I can't seem to be able to use exp() and sqrt() in CUDA.
I get:
NVCClibavfilter/try_cuda.ptx
clang-11: warning: Unknown CUDA version. cuda.h: CUDA_VERSION=11030.
Assuming the latest supported version 10.1 [-Wunknown-cuda-version]
libavfilter/try_cuda.cu:39:19: error: use of undeclared
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> Lynne
> Sent: Samstag, 26. Juni 2021 01:29
> To: FFmpeg development discussions and patches de...@ffmpeg.org>
> Subject: Re: [FFmpeg-devel] Hardware purchase request
>
> Jun 25, 2021, 13:25 by t...@rothenpieler.org:
>
> > On
On 26.06.2021 01:28, Lynne wrote:
Jun 25, 2021, 13:25 by t...@rothenpieler.org:
On 25.06.2021 10:14, Lynne wrote:
The prices have dropped a little, but the biggest difference is that
stuff is *actually* available now.
Unfortunately, now is not a good time to build an entire system.
Socket
Jun 25, 2021, 13:25 by t...@rothenpieler.org:
> On 25.06.2021 10:14, Lynne wrote:
>
>> The prices have dropped a little, but the biggest difference is that
>> stuff is *actually* available now.
>>
>> Unfortunately, now is not a good time to build an entire system.
>> Socket AM4's finished, so if
On Fri, 25 Jun 2021, Hayden Myers wrote:
Some encoders send GET_PARAMETER requests as a keep-alive mechanism.
If the client doesn't reply with an OK message, the encoder will close
the session. This was encountered with the impath i5110 encoder, when
the RTSP Keep-Alive checkbox is enabled
Some encoders send GET_PARAMETER requests as a keep-alive mechanism.
If the client doesn't reply with an OK message, the encoder will close
the session. This was encountered with the impath i5110 encoder, when
the RTSP Keep-Alive checkbox is enabled under streaming settings.
Alternatively one may
On 6/25/2021 12:58 PM, Michael Niedermayer wrote:
On Fri, Jun 25, 2021 at 09:22:03AM -0300, James Almer wrote:
On 6/24/2021 5:57 PM, Michael Niedermayer wrote:
Fixes: signed integer overflow: 27886261755 + 71189412840
cannot be represented in type 'long'
Fixes:
On Fri, Jun 25, 2021 at 09:22:03AM -0300, James Almer wrote:
> On 6/24/2021 5:57 PM, Michael Niedermayer wrote:
> > Fixes: signed integer overflow: 27886261755 + 71189412840
> > cannot be represented in type 'long'
> > Fixes:
> >
Hi Alan,
On Fri, Jun 25, 2021 at 7:53 AM Alan Kelly wrote:
>
>
> On Fri, Jun 25, 2021 at 1:26 PM Ronald S. Bultje
> wrote:
>
>> Hi Alan,
>>
>> On Fri, Jun 25, 2021 at 3:59 AM Alan Kelly <
>> alankelly-at-google@ffmpeg.org> wrote:
>>
>>> These functions replace all ff_hscale8to15_*_ssse3
On 6/24/2021 5:57 PM, Michael Niedermayer wrote:
Fixes: signed integer overflow: 27886261755 + 71189412840
cannot be represented in type 'long'
Fixes:
35215/clusterfuzz-testcase-minimized-ffmpeg_dem_SBG_fuzzer-6123272247836672
Found-by: continuous fuzzing process
On Fri, Jun 25, 2021 at 1:26 PM Ronald S. Bultje wrote:
> Hi Alan,
>
> On Fri, Jun 25, 2021 at 3:59 AM Alan Kelly <
> alankelly-at-google@ffmpeg.org> wrote:
>
>> These functions replace all ff_hscale8to15_*_ssse3 when avx2 is available.
>>
>
> Re-asking a question I asked before in the other
Hi Alan,
On Fri, Jun 25, 2021 at 3:59 AM Alan Kelly <
alankelly-at-google@ffmpeg.org> wrote:
> These functions replace all ff_hscale8to15_*_ssse3 when avx2 is available.
>
Re-asking a question I asked before in the other thread:
Also, what is the cycle count of ssse3/avx2 implementation
On 25.06.2021 10:14, Lynne wrote:
The prices have dropped a little, but the biggest difference is that
stuff is *actually* available now.
Unfortunately, now is not a good time to build an entire system.
Socket AM4's finished, so if I build an AMD system, it'll be obsolete
within a year or so,
On Fri, Jun 25, 2021 at 10:40 AM Lynne wrote:
> Jun 25, 2021, 09:54 by alankelly-at-google@ffmpeg.org:
>
> > Broadwell and later and Zen3 and later have fast gather instructions.
> > ---
> > Gather requires between 9 and 12 cycles on Haswell, 5 to 7 on Broadwell,
> > and 2 to 5 on Skylake
Jun 25, 2021, 09:54 by alankelly-at-google@ffmpeg.org:
> Broadwell and later and Zen3 and later have fast gather instructions.
> ---
> Gather requires between 9 and 12 cycles on Haswell, 5 to 7 on Broadwell,
> and 2 to 5 on Skylake and newer. It is also slow on AMD before Zen 3.
>
Jun 24, 2021, 22:59 by stefa...@gmail.com:
> On Thu, Jun 24, 2021 at 2:04 PM Michael Niedermayer
> wrote:
> [...]
>
>> > iam also adding stefano to the CC as there was some communication between
>> > a few developers and SPI about this expense and iam not sure stefano is
>> > following
>> > the
These functions replace all ff_hscale8to15_*_ssse3 when avx2 is available.
---
libswscale/swscale_internal.h | 2 +
libswscale/utils.c| 37 +++
libswscale/x86/Makefile | 1 +
libswscale/x86/scale_avx2.asm | 112 ++
Broadwell and later and Zen3 and later have fast gather instructions.
---
Gather requires between 9 and 12 cycles on Haswell, 5 to 7 on Broadwell,
and 2 to 5 on Skylake and newer. It is also slow on AMD before Zen 3.
libavutil/cpu.h | 2 ++
libavutil/x86/cpu.c | 18 --
18 matches
Mail list logo
