[tex-live] strange discrepancy in running time of etex between TL2015 and TL2017

[jfbu]

> 
> http://www.bytereef.org/mpdecimal/
> 
> but even then
> 
> Power function:
> Results are not correctly rounded, even if the allcr context field is set to 1. This might change in a future release. The error of the function is less than 1ULP+t, where t has a maximum of 0.1ULP, but is almost always less than 0.01ULP.
> 
> (power function with integer or half-integer exponent in xint currently achieves better precision than that, more like 0.5ULP + epsilon)

Hi, sorry for noise, but just to correct myself,
as the mpdecimal allows the various rounding modes
from the IBM specification, 

http://www.bytereef.org/mpdecimal/doc/libmpdec/context.html#rounding

such as "round away from 0" it is normal that the error
is qualified as 1ULP+something. Probably when the rounding
mode is half_up, half_down, or half_even, then the library
achieves 0.5ULP + something, but the doc does not say so
in this quote,

I know I am drifting off topic, but I wanted
to correct my remark (people will read this 
in decades even centuries, so I have to tighten
the bolts)

Best,
Jean-Francois