add m128 byte shift functions

variable_byte_shift_m128 taken from pug-interpreter branch

diff --git a/CMakeLists.txt b/CMakeLists.txt
index a4973c0265504f26100efa1180f6d4a1527a6df2..0e118dd07ac473ec73bbfd7a976358fc90d46505 100644 (file)
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -521,6 +521,7 @@ set (hs_exec_SRCS
     src/util/shuffle_ssse3.h
     src/util/simd_utils.h
     src/util/simd_utils_ssse3.h
+    src/util/simd_utils_ssse3.c
     src/util/state_compress.h
     src/util/state_compress.c
     src/util/unaligned.h

diff --git a/src/util/simd_utils.h b/src/util/simd_utils.h
index e115aa7ae11e5047df7759fc3031e3b65210b0a4..99ad7ce595c8922ecbce44051509000fb191a2fc 100644 (file)
--- a/src/util/simd_utils.h
+++ b/src/util/simd_utils.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2015, Intel Corporation
+ * Copyright (c) 2015-2016, Intel Corporation
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
@@ -199,6 +199,7 @@ static really_inline m128 shiftLeft8Bits(m128 a) {
 }
 
 #define byteShiftRight128(a, count_immed) _mm_srli_si128(a, count_immed)
+#define byteShiftLeft128(a, count_immed) _mm_slli_si128(a, count_immed)
 
 #if !defined(__AVX2__)
 // TODO: this entire file needs restructuring - this carveout is awful

diff --git a/src/util/simd_utils_ssse3.c b/src/util/simd_utils_ssse3.c
new file mode 100644 (file)
index 0000000..50cbe00
--- /dev/null
+++ b/src/util/simd_utils_ssse3.c
@@ -0,0 +1,40 @@
+/*
+ * Copyright (c) 2016, Intel Corporation
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ *  * Redistributions of source code must retain the above copyright notice,
+ *    this list of conditions and the following disclaimer.
+ *  * Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *  * Neither the name of Intel Corporation nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
+ * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
+ * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
+ * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
+ * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
+ * POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "simd_utils_ssse3.h"
+
+const char vbs_mask_data[] ALIGN_CL_DIRECTIVE = {
+    0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0,
+    0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0,
+
+    0x00, 0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07,
+    0x08, 0x09, 0x0a, 0x0b, 0x0c, 0x0d, 0x0e, 0x0f,
+
+    0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0,
+    0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0, 0xf0,
+};

diff --git a/src/util/simd_utils_ssse3.h b/src/util/simd_utils_ssse3.h
index 23c77d225574d000aa6541fa3f769d3446e354aa..268bf42298602120a1349a4196b49c2f9f9b9afc 100644 (file)
--- a/src/util/simd_utils_ssse3.h
+++ b/src/util/simd_utils_ssse3.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2015, Intel Corporation
+ * Copyright (c) 2015-2016, Intel Corporation
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
@@ -77,6 +77,15 @@ m128 pshufb(m128 a, m128 b) {
     return result;
 }
 
+extern const char vbs_mask_data[];
+
+static really_inline
+m128 variable_byte_shift_m128(m128 in, s32 amount) {
+    assert(amount >= -16 && amount <= 16);
+    m128 shift_mask = loadu128(vbs_mask_data + 16 - amount);
+    return pshufb(in, shift_mask);
+}
+
 #if defined(__AVX2__)
 
 static really_inline

diff --git a/unit/internal/simd_utils.cpp b/unit/internal/simd_utils.cpp
index c0b2b2a0027f995fa36c834d1ebbd80f43ce413e..de0f1eea777ad8f9e4b1318b78a9c01760c886d8 100644 (file)
--- a/unit/internal/simd_utils.cpp
+++ b/unit/internal/simd_utils.cpp
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2015, Intel Corporation
+ * Copyright (c) 2015-2016, Intel Corporation
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions are met:
@@ -32,7 +32,7 @@
 #include "util/alloc.h"
 #include "util/make_unique.h"
 #include "util/simd_utils.h"
-
+#include "util/simd_utils_ssse3.h"
 
 using namespace std;
 using namespace ue2;
@@ -640,4 +640,58 @@ TEST(SimdUtilsTest, set2x128) {
 }
 #endif
 
+TEST(SimdUtilsTest, variableByteShift128) {
+    char base[] = "0123456789ABCDEF";
+    m128 in = loadu128(base);
+
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 0),
+                         variable_byte_shift_m128(in, 0)));
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 1),
+                         variable_byte_shift_m128(in, -1)));
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 2),
+                         variable_byte_shift_m128(in, -2)));
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 3),
+                         variable_byte_shift_m128(in, -3)));
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 4),
+                         variable_byte_shift_m128(in, -4)));
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 5),
+                         variable_byte_shift_m128(in, -5)));
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 6),
+                         variable_byte_shift_m128(in, -6)));
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 7),
+                         variable_byte_shift_m128(in, -7)));
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 8),
+                         variable_byte_shift_m128(in, -8)));
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 9),
+                         variable_byte_shift_m128(in, -9)));
+    EXPECT_TRUE(!diff128(byteShiftRight128(in, 10),
+                         variable_byte_shift_m128(in, -10)));
+
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 0),
+                         variable_byte_shift_m128(in, 0)));
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 1),
+                         variable_byte_shift_m128(in, 1)));
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 2),
+                         variable_byte_shift_m128(in, 2)));
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 3),
+                         variable_byte_shift_m128(in, 3)));
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 4),
+                         variable_byte_shift_m128(in, 4)));
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 5),
+                         variable_byte_shift_m128(in, 5)));
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 6),
+                         variable_byte_shift_m128(in, 6)));
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 7),
+                         variable_byte_shift_m128(in, 7)));
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 8),
+                         variable_byte_shift_m128(in, 8)));
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 9),
+                         variable_byte_shift_m128(in, 9)));
+    EXPECT_TRUE(!diff128(byteShiftLeft128(in, 10),
+                         variable_byte_shift_m128(in, 10)));
+
+    EXPECT_TRUE(!diff128(zeroes128(), variable_byte_shift_m128(in, 16)));
+    EXPECT_TRUE(!diff128(zeroes128(), variable_byte_shift_m128(in, -16)));
+}
+
 } // namespace