[PATCH 1/1] log10: Use sw implementation from amd builtins

Runtimes OpenCL
1

Add missing table.
Fixes log10D CTS on carrizo.
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>

2

Reproduced the failures before, confirms CTS passes log10 after on my RX580

Tested-By: Aaron Watry <awatry@gmail.com>

--Aaron

3

Reproduced the failures before, confirms CTS passes log10 after on my RX580

It should also improve worst case ULP for the single precision log10
test.

Tested-By: Aaron Watry <awatry@gmail.com>

thanks. do you plan to take a closer look for review later? or should I
go ahead and push this patch?

thanks,
Jan

4

Reproduced the failures before, confirms CTS passes log10 after on my RX580

It should also improve worst case ULP for the single precision log10
test.

Tested-By: Aaron Watry <awatry@gmail.com>

thanks. do you plan to take a closer look for review later? or should I
go ahead and push this patch?

Just go ahead and push it. As far as I’m concerned it’s fine. It builds and passes CTS, so the only quibbles I’d likely find with it would be stylistic given the table-based nature of the implementation.

5

> > Reproduced the failures before, confirms CTS passes log10 after on my
>
> RX580
>
> It should also improve worst case ULP for the single precision log10
> test.
>
> > Tested-By: Aaron Watry <awatry@gmail.com>
>
> thanks. do you plan to take a closer look for review later? or should I
> go ahead and push this patch?
>

Just go ahead and push it. As far as I'm concerned it's fine. It builds
and passes CTS, so the only quibbles I'd likely find with it would be
stylistic given the table-based nature of the implementation.

Can I consider this acked by? I don't want to introduce precedent of
unreviewerd/acked commits now that were are close to clc 1.1
completion.

thanks,
Jan

6

Acked-by: Aaron Watry <awatry@gmail.com>
Tested-by: Aaron Watry <awatry@gmail.com>

--Aaron