为什么Clang不优化为浮点数增加0.0的优化？

发布于 2025-02-08 16:30:48 字数 864 浏览 1 评论 0原文

用clang 14.0.0（x86-64，-o3）编译以下代码

double f (double x)
{
    return x + 5.0 + 0;
}

结果

.LCPI0_0:
  .quad 0x4014000000000000 # double 5
f(double): # @f(double)
  addsd xmm0, qword ptr [rip + .LCPI0_0]
  xorpd xmm1, xmm1
  addsd xmm0, xmm1
  ret

（

在哪种情况下，该序列

  xorpd xmm1, xmm1
  addsd xmm0, xmm1

有任何区别？

虽然我知道没有-FFAST-MATH的float常数折叠是不可能的，但我看不出任何原因是Xord/addsd sequence需要序列：更改XMM0的位模式，我看不到它如何触发异常或具有任何其他副作用。

编辑：Clang的默认值为-fno-rounding-math（请参阅手册/a>）。因此，可以安全地假设x + 5.0永远不会在-0.0中导致+0.0可以被视为no-op。

原文

Compiling the following code with clang 14.0.0 (x86-64, -O3)

double f (double x)
{
    return x + 5.0 + 0;
}

results in

.LCPI0_0:
  .quad 0x4014000000000000 # double 5
f(double): # @f(double)
  addsd xmm0, qword ptr [rip + .LCPI0_0]
  xorpd xmm1, xmm1
  addsd xmm0, xmm1
  ret

(Godbolt).

In which situation does the sequence

  xorpd xmm1, xmm1
  addsd xmm0, xmm1

make any difference?

While I am aware that float constant folding without -ffast-math is no possible in general, I cannot see any reason why the xord/addsd sequence is needed: It does not change the bit pattern in xmm0, I cannot see how it could trigger an exception or have any other side effect.

Edit: clang's default is -fno-rounding-math (see manual). So it is safe to assume x + 5.0 never results in -0.0 and thus +0.0 can be considered a no-op.

分享到QQ

分享到微博