This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[nvptx, PATCH, 3/3] Add v2di support


Hi,

this patch adds v2di support to the nvptx target. This allows us to generate 128-bit loads and stores.

Tested in nvptx mainkernel mode and x86_64 accelerator mode.

OK for trunk?

Thanks,
- Tom
Add v2di support

2017-06-06  Tom de Vries  <tom@codesourcery.com>

	* config/nvptx/nvptx-modes.def: Add V2DImode.
	* config/nvptx/nvptx-protos.h (nvptx_data_alignment): Declare.
	* config/nvptx/nvptx.c (nvptx_ptx_type_from_mode): Handle V2DImode.
	(nvptx_output_mov_insn): Handle lack of mov.b128.
	(nvptx_print_operand): Handle 'H' and 'L' codes.
	(nvptx_vector_mode_supported): Allow V2DImode.
	(nvptx_preferred_simd_mode): New function.
	(nvptx_data_alignment): New function.
	(TARGET_VECTORIZE_PREFERRED_SIMD_MODE): Redefine to
	nvptx_preferred_simd_mode.
	* config/nvptx/nvptx.h (STACK_BOUNDARY, BIGGEST_ALIGNMENT): Change from
	64 to 128 bits.
	(DATA_ALIGNMENT): Define.  Set to nvptx_data_alignment.

	* config/nvptx/nvptx.md (VECIM): Add V2DI.

	* gcc.target/nvptx/decl-init.c: Update alignment.
	* gcc.target/nvptx/slp-2-run.c: New test.
	* gcc.target/nvptx/slp-2.c: New test.
	* gcc.target/nvptx/v2di.c: New test.

	* testsuite/libgomp.oacc-c/vec.c: New test.

---
 gcc/config/nvptx/nvptx-modes.def           |  2 +
 gcc/config/nvptx/nvptx-protos.h            |  1 +
 gcc/config/nvptx/nvptx.c                   | 68 +++++++++++++++++++++++++++++-
 gcc/config/nvptx/nvptx.h                   |  6 ++-
 gcc/config/nvptx/nvptx.md                  |  2 +-
 gcc/testsuite/gcc.target/nvptx/decl-init.c |  2 +-
 gcc/testsuite/gcc.target/nvptx/slp-2-run.c | 23 ++++++++++
 gcc/testsuite/gcc.target/nvptx/slp-2.c     | 25 +++++++++++
 gcc/testsuite/gcc.target/nvptx/v2di.c      | 12 ++++++
 libgomp/testsuite/libgomp.oacc-c/vec.c     | 48 +++++++++++++++++++++
 10 files changed, 183 insertions(+), 6 deletions(-)

diff --git a/gcc/config/nvptx/nvptx-modes.def b/gcc/config/nvptx/nvptx-modes.def
index d49429c..ff61b36 100644
--- a/gcc/config/nvptx/nvptx-modes.def
+++ b/gcc/config/nvptx/nvptx-modes.def
@@ -1 +1,3 @@
 VECTOR_MODE (INT, SI, 2);  /* V2SI */
+
+VECTOR_MODE (INT, DI, 2);  /* V2DI */
diff --git a/gcc/config/nvptx/nvptx-protos.h b/gcc/config/nvptx/nvptx-protos.h
index 16b316f..c3e3b84 100644
--- a/gcc/config/nvptx/nvptx-protos.h
+++ b/gcc/config/nvptx/nvptx-protos.h
@@ -41,6 +41,7 @@ extern void nvptx_function_end (FILE *);
 extern void nvptx_output_skip (FILE *, unsigned HOST_WIDE_INT);
 extern void nvptx_output_ascii (FILE *, const char *, unsigned HOST_WIDE_INT);
 extern void nvptx_register_pragmas (void);
+extern unsigned int nvptx_data_alignment (const_tree, unsigned int);
 
 #ifdef RTX_CODE
 extern void nvptx_expand_oacc_fork (unsigned);
diff --git a/gcc/config/nvptx/nvptx.c b/gcc/config/nvptx/nvptx.c
index d513ddb..1c84b1b 100644
--- a/gcc/config/nvptx/nvptx.c
+++ b/gcc/config/nvptx/nvptx.c
@@ -236,6 +236,8 @@ nvptx_ptx_type_from_mode (machine_mode mode, bool promote)
 
     case V2SImode:
       return ".v2.u32";
+    case V2DImode:
+      return ".v2.u64";
 
     default:
       gcc_unreachable ();
@@ -2181,7 +2183,20 @@ nvptx_output_mov_insn (rtx dst, rtx src)
 	    ? "%.\tmov%t0\t%0, %1;" : "%.\tmov.b%T0\t%0, %1;");
 
   if (GET_MODE_SIZE (dst_inner) == GET_MODE_SIZE (src_inner))
-    return "%.\tmov.b%T0\t%0, %1;";
+    {
+      if (GET_MODE_BITSIZE (dst_mode) == 128
+	  && GET_MODE_BITSIZE (GET_MODE (src)) == 128)
+	{
+	  /* mov.b128 is not supported.  */
+	  if (dst_inner == V2DImode && src_inner == TImode)
+	    return "%.\tmov.u64\t%0.x, %L1;\n\t%.\tmov.u64\t%0.y, %H1;";
+	  else if (dst_inner == TImode && src_inner == V2DImode)
+	    return "%.\tmov.u64\t%L0, %1.x;\n\t%.\tmov.u64\t%H0, %1.y;";
+
+	  gcc_unreachable ();
+	}
+      return "%.\tmov.b%T0\t%0, %1;";
+    }
 
   return "%.\tcvt%t0%t1\t%0, %1;";
 }
@@ -2419,6 +2434,20 @@ nvptx_print_operand (FILE *file, rtx x, int code)
       fprintf (file, "%s", nvptx_ptx_type_from_mode (mode, code == 't'));
       break;
 
+    case 'H':
+    case 'L':
+      {
+	rtx inner_x = SUBREG_REG (x);
+	machine_mode inner_mode = GET_MODE (inner_x);
+	machine_mode split = maybe_split_mode (inner_mode);
+
+	output_reg (file, REGNO (inner_x), split,
+		    (code == 'H'
+		     ? GET_MODE_SIZE (inner_mode) / 2
+		     : 0));
+      }
+      break;
+
     case 'S':
       {
 	nvptx_shuffle_kind kind = (nvptx_shuffle_kind) UINTVAL (x);
@@ -5363,7 +5392,38 @@ nvptx_goacc_reduction (gcall *call)
 static bool
 nvptx_vector_mode_supported (machine_mode mode)
 {
-  return mode == V2SImode;
+  return (mode == V2SImode
+	  || mode == V2DImode);
+}
+
+/* Return the preferred mode for vectorizing scalar MODE.  */
+
+static machine_mode
+nvptx_preferred_simd_mode (machine_mode mode)
+{
+  switch (mode)
+    {
+    case DImode:
+      return V2DImode;
+    case SImode:
+      return V2SImode;
+
+    default:
+      return default_preferred_simd_mode (mode);
+    }
+}
+
+unsigned int
+nvptx_data_alignment (const_tree type, unsigned int basic_align)
+{
+  if (TREE_CODE (type) == INTEGER_TYPE)
+    {
+      unsigned HOST_WIDE_INT size = tree_to_uhwi (TYPE_SIZE_UNIT (type));
+      if (size == GET_MODE_SIZE (TImode))
+	return GET_MODE_BITSIZE (maybe_split_mode (TImode));
+    }
+
+  return basic_align;
 }
 
 #undef TARGET_OPTION_OVERRIDE
@@ -5483,6 +5543,10 @@ nvptx_vector_mode_supported (machine_mode mode)
 #undef TARGET_VECTOR_MODE_SUPPORTED_P
 #define TARGET_VECTOR_MODE_SUPPORTED_P nvptx_vector_mode_supported
 
+#undef TARGET_VECTORIZE_PREFERRED_SIMD_MODE
+#define TARGET_VECTORIZE_PREFERRED_SIMD_MODE \
+    nvptx_preferred_simd_mode
+
 struct gcc_target targetm = TARGET_INITIALIZER;
 
 #include "gt-nvptx.h"
diff --git a/gcc/config/nvptx/nvptx.h b/gcc/config/nvptx/nvptx.h
index 0a000a7..4224acf 100644
--- a/gcc/config/nvptx/nvptx.h
+++ b/gcc/config/nvptx/nvptx.h
@@ -52,13 +52,15 @@
 
 /* Alignments in bits.  */
 #define PARM_BOUNDARY 32
-#define STACK_BOUNDARY 64
+#define STACK_BOUNDARY 128
 #define FUNCTION_BOUNDARY 32
-#define BIGGEST_ALIGNMENT 64
+#define BIGGEST_ALIGNMENT 128
 #define STRICT_ALIGNMENT 1
 
 #define MAX_STACK_ALIGNMENT (1024 * 8)
 
+#define DATA_ALIGNMENT nvptx_data_alignment
+
 /* Copied from elf.h and other places.  We'd otherwise use
    BIGGEST_ALIGNMENT and fail a number of testcases.  */
 #define MAX_OFILE_ALIGNMENT (32768 * 8)
diff --git a/gcc/config/nvptx/nvptx.md b/gcc/config/nvptx/nvptx.md
index ba0567c..dff7cc0 100644
--- a/gcc/config/nvptx/nvptx.md
+++ b/gcc/config/nvptx/nvptx.md
@@ -184,7 +184,7 @@
 (define_mode_iterator SDCM [SC DC])
 (define_mode_iterator BITS [SI SF])
 (define_mode_iterator BITD [DI DF])
-(define_mode_iterator VECIM [V2SI])
+(define_mode_iterator VECIM [V2SI V2DI])
 
 ;; This mode iterator allows :P to be used for patterns that operate on
 ;; pointer-sized quantities.  Exactly one of the two alternatives will match.
diff --git a/gcc/testsuite/gcc.target/nvptx/decl-init.c b/gcc/testsuite/gcc.target/nvptx/decl-init.c
index e9af907..23008fb 100644
--- a/gcc/testsuite/gcc.target/nvptx/decl-init.c
+++ b/gcc/testsuite/gcc.target/nvptx/decl-init.c
@@ -37,7 +37,7 @@ struct five five2[2] = {{12, 13}, {14, 15}};
 /* { dg-final { scan-assembler ".align 1 .u8 five2\\\[10\\\] = { 12, 13, 0, 0, 0, 14, 15, 0, 0, 0 };" } } */
 
 int __attribute__((vector_size(16))) vi = {16, 17, 18, 19};
-/* { dg-final { scan-assembler ".align 8 .u32 vi\\\[4\\\] = { 16, 17, 18, 19 };" } } */
+/* { dg-final { scan-assembler ".align 16 .u32 vi\\\[4\\\] = { 16, 17, 18, 19 };" } } */
 
 typedef int __attribute ((mode(TI))) ti_t;
 
diff --git a/gcc/testsuite/gcc.target/nvptx/slp-2-run.c b/gcc/testsuite/gcc.target/nvptx/slp-2-run.c
new file mode 100644
index 0000000..f9841a6
--- /dev/null
+++ b/gcc/testsuite/gcc.target/nvptx/slp-2-run.c
@@ -0,0 +1,23 @@
+/* { dg-do run } */
+/* { dg-options "-O2 -ftree-slp-vectorize" } */
+
+#include "slp-2.c"
+
+int
+main(void)
+{
+  unsigned int i;
+  for (i = 0; i < 1000; i += 1)
+    {
+      p[i] = i;
+      p2[i] = 0;
+    }
+
+  foo ();
+
+  for (i = 0; i < 1000; i += 1)
+    if (p2[i] != i)
+      return 1;
+
+  return 0;
+}
diff --git a/gcc/testsuite/gcc.target/nvptx/slp-2.c b/gcc/testsuite/gcc.target/nvptx/slp-2.c
new file mode 100644
index 0000000..6639491
--- /dev/null
+++ b/gcc/testsuite/gcc.target/nvptx/slp-2.c
@@ -0,0 +1,25 @@
+/* { dg-do assemble } */
+/* { dg-options "-O2 -ftree-slp-vectorize -save-temps" } */
+
+long long int p[1000] __attribute__((aligned(16)));
+long long int p2[1000] __attribute__((aligned(16)));
+
+void __attribute__((noinline, noclone))
+foo ()
+{
+  long long int a, b;
+
+  unsigned int i;
+  for (i = 0; i < 1000; i += 2)
+    {
+      a = p[i];
+      b = p[i+1];
+
+      p2[i] = a;
+      p2[i+1] = b;
+    }
+}
+
+/* { dg-final { scan-assembler "ld.v2.u64" } } */
+/* { dg-final { scan-assembler "st.v2.u64" } } */
+
diff --git a/gcc/testsuite/gcc.target/nvptx/v2di.c b/gcc/testsuite/gcc.target/nvptx/v2di.c
new file mode 100644
index 0000000..f63ad35
--- /dev/null
+++ b/gcc/testsuite/gcc.target/nvptx/v2di.c
@@ -0,0 +1,12 @@
+/* { dg-do assemble } */
+/* { dg-options "-O2 -save-temps" } */
+
+typedef long long int __v2di __attribute__((__vector_size__(16)));
+
+#define TYPE __v2di
+#include "vec.inc"
+
+/* { dg-final { scan-assembler ".reg\\.v2\\.u64" } } */
+/* { dg-final { scan-assembler "ld\\.v2\\.u64" } } */
+/* { dg-final { scan-assembler "st\\.v2\\.u64" } } */
+/* { dg-final { scan-assembler "mov\\.v2\\.u64.*\\{ 1, 2 \\}" } } */
diff --git a/libgomp/testsuite/libgomp.oacc-c/vec.c b/libgomp/testsuite/libgomp.oacc-c/vec.c
new file mode 100644
index 0000000..79c1c17
--- /dev/null
+++ b/libgomp/testsuite/libgomp.oacc-c/vec.c
@@ -0,0 +1,48 @@
+/* { dg-skip-if "" { *-*-* } { "-O0" } { "" } } */
+/* { dg-skip-if "" { *-*-* } { "*" } { "-DACC_MEM_SHARED=0" } } */
+/* { dg-additional-options "-std=c99 -ftree-slp-vectorize -foffload=-ftree-slp-vectorize -foffload=-fdump-tree-slp1 -foffload=-save-temps -save-temps" } */
+
+#include <stdio.h>
+#include <sys/time.h>
+
+long long int p[32 *1000] __attribute__((aligned(16)));
+long long int p2[32 *1000] __attribute__((aligned(16)));
+
+int
+main (void)
+{
+#pragma acc parallel num_gangs(1) num_workers(1) vector_length(32)
+  {
+    if (((unsigned long int)p & (0xfULL)) != 0)
+      __builtin_abort ();
+    if (((unsigned long int)p2 & (0xfULL)) != 0)
+      __builtin_abort ();
+
+    for (unsigned int k = 0; k < 10000; k += 1)
+      {
+#pragma acc loop vector
+	for (unsigned long long int j = 0; j < 32; j += 1)
+	  {
+	    unsigned long long a, b;
+	    unsigned long long *p3, *p4;
+	    p3 = (unsigned long long *)((unsigned long long int)p & (~0xfULL));
+	    p4 = (unsigned long long *)((unsigned long long int)p2 & (~0xfULL));
+
+	    for (unsigned int i = 0; i < 1000; i += 2)
+	      {
+		a = p3[j * 1000 + i];
+		b = p3[j * 1000 + i + 1];
+		
+		p4[j * 1000 + i] = a;
+		p4[j * 1000 + i + 1] = b;
+	      }
+	  }
+      }
+  }
+
+  return 0;
+}
+
+/* Todo: make a scan-tree-dump variant that scans vec.o instead.  */
+/* { dg-final { file copy -force [glob vec.o.*] [regsub \.o\. [glob vec.o.*] \.c\.] } } */
+/* { dg-final { scan-tree-dump "vector\\(2\\) long long unsigned int" "slp1" } } */

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]