[Beignet] Building beignet with OpenCL 2.0 support in distros

[Bruno Pagani]

Le 21/01/2017 à 18:01, Rebecca N. Palmer a écrit :

> I have a first attempt at "enable/disable 2.0 at run time" written,
> but haven't yet tested it.

OK. I have both an Haswell and Skylake machine at hand if this can help
in testing.

> On 21/01/17 16:08, Bruno Pagani wrote:
>> Le 21/01/2017 à 16:55, Rebecca N. Palmer a écrit :
>>
>>> On 21/01/17 15:40, Bruno Pagani wrote:
>>>> Le 21/01/2017 à 16:20, Rebecca N. Palmer a écrit :
>>>>
>>>>>> is there any downside in compiling with
>>>>>> OpenCL 2.0 support,
>>>>> Yes - on older (Ivybridge/Haswell - no emitUntypedReadA64Instruction)
>>>>> hardware, a 2.0-enabled beignet won't work, at all.
>>>>
>>>> That’s what I’ve feared but after trying it here I haven’t encountered
>>>> more issues than with a non-2.0-enabled beignet.
>>>
>>> Trying it on what hardware?  There might be hardware (all of gen8 if
>>> this assert(0) is the only failure point) where a 2.0 build doesn't
>>> crash outright but also doesn't have working 2.0.
>>
>> Haswell (HD4600).
>
> Weird - are you sure darktable loads OpenCL by default?

Not in the default configuration, but does in mine. It also sucessfully
executed some tasks using OpenCL.

>   For me (Ivybridge M GT2), everything that does immediately crashes with
>
> clinfo:
> /home/rnpalmer/Debian/builds/stackbuild/beignet/backend/src/backend/gen_context.cpp:2259:
> virtual void gbe::GenContext::emitUntypedReadA64Instruction(const
> gbe::SelectionInstruction&): Assertion `0' failed.
> Aborted
>
> and that assert(0) only gets replaced by a real implementation in gen8
> and above, not in Haswell (gen75).

I’ll try again tomorrow and join the opencl debug log of darktable if
needed.

Bruno

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 520 bytes
Desc: OpenPGP digital signature
URL: <https://lists.freedesktop.org/archives/beignet/attachments/20170121/f6e1de1e/attachment.sig>