Bug found in nightbuilds

[bodrato at mail.dm.unipi.it]

Ciao,

Il Dom, 16 Febbraio 2014 5:45 pm, Torbjorn Granlund ha scritto:
> bodrato at mail.dm.unipi.it writes:

>   I disallowed zero DIV_QR_1_NORM_THRESHOLD with this patch:

> Since we control these thresholds at the source level, if it is
> desirable to keep DIV_QR_1_UNNORM_THRESHOLD >= 1, that should be done in
> tuneup.

I thought that this constant was used elsewhere in the code. My goal was,
not changing its meaning, only its use in this context.

> Your patch adds some overhead when tuning, since then
> DIV_QR_1_UNNORM_THRESHOLD is not really constant.

Does tuning this constant work at all? This code is not suited to intract
correctly with tuneup...

The function one() in tuneup compares the timings of a function with an
argument $n$ and thresholds $n$ and $n+1$, but here $n-1$ is compared with
the thresholds, and it's BELOW both... tuneup can measure only noise, I
fear.

Regards,
m

-- 
http://bodrato.it/toom-cook/