On Wed, 5 Dec 2018 at 15:34, Roger Cuypers rcuypers@arcor.de wrote:

Greetings,

I’m trying to have the Aarch64 gcc optimize a single function using the O3 optimization in this manner:

void __attribute__ ((optimize ("-O3", "-ftree-vectorize" ))) However, when examining the assembler code, there is no trace of any optimization beyond the project default. The only way to successfully enforce the optimization is using -O3 in the gcc command line. Then the compiler produces ARM neon instructions.

Do you know any issues regarding function level optimizations and the Aarch64 gcc?

Hello,

Which version of GCC are you using? I've just tried with gcc-arm-8.2, with this sample code: ============================================== #define SIZE 100 int a[SIZE]; int b[SIZE];

//int myfunc() __attribute__ ((optimize ("-O3", "-ftree-vectorize" ))); int myfunc() { int i; int s=0; for(i=0; i<SIZE; i++) s+=a[i]+b[i]; return s; } ==============================================

Compiled at -O2, the loop contains: ldr w2, [x5, x1] ldr w3, [x4, x1] add x1, x1, 4 cmp x1, 400 add w2, w2, w3 add w0, w0, w2

If I un-comment the attribute line, I get: ldr q0, [x2, x0] ldr q2, [x1, x0] add x0, x0, 16 cmp x0, 400 add v0.4s, v0.4s, v2.4s add v1.4s, v1.4s, v0.4s

In addition, -ftree-vectorize is not needed, it's implied by -O3.

Thanks,

Christophe

Regards _______________________________________________ linaro-toolchain mailing list linaro-toolchain@lists.linaro.org https://lists.linaro.org/mailman/listinfo/linaro-toolchain