The two fundamental laws of optimization:
1. don't do it
2. don't do it *yet*
[edited by - matias suarez on January 20, 2004 3:10:04 AM]
Reducing the penalty for branching?
i think i prefer 1. by now. all those "hidden" things in compiler and cpu seem to make it a pain to figure out whats actually happening and causing the numbers you see. at least its just a way to kill some time, so i can afford doing pointless micro optimizations on functions that will most likely never ever be a problem. hoped it would be a good way to figure out how a few things work.. but only showed me how hard it is to know/guess whats really going on. thats making optimizing a function a bit of a trial and error thing.. change something, check the time, change it again, check the time, notice that the same version varies by 200% with every time you run the program and then just decide its not worth it. one more reason to optimize algorithms rather than code.
quote:Original post by TriencoAnd when you start changing algorithms, you may end up writing 100 times more code than the original, and it''ll be faster.
thats what i hate about programming. the shortest line doesnt have to produce the fastest code.
From here:
---
Our code is expertly written in hand-optimized assembly language. It is thoroughly documented and virtually bug-free.
---
from the benchmark page:
---
...is faster than Direct3D software rendering. In some cases, more than 1000% faster.
---
Hand made optimization can really make a difference, compiler optimizations are still really far from producing ''the best possible optimized code'', they just produce ''an acceptable result on a general purpose way''.
This one is just one of the many examples were hand made optimizations are present as of today.
Anyway sometimes is better to focus more on content than speed
-Mat
---
Our code is expertly written in hand-optimized assembly language. It is thoroughly documented and virtually bug-free.
---
from the benchmark page:
---
...is faster than Direct3D software rendering. In some cases, more than 1000% faster.
---
Hand made optimization can really make a difference, compiler optimizations are still really far from producing ''the best possible optimized code'', they just produce ''an acceptable result on a general purpose way''.
This one is just one of the many examples were hand made optimizations are present as of today.
Anyway sometimes is better to focus more on content than speed
-Mat
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement