[libre-riscv-dev] [Bug 74] preliminary exploratory software emulation of FP SQRT

bugzilla-daemon at libre-riscv.org bugzilla-daemon at libre-riscv.org
Sun Apr 28 18:09:03 BST 2019


--- Comment #24 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
(In reply to Aleksandar Kostovic from comment #18)
> > great.  so... bye bye sqrt for now.  can always track that down later.
> I think you misunderstood. Its the sqrtsimple function that is failing:
> File "/home/aleksandar/projects/ieee754fpu/src/add/fsqrt.py", line 88, in
> <module>
>     assert int(Q**0.5) == sqrtsimple(Q), "Q sqrtsimpl fail %d" % Q
> So we use the regular sqrt func.
> >so, next stage, you want to write a test for-loop around the main() function?
> Will try to do it now

looks great, works well.

> >btw do you already have sfpy installed and built?  if so, we can move
> >to the next phase of "from sfpy import Float32" and use that to mangle/de-mangle >FP numbers into sign/mantissa/exponent, using sfpy.Float32.get_bits
> Yes i have it installed

ok great, 

so next phase will be to use "from sfpy import Float32" and create some
arbitrary numbers, say... oh i dunno, pick one! 753.415926535897932

x = Float32(fred_the_number)
xbits = x.bits

then use the function decode_fp32(xbits) to get the s, m, e

and from there you can throw those into the function main() and see what

theeeeennn... you can do sq_test = x.sqrt() above... and on the answer,
use the function decodefp32(sq_test)...

and print those out next to the answer from using main().

if those are even remotely close we have a winner!

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

More information about the libre-riscv-dev mailing list