guibuilder_pel7ppc64lebuilder0
5 years ago
8 changed files with 809 additions and 0 deletions
@ -0,0 +1,35 @@
@@ -0,0 +1,35 @@
|
||||
From: Adam Sampson <unknown@bitbucket> |
||||
Date: Sun, 14 Oct 2018 14:04:18 +0200 |
||||
Subject: detect512 is needed on all architectures |
||||
|
||||
--- |
||||
source/common/cpu.cpp | 9 +++++---- |
||||
1 file changed, 5 insertions(+), 4 deletions(-) |
||||
|
||||
diff --git a/source/common/cpu.cpp b/source/common/cpu.cpp |
||||
index 0681ff5..fa687da 100644 |
||||
--- a/source/common/cpu.cpp |
||||
+++ b/source/common/cpu.cpp |
||||
@@ -110,6 +110,11 @@ const cpu_name_t cpu_names[] = |
||||
{ "", 0 }, |
||||
}; |
||||
|
||||
+bool detect512() |
||||
+{ |
||||
+ return(enable512); |
||||
+} |
||||
+ |
||||
#if X265_ARCH_X86 |
||||
|
||||
extern "C" { |
||||
@@ -123,10 +128,6 @@ uint64_t PFX(cpu_xgetbv)(int xcr); |
||||
#pragma warning(disable: 4309) // truncation of constant value |
||||
#endif |
||||
|
||||
-bool detect512() |
||||
-{ |
||||
- return(enable512); |
||||
-} |
||||
uint32_t cpu_detect(bool benableavx512 ) |
||||
{ |
||||
|
||||
@ -0,0 +1,347 @@
@@ -0,0 +1,347 @@
|
||||
--- ./source/common/arm/asm-primitives.cpp.orig 2018-05-21 02:33:10.000000000 -0600 |
||||
+++ ./source/common/arm/asm-primitives.cpp 2018-05-28 20:38:37.302378303 -0600 |
||||
@@ -48,77 +48,77 @@ void setupAssemblyPrimitives(EncoderPrim |
||||
p.ssim_4x4x2_core = PFX(ssim_4x4x2_core_neon); |
||||
|
||||
// addAvg |
||||
- p.pu[LUMA_4x4].addAvg = PFX(addAvg_4x4_neon); |
||||
- p.pu[LUMA_4x8].addAvg = PFX(addAvg_4x8_neon); |
||||
- p.pu[LUMA_4x16].addAvg = PFX(addAvg_4x16_neon); |
||||
- p.pu[LUMA_8x4].addAvg = PFX(addAvg_8x4_neon); |
||||
- p.pu[LUMA_8x8].addAvg = PFX(addAvg_8x8_neon); |
||||
- p.pu[LUMA_8x16].addAvg = PFX(addAvg_8x16_neon); |
||||
- p.pu[LUMA_8x32].addAvg = PFX(addAvg_8x32_neon); |
||||
- p.pu[LUMA_12x16].addAvg = PFX(addAvg_12x16_neon); |
||||
- p.pu[LUMA_16x4].addAvg = PFX(addAvg_16x4_neon); |
||||
- p.pu[LUMA_16x8].addAvg = PFX(addAvg_16x8_neon); |
||||
- p.pu[LUMA_16x12].addAvg = PFX(addAvg_16x12_neon); |
||||
- p.pu[LUMA_16x16].addAvg = PFX(addAvg_16x16_neon); |
||||
- p.pu[LUMA_16x32].addAvg = PFX(addAvg_16x32_neon); |
||||
- p.pu[LUMA_16x64].addAvg = PFX(addAvg_16x64_neon); |
||||
- p.pu[LUMA_24x32].addAvg = PFX(addAvg_24x32_neon); |
||||
- p.pu[LUMA_32x8].addAvg = PFX(addAvg_32x8_neon); |
||||
- p.pu[LUMA_32x16].addAvg = PFX(addAvg_32x16_neon); |
||||
- p.pu[LUMA_32x24].addAvg = PFX(addAvg_32x24_neon); |
||||
- p.pu[LUMA_32x32].addAvg = PFX(addAvg_32x32_neon); |
||||
- p.pu[LUMA_32x64].addAvg = PFX(addAvg_32x64_neon); |
||||
- p.pu[LUMA_48x64].addAvg = PFX(addAvg_48x64_neon); |
||||
- p.pu[LUMA_64x16].addAvg = PFX(addAvg_64x16_neon); |
||||
- p.pu[LUMA_64x32].addAvg = PFX(addAvg_64x32_neon); |
||||
- p.pu[LUMA_64x48].addAvg = PFX(addAvg_64x48_neon); |
||||
- p.pu[LUMA_64x64].addAvg = PFX(addAvg_64x64_neon); |
||||
+ p.pu[LUMA_4x4].addAvg[ALIGNED] = PFX(addAvg_4x4_neon); |
||||
+ p.pu[LUMA_4x8].addAvg[ALIGNED] = PFX(addAvg_4x8_neon); |
||||
+ p.pu[LUMA_4x16].addAvg[ALIGNED] = PFX(addAvg_4x16_neon); |
||||
+ p.pu[LUMA_8x4].addAvg[ALIGNED] = PFX(addAvg_8x4_neon); |
||||
+ p.pu[LUMA_8x8].addAvg[ALIGNED] = PFX(addAvg_8x8_neon); |
||||
+ p.pu[LUMA_8x16].addAvg[ALIGNED] = PFX(addAvg_8x16_neon); |
||||
+ p.pu[LUMA_8x32].addAvg[ALIGNED] = PFX(addAvg_8x32_neon); |
||||
+ p.pu[LUMA_12x16].addAvg[ALIGNED] = PFX(addAvg_12x16_neon); |
||||
+ p.pu[LUMA_16x4].addAvg[ALIGNED] = PFX(addAvg_16x4_neon); |
||||
+ p.pu[LUMA_16x8].addAvg[ALIGNED] = PFX(addAvg_16x8_neon); |
||||
+ p.pu[LUMA_16x12].addAvg[ALIGNED] = PFX(addAvg_16x12_neon); |
||||
+ p.pu[LUMA_16x16].addAvg[ALIGNED] = PFX(addAvg_16x16_neon); |
||||
+ p.pu[LUMA_16x32].addAvg[ALIGNED] = PFX(addAvg_16x32_neon); |
||||
+ p.pu[LUMA_16x64].addAvg[ALIGNED] = PFX(addAvg_16x64_neon); |
||||
+ p.pu[LUMA_24x32].addAvg[ALIGNED] = PFX(addAvg_24x32_neon); |
||||
+ p.pu[LUMA_32x8].addAvg[ALIGNED] = PFX(addAvg_32x8_neon); |
||||
+ p.pu[LUMA_32x16].addAvg[ALIGNED] = PFX(addAvg_32x16_neon); |
||||
+ p.pu[LUMA_32x24].addAvg[ALIGNED] = PFX(addAvg_32x24_neon); |
||||
+ p.pu[LUMA_32x32].addAvg[ALIGNED] = PFX(addAvg_32x32_neon); |
||||
+ p.pu[LUMA_32x64].addAvg[ALIGNED] = PFX(addAvg_32x64_neon); |
||||
+ p.pu[LUMA_48x64].addAvg[ALIGNED] = PFX(addAvg_48x64_neon); |
||||
+ p.pu[LUMA_64x16].addAvg[ALIGNED] = PFX(addAvg_64x16_neon); |
||||
+ p.pu[LUMA_64x32].addAvg[ALIGNED] = PFX(addAvg_64x32_neon); |
||||
+ p.pu[LUMA_64x48].addAvg[ALIGNED] = PFX(addAvg_64x48_neon); |
||||
+ p.pu[LUMA_64x64].addAvg[ALIGNED] = PFX(addAvg_64x64_neon); |
||||
|
||||
// chroma addAvg |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_4x2].addAvg = PFX(addAvg_4x2_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_4x4].addAvg = PFX(addAvg_4x4_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_4x8].addAvg = PFX(addAvg_4x8_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_4x16].addAvg = PFX(addAvg_4x16_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_6x8].addAvg = PFX(addAvg_6x8_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_8x2].addAvg = PFX(addAvg_8x2_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_8x4].addAvg = PFX(addAvg_8x4_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_8x6].addAvg = PFX(addAvg_8x6_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_8x8].addAvg = PFX(addAvg_8x8_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_8x16].addAvg = PFX(addAvg_8x16_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_8x32].addAvg = PFX(addAvg_8x32_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_12x16].addAvg = PFX(addAvg_12x16_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_16x4].addAvg = PFX(addAvg_16x4_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_16x8].addAvg = PFX(addAvg_16x8_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_16x12].addAvg = PFX(addAvg_16x12_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_16x16].addAvg = PFX(addAvg_16x16_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_16x32].addAvg = PFX(addAvg_16x32_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_24x32].addAvg = PFX(addAvg_24x32_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_32x8].addAvg = PFX(addAvg_32x8_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_32x16].addAvg = PFX(addAvg_32x16_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_32x24].addAvg = PFX(addAvg_32x24_neon); |
||||
- p.chroma[X265_CSP_I420].pu[CHROMA_420_32x32].addAvg = PFX(addAvg_32x32_neon); |
||||
- |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_4x8].addAvg = PFX(addAvg_4x8_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_4x16].addAvg = PFX(addAvg_4x16_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_4x32].addAvg = PFX(addAvg_4x32_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_6x16].addAvg = PFX(addAvg_6x16_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_8x4].addAvg = PFX(addAvg_8x4_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_8x8].addAvg = PFX(addAvg_8x8_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_8x12].addAvg = PFX(addAvg_8x12_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_8x16].addAvg = PFX(addAvg_8x16_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_8x32].addAvg = PFX(addAvg_8x32_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_8x64].addAvg = PFX(addAvg_8x64_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_12x32].addAvg = PFX(addAvg_12x32_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_16x8].addAvg = PFX(addAvg_16x8_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_16x16].addAvg = PFX(addAvg_16x16_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_16x24].addAvg = PFX(addAvg_16x24_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_16x32].addAvg = PFX(addAvg_16x32_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_16x64].addAvg = PFX(addAvg_16x64_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_24x64].addAvg = PFX(addAvg_24x64_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_32x16].addAvg = PFX(addAvg_32x16_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_32x32].addAvg = PFX(addAvg_32x32_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_32x48].addAvg = PFX(addAvg_32x48_neon); |
||||
- p.chroma[X265_CSP_I422].pu[CHROMA_422_32x64].addAvg = PFX(addAvg_32x64_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_4x2].addAvg[ALIGNED] = PFX(addAvg_4x2_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_4x4].addAvg[ALIGNED] = PFX(addAvg_4x4_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_4x8].addAvg[ALIGNED] = PFX(addAvg_4x8_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_4x16].addAvg[ALIGNED] = PFX(addAvg_4x16_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_6x8].addAvg[ALIGNED] = PFX(addAvg_6x8_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_8x2].addAvg[ALIGNED] = PFX(addAvg_8x2_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_8x4].addAvg[ALIGNED] = PFX(addAvg_8x4_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_8x6].addAvg[ALIGNED] = PFX(addAvg_8x6_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_8x8].addAvg[ALIGNED] = PFX(addAvg_8x8_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_8x16].addAvg[ALIGNED] = PFX(addAvg_8x16_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_8x32].addAvg[ALIGNED] = PFX(addAvg_8x32_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_12x16].addAvg[ALIGNED] = PFX(addAvg_12x16_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_16x4].addAvg[ALIGNED] = PFX(addAvg_16x4_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_16x8].addAvg[ALIGNED] = PFX(addAvg_16x8_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_16x12].addAvg[ALIGNED] = PFX(addAvg_16x12_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_16x16].addAvg[ALIGNED] = PFX(addAvg_16x16_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_16x32].addAvg[ALIGNED] = PFX(addAvg_16x32_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_24x32].addAvg[ALIGNED] = PFX(addAvg_24x32_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_32x8].addAvg[ALIGNED] = PFX(addAvg_32x8_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_32x16].addAvg[ALIGNED] = PFX(addAvg_32x16_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_32x24].addAvg[ALIGNED] = PFX(addAvg_32x24_neon); |
||||
+ p.chroma[X265_CSP_I420].pu[CHROMA_420_32x32].addAvg[ALIGNED] = PFX(addAvg_32x32_neon); |
||||
+ |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_4x8].addAvg[ALIGNED] = PFX(addAvg_4x8_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_4x16].addAvg[ALIGNED] = PFX(addAvg_4x16_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_4x32].addAvg[ALIGNED] = PFX(addAvg_4x32_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_6x16].addAvg[ALIGNED] = PFX(addAvg_6x16_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_8x4].addAvg[ALIGNED] = PFX(addAvg_8x4_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_8x8].addAvg[ALIGNED] = PFX(addAvg_8x8_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_8x12].addAvg[ALIGNED] = PFX(addAvg_8x12_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_8x16].addAvg[ALIGNED] = PFX(addAvg_8x16_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_8x32].addAvg[ALIGNED] = PFX(addAvg_8x32_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_8x64].addAvg[ALIGNED] = PFX(addAvg_8x64_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_12x32].addAvg[ALIGNED] = PFX(addAvg_12x32_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_16x8].addAvg[ALIGNED] = PFX(addAvg_16x8_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_16x16].addAvg[ALIGNED] = PFX(addAvg_16x16_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_16x24].addAvg[ALIGNED] = PFX(addAvg_16x24_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_16x32].addAvg[ALIGNED] = PFX(addAvg_16x32_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_16x64].addAvg[ALIGNED] = PFX(addAvg_16x64_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_24x64].addAvg[ALIGNED] = PFX(addAvg_24x64_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_32x16].addAvg[ALIGNED] = PFX(addAvg_32x16_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_32x32].addAvg[ALIGNED] = PFX(addAvg_32x32_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_32x48].addAvg[ALIGNED] = PFX(addAvg_32x48_neon); |
||||
+ p.chroma[X265_CSP_I422].pu[CHROMA_422_32x64].addAvg[ALIGNED] = PFX(addAvg_32x64_neon); |
||||
|
||||
// quant |
||||
p.quant = PFX(quant_neon); |
||||
@@ -402,7 +402,7 @@ void setupAssemblyPrimitives(EncoderPrim |
||||
p.scale2D_64to32 = PFX(scale2D_64to32_neon); |
||||
|
||||
// scale1D_128to64 |
||||
- p.scale1D_128to64 = PFX(scale1D_128to64_neon); |
||||
+ p.scale1D_128to64[ALIGNED] = PFX(scale1D_128to64_neon); |
||||
|
||||
// copy_count |
||||
p.cu[BLOCK_4x4].copy_cnt = PFX(copy_cnt_4_neon); |
||||
@@ -411,37 +411,37 @@ void setupAssemblyPrimitives(EncoderPrim |
||||
p.cu[BLOCK_32x32].copy_cnt = PFX(copy_cnt_32_neon); |
||||
|
||||
// filterPixelToShort |
||||
- p.pu[LUMA_4x4].convert_p2s = PFX(filterPixelToShort_4x4_neon); |
||||
- p.pu[LUMA_4x8].convert_p2s = PFX(filterPixelToShort_4x8_neon); |
||||
- p.pu[LUMA_4x16].convert_p2s = PFX(filterPixelToShort_4x16_neon); |
||||
- p.pu[LUMA_8x4].convert_p2s = PFX(filterPixelToShort_8x4_neon); |
||||
- p.pu[LUMA_8x8].convert_p2s = PFX(filterPixelToShort_8x8_neon); |
||||
- p.pu[LUMA_8x16].convert_p2s = PFX(filterPixelToShort_8x16_neon); |
||||
- p.pu[LUMA_8x32].convert_p2s = PFX(filterPixelToShort_8x32_neon); |
||||
- p.pu[LUMA_12x16].convert_p2s = PFX(filterPixelToShort_12x16_neon); |
||||
- p.pu[LUMA_16x4].convert_p2s = PFX(filterPixelToShort_16x4_neon); |
||||
- p.pu[LUMA_16x8].convert_p2s = PFX(filterPixelToShort_16x8_neon); |
||||
- p.pu[LUMA_16x12].convert_p2s = PFX(filterPixelToShort_16x12_neon); |
||||
- p.pu[LUMA_16x16].convert_p2s = PFX(filterPixelToShort_16x16_neon); |
||||
- p.pu[LUMA_16x32].convert_p2s = PFX(filterPixelToShort_16x32_neon); |
||||
- p.pu[LUMA_16x64].convert_p2s = PFX(filterPixelToShort_16x64_neon); |
||||
- p.pu[LUMA_24x32].convert_p2s = PFX(filterPixelToShort_24x32_neon); |
||||
- p.pu[LUMA_32x8].convert_p2s = PFX(filterPixelToShort_32x8_neon); |
||||
- p.pu[LUMA_32x16].convert_p2s = PFX(filterPixelToShort_32x16_neon); |
||||
- p.pu[LUMA_32x24].convert_p2s = PFX(filterPixelToShort_32x24_neon); |
||||
- p.pu[LUMA_32x32].convert_p2s = PFX(filterPixelToShort_32x32_neon); |
||||
- p.pu[LUMA_32x64].convert_p2s = PFX(filterPixelToShort_32x64_neon); |
||||
- p.pu[LUMA_48x64].convert_p2s = PFX(filterPixelToShort_48x64_neon); |
||||
- p.pu[LUMA_64x16].convert_p2s = PFX(filterPixelToShort_64x16_neon); |
||||
- p.pu[LUMA_64x32].convert_p2s = PFX(filterPixelToShort_64x32_neon); |
||||
- p.pu[LUMA_64x48].convert_p2s = PFX(filterPixelToShort_64x48_neon); |
||||
- p.pu[LUMA_64x64].convert_p2s = PFX(filterPixelToShort_64x64_neon); |
||||
+ p.pu[LUMA_4x4].convert_p2s[ALIGNED] = PFX(filterPixelToShort_4x4_neon); |
||||
+ p.pu[LUMA_4x8].convert_p2s[ALIGNED] = PFX(filterPixelToShort_4x8_neon); |
||||
+ p.pu[LUMA_4x16].convert_p2s[ALIGNED] = PFX(filterPixelToShort_4x16_neon); |
||||
+ p.pu[LUMA_8x4].convert_p2s[ALIGNED] = PFX(filterPixelToShort_8x4_neon); |
||||
+ p.pu[LUMA_8x8].convert_p2s[ALIGNED] = PFX(filterPixelToShort_8x8_neon); |
||||
+ p.pu[LUMA_8x16].convert_p2s[ALIGNED] = PFX(filterPixelToShort_8x16_neon); |
||||
+ p.pu[LUMA_8x32].convert_p2s[ALIGNED] = PFX(filterPixelToShort_8x32_neon); |
||||
+ p.pu[LUMA_12x16].convert_p2s[ALIGNED] = PFX(filterPixelToShort_12x16_neon); |
||||
+ p.pu[LUMA_16x4].convert_p2s[ALIGNED] = PFX(filterPixelToShort_16x4_neon); |
||||
+ p.pu[LUMA_16x8].convert_p2s[ALIGNED] = PFX(filterPixelToShort_16x8_neon); |
||||
+ p.pu[LUMA_16x12].convert_p2s[ALIGNED] = PFX(filterPixelToShort_16x12_neon); |
||||
+ p.pu[LUMA_16x16].convert_p2s[ALIGNED] = PFX(filterPixelToShort_16x16_neon); |
||||
+ p.pu[LUMA_16x32].convert_p2s[ALIGNED] = PFX(filterPixelToShort_16x32_neon); |
||||
+ p.pu[LUMA_16x64].convert_p2s[ALIGNED] = PFX(filterPixelToShort_16x64_neon); |
||||
+ p.pu[LUMA_24x32].convert_p2s[ALIGNED] = PFX(filterPixelToShort_24x32_neon); |
||||
+ p.pu[LUMA_32x8].convert_p2s[ALIGNED] = PFX(filterPixelToShort_32x8_neon); |
||||
+ p.pu[LUMA_32x16].convert_p2s[ALIGNED] = PFX(filterPixelToShort_32x16_neon); |
||||
+ p.pu[LUMA_32x24].convert_p2s[ALIGNED] = PFX(filterPixelToShort_32x24_neon); |
||||
+ p.pu[LUMA_32x32].convert_p2s[ALIGNED] = PFX(filterPixelToShort_32x32_neon); |
||||
+ p.pu[LUMA_32x64].convert_p2s[ALIGNED] = PFX(filterPixelToShort_32x64_neon); |
||||
+ p.pu[LUMA_48x64].convert_p2s[ALIGNED] = PFX(filterPixelToShort_48x64_neon); |
||||
+ p.pu[LUMA_64x16].convert_p2s[ALIGNED] = PFX(filterPixelToShort_64x16_neon); |
||||
+ p.pu[LUMA_64x32].convert_p2s[ALIGNED] = PFX(filterPixelToShort_64x32_neon); |
||||
+ p.pu[LUMA_64x48].convert_p2s[ALIGNED] = PFX(filterPixelToShort_64x48_neon); |
||||
+ p.pu[LUMA_64x64].convert_p2s[ALIGNED] = PFX(filterPixelToShort_64x64_neon); |
||||
|
||||
// Block_fill |
||||
- p.cu[BLOCK_4x4].blockfill_s = PFX(blockfill_s_4x4_neon); |
||||
- p.cu[BLOCK_8x8].blockfill_s = PFX(blockfill_s_8x8_neon); |
||||
- p.cu[BLOCK_16x16].blockfill_s = PFX(blockfill_s_16x16_neon); |
||||
- p.cu[BLOCK_32x32].blockfill_s = PFX(blockfill_s_32x32_neon); |
||||
+ p.cu[BLOCK_4x4].blockfill_s[ALIGNED] = PFX(blockfill_s_4x4_neon); |
||||
+ p.cu[BLOCK_8x8].blockfill_s[ALIGNED] = PFX(blockfill_s_8x8_neon); |
||||
+ p.cu[BLOCK_16x16].blockfill_s[ALIGNED] = PFX(blockfill_s_16x16_neon); |
||||
+ p.cu[BLOCK_32x32].blockfill_s[ALIGNED] = PFX(blockfill_s_32x32_neon); |
||||
|
||||
// Blockcopy_ss |
||||
p.cu[BLOCK_4x4].copy_ss = PFX(blockcopy_ss_4x4_neon); |
||||
@@ -495,21 +495,21 @@ void setupAssemblyPrimitives(EncoderPrim |
||||
p.chroma[X265_CSP_I422].cu[BLOCK_422_32x64].copy_sp = PFX(blockcopy_sp_32x64_neon); |
||||
|
||||
// pixel_add_ps |
||||
- p.cu[BLOCK_4x4].add_ps = PFX(pixel_add_ps_4x4_neon); |
||||
- p.cu[BLOCK_8x8].add_ps = PFX(pixel_add_ps_8x8_neon); |
||||
- p.cu[BLOCK_16x16].add_ps = PFX(pixel_add_ps_16x16_neon); |
||||
- p.cu[BLOCK_32x32].add_ps = PFX(pixel_add_ps_32x32_neon); |
||||
- p.cu[BLOCK_64x64].add_ps = PFX(pixel_add_ps_64x64_neon); |
||||
+ p.cu[BLOCK_4x4].add_ps[ALIGNED] = PFX(pixel_add_ps_4x4_neon); |
||||
+ p.cu[BLOCK_8x8].add_ps[ALIGNED] = PFX(pixel_add_ps_8x8_neon); |
||||
+ p.cu[BLOCK_16x16].add_ps[ALIGNED] = PFX(pixel_add_ps_16x16_neon); |
||||
+ p.cu[BLOCK_32x32].add_ps[ALIGNED] = PFX(pixel_add_ps_32x32_neon); |
||||
+ p.cu[BLOCK_64x64].add_ps[ALIGNED] = PFX(pixel_add_ps_64x64_neon); |
||||
|
||||
// chroma add_ps |
||||
- p.chroma[X265_CSP_I420].cu[BLOCK_420_4x4].add_ps = PFX(pixel_add_ps_4x4_neon); |
||||
- p.chroma[X265_CSP_I420].cu[BLOCK_420_8x8].add_ps = PFX(pixel_add_ps_8x8_neon); |
||||
- p.chroma[X265_CSP_I420].cu[BLOCK_420_16x16].add_ps = PFX(pixel_add_ps_16x16_neon); |
||||
- p.chroma[X265_CSP_I420].cu[BLOCK_420_32x32].add_ps = PFX(pixel_add_ps_32x32_neon); |
||||
- p.chroma[X265_CSP_I422].cu[BLOCK_422_4x8].add_ps = PFX(pixel_add_ps_4x8_neon); |
||||
- p.chroma[X265_CSP_I422].cu[BLOCK_422_8x16].add_ps = PFX(pixel_add_ps_8x16_neon); |
||||
- p.chroma[X265_CSP_I422].cu[BLOCK_422_16x32].add_ps = PFX(pixel_add_ps_16x32_neon); |
||||
- p.chroma[X265_CSP_I422].cu[BLOCK_422_32x64].add_ps = PFX(pixel_add_ps_32x64_neon); |
||||
+ p.chroma[X265_CSP_I420].cu[BLOCK_420_4x4].add_ps[ALIGNED] = PFX(pixel_add_ps_4x4_neon); |
||||
+ p.chroma[X265_CSP_I420].cu[BLOCK_420_8x8].add_ps[ALIGNED] = PFX(pixel_add_ps_8x8_neon); |
||||
+ p.chroma[X265_CSP_I420].cu[BLOCK_420_16x16].add_ps[ALIGNED] = PFX(pixel_add_ps_16x16_neon); |
||||
+ p.chroma[X265_CSP_I420].cu[BLOCK_420_32x32].add_ps[ALIGNED] = PFX(pixel_add_ps_32x32_neon); |
||||
+ p.chroma[X265_CSP_I422].cu[BLOCK_422_4x8].add_ps[ALIGNED] = PFX(pixel_add_ps_4x8_neon); |
||||
+ p.chroma[X265_CSP_I422].cu[BLOCK_422_8x16].add_ps[ALIGNED] = PFX(pixel_add_ps_8x16_neon); |
||||
+ p.chroma[X265_CSP_I422].cu[BLOCK_422_16x32].add_ps[ALIGNED] = PFX(pixel_add_ps_16x32_neon); |
||||
+ p.chroma[X265_CSP_I422].cu[BLOCK_422_32x64].add_ps[ALIGNED] = PFX(pixel_add_ps_32x64_neon); |
||||
|
||||
// cpy2Dto1D_shr |
||||
p.cu[BLOCK_4x4].cpy2Dto1D_shr = PFX(cpy2Dto1D_shr_4x4_neon); |
||||
@@ -518,10 +518,10 @@ void setupAssemblyPrimitives(EncoderPrim |
||||
p.cu[BLOCK_32x32].cpy2Dto1D_shr = PFX(cpy2Dto1D_shr_32x32_neon); |
||||
|
||||
// ssd_s |
||||
- p.cu[BLOCK_4x4].ssd_s = PFX(pixel_ssd_s_4x4_neon); |
||||
- p.cu[BLOCK_8x8].ssd_s = PFX(pixel_ssd_s_8x8_neon); |
||||
- p.cu[BLOCK_16x16].ssd_s = PFX(pixel_ssd_s_16x16_neon); |
||||
- p.cu[BLOCK_32x32].ssd_s = PFX(pixel_ssd_s_32x32_neon); |
||||
+ p.cu[BLOCK_4x4].ssd_s[ALIGNED] = PFX(pixel_ssd_s_4x4_neon); |
||||
+ p.cu[BLOCK_8x8].ssd_s[ALIGNED] = PFX(pixel_ssd_s_8x8_neon); |
||||
+ p.cu[BLOCK_16x16].ssd_s[ALIGNED] = PFX(pixel_ssd_s_16x16_neon); |
||||
+ p.cu[BLOCK_32x32].ssd_s[ALIGNED] = PFX(pixel_ssd_s_32x32_neon); |
||||
|
||||
// sse_ss |
||||
p.cu[BLOCK_4x4].sse_ss = PFX(pixel_sse_ss_4x4_neon); |
||||
@@ -548,10 +548,10 @@ void setupAssemblyPrimitives(EncoderPrim |
||||
p.chroma[X265_CSP_I422].cu[BLOCK_422_32x64].sub_ps = PFX(pixel_sub_ps_32x64_neon); |
||||
|
||||
// calc_Residual |
||||
- p.cu[BLOCK_4x4].calcresidual = PFX(getResidual4_neon); |
||||
- p.cu[BLOCK_8x8].calcresidual = PFX(getResidual8_neon); |
||||
- p.cu[BLOCK_16x16].calcresidual = PFX(getResidual16_neon); |
||||
- p.cu[BLOCK_32x32].calcresidual = PFX(getResidual32_neon); |
||||
+ p.cu[BLOCK_4x4].calcresidual[ALIGNED] = PFX(getResidual4_neon); |
||||
+ p.cu[BLOCK_8x8].calcresidual[ALIGNED] = PFX(getResidual8_neon); |
||||
+ p.cu[BLOCK_16x16].calcresidual[ALIGNED] = PFX(getResidual16_neon); |
||||
+ p.cu[BLOCK_32x32].calcresidual[ALIGNED] = PFX(getResidual32_neon); |
||||
|
||||
// sse_pp |
||||
p.cu[BLOCK_4x4].sse_pp = PFX(pixel_sse_pp_4x4_neon); |
||||
@@ -722,31 +722,31 @@ void setupAssemblyPrimitives(EncoderPrim |
||||
p.pu[LUMA_64x64].sad_x4 = PFX(sad_x4_64x64_neon); |
||||
|
||||
// pixel_avg_pp |
||||
- p.pu[LUMA_4x4].pixelavg_pp = PFX(pixel_avg_pp_4x4_neon); |
||||
- p.pu[LUMA_4x8].pixelavg_pp = PFX(pixel_avg_pp_4x8_neon); |
||||
- p.pu[LUMA_4x16].pixelavg_pp = PFX(pixel_avg_pp_4x16_neon); |
||||
- p.pu[LUMA_8x4].pixelavg_pp = PFX(pixel_avg_pp_8x4_neon); |
||||
- p.pu[LUMA_8x8].pixelavg_pp = PFX(pixel_avg_pp_8x8_neon); |
||||
- p.pu[LUMA_8x16].pixelavg_pp = PFX(pixel_avg_pp_8x16_neon); |
||||
- p.pu[LUMA_8x32].pixelavg_pp = PFX(pixel_avg_pp_8x32_neon); |
||||
- p.pu[LUMA_12x16].pixelavg_pp = PFX(pixel_avg_pp_12x16_neon); |
||||
- p.pu[LUMA_16x4].pixelavg_pp = PFX(pixel_avg_pp_16x4_neon); |
||||
- p.pu[LUMA_16x8].pixelavg_pp = PFX(pixel_avg_pp_16x8_neon); |
||||
- p.pu[LUMA_16x12].pixelavg_pp = PFX(pixel_avg_pp_16x12_neon); |
||||
- p.pu[LUMA_16x16].pixelavg_pp = PFX(pixel_avg_pp_16x16_neon); |
||||
- p.pu[LUMA_16x32].pixelavg_pp = PFX(pixel_avg_pp_16x32_neon); |
||||
- p.pu[LUMA_16x64].pixelavg_pp = PFX(pixel_avg_pp_16x64_neon); |
||||
- p.pu[LUMA_24x32].pixelavg_pp = PFX(pixel_avg_pp_24x32_neon); |
||||
- p.pu[LUMA_32x8].pixelavg_pp = PFX(pixel_avg_pp_32x8_neon); |
||||
- p.pu[LUMA_32x16].pixelavg_pp = PFX(pixel_avg_pp_32x16_neon); |
||||
- p.pu[LUMA_32x24].pixelavg_pp = PFX(pixel_avg_pp_32x24_neon); |
||||
- p.pu[LUMA_32x32].pixelavg_pp = PFX(pixel_avg_pp_32x32_neon); |
||||
- p.pu[LUMA_32x64].pixelavg_pp = PFX(pixel_avg_pp_32x64_neon); |
||||
- p.pu[LUMA_48x64].pixelavg_pp = PFX(pixel_avg_pp_48x64_neon); |
||||
- p.pu[LUMA_64x16].pixelavg_pp = PFX(pixel_avg_pp_64x16_neon); |
||||
- p.pu[LUMA_64x32].pixelavg_pp = PFX(pixel_avg_pp_64x32_neon); |
||||
- p.pu[LUMA_64x48].pixelavg_pp = PFX(pixel_avg_pp_64x48_neon); |
||||
- p.pu[LUMA_64x64].pixelavg_pp = PFX(pixel_avg_pp_64x64_neon); |
||||
+ p.pu[LUMA_4x4].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_4x4_neon); |
||||
+ p.pu[LUMA_4x8].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_4x8_neon); |
||||
+ p.pu[LUMA_4x16].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_4x16_neon); |
||||
+ p.pu[LUMA_8x4].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_8x4_neon); |
||||
+ p.pu[LUMA_8x8].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_8x8_neon); |
||||
+ p.pu[LUMA_8x16].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_8x16_neon); |
||||
+ p.pu[LUMA_8x32].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_8x32_neon); |
||||
+ p.pu[LUMA_12x16].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_12x16_neon); |
||||
+ p.pu[LUMA_16x4].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_16x4_neon); |
||||
+ p.pu[LUMA_16x8].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_16x8_neon); |
||||
+ p.pu[LUMA_16x12].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_16x12_neon); |
||||
+ p.pu[LUMA_16x16].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_16x16_neon); |
||||
+ p.pu[LUMA_16x32].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_16x32_neon); |
||||
+ p.pu[LUMA_16x64].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_16x64_neon); |
||||
+ p.pu[LUMA_24x32].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_24x32_neon); |
||||
+ p.pu[LUMA_32x8].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_32x8_neon); |
||||
+ p.pu[LUMA_32x16].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_32x16_neon); |
||||
+ p.pu[LUMA_32x24].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_32x24_neon); |
||||
+ p.pu[LUMA_32x32].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_32x32_neon); |
||||
+ p.pu[LUMA_32x64].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_32x64_neon); |
||||
+ p.pu[LUMA_48x64].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_48x64_neon); |
||||
+ p.pu[LUMA_64x16].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_64x16_neon); |
||||
+ p.pu[LUMA_64x32].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_64x32_neon); |
||||
+ p.pu[LUMA_64x48].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_64x48_neon); |
||||
+ p.pu[LUMA_64x64].pixelavg_pp[ALIGNED] = PFX(pixel_avg_pp_64x64_neon); |
||||
|
||||
// planecopy |
||||
p.planecopy_cp = PFX(pixel_planecopy_cp_neon); |
||||
@ -0,0 +1,60 @@
@@ -0,0 +1,60 @@
|
||||
diff -up x265_2.7/source/CMakeLists.txt.cflags x265_2.7/source/CMakeLists.txt |
||||
--- x265_2.7/source/CMakeLists.txt.cflags 2018-02-27 20:19:33.328932385 +0100 |
||||
+++ x265_2.7/source/CMakeLists.txt 2018-02-27 20:29:14.808956583 +0100 |
||||
@@ -233,18 +233,6 @@ if(GCC) |
||||
endif() |
||||
endif() |
||||
endif() |
||||
- if(ARM AND CROSS_COMPILE_ARM) |
||||
- set(ARM_ARGS -march=armv6 -mfloat-abi=soft -mfpu=vfp -marm -fPIC) |
||||
- elseif(ARM) |
||||
- find_package(Neon) |
||||
- if(CPU_HAS_NEON) |
||||
- set(ARM_ARGS -mcpu=native -mfloat-abi=hard -mfpu=neon -marm -fPIC) |
||||
- add_definitions(-DHAVE_NEON) |
||||
- else() |
||||
- set(ARM_ARGS -mcpu=native -mfloat-abi=hard -mfpu=vfp -marm) |
||||
- endif() |
||||
- endif() |
||||
- add_definitions(${ARM_ARGS}) |
||||
if(FPROFILE_GENERATE) |
||||
if(INTEL_CXX) |
||||
add_definitions(-prof-gen -prof-dir="${CMAKE_CURRENT_BINARY_DIR}") |
||||
@@ -517,7 +505,7 @@ if((MSVC_IDE OR XCODE OR GCC) AND ENABLE |
||||
add_custom_command( |
||||
OUTPUT ${ASM}.${SUFFIX} |
||||
COMMAND ${CMAKE_CXX_COMPILER} |
||||
- ARGS ${ARM_ARGS} -c ${ASM_SRC} -o ${ASM}.${SUFFIX} |
||||
+ ARGS ${CFLAGS} -c ${ASM_SRC} -o ${ASM}.${SUFFIX} |
||||
DEPENDS ${ASM_SRC}) |
||||
endforeach() |
||||
elseif(X86) |
||||
diff -up x265_2.7/source/dynamicHDR10/CMakeLists.txt.cflags x265_2.7/source/dynamicHDR10/CMakeLists.txt |
||||
--- x265_2.7/source/dynamicHDR10/CMakeLists.txt.cflags 2018-02-21 09:55:56.000000000 +0100 |
||||
+++ x265_2.7/source/dynamicHDR10/CMakeLists.txt 2018-02-27 20:29:45.377062994 +0100 |
||||
@@ -42,18 +42,6 @@ if(GCC) |
||||
endif() |
||||
endif() |
||||
endif() |
||||
- if(ARM AND CROSS_COMPILE_ARM) |
||||
- set(ARM_ARGS -march=armv6 -mfloat-abi=soft -mfpu=vfp -marm -fPIC) |
||||
- elseif(ARM) |
||||
- find_package(Neon) |
||||
- if(CPU_HAS_NEON) |
||||
- set(ARM_ARGS -mcpu=native -mfloat-abi=hard -mfpu=neon -marm -fPIC) |
||||
- add_definitions(-DHAVE_NEON) |
||||
- else() |
||||
- set(ARM_ARGS -mcpu=native -mfloat-abi=hard -mfpu=vfp -marm) |
||||
- endif() |
||||
- endif() |
||||
- add_definitions(${ARM_ARGS}) |
||||
if(FPROFILE_GENERATE) |
||||
if(INTEL_CXX) |
||||
add_definitions(-prof-gen -prof-dir="${CMAKE_CURRENT_BINARY_DIR}") |
||||
@@ -150,4 +138,4 @@ set(BIN_INSTALL_DIR bin CACHE STRING "In |
||||
option(ENABLE_SHARED "Build shared library" OFF) |
||||
|
||||
install(FILES hdr10plus.h DESTINATION include) |
||||
-endif() |
||||
\ Pas de fin de ligne à la fin du fichier |
||||
+endif() |
||||
@ -0,0 +1,59 @@
@@ -0,0 +1,59 @@
|
||||
diff -up x265_2.2/source/CMakeLists.txt.orig x265_2.2/source/CMakeLists.txt |
||||
--- x265_2.2/source/CMakeLists.txt.orig 2016-12-23 06:57:39.000000000 +0100 |
||||
+++ x265_2.2/source/CMakeLists.txt 2017-01-03 11:18:34.773738470 +0100 |
||||
@@ -41,9 +41,11 @@ SET(CMAKE_MODULE_PATH "${PROJECT_SOURCE_ |
||||
# System architecture detection |
||||
string(TOLOWER "${CMAKE_SYSTEM_PROCESSOR}" SYSPROC) |
||||
set(X86_ALIASES x86 i386 i686 x86_64 amd64) |
||||
-set(ARM_ALIASES armv6l armv7l) |
||||
+set(ARMv6_ALIASES armv6l) |
||||
+set(ARMv7_ALIASES armv7l) |
||||
list(FIND X86_ALIASES "${SYSPROC}" X86MATCH) |
||||
-list(FIND ARM_ALIASES "${SYSPROC}" ARMMATCH) |
||||
+list(FIND ARMv6_ALIASES "${SYSPROC}" ARMv6MATCH) |
||||
+list(FIND ARMv7_ALIASES "${SYSPROC}" ARMv7MATCH) |
||||
set(POWER_ALIASES ppc64 ppc64le) |
||||
list(FIND POWER_ALIASES "${SYSPROC}" POWERMATCH) |
||||
if("${SYSPROC}" STREQUAL "" OR X86MATCH GREATER "-1") |
||||
@@ -65,15 +67,24 @@ elseif(POWERMATCH GREATER "-1") |
||||
add_definitions(-DPPC64=1) |
||||
message(STATUS "Detected POWER PPC64 target processor") |
||||
endif() |
||||
-elseif(ARMMATCH GREATER "-1") |
||||
+elseif(ARMv6MATCH GREATER "-1") |
||||
if(CROSS_COMPILE_ARM) |
||||
message(STATUS "Cross compiling for ARM arch") |
||||
else() |
||||
set(CROSS_COMPILE_ARM 0) |
||||
endif() |
||||
- message(STATUS "Detected ARM target processor") |
||||
+ message(STATUS "Detected ARMv6 target processor") |
||||
set(ARM 1) |
||||
add_definitions(-DX265_ARCH_ARM=1 -DHAVE_ARMV6=1) |
||||
+elseif(ARMv7MATCH GREATER "-1") |
||||
+ if(CROSS_COMPILE_ARM) |
||||
+ message(STATUS "Cross compiling for ARM arch") |
||||
+ else() |
||||
+ set(CROSS_COMPILE_ARM 0) |
||||
+ endif() |
||||
+ message(STATUS "Detected ARMv7 target processor") |
||||
+ set(ARM 1) |
||||
+ add_definitions(-DX265_ARCH_ARM=1 -DHAVE_ARMV7=1) |
||||
else() |
||||
message(STATUS "CMAKE_SYSTEM_PROCESSOR value `${CMAKE_SYSTEM_PROCESSOR}` is unknown") |
||||
message(STATUS "Please add this value near ${CMAKE_CURRENT_LIST_FILE}:${CMAKE_CURRENT_LIST_LINE}") |
||||
--- x265_v2.6/source/test/testharness.h.orig 2017-12-30 22:27:49.827620181 +0000 |
||||
+++ x265_v2.6/source/test/testharness.h 2017-12-30 22:30:53.239500941 +0000 |
||||
@@ -70,9 +70,10 @@ protected: |
||||
#include <intrin.h> |
||||
#elif (!defined(__APPLE__) && (defined (__GNUC__) && (defined(__x86_64__) || defined(__i386__)))) |
||||
#include <x86intrin.h> |
||||
-#elif ( !defined(__APPLE__) && defined (__GNUC__) && defined(__ARM_NEON__)) |
||||
-#include <arm_neon.h> |
||||
#elif defined(__GNUC__) && (!defined(__clang__) || __clang_major__ < 4) |
||||
+#if ( !defined(__APPLE__) && defined(__ARM_NEON__)) |
||||
+#include <arm_neon.h> |
||||
+#endif |
||||
/* fallback for older GCC/MinGW */ |
||||
static inline uint32_t __rdtsc(void) |
||||
{ |
||||
@ -0,0 +1,32 @@
@@ -0,0 +1,32 @@
|
||||
diff -Naur x265_2.2.old/source/CMakeLists.txt x265_2.2/source/CMakeLists.txt |
||||
--- x265_2.2.old/source/CMakeLists.txt 2017-04-11 12:14:19.717605457 +0200 |
||||
+++ x265_2.2/source/CMakeLists.txt 2017-04-11 12:14:27.499743262 +0200 |
||||
@@ -548,7 +548,15 @@ |
||||
if(MSVC) |
||||
set_target_properties(x265-shared PROPERTIES OUTPUT_NAME libx265) |
||||
else() |
||||
- set_target_properties(x265-shared PROPERTIES OUTPUT_NAME x265) |
||||
+ if(HIGH_BIT_DEPTH) |
||||
+ if(MAIN12) |
||||
+ set_target_properties(x265-shared PROPERTIES OUTPUT_NAME x265_main12) |
||||
+ else() |
||||
+ set_target_properties(x265-shared PROPERTIES OUTPUT_NAME x265_main10) |
||||
+ endif() |
||||
+ else() |
||||
+ set_target_properties(x265-shared PROPERTIES OUTPUT_NAME x265) |
||||
+ endif(HIGH_BIT_DEPTH) |
||||
endif() |
||||
if(UNIX) |
||||
set_target_properties(x265-shared PROPERTIES VERSION ${X265_BUILD}) |
||||
diff -Naur x265_2.2.old/source/encoder/api.cpp x265_2.2/source/encoder/api.cpp |
||||
--- x265_2.2.old/source/encoder/api.cpp 2017-04-11 12:14:19.727605634 +0200 |
||||
+++ x265_2.2/source/encoder/api.cpp 2017-04-11 12:21:58.928758432 +0200 |
||||
@@ -356,7 +356,7 @@ |
||||
#define ext ".dylib" |
||||
#else |
||||
#include <dlfcn.h> |
||||
-#define ext ".so" |
||||
+#define ext ".so." xstr(X265_BUILD) |
||||
#endif |
||||
|
||||
static int g_recursion /* = 0 */; |
||||
@ -0,0 +1,12 @@
@@ -0,0 +1,12 @@
|
||||
--- a/source/CMakeLists.txt |
||||
+++ b/source/CMakeLists.txt |
||||
@@ -196,7 +196,7 @@ |
||||
add_definitions(-std=gnu++98) |
||||
endif() |
||||
if(ENABLE_PIC) |
||||
- add_definitions(-fPIC) |
||||
+ add_definitions(-fPIC -DPIC) |
||||
endif(ENABLE_PIC) |
||||
if(NATIVE_BUILD) |
||||
if(INTEL_CXX) |
||||
|
||||
@ -0,0 +1,11 @@
@@ -0,0 +1,11 @@
|
||||
--- a/source/x265.pc.in |
||||
+++ b/source/x265.pc.in |
||||
@@ -1,6 +1,6 @@ |
||||
prefix=@CMAKE_INSTALL_PREFIX@ |
||||
exec_prefix=${prefix} |
||||
-libdir=${exec_prefix}/@LIB_INSTALL_DIR@ |
||||
+libdir=@LIB_INSTALL_DIR@ |
||||
includedir=${prefix}/include |
||||
|
||||
Name: @CMAKE_PROJECT_NAME@ |
||||
|
||||
@ -0,0 +1,253 @@
@@ -0,0 +1,253 @@
|
||||
%global _so_version 165 |
||||
|
||||
Summary: H.265/HEVC encoder |
||||
Name: x265 |
||||
Version: 2.9 |
||||
Release: 3%{?dist} |
||||
URL: http://x265.org/ |
||||
# source/Lib/TLibCommon - BSD |
||||
# source/Lib/TLibEncoder - BSD |
||||
# everything else - GPLv2+ |
||||
License: GPLv2+ and BSD |
||||
|
||||
Source0: https://bitbucket.org/multicoreware/%{name}/downloads/%{name}_%{version}.tar.gz |
||||
|
||||
# fix building as PIC |
||||
Patch0: x265-pic.patch |
||||
Patch1: x265-high-bit-depth-soname.patch |
||||
Patch2: x265-detect_cpu_armhfp.patch |
||||
Patch3: x265-arm-cflags.patch |
||||
Patch4: x265-pkgconfig_path_fix.patch |
||||
Patch5: x265-2.8-asm-primitives.patch |
||||
Patch6: https://sources.debian.org/data/main/x/x265/2.9-3/debian/patches/0003-detect512-is-needed-on-all-architectures.patch |
||||
|
||||
BuildRequires: gcc-c++ |
||||
BuildRequires: cmake3 |
||||
%{?el7:BuildRequires: epel-rpm-macros} |
||||
BuildRequires: nasm |
||||
BuildRequires: ninja-build |
||||
|
||||
%ifnarch armv7hl armv7hnl s390 s390x |
||||
BuildRequires: numactl-devel |
||||
%endif |
||||
|
||||
%description |
||||
The primary objective of x265 is to become the best H.265/HEVC encoder |
||||
available anywhere, offering the highest compression efficiency and the highest |
||||
performance on a wide variety of hardware platforms. |
||||
|
||||
This package contains the command line encoder. |
||||
|
||||
%package libs |
||||
Summary: H.265/HEVC encoder library |
||||
|
||||
%description libs |
||||
The primary objective of x265 is to become the best H.265/HEVC encoder |
||||
available anywhere, offering the highest compression efficiency and the |
||||
highest performance on a wide variety of hardware platforms. |
||||
|
||||
This package contains the shared library. |
||||
|
||||
%package devel |
||||
Summary: H.265/HEVC encoder library development files |
||||
Requires: %{name}-libs%{?_isa} = %{version}-%{release} |
||||
|
||||
%description devel |
||||
The primary objective of x265 is to become the best H.265/HEVC encoder |
||||
available anywhere, offering the highest compression efficiency and the highest |
||||
performance on a wide variety of hardware platforms. |
||||
|
||||
This package contains the shared library development files. |
||||
|
||||
%prep |
||||
%autosetup -p1 -n %{name}_%{version} |
||||
|
||||
%build |
||||
# High depth libraries (from source/h265.h): |
||||
# If the requested bitDepth is not supported by the linked libx265, |
||||
# it will attempt to dynamically bind x265_api_get() from a shared |
||||
# library with an appropriate name: |
||||
# 8bit: libx265_main.so |
||||
# 10bit: libx265_main10.so |
||||
|
||||
build() { |
||||
%cmake3 -Wno-dev -G "Ninja" \ |
||||
-DCMAKE_POSITION_INDEPENDENT_CODE:BOOL=ON \ |
||||
-DCMAKE_SKIP_RPATH:BOOL=YES \ |
||||
-DENABLE_PIC:BOOL=ON \ |
||||
-DENABLE_TESTS:BOOL=ON \ |
||||
$* \ |
||||
../source |
||||
%ninja_build |
||||
} |
||||
|
||||
# High depth 10/12 bit libraries are supported only on 64 bit. They require |
||||
# disabled AltiVec instructions for building on ppc64/ppc64le. |
||||
%ifarch x86_64 aarch64 ppc64 ppc64le |
||||
mkdir 10bit; pushd 10bit |
||||
build -DENABLE_CLI=OFF -DENABLE_ALTIVEC=OFF -DHIGH_BIT_DEPTH=ON |
||||
popd |
||||
|
||||
mkdir 12bit; pushd 12bit |
||||
build -DENABLE_CLI=OFF -DENABLE_ALTIVEC=OFF -DHIGH_BIT_DEPTH=ON -DMAIN12=ON |
||||
popd |
||||
%endif |
||||
|
||||
# 8 bit base library + encoder |
||||
mkdir 8bit; pushd 8bit |
||||
build |
||||
popd |
||||
|
||||
%install |
||||
for i in 8 10 12; do |
||||
if [ -d ${i}bit ]; then |
||||
pushd ${i}bit |
||||
%ninja_install |
||||
# Remove unversioned library, should not be linked to |
||||
rm -f %{buildroot}%{_libdir}/libx265_main${i}.so |
||||
popd |
||||
fi |
||||
done |
||||
|
||||
find %{buildroot} -name "*.a" -delete |
||||
|
||||
%check |
||||
for i in 8 10 12; do |
||||
if [ -d ${i}bit ]; then |
||||
pushd ${i}bit |
||||
test/TestBench || : |
||||
popd |
||||
fi |
||||
done |
||||
|
||||
%ldconfig_scriptlets libs |
||||
|
||||
%files |
||||
%{_bindir}/x265 |
||||
|
||||
%files libs |
||||
%license COPYING |
||||
%{_libdir}/libx265.so.%{_so_version} |
||||
%ifarch x86_64 aarch64 ppc64 ppc64le |
||||
%{_libdir}/libx265_main10.so.%{_so_version} |
||||
%{_libdir}/libx265_main12.so.%{_so_version} |
||||
%endif |
||||
|
||||
%files devel |
||||
%doc doc/* |
||||
%{_includedir}/x265.h |
||||
%{_includedir}/x265_config.h |
||||
%{_libdir}/libx265.so |
||||
%{_libdir}/pkgconfig/x265.pc |
||||
|
||||
%changelog |
||||
* Sun Dec 30 2018 Leigh Scott <leigh123linux@googlemail.com> - 2.9-3 |
||||
- Rebuild against newer nasm on el7 (rfbz #5128) |
||||
|
||||
* Wed Nov 21 2018 Antonio Trande <sagitter@fedoraproject.org> - 2.9-2 |
||||
- Rebuild for ffmpeg-3.* on el7 |
||||
|
||||
* Sun Nov 18 2018 Leigh Scott <leigh123linux@googlemail.com> - 2.9-1 |
||||
- Update to 2.9 |
||||
|
||||
* Thu Oct 04 2018 Sérgio Basto <sergio@serjux.com> - 2.8-1 |
||||
- Update to 2.8 more 2 patches to fix builds on non-x86 and arm |
||||
https://bitbucket.org/multicoreware/x265/issues/404/28-fails-to-build-on-ppc64le-gnu-linux |
||||
https://bitbucket.org/multicoreware/x265/issues/406/arm-assembly-fail-to-compile-on-18 |
||||
|
||||
* Sun Aug 19 2018 Leigh Scott <leigh123linux@googlemail.com> - 2.7-5 |
||||
- Rebuilt for Fedora 29 Mass Rebuild binutils issue |
||||
|
||||
* Fri Jul 27 2018 RPM Fusion Release Engineering <leigh123linux@gmail.com> - 2.7-4 |
||||
- Rebuilt for https://fedoraproject.org/wiki/Fedora_29_Mass_Rebuild |
||||
|
||||
* Sun Apr 08 2018 Leigh Scott <leigh123linux@googlemail.com> - 2.7-3 |
||||
- Fix pkgconfig file (rfbz #4853) |
||||
|
||||
* Tue Feb 27 2018 Nicolas Chauvet <kwizart@gmail.com> - 2.7-2 |
||||
- Fix CFLAGS on ARM |
||||
|
||||
* Tue Feb 27 2018 Leigh Scott <leigh123linux@googlemail.com> - 2.7-1 |
||||
- update to 2.7 |
||||
- Drop shared test patch as it causes nasm build to fail |
||||
- Fix scriptlets |
||||
- Use ninja to build |
||||
|
||||
* Sat Dec 30 2017 Sérgio Basto <sergio@serjux.com> - 2.6-1 |
||||
- Update x265 to 2.6 |
||||
|
||||
* Mon Oct 16 2017 Leigh Scott <leigh123linux@googlemail.com> - 2.5-1 |
||||
- update to 2.5 |
||||
|
||||
* Thu Aug 31 2017 RPM Fusion Release Engineering <kwizart@rpmfusion.org> - 2.4-2 |
||||
- Rebuilt for https://fedoraproject.org/wiki/Fedora_27_Mass_Rebuild |
||||
|
||||
* Sat Apr 29 2017 Leigh Scott <leigh123linux@googlemail.com> - 2.4-1 |
||||
- update to 2.4 |
||||
|
||||
* Mon Apr 10 2017 Simone Caronni <negativo17@gmail.com> - 2.2-3 |
||||
- Use source from multicoreware website. |
||||
- Clean up SPEC file a bit (formatting, 80 char wide descriptions). |
||||
- Enable shared 10/12 bit libraries on 64 bit architectures. |
||||
|
||||
* Mon Mar 20 2017 RPM Fusion Release Engineering <kwizart@rpmfusion.org> - 2.2-2 |
||||
- Rebuilt for https://fedoraproject.org/wiki/Fedora_26_Mass_Rebuild |
||||
|
||||
* Tue Jan 03 2017 Dominik Mierzejewski <rpm@greysector.net> - 2.2-1 |
||||
- update to 2.2 |
||||
- spell out SO version in file list |
||||
- fix typo in patch |
||||
|
||||
* Mon Nov 07 2016 Sérgio Basto <sergio@serjux.com> - 2.1-1 |
||||
- Update to 2.1 |
||||
|
||||
* Thu Aug 18 2016 Sérgio Basto <sergio@serjux.com> - 1.9-3 |
||||
- Clean spec, Vascom patches series, rfbz #4199, add license tag |
||||
|
||||
* Tue Jul 19 2016 Dominik Mierzejewski <rpm@greysector.net> - 1.9-2 |
||||
- use https for source URL |
||||
- enable NUMA support |
||||
- make sure Fedora compiler flags are used on ARM |
||||
|
||||
* Fri Apr 08 2016 Adrian Reber <adrian@lisas.de> - 1.9-1 |
||||
- Update to 1.9 |
||||
|
||||
* Sun Oct 25 2015 Dominik Mierzejewski <rpm@greysector.net> 1.8-2 |
||||
- fix building as PIC |
||||
- update SO version in file list |
||||
|
||||
* Sat Oct 24 2015 Nicolas Chauvet <kwizart@gmail.com> - 1.8-1 |
||||
- Update to 1.8 |
||||
- Avoid tests for now |
||||
|
||||
* Wed Apr 15 2015 Dominik Mierzejewski <rpm@greysector.net> 1.6-1 |
||||
- update to 1.6 (ABI bump, rfbz#3593) |
||||
- release tarballs are now hosted on videolan.org |
||||
- drop obsolete patches |
||||
|
||||
* Thu Dec 18 2014 Dominik Mierzejewski <rpm@greysector.net> 1.2-6 |
||||
- fix build on armv7l arch (partially fix rfbz#3361, patch by Nicolas Chauvet) |
||||
- don't run tests on ARM for now (rfbz#3361) |
||||
|
||||
* Sun Aug 17 2014 Dominik Mierzejewski <rpm@greysector.net> 1.2-5 |
||||
- don't include contributor agreement in doc |
||||
- make sure /usr/share/doc/x265 is owned |
||||
- add a comment noting which files are BSD-licenced |
||||
|
||||
* Fri Aug 08 2014 Dominik Mierzejewski <rpm@greysector.net> 1.2-4 |
||||
- don't create bogus soname (patch by Xavier) |
||||
|
||||
* Thu Jul 17 2014 Dominik Mierzejewski <rpm@greysector.net> 1.2-3 |
||||
- fix tr call to remove DOS EOL |
||||
- build the library with -fPIC on arm and i686, too |
||||
|
||||
* Sun Jul 13 2014 Dominik Mierzejewski <rpm@greysector.net> 1.2-2 |
||||
- use version in source URL |
||||
- update License tag |
||||
- fix EOL in drag-uncrustify.bat |
||||
- don't link test binaries with shared binary on x86 (segfault) |
||||
|
||||
* Thu Jul 10 2014 Dominik Mierzejewski <rpm@greysector.net> 1.2-1 |
||||
- initial build |
||||
- fix pkgconfig file install location |
||||
- link test binaries with shared library |
||||
Loading…
Reference in new issue