Jump to content

  • Log In with Google      Sign In   
  • Create Account

Interested in a FREE copy of HTML5 game maker Construct 2?

We'll be giving away three Personal Edition licences in next Tuesday's GDNet Direct email newsletter!

Sign up from the right-hand sidebar on our homepage and read Tuesday's newsletter for details!


We're also offering banner ads on our site from just $5! 1. Details HERE. 2. GDNet+ Subscriptions HERE. 3. Ad upload HERE.


SLMATH library and SSE optimisation problem.


Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

  • You cannot reply to this topic
5 replies to this topic

#1 RobinsonUK   Members   -  Reputation: 108

Like
0Likes
Like

Posted 02 March 2012 - 07:31 AM

I have a problem with the SLMATH library. Not sure if anyone uses it or has used it before? Anyway, the issue is that when I compile with SSE optimisation enabled (in VS 2010), I obviously have to provide a container that has the correct byte alignment for SSE type objects. This is OK because there's a little class in SLMATH that's an aligned vector; it aligns the vector allocation on an 8 byte boundary (i.e. I do not use std::vector<>).

Now the problem is that it appears any structure or class that contains something like slm::mat4 must also be aligned on such a boundary too, before it's put into a collection. So, for example, I used an aligned vector to create an array of slm::mat4, but if I create a class called Mesh, and Mesh contains an slm::mat4 and I want to put Mesh into a std::vector, well, I get strange memory errors whilst debugging.

So given the documentation is very sparse indeed, can anyone who's used this library tell me what, precisely, I have to do to use it with SSE optimisation? I mean I don't like the idea of having to use aligned vectors absolutely everywhere in place of std::vector just in case an slm:: component ends up being encapsulated into a class or structure somehow.

Alternatively, a fast vector/matrix/graphics math library as good as SLMATH would be great if there's on around.

Thanks for any advice you can offer.



Robin

Sponsor:

#2 RobinsonUK   Members   -  Reputation: 108

Like
1Likes
Like

Posted 02 March 2012 - 07:53 AM

Actually this is a general problem with putting aligned objects into containers, i.e. a very simple repro case:


#include <vector>

class Item
{

public:

    __declspec(align(8))
    struct {

	    float a, b, c, d;

    } Aligned;
};


int main()
{
    // Error - won't compile.

    std::vector<Item> myItems;
}


#3 Ripiz   Members   -  Reputation: 529

Like
1Likes
Like

Posted 02 March 2012 - 08:13 AM

std::vector<> doesn't support aligned structures/classes.
Possible hack/workaround is to edit STL code, and add reference in resize(), but that's not very good idea.
You might need to create own container for this.

Sidenote: I think SSE requires 16 byte alignment, not 8 byte.

#4 RobinsonUK   Members   -  Reputation: 108

Like
0Likes
Like

Posted 02 March 2012 - 08:32 AM

Thanks. It does, yes. It's a giant pain, so I think I'll just switch off SSE in the project settings and live without the performance boost.

#5 Dave Eberly   Members   -  Reputation: 1161

Like
1Likes
Like

Posted 06 March 2012 - 01:51 AM

Thanks. It does, yes. It's a giant pain, so I think I'll just switch off SSE in the project settings and live without the performance boost.


std::vector should be able to support alignment through custom allocators. However, if you are using MSVS 2010, the dinkumware STL they use has a bug in that the std::vector resize does not do the right thing (fixed in MSVS 2011). For MSVS 2010, you'll have to roll your own std::vector (maybe copy what dinkumware does and "fix" the resize).

#6 jkajala   Members   -  Reputation: 122

Like
0Likes
Like

Posted 08 March 2012 - 07:38 PM

Nice to hear someone finds the lib useful. In the 2.4.1 version there is vector_simd<T> (in vector_simd.h) which does this:


/**
* Very minimal but very efficient std::vector clone for plain data SSE/SIMD contents like vec4.
* This is useful for using SSE support on 32-bit Visual Studio builds,
* which suffer from the memory allocation alignment problem (std::vector memory not aligned to vec4).
* NOTE: vector_simd does NOT call constructors/destructors of the contained elements correctly,
* so it is NOT suitable to store anything else than "plain old datastructures" to these.
* If more correct semantics of constructors/destructors are needed then std::vector should be used.
* For documentation of the methods, please see std::vector.
*
* @ingroup vec_util
*/
template <class T> class vector_simd

Hope this helps!


Bests,
Jani
(author of slmath)

I have a problem with the SLMATH library. Not sure if anyone uses it or has used it before? Anyway, the issue is that when I compile with SSE optimisation enabled (in VS 2010), I obviously have to provide a container that has the correct byte alignment for SSE type objects. This is OK because there's a little class in SLMATH that's an aligned vector; it aligns the vector allocation on an 8 byte boundary (i.e. I do not use std::vector<>).






Old topic!
Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.



PARTNERS