Re: Floating-point differences between ARM processors
Re: Floating-point differences between ARM processors
- Subject: Re: Floating-point differences between ARM processors
- From: vincent habchi <email@hidden>
- Date: Mon, 07 Jan 2013 21:50:38 +0100
On 7 janv. 2013, at 20:22, Greg Parker <email@hidden> wrote:
> IEEE 754 guarantees exact results for + - * / sqrt. Everything else is implementation-defined.
That’s why I suggested a mixed approach combining exact table lookup and a refinement via only multiplications and divisions. It should give, if not exact, at least consistent results whatever processor is involved.
> Note also that physics simulations will always need to be careful with the error inherent to finite precision floating-point arithmetic. IEEE specification of exact results for every operation wouldn't […]
I assumed the original poster had a knowledge of both rounding error and stability theory in algorithms, including use of ad-hoc band-aids like preconditioning. Writing a simulation code w/out such knowledge would anyhow be somewhat vain.
Cheers!
Vincent
_______________________________________________
Cocoa-dev mailing list (email@hidden)
Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden