Simple sample data generator.
I can't see code or attachment to your post,but here is 2 files.
I think they are correct ... but check.
could you also paste proper answers for those test files?
I don't guarantee these results are correct but here's what our program outputs so far:"3 4 17 14" for the second one and "156 134 424 290" for the first. Is this what you also get?
I can confirm these outputs, I get them too.
Could you also post what you get doing the sum of this square (if you can get it easily) ?
Well I could... But you should do it. You just need to do two for loops and you can find out what the exact sum is.
No what I meant is I find 0 11 19 14 for a total of 252692. And I wondered if you could tell me what you find for your square, so I could compare to what my algorithm sums for it (which is 363). And eventually find out what is going on :P
Nevermind, little found it. Little mixup between declaring elements in int16_t and sums in int32_t. Anyway, thanks for making me notice :)
for others who may want these results, I have :
result : 156 134 424 290
maxSum : 10494167
result : 3 4 17 14
maxSum : 393579
maybe you could post what processing time you had for the example data?
First post on the forum, here to share some results!
So here is what I get:
$ time ./run 40 example_input_20-20.txt3 4 17 14 = 393579./run 40 example_input_20-20.txt 0.07s user 0.12s system 676% cpu 0.028 total
$ time ./run 40 example_input_500-300.txt
156 134 424 290 = 10494167./run 40 example_input_500-300.txt 0.66s user 0.17s system 1300% cpu 0.064 total
With a bigger input (generated):$ time ./run 40 example_input_5000-5000.txt607 1376 4999 4939 = 117784684./run 40 example_input_5000-5000.txt 404.06s user 8.40s system 5957% cpu 6.923 total
What are your numbers?
Just a question, were those numbers got on the MTL ?
Anywho I got on my virtual machine (without parallelization, not done yet) :
for the 20 by 20
for the 500 by 300
I test only the second one : time ./run 1 example_input_500-300.txtI got 63 63 421 236 Max value : 12028820I have done some correction on our program. D:max_value : 10494167156 134 424 290
0.401ms for 500x300? That's fast!
My results are for 20-20 example
from (13,4) to (16,18)
Is anyone have these results?
Thanks in advance
@dina: if you ask about confirming if this is a correct result, scroll up and check a few post above. It is answered already
@buffer, I think you are misleading us, because the real time should always be larger than the user time (without parallelization). And is that 401 ms or 0.401 ms ?
The real time is larger than the user time in the results posted. 0.401>0.270 & 0.036>0.003.
Anywho made others test with the same files :http://pix.toile-libre.org/upload/original/1320083187.png
As you can see it's 0.401s or 401ms. Got even better :)
PS: Those time were obtained on my virtual Fedora, on a Windows 7 Acer laptop.
What's your laptop configuration?
Can you provide us with some more details about test generation?
Are all tests cases random generated, or will there be manually generated test cases?
Also, some upper bound on matrix size would be quite nice to have to see how to handle memory constraints.
Acer Aspire 7738G :
Intel Core 2 Duo P8700 2.53GHz
NVidia GeForce GT 130M
I'll try to make some test on the MTL, but I currently don't have that kind of time ;)