📄 prefetch-5.c
字号:
/* { dg-do compile { target i?86-*-* x86_64-*-* } } *//* { dg-require-effective-target ilp32 } *//* { dg-options "-O2 -fprefetch-loop-arrays -march=athlon -fdump-tree-aprefetch-details" } *//* These are common idioms for writing variable-length arrays at the end of structures. We should not deduce anything about the number of iterations of the loops from them. */struct tail0{ int xxx; int yyy[0];};int loop0 (int n, struct tail0 *x){ int i, s = 0; for (i = 0; i < n; i++) s += x->yyy[i]; return s;}struct tail1{ int xxx; int yyy[1];};int loop1 (int n, struct tail1 *x){ int i, s = 0; for (i = 0; i < n; i++) s += x->yyy[i]; return s;}/* It is unlikely that this should be a tail array. We may deduce that most likely, the loop iterates about 5 times, and the array is not worth prefetching. */struct tail5{ int xxx; int yyy[5];};int loop5 (int n, struct tail5 *x){ int i, s = 0; for (i = 0; i < n; i++) s += x->yyy[i]; return s;}/* { dg-final { scan-tree-dump-times "Issued prefetch" 2 "aprefetch" } } *//* { dg-final { scan-tree-dump-times "Not prefetching" 1 "aprefetch" } } *//* { dg-final { cleanup-tree-dump "aprefetch" } } */
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -