Hi,
I am using gcc-4.2 to "compile" my ARMv7 assembly language .s file as
part of an XCode3.2 project on Mac OSX 10.6. I am using ARM NEON
assembly language instructions, but gcc doesn't allow me to set the
memory alignment for ARM NEON loads. For example, the following
assembly language instruction should tell the ARM NEON coprocessor
that it should load a chunk of memory from memory that is aligned to
64 bytes:
vld1.8 {q0}, [r0@64]
This should run execute faster than the instruction for unknown memory
alignment:
vld1.8 {q0}, [r0]
Unfortunately, gcc-4.2 gives me an error when it sees the @ symbol. Is
there any way to specify the memory alignment in assembly code (for
NEON load and store instructions)?
Cheers,
Shervin Emami.
http://www.shervinemami.co.cc/
