[PATCH] D103903: [ARM] Transform a fixed-point to floating-point conversion into a VCVT_fix

[Sam Tebbs via Phabricator via llvm-commits]

samtebbs marked an inline comment as done.
samtebbs added inline comments.


================
Comment at: llvm/test/CodeGen/ARM/arm_q15_to_float_autovec.ll:434
+  %2 = sitofp <4 x i32> %0 to <4 x float>
+  %3 = fmul <4 x float> %2, <float 0xBE00000000000000, float 0xBE00000000000000, float 0xBE00000000000000, float 0xBE00000000000000>
+  ret <4 x float> %3
----------------
dmgreen wrote:
> This kind of looks negative ;)
It does! I have added a check for fracbits == scalarbits - 1 when not unsigned.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D103903/new/

https://reviews.llvm.org/D103903