mirror/gcc
2
0
Fork 0
mirror of git://gcc.gnu.org/git/gcc.git synced 2025-03-25 10:30:43 +08:00
Code Issues Packages Projects Releases Wiki Activity

[AArch64] Set ctz rtx_cost (PR93565)

Browse Source 
Combine sometimes behaves oddly and duplicates ctz to remove an unnecessary
sign extension.  Avoid this by setting the cost for ctz to be higher than
that of a simple ALU instruction.  Deepsjeng performance improves by ~0.6%.

gcc/
	PR rtl-optimization/93565
	* config/aarch64/aarch64.c (aarch64_rtx_costs): Add CTZ costs.

testsuite/
	PR rtl-optimization/93565
	* gcc.target/aarch64/pr93565.c: New test.
This commit is contained in:
Wilco Dijkstra 2020-02-12 18:23:21 +00:00
parent 9921bbf9b2
commit 5bfc8303ff
4 changed files with 51 additions and 0 deletions
gcc
ChangeLog
config/aarch64
aarch64.c
testsuite
ChangeLog
gcc.target/aarch64
pr93565.c

5
gcc/ChangeLog

@ -1,3 +1,8 @@
2020-02-12 Wilco Dijkstra <wdijkstr@arm.com>
PR rtl-optimization/93565
* config/aarch64/aarch64.c (aarch64_rtx_costs): Add CTZ costs.
2020-02-12 Wilco Dijkstra <wdijkstr@arm.com>
* config/aarch64/aarch64-simd.md

7
gcc/config/aarch64/aarch64.c

@ -11507,6 +11507,13 @@ aarch64_rtx_costs (rtx x, machine_mode mode, int outer ATTRIBUTE_UNUSED,
return false;
case CTZ:
*cost = COSTS_N_INSNS (2);
if (speed)
*cost += extra_cost->alu.clz + extra_cost->alu.rev;
return false;
case COMPARE:
op0 = XEXP (x, 0);
op1 = XEXP (x, 1);

5
gcc/testsuite/ChangeLog

@ -1,3 +1,8 @@
2020-02-12 Wilco Dijkstra <wdijkstr@arm.com>
PR rtl-optimization/93565
* gcc.target/aarch64/pr93565.c: New test.
2020-02-12 Wilco Dijkstra <wdijkstr@arm.com>
* gcc.target/aarch64/popcnt2.c: New test.

34
gcc/testsuite/gcc.target/aarch64/pr93565.c Normal file

@ -0,0 +1,34 @@
/* { dg-do compile } */
/* { dg-options "-O2" } */
static const unsigned long long magic = 0x03f08c5392f756cdULL;
static const char table[64] = {
0, 1, 12, 2, 13, 22, 17, 3,
14, 33, 23, 36, 18, 58, 28, 4,
62, 15, 34, 26, 24, 48, 50, 37,
19, 55, 59, 52, 29, 44, 39, 5,
63, 11, 21, 16, 32, 35, 57, 27,
61, 25, 47, 49, 54, 51, 43, 38,
10, 20, 31, 56, 60, 46, 53, 42,
9, 30, 45, 41, 8, 40, 7, 6,
};
static inline int ctz1 (unsigned long b)
{
unsigned long lsb = b & -b;
return table[(lsb * magic) >> 58];
}
void f (unsigned long x, int *p)
{
if (x != 0)
{
int a = ctz1 (x);
*p = a | p[a];
}
}
/* { dg-final { scan-assembler-times "rbit\t" 1 } } */
/* { dg-final { scan-assembler-times "clz\t" 1 } } */