Re: pmiostat qa failures

["Frank Ch. Eigler" <fche@xxxxxxxxxx>]

Hi -

On Mon, Sep 08, 2014 at 11:06:25PM +1000, Mark Goodwin wrote:
> [...]
> The time delta calculations are all supposed to be done in double
> precision arithmetic, so I'd agree if the mantissa are within ~60 bits ...

(Well, 64-bit doubles have a 52-bit mantissa.)

> [...] The following one character patch in libpcp promotes it to
> double precision, [...]

> -    dbl += (double)val->tv_usec / 1000000.0;
> +    dbl += (double)val->tv_usec / 1000000.0L;

Please note that in the L suffix means "long double".
Unsuffixed already means "double".

It's also not hard to get FP loss of precision by subtracting
similar-sized numbers.  I've had to play with this a bit during the
pmwebd graphite work, which does plenty of scaling transforms.

- FChE