我试图找到一种方法来击败H / W预取器以检测流模式并以随机顺序访问4KB数据 这样它就不会被H / W预取器检测到并预取。
最初我想以随机模式访问所有偶数索引数据，因为H / w预取器预取下一个缓存行 总是（所以当我访问索引时，下一个奇数索引数据已被预取）。
我编写代码以随机模式访问所有偶数索引数据，但结果表明预取程序检测到模式 （不知道怎么样？没有固定的步幅，都是随机的步幅）
我正在调查原因 - 为什么会发生这种情况，然后我在英特尔找到了这篇文章; https://software.intel.com/en-us/forums/topic/473493
根据John D. McCalpin博士的说法，＆＃34;博士。带宽，
＆＃34;英特尔64和IA-32架构优化的2.2.5.4节 参考手册＆＃34; （文件248966 - 028，2013年7月），它声明，
流媒体预取器＆＃34; [d]检测并维护多达32个数据流 访问。对于每个4K字节页面，您可以保持一个前进和一个 可以维持后向流。这意味着L2硬件预取器跟踪16个4KiB页面 最近访问并记住了足够的访问模式 那些跟踪一个前向流和一个后向流的页面。所以 使用＆＃34;随机＆＃34;击败L2流媒体预取器简单地说 确保在制作之前访问超过15个其他4 KiB页面 对先前引用的页面的第二个引用。所以＆＃34;随机＆＃34; 提取序列可以由更多的随机排列组成 超过16 4 KiB页码，每页内有随机偏移量。 （一世 通常在我的排列列表中使用至少32页。）
所以这意味着在访问相同4KB页面的两个不同随机索引之间，我们需要访问至少16个4KB页面来击败H / W预取器。
我已经实施了John D. McCalpin建议的概念，但结果再次表明h / w预取器没有被击败。它能够检测一些模式和预取数据（参见示例输出）。我有20-40个4KB页面中不同数量的访问页面，但结果没有改进/变化。
这是我的代码：
#define _GNU_SOURCE /* See feature_test_macros(7) */
#include <stdlib.h>
#include <stdio.h>
#include <unistd.h>
#include <sched.h>
#ifndef _POSIX_THREAD_PROCESS_SHARED
#error This system does not support process shared mutex
#endif
#define MAX_COUNT 3000
#define INDEX (40*1024) // size of DUMMY 40 4KB pages
inline void clflush(volatile void *p)
{
asm volatile ("clflush (%0)" :: "r"(p));
}
unsigned long probe(char *adrs) {
volatile unsigned long time;
asm __volatile__ (
" mfence \n"
" lfence \n"
" rdtsc \n"
" lfence \n"
" movl %%eax, %%esi \n"
" movl (%1), %%eax \n"
" lfence \n"
" rdtsc \n"
" subl %%esi, %%eax \n"
" clflush 0(%1) \n"
: "=a" (time)
: "c" (adrs)
: "%esi", "%edx");
return time;
}
void shuffle(int *arr, size_t n)
{
if (n > 1)
{
size_t i;
srand(time(NULL));
for (i = 0; i < n - 1; i++)
{
size_t j = i + rand() / (RAND_MAX / (n - i) + 1);
int t = arr[j];
arr[j] = arr[i];
arr[i] = t;
}
}
}
static const int DATA[1024]={0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131,132,133,134,135,136,137,138,139,140,141,142,143,144,145,146,147,148,149,150,151,152,153,154,155,156,157,158,159,160,161,162,163,164,165,166,167,168,169,170,171,172,173,174,175,176,177,178,179,180,181,182,183,184,185,186,187,188,189,190,191,192,193,194,195,196,197,198,199,200,201,202,203,204,205,206,207,208,209,210,211,212,213,214,215,216,217,218,219,220,221,222,223,224,225,226,227,228,229,230,231,232,233,234,235,236,237,238,239,240,241,242,243,244,245,246,247,248,249,250,251,252,253,254,255,256,257,258,259,260,261,262,263,264,265,266,267,268,269,270,271,272,273,274,275,276,277,278,279,280,281,282,283,284,285,286,287,288,289,290,291,292,293,294,295,296,297,298,299,300,301,302,303,304,305,306,307,308,309,310,311,312,313,314,315,316,317,318,319,320,321,322,323,324,325,326,327,328,329,330,331,332,333,334,335,336,337,338,339,340,341,342,343,344,345,346,347,348,349,350,351,352,353,354,355,356,357,358,359,360,361,362,363,364,365,366,367,368,369,370,371,372,373,374,375,376,377,378,379,380,381,382,383,384,385,386,387,388,389,390,391,392,393,394,395,396,397,398,399,400,401,402,403,404,405,406,407,408,409,410,411,412,413,414,415,416,417,418,419,420,421,422,423,424,425,426,427,428,429,430,431,432,433,434,435,436,437,438,439,440,441,442,443,444,445,446,447,448,449,450,451,452,453,454,455,456,457,458,459,460,461,462,463,464,465,466,467,468,469,470,471,472,473,474,475,476,477,478,479,480,481,482,483,484,485,486,487,488,489,490,491,492,493,494,495,496,497,498,499,500,501,502,503,504,505,506,507,508,509,510,511,512,513,514,515,516,517,518,519,520,521,522,523,524,525,526,527,528,529,530,531,532,533,534,535,536,537,538,539,540,541,542,543,544,545,546,547,548,549,550,551,552,553,554,555,556,557,558,559,560,561,562,563,564,565,566,567,568,569,570,571,572,573,574,575,576,577,578,579,580,581,582,583,584,585,586,587,588,589,590,591,592,593,594,595,596,597,598,599,600,601,602,603,604,605,606,607,608,609,610,611,612,613,614,615,616,617,618,619,620,621,622,623,624,625,626,627,628,629,630,631,632,633,634,635,636,637,638,639,640,641,642,643,644,645,646,647,648,649,650,651,652,653,654,655,656,657,658,659,660,661,662,663,664,665,666,667,668,669,670,671,672,673,674,675,676,677,678,679,680,681,682,683,684,685,686,687,688,689,690,691,692,693,694,695,696,697,698,699,700,701,702,703,704,705,706,707,708,709,710,711,712,713,714,715,716,717,718,719,720,721,722,723,724,725,726,727,728,729,730,731,732,733,734,735,736,737,738,739,740,741,742,743,744,745,746,747,748,749,750,751,752,753,754,755,756,757,758,759,760,761,762,763,764,765,766,767,768,769,770,771,772,773,774,775,776,777,778,779,780,781,782,783,784,785,786,787,788,789,790,791,792,793,794,795,796,797,798,799,800,801,802,803,804,805,806,807,808,809,810,811,812,813,814,815,816,817,818,819,820,821,822,823,824,825,826,827,828,829,830,831,832,833,834,835,836,837,838,839,840,841,842,843,844,845,846,847,848,849,850,851,852,853,854,855,856,857,858,859,860,861,862,863,864,865,866,867,868,869,870,871,872,873,874,875,876,877,878,879,880,881,882,883,884,885,886,887,888,889,890,891,892,893,894,895,896,897,898,899,900,901,902,903,904,905,906,907,908,909,910,911,912,913,914,915,916,917,918,919,920,921,922,923,924,925,926,927,928,929,930,931,932,933,934,935,936,937,938,939,940,941,942,943,944,945,946,947,948,949,950,951,952,953,954,955,956,957,958,959,960,961,962,963,964,965,966,967,968,969,970,971,972,973,974,975,976,977,978,979,980,981,982,983,984,985,986,987,988,989,990,991,992,993,994,995,996,997,998,999,1000,1001,1002,1003,1004,1005,1006,1007,1008,1009,1010,1011,1012,1013,1014,1015,1016,1017,1018,1019,1020,1021,1022,1023};
int main(int argc, char *argv[])
{
int counter=0,k=0;
unsigned long Access_Time[MAX_COUNT][64]={0};
int DUMMY[INDEX];// dummy array of 40 * 4KB ;
//Initialize
for(k=0;k<INDEX;k++)
DUMMY[k]=k;
//access it to check segmentation fault is happening or not
for(k=0;k<INDEX;k++)
DUMMY[k]+=k;
// even index in random order
int index[32]={4,8,16,32,54,34,62,50,26,52,30,60,46,18,36,58,42,10,20,40,6,12,24,48,22,44,14,28,56,38,2,0};
int TOTAL_RANDOM_PAGE=40;
int i,PAGE[TOTAL_RANDOM_PAGE]; // PAGE will contain page no of 40 pages which will be accessed in random order to defeat prefetcher
for (i=0; i<TOTAL_RANDOM_PAGE; i++)
{
PAGE[i] = i;
}
shuffle(PAGE, TOTAL_RANDOM_PAGE); // PAGE now have page no in random order
FILE *fp2;
int s,s1;
int random_index=0,sum=0;
const int *p0=&DATA[0];
for (s=0;s<64;s++)
{
clflush((void *)(p0+s*16));
}
while(counter<MAX_COUNT)
{
// Find Access time for Even Index
for (s=0;s<32;s++)
{
// Access a random index
Access_Time[counter][index[s]]=probe((char *)(p0+16*index[s]));
//Now, access 40 other indexes belong to other 40 4KB page
shuffle(PAGE, TOTAL_RANDOM_PAGE); // random orderpage
for(random_index=0;random_index<TOTAL_RANDOM_PAGE;random_index++)
{
DUMMY[1024*PAGE[random_index]+16*PAGE[random_index]]=2*DUMMY[1024*PAGE[random_index]+16*PAGE[random_index]];
}
}// end of for loop
// Flush all DATA from cache
for (s1=0;s1<64;s1++)
{
clflush((void *)(p0+s1*16));
}
counter++;
}// end of while loop
fp2=fopen("All_access_time.txt","a");
int index4;
for(counter=0;counter<MAX_COUNT;counter++)
{
for (index4=0;index4<64;index4++)
{
if(Access_Time[counter][index4]>0 && Access_Time[counter][index4]<200)
fprintf(fp2,"%d,%d,%lu\n",counter,index4,Access_Time[counter][index4]);
}
}
return 1;
}
另一个有趣的观察是，预取的随机索引的访问时间具有大约35-70个滴答的访问时间。 （见样本输出）
在我的系统中，L1访问时间36-44滴答，L2访问时间50-70滴答，L3访问时间= 90-120滴答。
在Intel（R）Core（TM）i3-2100 CPU @ 3.10GHz和Intel（R）Core（TM）i7-3770 CPU @ 3.40GHz上进行了实验，但结果相似。
系统的内部细节很少，
L1-D = 32KB, ways_of_associative=8 L1-I = 32KB, ways_of_associative=8 L2 = 256KB, ways_of_associative=8 L3 = 3072KB (Core-i3), ways_of_associative=12 L3 = 8192KB (Core-i7), ways_of_associative=16 Cache line size=64Bytes
你能帮我理解为什么H / W预取器能够检测到我的随机模式吗？我在哪里犯错误？
如何进行编码以便我可以打败prefetcher和h / w预取器无法预取我的数据？
注意：我在使用gcc使用-O0选项进行编译时禁用了s / w预取器优化。
示例输出：
(counter,index,access_time)
30,8,56
30,18,72
30,20,52
30,28,72
30,34,72
30,36,72
30,38,72
30,40,72
30,42,72
31,8,52
31,18,56
31,20,52
31,28,72
31,34,52
31,36,72
31,38,56
31,40,72
31,42,52
31,60,56
32,8,52
32,18,72
32,20,52
32,28,52
32,34,72
32,36,52
32,38,72
32,40,52
32,42,52
32,48,52
33,8,56
33,18,72
33,20,52
33,28,72
33,34,52
33,36,72
33,38,72
33,40,52
33,42,72
34,8,72
34,18,52
34,20,72
34,28,72
34,34,72
34,36,52
34,38,76
34,40,72
34,42,76
34,60,72
答案 0 :(得分：4)
如果你足够勇敢地编写内核模块，你可以做你想做的事。
由于核心CPU的几乎所有功能都可以禁用硬件预取逻辑以进行调试。
硬件预取由模型特定寄存器IA32_MISC_ENABLE （0x1a0）控制。只需设置该寄存器的第9位，预取器即可关闭。
有关详细信息，请查看“英特尔®64和IA-32架构” 软件开发人员手册“。搜索IA32_MISC_ENABLE将带您进入正确的章节。
同样在Linux内核源代码上搜索相同的关键字会产生一些命中。它们与预取无关，但与其他东西无关，但代码看起来像一个很好的样板，因为它显示了如何从内核读取和写入IA32_MISC_ENABLE寄存器。
如果你这样做，加倍和三重检查你正在做什么。您不希望意外禁用热监视器。它们也位于MISC_ENABLE中： - ）