04 | kay2 | I mean |
| 16:13.10 | kay2 | per clock pulse |
| 16:13.24 | kay2 | it's just piped |
| 16:13.41 | kay2 | and it does it in 24 clock pulses |
| 16:14.00 | kay2 | which means after 24 pulses you get the first pwd, then each pulse generates one pwd |
| 16:14.41 | kay2 | and even if it was 200Mhz instead of 450 |
| 16:14.57 | kay2 | it would mean that I need 6 days instead of 3 |
| 16:15.25 | kay2 | It means that the system used in Credit card today are really not complicated |
| 17:10.50 | *** join/#openjtag p2-mate (n=p2@NAT.office.mind.be) |
| 17:55.36 | *** join/#openjtag dwery (n=dwery@nslu2-linux/dwery) |
| 18:12.40 | *** join/#openjtag bullet (n=bullet@100.22.79.83.cust.bluewin.ch) |
| 18:51.07 | *** join/#openjtag AchiestDragon (n=dave@whipy.demon.co.uk) |
| 22:28.20 | *** join/#openjtag pb4 (n=pb@brother.ludd.ltu.se) |