[ODE] Some SSE in Quick step

Billy Zelsnack billy_zelsnack at yahoo.com
Mon May 24 10:29:12 MST 2004


 >Nguyen Binh wrote:
 >Hi , As recommend by Russ, I SIMDize some part
 >of QuickStep. I expect some performance gain but
 >have not profiled it yet. I remember the list have a
 >person who provide many performance comparison.
 >Now please re-run the tests... ;)

Here you go..

7pyramid at 60hz

quickstep from the 19th

      o1p4i10 o2p4i20 o2p4i10 o2p4i5  o2p2i20 o2p2i10 o2p2i5  tokamak
newton
 0 :  0.21ms  0.09ms  0.16ms  0.08ms  0.08ms  0.08ms  0.08ms  0.12ms 
0.14ms
 1 :  0.21ms  0.09ms  0.09ms  0.09ms  0.09ms  0.09ms  0.10ms  0.07ms 
0.05ms
 2 :  0.20ms  0.09ms  0.09ms  0.09ms  0.09ms  0.09ms  0.09ms  0.06ms 
0.05ms
 3 :  0.85ms  0.36ms  0.28ms  0.21ms  0.23ms  0.22ms  0.18ms  0.15ms 
0.24ms
 4 :  2.03ms  0.98ms  0.56ms  0.41ms  0.52ms  0.37ms  0.29ms  0.17ms 
0.17ms
 5 :  3.10ms  0.76ms  0.50ms  0.39ms  0.46ms  0.35ms  0.33ms  0.33ms 
0.63ms
 6 :  3.93ms  1.71ms  1.06ms  0.73ms  0.94ms  0.62ms  0.47ms  0.45ms 
1.33ms
 7 :  4.72ms  1.57ms  1.00ms  0.88ms  0.86ms  0.66ms  0.51ms  0.55ms 
1.90ms
 8 :  5.28ms  2.13ms  1.29ms  0.90ms  1.17ms  0.78ms  0.56ms  0.67ms 
3.11ms
 9 :  7.58ms  2.29ms  1.32ms  0.92ms  1.21ms  0.81ms  0.58ms  0.70ms 
3.64ms
10 :  5.43ms  2.27ms  1.34ms  0.92ms  1.22ms  0.79ms  0.58ms  0.71ms 
3.75ms
11 :  5.30ms  2.14ms  1.39ms  0.95ms  1.20ms  0.80ms  0.58ms  0.73ms 
3.65ms
12 :  5.38ms  2.12ms  1.32ms  0.92ms  1.18ms  0.78ms  0.58ms  0.74ms 
3.26ms
13 :  5.19ms  2.00ms  1.30ms  1.04ms  1.16ms  0.78ms  0.58ms  0.72ms 
3.90ms
14 :  5.32ms  1.91ms  1.27ms  0.93ms  1.16ms  0.77ms  0.57ms  0.73ms 
4.34ms
15 :  5.37ms  1.87ms  1.30ms  0.91ms  1.13ms  0.75ms  0.58ms  0.71ms 
3.93ms
16 :  5.20ms  1.92ms  1.25ms  0.94ms  1.25ms  0.77ms  0.58ms  0.71ms 
3.83ms
17 :  5.33ms  1.88ms  1.21ms  0.95ms  1.14ms  0.79ms  0.57ms  0.73ms 
3.95ms
18 :  5.40ms  2.10ms  1.23ms  0.91ms  1.13ms  0.77ms  0.57ms  0.70ms 
3.73ms
19 :  5.14ms  1.88ms  1.24ms  0.93ms  1.28ms  0.77ms  0.57ms  0.69ms 
3.61ms
20 :  5.29ms  2.04ms  1.26ms  0.90ms  1.18ms  0.79ms  0.61ms  0.71ms 
4.81ms
21 :  7.22ms  1.90ms  1.22ms  0.87ms  1.16ms  0.78ms  0.57ms  0.71ms 
3.00ms
22 :  5.14ms  1.89ms  1.26ms  1.08ms  1.27ms  0.78ms  0.57ms  0.69ms 
4.25ms
23 :  5.27ms  1.89ms  1.22ms  0.94ms  1.15ms  0.79ms  0.59ms  0.70ms 
4.58ms
24 :  5.34ms  1.74ms  1.24ms  0.92ms  1.16ms  0.77ms  0.58ms  0.77ms 
3.90ms
25 :  5.17ms  1.80ms  1.27ms  0.94ms  1.20ms  0.76ms  0.58ms  0.71ms 
4.09ms
26 :  5.27ms  1.83ms  1.27ms  0.90ms  1.16ms  0.78ms  0.58ms  0.71ms 
4.35ms
27 :  5.35ms  1.80ms  1.20ms  0.92ms  1.15ms  0.75ms  0.57ms  0.69ms 
3.90ms
28 :  5.16ms  1.81ms  1.29ms  1.04ms  1.14ms  0.75ms  0.57ms  0.69ms 
4.32ms
29 :  5.26ms  1.77ms  1.21ms  0.92ms  1.18ms  0.77ms  0.60ms  0.70ms 
3.96ms
30 :  5.33ms  1.78ms  1.18ms  0.90ms  1.12ms  0.75ms  0.58ms  0.69ms 
3.96ms
31 :  5.10ms  1.82ms  1.23ms  0.88ms  1.17ms  0.90ms  0.59ms  0.69ms 
3.70ms
32 :  5.45ms  1.72ms  1.18ms  0.89ms  1.19ms  0.77ms  0.60ms  0.71ms 
4.20ms
33 :  5.20ms  1.86ms  1.18ms  0.90ms  1.30ms  0.84ms  0.61ms  0.71ms 
3.56ms
34 :  6.76ms  1.81ms  1.29ms  0.92ms  1.30ms  0.78ms  0.58ms  0.70ms 
4.02ms
35 :  5.31ms  1.81ms  1.23ms  0.89ms  1.21ms  0.77ms  0.60ms  0.72ms 
4.39ms
36 :  5.22ms  1.90ms  1.23ms  0.90ms  1.12ms  0.77ms  0.59ms  0.70ms 
3.77ms
37 :  5.15ms  1.76ms  1.28ms  0.91ms  1.27ms  0.77ms  0.59ms  0.70ms 
4.28ms
38 :  5.29ms  1.84ms  1.25ms  0.89ms  1.17ms  0.77ms  0.61ms  0.72ms 
4.10ms
39 :  5.22ms  2.08ms  1.26ms  0.88ms  1.16ms  0.75ms  0.59ms  0.69ms 
5.08ms
40 :  5.31ms  1.92ms  1.23ms  1.02ms  1.17ms  0.75ms  0.59ms  0.69ms 
4.07ms
41 :  5.29ms  1.81ms  1.24ms  0.86ms  1.20ms  0.76ms  0.61ms  0.71ms 
4.27ms
42 :  5.20ms  2.00ms  1.30ms  0.86ms  1.16ms  0.75ms  0.59ms  0.69ms 
5.04ms
43 :  5.40ms  1.90ms  1.19ms  0.97ms  1.18ms  0.77ms  0.60ms  0.68ms 
3.76ms
44 :  5.10ms  1.86ms  1.22ms  0.91ms  1.33ms  0.79ms  0.60ms  0.68ms 
4.55ms
45 :  5.31ms  2.12ms  1.32ms  0.85ms  1.16ms  0.75ms  0.58ms  0.68ms 
4.39ms
46 :  5.12ms  1.83ms  1.29ms  0.99ms  1.15ms  0.78ms  0.59ms  0.68ms 
3.93ms
47 :  6.86ms  1.88ms  1.30ms  0.89ms  1.21ms  0.79ms  0.73ms  0.69ms 
3.80ms
48 :  5.30ms  1.84ms  1.29ms  0.88ms  1.13ms  0.75ms  0.61ms  0.70ms 
4.20ms
49 :  5.14ms  1.83ms  1.24ms  0.86ms  1.15ms  0.77ms  0.59ms  0.66ms 
4.48ms

sse quickstep

      o1p4i10 o2p4i20 o2p4i10 o2p4i5  o2p2i20 o2p2i10 o2p2i5  tokamak
newton
 0 :  0.21ms  0.09ms  0.08ms  0.08ms  0.08ms  0.08ms  0.08ms  0.12ms 
0.16ms
 1 :  0.20ms  0.10ms  0.11ms  0.09ms  0.09ms  0.08ms  0.09ms  0.06ms 
0.05ms
 2 :  0.20ms  0.09ms  0.16ms  0.09ms  0.09ms  0.08ms  0.09ms  0.06ms 
0.05ms
 3 :  0.84ms  0.44ms  0.30ms  0.23ms  0.27ms  0.21ms  0.20ms  0.23ms 
0.24ms
 4 :  2.06ms  1.21ms  0.80ms  0.53ms  0.66ms  0.45ms  0.37ms  0.16ms 
0.17ms
 5 :  3.10ms  1.03ms  0.65ms  0.50ms  0.59ms  0.43ms  0.42ms  0.33ms 
0.66ms
 6 :  3.99ms  2.35ms  1.47ms  0.97ms  1.23ms  0.78ms  0.57ms  0.45ms 
1.20ms
 7 :  4.83ms  2.13ms  1.34ms  1.02ms  1.19ms  0.90ms  0.64ms  0.56ms 
2.11ms
 8 :  5.19ms  3.08ms  1.82ms  1.20ms  1.55ms  1.01ms  0.77ms  0.67ms 
3.22ms
 9 :  5.53ms  3.13ms  1.83ms  1.25ms  1.63ms  1.09ms  0.76ms  0.72ms 
3.79ms
10 :  5.46ms  2.99ms  1.83ms  1.24ms  1.66ms  1.07ms  0.82ms  0.71ms 
3.75ms
11 :  5.36ms  2.99ms  1.83ms  1.27ms  1.63ms  1.07ms  0.85ms  0.71ms 
3.60ms
12 :  5.46ms  2.85ms  1.87ms  1.29ms  1.63ms  1.11ms  0.74ms  0.71ms 
3.18ms
13 :  5.33ms  2.88ms  1.76ms  1.28ms  1.60ms  1.05ms  0.82ms  0.72ms 
3.80ms
14 :  5.44ms  2.64ms  1.77ms  1.41ms  1.60ms  1.04ms  0.74ms  0.70ms 
4.17ms
15 :  5.31ms  2.65ms  1.77ms  1.29ms  1.67ms  1.00ms  0.73ms  0.71ms 
4.17ms
16 :  5.69ms  2.65ms  1.77ms  1.27ms  1.70ms  1.01ms  0.76ms  0.71ms 
3.82ms
17 :  5.23ms  2.72ms  1.73ms  1.27ms  1.68ms  0.99ms  0.91ms  0.71ms 
3.84ms
18 :  5.15ms  2.72ms  1.76ms  1.25ms  1.67ms  1.00ms  0.74ms  0.70ms 
3.68ms
19 :  5.15ms  2.63ms  1.67ms  1.32ms  1.53ms  1.02ms  0.75ms  0.75ms 
3.63ms
20 :  5.19ms  2.59ms  1.71ms  1.25ms  1.69ms  1.00ms  0.72ms  0.69ms 
4.68ms
21 :  6.83ms  2.69ms  1.82ms  1.20ms  1.61ms  1.05ms  0.74ms  0.68ms 
3.00ms
22 :  5.18ms  2.80ms  1.81ms  1.18ms  1.55ms  0.99ms  0.73ms  0.69ms 
4.33ms
23 :  5.19ms  2.68ms  1.83ms  1.21ms  1.55ms  1.10ms  0.74ms  0.69ms 
4.46ms
24 :  5.19ms  2.59ms  1.84ms  1.19ms  1.57ms  1.02ms  0.73ms  0.69ms 
3.93ms
25 :  5.23ms  2.57ms  1.82ms  1.23ms  1.55ms  1.16ms  0.74ms  0.70ms 
4.17ms
26 :  5.19ms  2.50ms  1.85ms  1.22ms  1.57ms  1.01ms  0.72ms  0.69ms 
4.32ms
27 :  5.22ms  2.68ms  1.61ms  1.21ms  1.61ms  1.09ms  0.75ms  0.71ms 
3.85ms
28 :  5.21ms  2.69ms  1.66ms  1.18ms  1.60ms  1.02ms  0.73ms  0.68ms 
4.54ms
29 :  5.26ms  2.71ms  1.64ms  1.18ms  1.56ms  1.20ms  0.75ms  0.70ms 
3.88ms
30 :  5.23ms  2.56ms  1.68ms  1.19ms  1.56ms  1.02ms  0.72ms  0.71ms 
3.96ms
31 :  5.23ms  2.66ms  1.60ms  1.13ms  1.53ms  1.02ms  0.82ms  0.70ms 
3.70ms
32 :  5.19ms  2.69ms  1.62ms  1.15ms  1.54ms  1.00ms  0.74ms  0.71ms 
4.08ms
33 :  5.22ms  2.59ms  1.67ms  1.16ms  1.54ms  1.02ms  0.72ms  0.79ms 
5.41ms
34 :  5.24ms  2.73ms  1.70ms  1.20ms  1.49ms  1.00ms  0.74ms  0.72ms 
4.22ms
35 :  5.23ms  2.75ms  1.65ms  1.16ms  1.49ms  1.02ms  1.07ms  0.73ms 
4.20ms
36 :  5.20ms  2.65ms  1.65ms  1.15ms  1.54ms  1.04ms  0.76ms  0.69ms 
4.07ms
37 :  5.28ms  2.70ms  1.69ms  1.15ms  1.59ms  1.04ms  0.83ms  0.70ms 
4.48ms
38 :  5.24ms  2.78ms  1.66ms  1.14ms  1.52ms  1.02ms  0.76ms  0.69ms 
4.10ms
39 :  5.17ms  2.76ms  1.59ms  1.15ms  1.57ms  1.04ms  0.76ms  0.70ms 
4.97ms
40 :  5.31ms  2.55ms  1.55ms  1.15ms  1.59ms  1.03ms  0.76ms  0.68ms 
4.25ms
41 :  5.25ms  2.68ms  1.59ms  1.15ms  1.55ms  1.02ms  0.89ms  0.69ms 
4.11ms
42 :  5.10ms  2.72ms  1.56ms  1.13ms  1.64ms  0.99ms  0.74ms  0.68ms 
4.95ms
43 :  5.19ms  2.72ms  1.60ms  1.14ms  1.55ms  1.00ms  0.74ms  0.74ms 
3.73ms
44 :  5.17ms  2.83ms  1.67ms  1.15ms  1.56ms  1.01ms  0.74ms  0.68ms 
4.36ms
45 :  5.10ms  2.80ms  1.72ms  1.15ms  1.60ms  1.02ms  0.72ms  0.67ms 
4.60ms
46 :  5.09ms  2.75ms  1.76ms  1.18ms  1.58ms  2.63ms  0.74ms  0.68ms 
3.84ms
47 :  5.13ms  2.83ms  1.67ms  1.18ms  1.65ms  1.01ms  0.82ms  0.68ms 
3.78ms
48 :  5.10ms  2.72ms  1.74ms  1.18ms  1.58ms  1.00ms  0.72ms  0.67ms 
3.98ms
49 :  5.10ms  2.79ms  1.78ms  1.15ms  1.58ms  0.99ms  0.71ms  0.66ms 
4.54ms



	
		
__________________________________
Do you Yahoo!?
Yahoo! Domains – Claim yours for only $14.70/year
http://smallbusiness.promotions.yahoo.com/offer 


More information about the ODE mailing list