[libre-riscv-dev] [Bug 208] implement CORDIC in a general way sufficient to do transcendentals

bugzilla-daemon at libre-riscv.org bugzilla-daemon at libre-riscv.org
Sun Mar 8 00:28:09 GMT 2020


http://bugs.libre-riscv.org/show_bug.cgi?id=208

--- Comment #1 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
btw michael, we did FPDIV in separate pgases as well.  jacob researched
algorithms that would cover div, sqrt and rsqrt, and did them as a simple
integer only pipeline.  then i hooked that up to the FP infrastructure and it
worked well.

i can recommend an incremental approach here as well: start from something
small such as converting the myhdl code to nmigen in order to avoid bogging
down with the FP infrastructure *and* with the many variations of CORDIC.

it is only around 70 lines of code so very basic: don't even pipeline it at
first, make it a FSM with a counter. then once that is achieved move on to the
next small incremental improvement.

-- 
You are receiving this mail because:
You are on the CC list for the bug.


More information about the libre-riscv-dev mailing list