|
-
November 10th, 2010, 05:12 AM
#9
Re: Problems designing my vectorized math library
 Originally Posted by Chris_F
I don't see the point. I looked at a dissassembly and the code is horribly mangled and in addition I did some basic tests and timed them. If you want to call my methods of testing inaccurate, go ahead,
No need to get defensive. I didn't say your method of testing is flawed.
However, others could benefit from being able to run the test you made, maybe even try to optimize the original code.
 Originally Posted by Chris_F
That's exactly how I feel too, but given the results I got, I can't really argue with it unless someone can prove it all wrong. He gets good results with the Intel compiler, but I'm stuck with MSVC.
...
Believe me, I rather not, after all I tried to do things "the C++ way" the first time around.
Even if you need to abandon C++ constructs in some parts of the code for the sake of performance, it doesn't mean you have to resort to C style for everything. IMO, if you observe that some code works faster than some other code, but don't understand why, you should consider it a special circumstance and not a general rule. Only if you understand why one thing is faster than another (maybe not in detail, but at least it shouldn't be magic), you are able to form a general rule.
Cheers, D Drmmr
Please put [code][/code] tags around your code to preserve indentation and make it more readable.
As long as man ascribes to himself what is merely a posibility, he will not work for the attainment of it. - P. D. Ouspensky
Posting Permissions
- You may not post new threads
- You may not post replies
- You may not post attachments
- You may not edit your posts
-
Forum Rules
|
Click Here to Expand Forum to Full Width
|