[PATCH] D27334: [OpenCL] Ambiguous function call.

[Anastasia Stulova via cfe-commits]

> Perhaps that is the problem (that there are two modes that do different things)? Could we make the double overload be present but unselectable to diagnose this problem in that mode too?

If we could resolve the overload candidate to prefer ‘int -> float’ than ‘int->double’, it would work best.

From: metafoo at gmail.com [mailto:metafoo at gmail.com] On Behalf Of Richard Smith
Sent: 05 December 2016 17:53
To: Bruno Cardoso Lopes via Phabricator; reviews+D27334+public+f2c5a66032c4c02e at reviews.llvm.org
Cc: alexey.bader at intel.com; egor.churaev at gmail.com; yaxun.liu at amd.com; Anastasia Stulova; cfe-commits at lists.llvm.org
Subject: Re: [PATCH] D27334: [OpenCL] Ambiguous function call.

On 5 Dec 2016 9:42 am, "Anastasia Stulova via Phabricator via cfe-commits" <cfe-commits at lists.llvm.org<mailto:cfe-commits at lists.llvm.org>> wrote:
Anastasia added a comment.

In https://reviews.llvm.org/D27334#612858, @bader wrote:

> In https://reviews.llvm.org/D27334#611703, @Anastasia wrote:
>
> > This change seems to modify normal C behavior again. Is there any strong motivation for doing this and if yes could it be done generically with C?
>
>
> Motivation:
>
>   // Non-portable OpenCL 1.2 code
>   __kernel void foo(global float* out) {
>     out[get_global_id(0)] = sin(get_global_id(0));
>   }
>
>
> This program compiles fine on OpenCL platform w/o doubles support

Perhaps that is the problem (that there are two modes that do different things)? Could we make the double overload be present but unselectable to diagnose this problem in that mode too?

and fails otherwise.
>  If OpenCL driver supports doubles it provides at least two versions of 'sin' built-in math function and compiler will not be able to choose the right one for 'size_t' argument.

Do you have a real example? If someone passes an integer to 'sin', I think there's a better warning we can give them :)

>  The goal of this warning is to let OpenCL developer know about potential issues with code portability.

I would argue this improves the portability much as it can also be misleading in some situations (because it refers to a potentially hypothetical problem). For example there can be builtin functions that only have a float parameter (without a double version of it). This is for example the case with read_image functions that take a float coordinate value between 0 and 1. Unfortunately this warning won't be triggered on read_image functions because there is an overload candidate with an int type of the same parameter too. But we can't exclude this situations to appear in the future or from some vendor extensions or even custom OpenCL code.


https://reviews.llvm.org/D27334



_______________________________________________
cfe-commits mailing list
cfe-commits at lists.llvm.org<mailto:cfe-commits at lists.llvm.org>
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20161205/9d2f8d43/attachment.html>