为了测量程序中缓存未命中的影响，我想要计算缓存未命中对用于实际计算的周期造成的延迟。\n我用它来perf stat测量周期、L1 负载、L1 未命中、LLC 负载和 LLC - 我的程序中遗漏了。这是一个示例输出：
467\xe2\x80\xaf769,70 msec task-clock # 1,000 CPUs utilized \n 1\xe2\x80\xaf234\xe2\x80\xaf063\xe2\x80\xaf672\xe2\x80\xaf432 cycles # 2,638 GHz (62,50%)\n 572\xe2\x80\xaf761\xe2\x80\xaf379\xe2\x80\xaf098 instructions # 0,46 insn per cycle (75,00%)\n 129\xe2\x80\xaf143\xe2\x80\xaf035\xe2\x80\xaf219 branches # 276,083 M/sec (75,00%)\n 6\xe2\x80\xaf457\xe2\x80\xaf141\xe2\x80\xaf079 branch-misses # 5,00% of all branches (75,00%)\n 195\xe2\x80\xaf360\xe2\x80\xaf583\xe2\x80\xaf052 L1-dcache-loads # 417,643 M/sec (75,00%)\n 33\xe2\x80\xaf224\xe2\x80\xaf066\xe2\x80\xaf301 L1-dcache-load-misses # 17,01% of all L1-dcache hits (75,00%)\n 20\xe2\x80\xaf620\xe2\x80\xaf655\xe2\x80\xaf322 LLC-loads # 44,083 M/sec (50,00%)\n 6\xe2\x80\xaf030\xe2\x80\xaf530\xe2\x80\xaf728 LLC-load-misses # 29,25% of all LL-cache hits (50,00%)\n那么我的问题是：\n如何将缓存未命中数转换为“丢失”时钟周期数？ …