maytonЗатестил. Пока результат .. не очень. Складывается ощущение что где-то идет жесткая конкуренция
между вычислительными потоками. Приложение в 1 поток отработало быстрее чем многопоточные кейсы.
Я слегка изменил скриптик добавив 3-й опциональный параметр от 1.. до 12
1.
2.
3.
4.
5.
6.
7.
#!/bin/bash -v
for i in {1..12}
do
echo "Start with $i thread(s)"
time ./card-raytracer-cpp-mt.exe $i.ppm $i
done
Щас приаттачу лог и цифирки.
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.
#!/bin/bash -v
for i in {1..12}
do
echo "Start with $i thread(s)"
time ./card-raytracer-cpp-mt.exe $i.ppm $i
done
Start with 1 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 1 threads to file 1.ppm ...
Init end: 36 msec
------- STAT -------
thread_max 1
parallel_run 1
thread_create 1
thread_wake_up 1
try_wake_up 243
msg_create 262144
actor_create 3
actor_get 0
actor_find 4
actor_not_run 0
cache_found 2
cache_bad 0
cache_full 524529
resource_lock 1
msg_send 786432
msg_send/sec 79889
Time: 9844 msec
real 0m9.851s
user 0m9.868s
sys 0m0.016s
Start with 2 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 2 threads to file 2.ppm ...
Init end: 53 msec
------- STAT -------
thread_max 2
parallel_run 2
thread_create 2
thread_wake_up 2
try_wake_up 49
msg_create 262144
actor_create 3
actor_get 0
actor_find 8
actor_not_run 0
cache_found 1
cache_bad 0
cache_full 262195
resource_lock 2
msg_send 786432
msg_send/sec 40544
Time: 19397 msec
real 0m19.403s
user 0m20.665s
sys 0m13.215s
Start with 3 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 3 threads to file 3.ppm ...
Init end: 69 msec
------- STAT -------
thread_max 5
parallel_run 4
thread_create 5
thread_wake_up 3
try_wake_up 172
msg_create 262144
actor_create 3
actor_get 0
actor_find 14
actor_not_run 0
cache_found 1
cache_bad 0
cache_full 262601
resource_lock 3
msg_send 786432
msg_send/sec 44193
Time: 17795 msec
real 0m17.801s
user 0m21.843s
sys 0m23.147s
Start with 4 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 4 threads to file 4.ppm ...
Init end: 70 msec
------- STAT -------
thread_max 5
parallel_run 4
thread_create 5
thread_wake_up 4
try_wake_up 113
msg_create 262144
actor_create 3
actor_get 0
actor_find 15
actor_not_run 0
cache_found 1
cache_bad 0
cache_full 262299
resource_lock 4
msg_send 786432
msg_send/sec 38652
Time: 20346 msec
real 0m20.352s
user 0m26.864s
sys 0m43.069s
Start with 5 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 5 threads to file 5.ppm ...
Init end: 67 msec
------- STAT -------
thread_max 6
parallel_run 5
thread_create 6
thread_wake_up 5
try_wake_up 206
msg_create 262144
actor_create 3
actor_get 0
actor_find 18
actor_not_run 0
cache_found 1
cache_bad 0
cache_full 263025
resource_lock 5
msg_send 786432
msg_send/sec 28075
Time: 28011 msec
real 0m28.016s
user 0m34.124s
sys 1m26.774s
Start with 6 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 6 threads to file 6.ppm ...
Init end: 68 msec
------- STAT -------
thread_max 7
parallel_run 6
thread_create 7
thread_wake_up 6
try_wake_up 201
msg_create 262144
actor_create 3
actor_get 0
actor_find 21
actor_not_run 0
cache_found 1
cache_bad 0
cache_full 262507
resource_lock 6
msg_send 786432
msg_send/sec 26394
Time: 29795 msec
real 0m29.801s
user 0m37.429s
sys 1m59.716s
Start with 7 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 7 threads to file 7.ppm ...
Init end: 53 msec
------- STAT -------
thread_max 8
parallel_run 7
thread_create 8
thread_wake_up 7
try_wake_up 167
msg_create 262144
actor_create 3
actor_get 0
actor_find 24
actor_not_run 0
cache_found 1
cache_bad 0
cache_full 262445
resource_lock 7
msg_send 786432
msg_send/sec 26198
Time: 30018 msec
real 0m30.025s
user 0m39.843s
sys 2m27.361s
Start with 8 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 8 threads to file 8.ppm ...
Init end: 62 msec
------- STAT -------
thread_max 9
parallel_run 8
thread_create 9
thread_wake_up 8
try_wake_up 312
msg_create 262144
actor_create 3
actor_get 0
actor_find 27
actor_not_run 0
cache_found 1
cache_bad 0
cache_full 263302
resource_lock 8
msg_send 786432
msg_send/sec 26198
Time: 30018 msec
real 0m30.022s
user 0m41.379s
sys 2m53.991s
Start with 9 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 9 threads to file 9.ppm ...
Init end: 62 msec
------- STAT -------
thread_max 10
parallel_run 9
thread_create 10
thread_wake_up 9
try_wake_up 162
msg_create 262144
actor_create 3
actor_get 0
actor_find 29
actor_not_run 0
cache_found 2
cache_bad 0
cache_full 262347
resource_lock 9
msg_send 786432
msg_send/sec 26570
Time: 29598 msec
real 0m29.603s
user 0m42.097s
sys 3m19.522s
Start with 10 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 10 threads to file 10.ppm ...
Init end: 57 msec
------- STAT -------
thread_max 11
parallel_run 10
thread_create 11
thread_wake_up 10
try_wake_up 399
msg_create 262144
actor_create 3
actor_get 0
actor_find 33
actor_not_run 0
cache_found 1
cache_bad 0
cache_full 265511
resource_lock 10
msg_send 786432
msg_send/sec 26454
Time: 29728 msec
real 0m29.734s
user 0m43.622s
sys 3m47.234s
Start with 11 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 11 threads to file 11.ppm ...
Init end: 57 msec
------- STAT -------
thread_max 12
parallel_run 11
thread_create 12
thread_wake_up 11
try_wake_up 351
msg_create 262144
actor_create 3
actor_get 0
actor_find 36
actor_not_run 0
cache_found 1
cache_bad 0
cache_full 263099
resource_lock 11
msg_send 786432
msg_send/sec 26665
Time: 29493 msec
real 0m29.500s
user 0m44.572s
sys 4m12.562s
Start with 12 thread(s)
compile Apr 9 2019 20:27:37
lite_thread 12 threads to file 12.ppm ...
Init end: 65 msec
------- STAT -------
thread_max 13
parallel_run 12
thread_create 13
thread_wake_up 16
try_wake_up 185
msg_create 262144
actor_create 3
actor_get 0
actor_find 46
actor_not_run 0
cache_found 2
cache_bad 0
cache_full 262307
resource_lock 15
msg_send 786432
msg_send/sec 26488
Time: 29690 msec
real 0m29.695s
user 0m44.418s
sys 4m42.378s
Забыл вывод статистики отключить, запушил, обновись.
Результаты очень странные, свел в таблицу
Потоков Время мс19844219397317795420346528011629795730018830018929598 102972811294931229690
Время должно уменьшаться, а тут наоборот. Запусти еще с ключом 0, это исходный однопоточный код без акторов.
Затестил у себя на i7-3770К 4 ядра + HT
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.
C:\Test>card_raytracer.exe 1.ppm 1
C:\Test>card_raytracer.exe 1.ppm 0
compile Apr 10 2019 09:11:59
original code to file 1.ppm ...
Time: 13610 msec
compile Apr 10 2019 09:11:59
lite_thread 1 threads to file 1.ppm ...
Init end: 34 msec
Time: 13741 msec
C:\Test>card_raytracer.exe 1.ppm 2
compile Apr 10 2019 09:11:59
lite_thread 2 threads to file 1.ppm ...
Init end: 36 msec
Time: 6999 msec
C:\Test>card_raytracer.exe 1.ppm 3
compile Apr 10 2019 09:11:59
lite_thread 3 threads to file 1.ppm ...
Init end: 53 msec
Time: 5198 msec
C:\Test>card_raytracer.exe 1.ppm 4
compile Apr 10 2019 09:11:59
lite_thread 4 threads to file 1.ppm ...
Init end: 42 msec
Time: 4146 msec
C:\Test>card_raytracer.exe 1.ppm 5
compile Apr 10 2019 09:11:59
lite_thread 5 threads to file 1.ppm ...
Init end: 55 msec
Time: 3757 msec
C:\Test>card_raytracer.exe 1.ppm 6
compile Apr 10 2019 09:11:59
lite_thread 6 threads to file 1.ppm ...
Init end: 54 msec
Time: 3494 msec
C:\Test>card_raytracer.exe 1.ppm 7
compile Apr 10 2019 09:11:59
lite_thread 7 threads to file 1.ppm ...
Init end: 64 msec
Time: 3192 msec
C:\Test>card_raytracer.exe 1.ppm 8
compile Apr 10 2019 09:11:59
lite_thread 8 threads to file 1.ppm ...
Init end: 63 msec
Time: 3057 msec
Потоков Время мс Ускорение раз0136101137411.0269991.9351982.6441463.3537573.6634943.9731924.3830574.5
i7 6700K 4 ядра без HT
Потоков Время мс Ускорение раз0115301110061.0259811.9343182.7435573.2
Похоже этот тест не очень подходит для замера производительности одного ядра, но очень сильно удивил твой результат, медленнее не должно становиться. Если это косяк моей либы с акторами, то это очень плохо, я не тестил ее на многоядерном линуксе. Сделаю загрузочную флэшку с линуксом, потестю у себя.
