On 19/09/16 12:59, Diego Biurrun wrote:
> On Mon, Sep 19, 2016 at 12:53:33PM +0200, Alexandra Hájková wrote:
%if HAVE_AVX2_EXTERNAL
INIT_YMM avx2
IDCT_DC16, 2, 8
@@ -100,6 +804,12 @@ IDCT_DC_NL 8, 10
IDCT_DC16, 4, 10
IDCT_DC32, 16, 10
>> %if HAVE_AVX2_EXTERNAL
>> INIT_YMM avx2
>> IDCT_DC16, 2, 8
>> @@ -100,6 +804,12 @@ IDCT_DC_NL 8, 10
>> IDCT_DC16, 4, 10
>> IDCT_DC32, 16, 10
>>
>> +INIT_XMM avx
>> +IDCT_4x4 10
>> +IDCT_8x8 10
>> +IDCT_16x16 10
>> +IDCT_32x32 10
>
> It seems rather silly to duplicate
On Mon, Sep 19, 2016 at 12:53:33PM +0200, Alexandra Hájková wrote:
> >> %if HAVE_AVX2_EXTERNAL
> >> INIT_YMM avx2
> >> IDCT_DC16, 2, 8
> >> @@ -100,6 +804,12 @@ IDCT_DC_NL 8, 10
> >> IDCT_DC16, 4, 10
> >> IDCT_DC32, 16, 10
> >>
> >> +INIT_XMM avx
> >> +IDCT_4x4 10
> >>
Added support for P010 and YUV444P16 pixel formats to hwupload_cuda.
Thanks,
Yogender
---
This email message is for the sole use of the intended recipient(s) and may
contain
confidential information. Any
On Sat, Sep 17, 2016 at 9:28 AM, Luca Barbato wrote:
> From: Yogender Gupta
>
> Their order got swapped in SDK 7.
> ---
> configure | 4 ++--
> libavcodec/nvenc.c | 2 +-
> 2 files changed, 3 insertions(+), 3 deletions(-)
Shouldn't we still
On 19/09/16 18:59, Anton Khirnov wrote:
> ---
> The scalarproduct test now makes sure to generate such vectors that the result
> doesn't overflow int32. This seems to fix the PPC failures from Oracle.
> ---
Thank you, sounds good to me =)
___
---
The scalarproduct test now makes sure to generate such vectors that the result
doesn't overflow int32. This seems to fix the PPC failures from Oracle.
---
tests/checkasm/Makefile | 1 +
tests/checkasm/audiodsp.c | 146 ++
On 19/09/16 18:48, Vittorio Giovara wrote:
> On Sat, Sep 17, 2016 at 9:28 AM, Luca Barbato wrote:
>> From: Yogender Gupta
>>
>> Their order got swapped in SDK 7.
>> ---
>> configure | 4 ++--
>> libavcodec/nvenc.c | 2 +-
>> 2 files changed, 3
Quoting Diego Biurrun (2016-09-06 15:37:27)
> On Mon, Sep 05, 2016 at 01:02:36PM +0200, Anton Khirnov wrote:
> > --- /dev/null
> > +++ b/tests/checkasm/blockdsp.c
> > @@ -0,0 +1,68 @@
> > +/*
> > + * Copyright (c) 2015 Henrik Gramner
>
> Henrik?
It's mostly a trivial modification of the bswapdsp
Not a super-thorough review by any means, but anyway...
On Sun, Sep 18, 2016 at 7:35 PM, Alexandra Hájková
wrote:
[...]
> +SECTION_RODATA
Check if any of the constants are duplicates of already existing ones.
[...]
> +%macro TR_4x4 2
> +; interleaves src0
Quoting Alexandra Hájková (2016-09-18 19:35:25)
> ---
> libavcodec/x86/hevc_idct.asm | 710
> ++
> libavcodec/x86/hevcdsp_init.c | 21 ++
> 2 files changed, 731 insertions(+)
>
> diff --git a/libavcodec/x86/hevc_idct.asm b/libavcodec/x86/hevc_idct.asm
>
---
libavcodec/avcodec.h | 3 +++
1 file changed, 3 insertions(+)
diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
index c863e53..840222c 100644
--- a/libavcodec/avcodec.h
+++ b/libavcodec/avcodec.h
@@ -1190,6 +1190,9 @@ typedef struct AVCPBProperties {
* @{
*/
enum
---
tests/fate/screen.mak| 3 +++
tests/ref/fate/rscc-8bit | 2 ++
2 files changed, 5 insertions(+)
create mode 100644 tests/ref/fate/rscc-8bit
diff --git a/tests/fate/screen.mak b/tests/fate/screen.mak
index 7e91f19..a386fd3 100644
--- a/tests/fate/screen.mak
+++ b/tests/fate/screen.mak
@@
From: Carl Eugen Hoyos
Signed-off-by: Vittorio Giovara
---
libavcodec/rscc.c | 21 ++---
1 file changed, 18 insertions(+), 3 deletions(-)
diff --git a/libavcodec/rscc.c b/libavcodec/rscc.c
index 4a91783..91ffae2 100644
---
On Sun, Sep 18, 2016 at 02:07:52PM -0300, James Almer wrote:
> The guessed layout was being stored in the decoder context, which in the case
> of
> stream copy is unused.
>
> Signed-off-by: James Almer
> ---
> Didn't run FATE after this patch, but aside from
15 matches
Mail list logo