As of this week I'm now officially the vice-chair of the RISC-V Bitmanipulation Task Group within the RISC-V Foundation.
Many thanks to the people who made this possible. I am very excited about this opportunity.

-
-
Yes. This is the frequency analysis. I agree many of the instructions like bext/bdep have a lot of potential applications but not wide use in practice because they are only arriving on commodity silicon (even though they have been on previous architectures like parisc).
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
There is also the need for VBitManip. e.g. VPOPCNT for quantization on vectors, per lane then parallel sum, or whatever is required. I think if there is a BStd and BAdv, then BStd could have V equivalents. V regs could help with configuration state for arbitrary but permutations.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I imagine that the crypto people can establish the need for clmul; this seems to be one of the instructions that crypto people want which is also more generally applicable (not one of those "here's our inner loop, suck it up, guys" type instructions).
-
Yes CLMUL is essential. Also the question about how GF2 for 128-bit is done on RV64. The division between extension groups is interesting. Finite field field arithmetic is crypto/ecc. GF2 (CLMUL) is simple but wondered if the crypto folk were considering other prime fields.
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.