[x265] [PATCH] arm: Implement blockcopy_pp_16x16_neon. Modified include guards with ARM suffix

dnyaneshwar at multicorewareinc.com dnyaneshwar at multicorewareinc.com
Tue Feb 2 07:00:44 CET 2016


# HG changeset patch
# User Dnyaneshwar G <dnyaneshwar at multicorewareinc.com>
# Date 1454327470 -19800
#      Mon Feb 01 17:21:10 2016 +0530
# Node ID 894e0fce5d14844d3c85cdb2a287f302fc8cffca
# Parent  dc62b47dd0d98f732165345883edac55320baec1
arm: Implement blockcopy_pp_16x16_neon. Modified include guards with ARM suffix.

diff -r dc62b47dd0d9 -r 894e0fce5d14 source/CMakeLists.txt
--- a/source/CMakeLists.txt	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/CMakeLists.txt	Mon Feb 01 17:21:10 2016 +0530
@@ -182,9 +182,11 @@
         add_definitions(-march=i686)
     endif()
     if(ARM AND CROSS_COMPILE_ARM)
-        add_definitions(-march=armv6 -mfloat-abi=soft -mfpu=vfp)
+        set(ARM_ARGS -march=armv6 -mfloat-abi=soft -mfpu=vfp)
+        add_definitions(${ARM_ARGS})
     elseif(ARM)
-        add_definitions(-march=armv6 -mfloat-abi=hard -mfpu=vfp)
+        set(ARM_ARGS -march=armv6 -mfloat-abi=hard -mfpu=vfp)
+        add_definitions(${ARM_ARGS})
     endif()
     if(FPROFILE_GENERATE)
         if(INTEL_CXX)
@@ -418,7 +420,7 @@
 add_subdirectory(encoder)
 add_subdirectory(common)
 
-if((MSVC_IDE OR XCODE) AND ENABLE_ASSEMBLY)
+if((MSVC_IDE OR XCODE OR GCC) AND ENABLE_ASSEMBLY)
     # this is required because of this cmake bug
     # http://www.cmake.org/Bug/print_bug_page.php?bug_id=8170
     if(WIN32)
@@ -429,7 +431,17 @@
 
     if(ARM OR CROSS_COMPILE_ARM)
     # compile ARM arch asm files here
-
+        enable_language(ASM)
+        foreach(ASM ${ARM_ASMS})
+            set(YASM_SRC ${CMAKE_CURRENT_SOURCE_DIR}/common/arm/${ASM})
+            list(APPEND YASM_SRCS ${YASM_SRC})
+            list(APPEND YASM_OBJS ${ASM}.${SUFFIX})
+            add_custom_command(
+                OUTPUT ${ASM}.${SUFFIX}
+                COMMAND ${CMAKE_CXX_COMPILER}
+                ARGS ${ARM_ARGS} -c ${YASM_SRC} -o ${ASM}.${SUFFIX}
+                DEPENDS ${YASM_SRC})
+        endforeach()
     elseif(X86)
     # compile X86 arch asm files here
         foreach(ASM ${MSVC_ASMS})
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/CMakeLists.txt
--- a/source/common/CMakeLists.txt	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/CMakeLists.txt	Mon Feb 01 17:21:10 2016 +0530
@@ -89,9 +89,10 @@
     set(C_SRCS asm-primitives.cpp pixel.h mc.h ipfilter8.h blockcopy8.h dct8.h loopfilter.h)
 
     # add ARM assembly/intrinsic files here
-    set(A_SRCS)
+    set(ARM_SRCS asm.S cpu-a.S mc-a.S)
     set(VEC_PRIMITIVES)
 
+    set(ARM_ASMS "${ARM_SRCS}" CACHE INTERNAL "ARM Assembly Sources")
     foreach(SRC ${C_SRCS})
         set(ASM_PRIMITIVES ${ASM_PRIMITIVES} arm/${SRC})
     endforeach()
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/asm-primitives.cpp
--- a/source/common/arm/asm-primitives.cpp	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/arm/asm-primitives.cpp	Mon Feb 01 17:21:10 2016 +0530
@@ -29,12 +29,18 @@
 #include "x265.h"
 #include "cpu.h"
 
+extern "C" {
+#include "blockcopy8.h"
+}
 
 namespace X265_NS {
 // private x265 namespace
 
 void setupAssemblyPrimitives(EncoderPrimitives &p, int cpuMask)
 {
-
+    if (cpuMask & X265_CPU_NEON)
+    {
+        p.pu[LUMA_16x16].copy_pp = PFX(blockcopy_pp_16x16_neon);
+    }
 }
 } // namespace X265_NS
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/asm.S
--- a/source/common/arm/asm.S	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/arm/asm.S	Mon Feb 01 17:21:10 2016 +0530
@@ -25,8 +25,6 @@
  * For more information, contact us at license @ x265.com.
  *****************************************************************************/
 
-#include "x265_config.h"
-
 .syntax unified
 
 #if   HAVE_NEON
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/blockcopy8.h
--- a/source/common/arm/blockcopy8.h	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/arm/blockcopy8.h	Mon Feb 01 17:21:10 2016 +0530
@@ -23,7 +23,9 @@
  * For more information, contact us at license @ x265.com.
  *****************************************************************************/
 
-#ifndef X265_BLOCKCOPY8_H
-#define X265_BLOCKCOPY8_H
+#ifndef X265_BLOCKCOPY8_ARM_H
+#define X265_BLOCKCOPY8_ARM_H
 
-#endif // ifndef X265_I386_PIXEL_H
+void x265_blockcopy_pp_16x16_neon(pixel* dst, intptr_t dstStride, const pixel* src, intptr_t srcStride);
+
+#endif // ifndef X265_I386_PIXEL_ARM_H
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/dct8.h
--- a/source/common/arm/dct8.h	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/arm/dct8.h	Mon Feb 01 17:21:10 2016 +0530
@@ -22,7 +22,7 @@
  * For more information, contact us at license @ x265.com.
  *****************************************************************************/
 
-#ifndef X265_DCT8_H
-#define X265_DCT8_H
+#ifndef X265_DCT8_ARM_H
+#define X265_DCT8_ARM_H
 
-#endif // ifndef X265_DCT8_H
+#endif // ifndef X265_DCT8_ARM_H
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/intrapred.h
--- a/source/common/arm/intrapred.h	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/arm/intrapred.h	Mon Feb 01 17:21:10 2016 +0530
@@ -25,7 +25,7 @@
  * For more information, contact us at license @ x265.com.
  *****************************************************************************/
 
-#ifndef X265_INTRAPRED_H
-#define X265_INTRAPRED_H
+#ifndef X265_INTRAPRED_ARM_H
+#define X265_INTRAPRED_ARM_H
 
-#endif // ifndef X265_INTRAPRED_H
+#endif // ifndef X265_INTRAPRED_ARM_H
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/ipfilter8.h
--- a/source/common/arm/ipfilter8.h	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/arm/ipfilter8.h	Mon Feb 01 17:21:10 2016 +0530
@@ -22,7 +22,7 @@
  * For more information, contact us at license @ x265.com.
  *****************************************************************************/
 
-#ifndef X265_IPFILTER8_H
-#define X265_IPFILTER8_H
+#ifndef X265_IPFILTER8_ARM_H
+#define X265_IPFILTER8_ARM_H
 
-#endif // ifndef X265_IPFILTER8_H
+#endif // ifndef X265_IPFILTER8_ARM_H
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/loopfilter.h
--- a/source/common/arm/loopfilter.h	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/arm/loopfilter.h	Mon Feb 01 17:21:10 2016 +0530
@@ -23,7 +23,7 @@
  * For more information, contact us at license @ x265.com.
  *****************************************************************************/
 
-#ifndef X265_LOOPFILTER_H
-#define X265_LOOPFILTER_H
+#ifndef X265_LOOPFILTER_ARM_H
+#define X265_LOOPFILTER_ARM_H
 
-#endif // ifndef X265_LOOPFILTER_H
+#endif // ifndef X265_LOOPFILTER_ARM_H
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/mc-a.S
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/source/common/arm/mc-a.S	Mon Feb 01 17:21:10 2016 +0530
@@ -0,0 +1,102 @@
+/*****************************************************************************
+ * Copyright (C) 2016 x265 project
+ *
+ * Authors: Dnyaneshwar Gorade <dnyaneshwar at multicorewareinc.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02111, USA.
+ *
+ * This program is also available under a commercial proprietary license.
+ * For more information, contact us at license @ x265.com.
+ *****************************************************************************/
+
+#include "asm.S"
+
+.section .rodata
+
+.align 4
+
+.text
+
+/* blockcopy_pp_16x16(pixel* dst, intptr_t dstStride, const pixel* src, intptr_t srcStride)
+ *
+ * r0   - dst
+ * r1   - dstStride
+ * r2   - src
+ * d3   - srcStride */
+function x265_blockcopy_pp_16x16_neon
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    add             r2, r2, r3
+    add             r0, r0, r1
+    vld1.8          {q0}, [r2]
+    vst1.8          {q0}, [r0]
+    bx              lr
+endfunc
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/mc.h
--- a/source/common/arm/mc.h	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/arm/mc.h	Mon Feb 01 17:21:10 2016 +0530
@@ -21,7 +21,7 @@
  * For more information, contact us at license @ x265.com.
  *****************************************************************************/
 
-#ifndef X265_MC_H
-#define X265_MC_H
+#ifndef X265_MC_ARM_H
+#define X265_MC_ARM_H
 
-#endif // ifndef X265_MC_H
+#endif // ifndef X265_MC_ARM_H
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/pixel-util.h
--- a/source/common/arm/pixel-util.h	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/arm/pixel-util.h	Mon Feb 01 17:21:10 2016 +0530
@@ -22,7 +22,7 @@
  * For more information, contact us at license @ x265.com.
  *****************************************************************************/
 
-#ifndef X265_PIXEL_UTIL_H
-#define X265_PIXEL_UTIL_H
+#ifndef X265_PIXEL_UTIL_ARM_H
+#define X265_PIXEL_UTIL_ARM_H
 
-#endif // ifndef X265_PIXEL_UTIL_H
+#endif // ifndef X265_PIXEL_UTIL_ARM_H
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/arm/pixel.h
--- a/source/common/arm/pixel.h	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/arm/pixel.h	Mon Feb 01 17:21:10 2016 +0530
@@ -27,7 +27,7 @@
  * For more information, contact us at license @ x265.com.
  *****************************************************************************/
 
-#ifndef X265_I386_PIXEL_H
-#define X265_I386_PIXEL_H
+#ifndef X265_I386_PIXEL_ARM_H
+#define X265_I386_PIXEL_ARM_H
 
-#endif // ifndef X265_I386_PIXEL_H
+#endif // ifndef X265_I386_PIXEL_ARM_H
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/common/primitives.cpp
--- a/source/common/primitives.cpp	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/common/primitives.cpp	Mon Feb 01 17:21:10 2016 +0530
@@ -260,7 +260,10 @@
 void PFX(cpu_emms)(void) {}
 void PFX(cpu_cpuid)(uint32_t, uint32_t *eax, uint32_t *, uint32_t *, uint32_t *) { *eax = 0; }
 void PFX(cpu_xgetbv)(uint32_t, uint32_t *, uint32_t *) {}
+
+#if ENABLE_ASSEMBLY && X265_ARCH_ARM == 0
 void PFX(cpu_neon_test)(void) {}
 int PFX(cpu_fast_neon_mrc_test)(void) { return 0; }
+#endif
 }
 #endif
diff -r dc62b47dd0d9 -r 894e0fce5d14 source/test/testharness.h
--- a/source/test/testharness.h	Mon Jan 25 14:59:50 2016 +0530
+++ b/source/test/testharness.h	Mon Feb 01 17:21:10 2016 +0530
@@ -80,6 +80,9 @@
 #elif X265_ARCH_ARM
     // TOD-DO: verify following inline asm to get cpu Timestamp Counter for ARM arch
     // asm volatile("mrc p15, 0, %0, c9, c13, 0" : "=r"(a));
+
+    // TO-DO: replace clock() function with appropriate ARM cpu instructions
+    a = clock();
 #endif
     return a;
 }


More information about the x265-devel mailing list