Skip to content

N
nexus-am

OpenXiangShan / nexus-am
9 个月 前同步成功

通知 0
Star 21
Fork 25
N
nexus-am

前往新版Gitcode，体验更适合开发者的 AI 搜索 >>

提交 d9b43ff8 编写于 作者: W William Wang
浏览文件
操作

maprobe: let all test return result

上级 b27f2968
隐藏空白更改
内联 并排
Showing with 51 addition and 33 deletion
apps/maprobe/bandwidth-test.c
浏览文件 @ d9b43ff8
#include "maprobe.h"
void test_l1_load_bandwidth(uint64_t size, int iter, int to_csv)
float test_l1_load_bandwidth(uint64_t size, int iter, int to_csv)
{
// printf("stride %d linear access latency test\n", step);
// printf("range (B), read latency, iters, samples, cycles\n");
......@@ -23,17 +23,19 @@ void test_l1_load_bandwidth(uint64_t size, int iter, int to_csv)
_perf_end_timer();
// _perf_print_timer();
uint64_t total_access = size / _PERF_CACHELINE_SIZE_BYTE * 8 * iter;
float bandwidth = total_access * 8 * BYTE / (float)perf.cycle;
if (to_csv) {
printf("%ld, %f, %d, %ld, %ld\n", size, (float)perf.cycle / total_access, iter, total_access, perf.cycle);
} else {
printf("range %ldKB (%d iters) dcache linear (8Byte) read, latency %f, throughput %f B/cycle (%ld samples, %ld cycles), stride %dB\n",
size/KB, iter, (float)perf.cycle / total_access, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle, 8
size/KB, iter, (float)perf.cycle / total_access, bandwidth, total_access, perf.cycle, 8
);
}
_perf_g_total_samples += total_access;
return bandwidth;
}
void test_l1_store_bandwidth(uint64_t size, int iter, int to_csv)
float test_l1_store_bandwidth(uint64_t size, int iter, int to_csv)
{
// printf("stride %d linear access latency test\n", step);
// printf("range (B), read latency, iters, samples, cycles\n");
......@@ -56,17 +58,19 @@ void test_l1_store_bandwidth(uint64_t size, int iter, int to_csv)
_perf_end_timer();
// _perf_print_timer();
uint64_t total_access = size / _PERF_CACHELINE_SIZE_BYTE * 8 * iter;
float bandwidth = total_access * 8 * BYTE / (float)perf.cycle;
if (to_csv) {
printf("%ld, %f, %d, %ld, %ld\n", size, (float)perf.cycle / total_access, iter, total_access, perf.cycle);
} else {
printf("range %ldKB (%d iters) dcache linear (8Byte) store latency %f, throughput %f B/cycle (%ld samples, %ld cycles), stride %dB\n",
size/KB, iter, (float)perf.cycle / total_access, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle, 8
size/KB, iter, (float)perf.cycle / total_access, bandwidth, total_access, perf.cycle, 8
);
}
_perf_g_total_samples += total_access;
return bandwidth;
}
void test_l1_store_wcb_bandwidth(uint64_t size, int iter, int to_csv)
float test_l1_store_wcb_bandwidth(uint64_t size, int iter, int to_csv)
{
// printf("stride %d linear access latency test\n", step);
// printf("range (B), read latency, iters, samples, cycles\n");
......@@ -82,6 +86,7 @@ void test_l1_store_wcb_bandwidth(uint64_t size, int iter, int to_csv)
_perf_end_timer();
// _perf_print_timer();
uint64_t total_access = size / _PERF_CACHELINE_SIZE_BYTE * iter;
float bandwidth = total_access * _PERF_CACHELINE_SIZE_BYTE / (float)perf.cycle;
if (to_csv) {
printf("%ld, %f, %d, %ld, %ld\n", size, (float)perf.cycle / total_access, iter, total_access, perf.cycle);
} else {
......@@ -90,4 +95,5 @@ void test_l1_store_wcb_bandwidth(uint64_t size, int iter, int to_csv)
);
}
_perf_g_total_samples += total_access;
return bandwidth;
}
\ No newline at end of file
apps/maprobe/include/maprobe.h
浏览文件 @ d9b43ff8
......@@ -68,16 +68,16 @@ extern void _perf_blackhole(uint64_t value);
extern uint64_t setup_pointer_tracing_linklist(uint64_t base_addr, uint64_t end_addr, uint64_t step);
extern uint64_t read_pointer_tracing_linklist(uint64_t base_addr, uint64_t num_valid_node);
extern void latency_test_warmup(uint64_t base_addr, uint64_t end_addr);
extern void test_pointer_tracing_latency(uint64_t size, int step, int iter, int to_csv);
extern void test_linear_access_latency(uint64_t size, uint64_t step, int iter, int to_csv);
extern void test_random_access_latency(uint64_t num_access, uint64_t test_range, uint64_t test_align, int pregen_addr, int iter, int to_csv);
extern void test_same_address_load_latency(int iter, int to_csv);
extern void test_read_after_write_latency(int iter, int to_csv);
extern float test_pointer_tracing_latency(uint64_t size, int step, int iter, int to_csv);
extern float test_linear_access_latency(uint64_t size, uint64_t step, int iter, int to_csv);
extern float test_random_access_latency(uint64_t num_access, uint64_t test_range, uint64_t test_align, int pregen_addr, int iter, int to_csv);
extern float test_same_address_load_latency(int iter, int to_csv);
extern float test_read_after_write_latency(int iter, int to_csv);
// bandwidth test
extern void test_l1_load_bandwidth(uint64_t size, int iter, int to_csv);
extern void test_l1_store_bandwidth(uint64_t size, int iter, int to_csv);
extern void test_l1_store_wcb_bandwidth(uint64_t size, int iter, int to_csv);
extern float test_l1_load_bandwidth(uint64_t size, int iter, int to_csv);
extern float test_l1_store_bandwidth(uint64_t size, int iter, int to_csv);
extern float test_l1_store_wcb_bandwidth(uint64_t size, int iter, int to_csv);
extern void legacy_test_mem_throughput(uint64_t iter);
extern void legacy_test_mem_throughput_same_set(uint64_t iter);
......
apps/maprobe/latency-test.c
浏览文件 @ d9b43ff8
......@@ -42,7 +42,7 @@ void latency_test_warmup(uint64_t base_addr, uint64_t end_addr)
setup_pointer_tracing_linklist(base_addr, end_addr, _PERF_CACHELINE_SIZE_BYTE);
}
void test_pointer_tracing_latency(uint64_t size, int step, int iter, int to_csv)
float test_pointer_tracing_latency(uint64_t size, int step, int iter, int to_csv)
{
// printf("pointer tracing latency test\n");
// printf("range (B), read latency, iters, samples, cycles\n");
......@@ -59,19 +59,21 @@ void test_pointer_tracing_latency(uint64_t size, int step, int iter, int to_csv)
}
_perf_end_timer();
// _perf_print_timer();
float acpa = (float)perf.cycle / total_node; // average cycle per access
if (to_csv) {
printf("%ld, %f, %d, %ld, %ld\n", size, (float)perf.cycle / total_node, iter, total_node, perf.cycle);
printf("%ld, %f, %d, %ld, %ld\n", size, acpa, iter, total_node, perf.cycle);
} else {
printf("range %ldKB (%d iters) pointer tracing read latency %f, throughput %f B/cycle (%ld samples, %ld cycles)\n",
size/KB, iter, (float)perf.cycle / total_node, total_node * 8 * BYTE / (float)perf.cycle, total_node, perf.cycle
size/KB, iter, acpa, total_node * 8 * BYTE / (float)perf.cycle, total_node, perf.cycle
);
}
_perf_blackhole(result);
_perf_g_total_samples += total_node;
return acpa;
}
void test_same_address_load_latency(int iter, int to_csv)
float test_same_address_load_latency(int iter, int to_csv)
{
// printf("same address load latency test\n", step);
// printf("range (B), read latency, iters, samples, cycles\n");
......@@ -86,19 +88,21 @@ void test_same_address_load_latency(int iter, int to_csv)
_perf_end_timer();
// _perf_print_timer();
uint64_t total_access = iter;
float acpa = (float)perf.cycle / total_access; // average cycle per access
if (to_csv) {
printf("%ld, %f, %d, %ld, %ld\n", 0, (float)perf.cycle / total_access, iter, total_access, perf.cycle);
printf("%ld, %f, %d, %ld, %ld\n", 0, acpa, iter, total_access, perf.cycle);
} else {
printf("same address read latency %f, throughput %f B/cycle (%ld samples, %ld cycles)\n",
(float)perf.cycle / total_access, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle
acpa, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle
);
}
_perf_blackhole(result);
_perf_g_total_samples += total_access;
return acpa;
}
void test_read_after_write_latency(int iter, int to_csv)
float test_read_after_write_latency(int iter, int to_csv)
{
// printf("same address store-load latency test\n", step);
// printf("range (B), read latency, iters, samples, cycles\n");
......@@ -114,19 +118,21 @@ void test_read_after_write_latency(int iter, int to_csv)
_perf_end_timer();
// _perf_print_timer();
uint64_t total_access = iter;
float acpa = (float)perf.cycle / total_access; // average cycle per access
if (to_csv) {
printf("%ld, %f, %d, %ld, %ld\n", 0, (float)perf.cycle / total_access, iter, total_access, perf.cycle);
printf("%ld, %f, %d, %ld, %ld\n", 0, acpa, iter, total_access, perf.cycle);
} else {
printf("read after write latency %f, throughput %f B/cycle (%ld samples, %ld cycles)\n",
(float)perf.cycle / total_access, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle
acpa, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle
);
}
_perf_blackhole(result);
_perf_g_total_samples += total_access;
return acpa;
}
void test_linear_access_latency_simple(uint64_t size, uint64_t step, int iter, int to_csv)
float test_linear_access_latency_simple(uint64_t size, uint64_t step, int iter, int to_csv)
{
// printf("stride %d linear access latency test\n", step);
// printf("range (B), read latency, iters, samples, cycles\n");
......@@ -145,19 +151,21 @@ void test_linear_access_latency_simple(uint64_t size, uint64_t step, int iter, i
_perf_end_timer();
// _perf_print_timer();
uint64_t total_access = num_access * iter;
float acpa = (float)perf.cycle / total_access; // average cycle per access
if (to_csv) {
printf("%ld, %f, %d, %ld, %ld\n", size, (float)perf.cycle / total_access, iter, total_access, perf.cycle);
printf("%ld, %f, %d, %ld, %ld\n", size, acpa, iter, total_access, perf.cycle);
} else {
printf("range %ldKB (%d iters) simple linear read latency %f, throughput %f B/cycle (%ld samples, %ld cycles), stride %dB\n",
size/KB, iter, (float)perf.cycle / total_access, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle, step
size/KB, iter, acpa, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle, step
);
}
_perf_blackhole(result);
_perf_g_total_samples += total_access;
return acpa;
}
void test_linear_access_latency_batch8(uint64_t size, uint64_t step, int iter, int to_csv)
float test_linear_access_latency_batch8(uint64_t size, uint64_t step, int iter, int to_csv)
{
// printf("stride %d linear access latency test\n", step);
// printf("range (B), read latency, iters, samples, cycles\n");
......@@ -203,24 +211,26 @@ void test_linear_access_latency_batch8(uint64_t size, uint64_t step, int iter, i
_perf_end_timer();
// _perf_print_timer();
uint64_t total_access = num_access * iter;
float acpa = (float)perf.cycle / total_access; // average cycle per access
if (to_csv) {
printf("%ld, %f, %d, %ld, %ld\n", size, (float)perf.cycle / total_access, iter, total_access, perf.cycle);
printf("%ld, %f, %d, %ld, %ld\n", size, acpa, iter, total_access, perf.cycle);
} else {
printf("range %ldKB (%d iters) batch(8) linear read latency %f, throughput %f B/cycle (%ld samples, %ld cycles), stride %dB\n",
size/KB, iter, (float)perf.cycle / total_access, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle, step
size/KB, iter, acpa, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle, step
);
}
_perf_blackhole(result);
_perf_g_total_samples += total_access;
return acpa;
}
void test_linear_access_latency(uint64_t size, uint64_t step, int iter, int to_csv)
float test_linear_access_latency(uint64_t size, uint64_t step, int iter, int to_csv)
{
test_linear_access_latency_batch8(size, step, iter, to_csv);
return test_linear_access_latency_batch8(size, step, iter, to_csv);
}
void test_random_access_latency(uint64_t num_access, uint64_t test_range, uint64_t test_align, int pregen_addr, int iter, int to_csv)
float test_random_access_latency(uint64_t num_access, uint64_t test_range, uint64_t test_align, int pregen_addr, int iter, int to_csv)
{
// printf("align %d random access (cache line) latency test, %s\n",
// test_align, pregen_addr ? "use pregen addr array" : "gen rand addr at run time"
......@@ -255,17 +265,19 @@ void test_random_access_latency(uint64_t num_access, uint64_t test_range, uint64
}
// _perf_print_timer();
uint64_t total_access = num_access * iter;
float acpa = (float)perf.cycle / total_access; // average cycle per access
if (to_csv) {
printf("%ld, %f, %d, %ld, %ld\n", test_range, (float)perf.cycle / total_access, iter, total_access, perf.cycle);
printf("%ld, %f, %d, %ld, %ld\n", test_range, acpa, iter, total_access, perf.cycle);
} else {
printf("range %ldKB, access cover %ldKB (%d iters) random read latency %f, throughput %f B/cycle (%ld samples, %ld cycles), align %ldB, %s\n",
test_range/KB, total_access*8*_PERF_CACHELINE_SIZE_BYTE/KB, iter, (float)perf.cycle / (total_access), total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle, test_align,
test_range/KB, total_access*8*_PERF_CACHELINE_SIZE_BYTE/KB, iter, acpa, total_access * 8 * BYTE / (float)perf.cycle, total_access, perf.cycle, test_align,
pregen_addr ? "pregen addr" : "runtime addr"
);
}
_perf_blackhole(result);
_perf_g_total_samples += total_access;
return acpa;
}
void legacy_test_mem_throughput(uint64_t iter)
......
Markdown is supported
0% 或 .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑！
想要评论请 注册 或