Re: The market of ASICs (One GigaKey / Second?)
"Dan Oetting" <[email protected]>
| Newsgroups | gmane.comp.distributed-net.hardware |
|---|---|
| Message-ID | <[email protected]> |
On Aug 9, 2004, at 2:49 PM, Elektron wrote: > The AltiVec cores are more interesting, but completely over my head > (so are the integer calculations which somehow go with it). I'm not > sure why you can't do 5 keys at a time either (since if everything's a > vector, you can do some scalar things too). I do. But you can't just process an integer key in parallel with the vector keys. What you can do is get a head start by processing the first few stages of 4 parallel keys in the integer unit then transfer the work to the vector unit to finish them off. The whole thing gets wrapped around itself so the integer processing takes place in the dispatch holes of the vector work. Here is one written for the 7400 PowerPC. The main loop is indented to show the separate code streams. <http://www.users.qwest.net/%7Eoetting/RC5testbench/AltiVec/rc5loop.s> Some day I may have to get back into this. The benchmarks produced for the G5 look a little on the slow side. But it's all on the back burner because I can't convince myself that RC5-72 is worth pursuing. Without a major breakthrough the time to crack the code can simply be extrapolated from the previous results. And the cost for the electricity to run the CPUs alone is orders of magnitude greater than the reword. _______________________________________________ Hardware mailing list Hardware-Ra3b/QYEcJ3d140v2zMXi0fjHoOT/h/[email protected] http://lists.distributed.net/mailman/listinfo/hardware