RE: bug fix for hp-timing.h (aarch64)

["Tang, Jun" <juntangc@amazon.com>]

Zack, Wilco,

Thanks for the suggestions,

Hp-timing.h is used for glibc bench timing, and calculating in float is to avoid any future overflowing (the chance is probably very small). 

Regards,

Jun


-----Original Message-----
From: Wilco Dijkstra <Wilco.Dijkstra@arm.com> 
Sent: Wednesday, January 11, 2023 4:49 PM
To: zack@owlfolio.org; Tang, Jun <juntangc@amazon.com>
Cc: 'GNU C Library' <libc-alpha@sourceware.org>
Subject: [EXTERNAL] bug fix for hp-timing.h (aarch64)

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.



Hi Zack,

> -   (Diff) = ((End) - (Start)) * (UINT64_C(1000000000) / freq); \
> +   (Diff) = (((End) - (Start)) * UINT64_C(1000000000)) / freq; \
>
> This avoids using floating-point arithmetic but should still do the rescale correctly.
> (The multiplier and divisor are both guaranteed to be less than 2**32, 
> so, as long as we do the multiplication in 64-bit arithmetic, we 
> cannot lose bits via overflow.)

However this will overflow after about 16 seconds - and that's short even for basic benchmarking. Since this code is not performance critical, using floating point here is reasonable.

Cheers,
Wilco