diff mbox series

nvptx: : Add support for popcount and widening multiply instructions

Message ID 000c01d64fa8$78551430$68ff3c90$@nextmovesoftware.com
State New
Headers show
Series nvptx: : Add support for popcount and widening multiply instructions | expand

Commit Message

Roger Sayle July 1, 2020, 1:06 p.m. UTC 
The following patch adds support for the popc and mul.wide instructions to the nvptx backend.
I've a follow-up patch for supporting mul.hi instructions, but those changes require some minor
tweaks to GCC's middle-end, so I'll submit those pieces separately.

Tested by "make" and "make -k check" on --build=nvptx-none hosted on x86_64-pc-linux-gnu with
no new regressions.

2020-07-01  Roger Sayle  <roger@nextmovesoftware.com>

gcc/ChangeLog:
        * config/nvptx/nvptx.md (popcount<mode>2): New instructions.
        (mulhishi3, mulsidi3, umulhisi3, umulsidi3): New instructions.

gcc/testsuite/ChangeLog:
        * gcc.target/nvptx/popc-1.c: New test.
        * gcc.target/nvptx/popc-2.c: New test.
        * gcc.target/nvptx/popc-3.c: New test.
        * gcc.target/nvptx/mul-wide.c: New test.
        * gcc.target/nvptx/umul-wide.c: New test.


Ok for mainline?

Thanks in advance,
Roger
--
Roger Sayle
NextMove Software
Cambridge, UK

/* { dg-do compile } */
/* { dg-options "-O2" } */

unsigned int foo(unsigned int x)
{
  return __builtin_popcount(x);
}

/* { dg-final { scan-assembler-times "popc.b32" 1 } } */
/* { dg-do compile } */
/* { dg-options "-O2" } */

unsigned long foo(unsigned long x)
{
  return __builtin_popcountl(x);
}

/* { dg-final { scan-assembler-times "popc.b64" 1 } } */
/* { dg-final { scan-assembler-times "cvt.s64.s32" 1 } } */
/* { dg-do compile } */
/* { dg-options "-O2" } */

unsigned int foo(unsigned long x)
{
  return __builtin_popcountl(x);
}

/* { dg-final { scan-assembler-times "popc.b64" 1 } } */
/* { dg-final { scan-assembler-times "cvt.s64.s32" 0 } } */
/* { dg-do compile } */
/* { dg-options "-O2" } */

int mulhisi3(short x, short y)
{
  return (int)x * (int)y;
}

long mulsidi3(int x, int y)
{
  return (long)x * (long)y;
}

/* { dg-final { scan-assembler-times "mul.wide.s16" 1 } } */
/* { dg-final { scan-assembler-times "mul.wide.s32" 1 } } */
/* { dg-do compile } */
/* { dg-options "-O2" } */

unsigned int umulhisi3(unsigned short x, unsigned short y)
{
  return (unsigned int)x * (unsigned int)y;
}

unsigned long umulsidi3(unsigned int x, unsigned int y)
{
  return (unsigned long)x * (unsigned long)y;
}

/* { dg-final { scan-assembler-times "mul.wide.u16" 1 } } */
/* { dg-final { scan-assembler-times "mul.wide.u32" 1 } } */

Comments

Tom de Vries July 2, 2020, 1:28 p.m. UTC | #1 
On 7/1/20 3:06 PM, Roger Sayle wrote:
> 
> The following patch adds support for the popc and mul.wide instructions to the nvptx backend.
> I've a follow-up patch for supporting mul.hi instructions, but those changes require some minor
> tweaks to GCC's middle-end, so I'll submit those pieces separately.
> 
> Tested by "make" and "make -k check" on --build=nvptx-none hosted on x86_64-pc-linux-gnu with
> no new regressions.
> 
> 2020-07-01  Roger Sayle  <roger@nextmovesoftware.com>
> 
> gcc/ChangeLog:
>         * config/nvptx/nvptx.md (popcount<mode>2): New instructions.
>         (mulhishi3, mulsidi3, umulhisi3, umulsidi3): New instructions.
> 
> gcc/testsuite/ChangeLog:
>         * gcc.target/nvptx/popc-1.c: New test.
>         * gcc.target/nvptx/popc-2.c: New test.
>         * gcc.target/nvptx/popc-3.c: New test.
>         * gcc.target/nvptx/mul-wide.c: New test.
>         * gcc.target/nvptx/umul-wide.c: New test.
> 
> 
> Ok for mainline?
> 

Hi Roger,

LGTM, please apply.

[ Btw, can you next time add the new files to the patch.  That's
somewhat more convenient to apply. ]

Thanks
- Tom
diff mbox series

Patch

diff --git a/gcc/config/nvptx/nvptx.md b/gcc/config/nvptx/nvptx.md
index 089cdf0..5ceeac7 100644
--- a/gcc/config/nvptx/nvptx.md
+++ b/gcc/config/nvptx/nvptx.md
@@ -493,6 +493,50 @@ 
   DONE;
 })
 
+(define_insn "popcount<mode>2"
+  [(set (match_operand:SI 0 "nvptx_register_operand" "=R")
+	(popcount:SI (match_operand:SDIM 1 "nvptx_register_operand" "R")))]
+  ""
+  "%.\\tpopc.b%T1\\t%0, %1;")
+
+;; Multiplication variants
+
+(define_insn "mulhisi3"
+  [(set (match_operand:SI 0 "nvptx_register_operand" "=R")
+	(mult:SI (sign_extend:SI
+		  (match_operand:HI 1 "nvptx_register_operand" "R"))
+		 (sign_extend:SI
+		  (match_operand:HI 2 "nvptx_register_operand" "R"))))]
+  ""
+  "%.\\tmul.wide.s16\\t%0, %1, %2;")
+
+(define_insn "mulsidi3"
+  [(set (match_operand:DI 0 "nvptx_register_operand" "=R")
+	(mult:DI (sign_extend:DI
+		  (match_operand:SI 1 "nvptx_register_operand" "R"))
+		 (sign_extend:DI
+		  (match_operand:SI 2 "nvptx_register_operand" "R"))))]
+  ""
+  "%.\\tmul.wide.s32\\t%0, %1, %2;")
+
+(define_insn "umulhisi3"
+  [(set (match_operand:SI 0 "nvptx_register_operand" "=R")
+	(mult:SI (zero_extend:SI
+		  (match_operand:HI 1 "nvptx_register_operand" "R"))
+		 (zero_extend:SI
+		  (match_operand:HI 2 "nvptx_register_operand" "R"))))]
+  ""
+  "%.\\tmul.wide.u16\\t%0, %1, %2;")
+
+(define_insn "umulsidi3"
+  [(set (match_operand:DI 0 "nvptx_register_operand" "=R")
+	(mult:DI (zero_extend:DI
+		  (match_operand:SI 1 "nvptx_register_operand" "R"))
+		 (zero_extend:DI
+		  (match_operand:SI 2 "nvptx_register_operand" "R"))))]
+  ""
+  "%.\\tmul.wide.u32\\t%0, %1, %2;")
+
 ;; Shifts
 
 (define_insn "ashl<mode>3"