OpenGL texture mapping without duplicate vertices - optimization

I've been using OpenGL for some time now, and I am at a point where I'm doing optimizing of the various aspects of my application.
Here's the problem: In some cases I end up duplicating vertixes for texture mapping. I've been using this book a lot:
http://staff.fit.ac.cy/eng.ap/FALL2017/3d-game-development-with-lwjgl.pdf
You can go to pages 70-71 and see the issue and the suggested solution. Basically, it says that when a vertex needs to have a different texture coord for different faces - you've got to duplicate that vertex.
And that that's the only solution.
But I hesitate - is it? Lots of UV maps will have cases where the coords are cut apart a bit, for vertexes. I.e. vertexes end up having different coords. Duplicating them seems like a huge waste.
Is there a way to do it a different way?

A vertex is not a just a position. A vertex is the whole vector of the position, color, texture coordinate and whatever else attribute you're putting into it.
But I hesitate - is it?
Yes it is. It is also the most efficient way to do this, as is allows for easier caching.
Is there a way to do it a different way?
You could use different arrays for each attribute, then have a primary index array to address into secondary index arrays, one for each attribute, that then do the lookup for each attribute.
Two indirections and very little cache coherence. Overall the overhead of secondary index arrays will consume more memory, than the little overhead of a "few" duplicate position vertices. And you're avoiding double indirection.
Overall, the way some duplicated vertex positions is also the most performant one.

Related

Best data structure for insert/update/delete and search of 2d point

I have a infrastructure for massive entities on a GIS map - using the graphic card for drawing (WebGL).
Today, I'm using quad-tree for indexing the data and querying - for selection/drawing purposes.
Recently I've found a way to enable my users to update the location of the entities and draw the change very fast. for example updating 15000 locations, and redrawing took less then 0.05 ms.
The problem now is updating my data structure. It is very very slow.
I've ran over many data structures, such as R-tree, B-Tree and more.. but not yet found satisfied result.
My question is
what is the optimal data structure for 2d points from performances point of view, for inserting/updating and query (by distance from point, rectangle) ?
Maybe there is a web gl solution for this ?
what is the optimal data structure for 2d points from performances
point of view, for inserting/updating and query (by distance from
point, rectangle) ?
It's hard to find a data structure that satisfies all of these beautifully. Typically for searches to be fastest, insertion has to be slower, for insertion to be faster, searches have to be slower, so it's a balancing act.
However, given your needs, a simple NxM grid might actually be ideal. I come originally from a gaming background where sprites can be moving at every single frame. Quad-trees are often too expensive for such dynamic content that wants to update the structure every single frame. As a result, a simple NxM grid often works better for those cases.
The NxM grid can do a rectangular query potentially even faster than the quad-tree, as it simply has to determine which grid cells overlap the rectangle and then simply check the elements within the cells to see if the overlap/are inside the rectangle. Same goes for a search within a radial parameter (circle).
Insertion and removal is dirt cheap since we only need to update the cells in which the elements overlap. If you're dealing with points, a point only belongs in one cell, and it becomes an extremely simple constant-time operation. In fact, if each point stores a list pointer for the next element, moving it from one cell to another (along with removing it from a cell and inserting it to a cell) only requires changing pointers around -- no memory allocation required. The grid then becomes a grid of head pointers for a singly-linked list.
Another approach is that I imagine you're currently using one massive quad-tree for the entire world. You can use a hybrid of these techniques by using a quad-tree to represent a small region, e.g., with quad-trees tiled together to form a grid of quad-trees. This can significantly reduce the cost to update a quad-tree for a specific region since each quad-tree stores only the data for that region.

Titan vertex centric indices vs Neo4j labels

I was trying to make a comparison between these two technologies when approaching this and I was wondering if any of you already have some experience dealing with any or both of them?
I am mainly interested in performance numbers when dealing with similar use cases.
The difference between the two concepts is the difference between global and local indexing.
As I understand it, Neo4j vertex labels allow you to break up your index space by "categories" of vertices. In this way, a O(log(|V|)) lookup is now an O(log(|V|/c)), where c is the number of categories/labels you have over your vertex set and (the equation) assumes an equal number of vertices in each category. As such, vertex label aid in global index calls as this is a function of V.
Next, Titan's vertex-centric indices sort and index the incident edges of a vertex. The cost to find a particular edge by its label/properties incident to a vertex is O(log(inc(v))), where inc(v) is the size of the incident edge set to vertex v. As such, vertex-centric indices are local indices as this is a function of v.
As I understand it, Neo4j does not support vertex-centric indices. You see this concept currently in Titan, OrientDB, and TinkerGraph (…and RDF stores sort in this manner as well -- via spog pairings). Next, all known graph databases support global indices though, (I believe only Neo4j and OrientDB), support a vertex set partition via the concept of a label.
Again, assuming my assumptions are correct about the use of vertex labels in Neo4j, we are talking about two different use cases — global vs. local indexing. From the perspective of the supernode problem, global indices do not quell the issue of traversing through a large vertex, while this is the sole purpose of the local vertex-centric indices.
You can read about the supernode problem and vertex-centric indices here:
http://thinkaurelius.com/2012/10/25/a-solution-to-the-supernode-problem/
Agreeing with everything Marko said, one could take it further and argue that in the graph database world local indexes can (and even should) substitute global ones. In my opinion, the single greatest advantage of a graph data model is that it lets you encode your data model into the graph topology, gaining qualitative advantages in terms of flexibility, ease of evolution and performance. With this in mind, I'd argue that labels in Neo4j actually detract from all this; reifying a label into a node with adjacent edges pointing to the source having that label is much more in line with the "schema is the graph" philosophy.
Of course, if your engine lacks local indexes we are back at the supernode problem. But if you do have them (something which I'd say should be a requirement for something to be called a graph database), you can easily transform your label into a node L, and create relationships pointing to that node for those vertices which you want labeled with L
v -[L]-> L
meaning that v has label L. Now if you want this in Titan to behave like a Neo4j label, just make the -[L]-> relation to be "manyToOne" (see Titan cardinality constraints) and create a vertex-centric index. This pattern lets you get everything that you could with labels and much more; you can
effectively use this as a namespace for properties relating to that label
sort your elements inside one label
nest labels easily without losing performance (just use a composite key)
separate the declaration of a label L with how elements labeled with it are accessed
Labels may afford some design patterns that improve performance by de-densifying the graph. For example: they eliminate the need for type nodes, which can often get quite dense. Labels can optionally be associated with a unique index. Here, the ability to index a property isn't new, but the ability to constrain it uniquely is. If you were previously doing work in your application, you may experience some performance gains by letting the database handle this. (It's certainly much more convenient to do so.) Finally, if you don't assign a unique index to a label, it will still be indexed, in order to help performance for certain kinds of queries (e.g. "give me all of the nodes having label ")
All that said, while labels may help with performance in certain cases, they were introduced more with ease-of-use in mind. We're just getting started with Neo4j 2.1, which specifically addresses dense node performance (something I know you've been waiting for), along with other performance & scalability improvements... including removing (for all practical purposes eliminating) the upper size limits.
Philip

Computational complexity and shape nesting

I have SVG abirtrary paths which i need to pack as efficiently as possible within a given rectangle(as less waste of space as possible). After some research i found the bin packing algorithms which seems to be dealing with boxes and not curved random shapes(my SVG shapes are quite complex and include beziers etc.).
AFAIK, there is no deterministic algorithm for actually packing abstract shapes.
I wish to be proven wrong here which would be ideal(having a mathematical deterministic method for packing them). In case I am right however and there is not, what would be the best approach to this problem
The subject name is Shape Nesting, Nesting Problem or Nesting Process.
In Shape Nesting there is no single/uniform algorithm or mathematical method for nesting shapes and getting the least space waste possible.
The 1st method is the packing algorithm(creates an imaginary bounding
box for each shape and uses a rectangular 2D algorithm to pack the
bounding boxes).
This method is fast but the least efficient in regards to space
waste.
The 2nd method is some kind of incremental rotation. The algorithm
rotates the shape at incremental steps and checks if it fits in the
space. This is better than the packing method in regards to space
waste but it is painstakingly slow,
What are some other classroom examples for achieving a solution to this problem?
[Edit1] new answer
as mentioned before bin-packing is NP complete (hard) so forget about algebraic solution
known approaches are:
generate and test
either you test all possibility of the problem and remember the best solution or incrementally add items (not all at once) one by one with the same way. It is basically what you are doing now without proper heuristic is unusably slow. But has the best space efficiency (the first one is much better but much slower) O(N!)
take advantage of sorting items by size
something like this it is much faster almost O(N.log(N)) (according to used sorting algorithm). Space efficiency strongly depends on the items size range and count. For rectangular shapes is this the best approach (fastest and usable even for N>1000). For complex shapes is this not a good way but look at it anyway maybe you get some idea ...
use of Neural network
This is extremly vague approach without any warrant of solution but possible best space efficiency/runtime ratio
I think there could be some field approach out there
I sow a few for generating graph layouts. All items create fields (booth attractive and repulsive) so they are moving to semi-stable state.
At first all items are at random locations
When the movement stop remember best solution and shake all items a little or randomize their position again.
Cycle this few times
This approach is much faster then genere and test and can provide very close solution to it but it can hang in local min/max or oscillate if the fields are not optimally choosed. For example all items can have constant attractive force to each other and repulsive force getting stronger only when the items are very close. You have to prevent overlapping of items (either by stronger repulsion or by collision tests). You have also to create some rotation moment for example with that repulsive force. It differs on any vertex so it creates a rotation moment (that can automatically align similar sides closer together). Also you can have semi-stable state with big distances between items and after finding best solution just turn off repulsion fields so they stick together. Sometimes it can have better results some times not ... here is nice example for graph layout computation
Logic to strategically place items in a container with minimum overlapping connections
Demo from the same QA
And here solver for placing sliders in 2D:
How to implement a constraint solver for 2-D geometry?
[Edit0] old answer before reformulating the question
I am not clear what you want to achieve.
have SVG picture and want to separate its parts to rectangular regions
as filled as can be
least empty space in them
no shape change in picture
have svg picture and want to change its shapes according to some purpose
if this is the case some additional info is needed
[solution for 1]
create a list of points for whole SVG in global SVG space (all points are transformed)
for line you need add 2 points
for rectangles 4 points
circle/elipse/bezier/eliptic arc 8 points
find local centres of mass
use classical approach
or can speed things up by computing the average density of points per x,y axis separately and after that just check all combinations of found positions of local max of densities if they really are sub cluster center or not.
all sub cluster center is the center of your region
now find the most far points which are still part of your cluster (the are close enough to neighbour points)
create rectangular area that cover all points from sub cluster.
you also can remove all used points from list
repeat fro all valid sub clusters
until all points are used
another not precise but simpler approach is:
find SVG size
create planar map of svg with some precision for example int map[256][256].
size of map can be constant or with the same aspect as SVG
clear map with 0
for any point of SVG set related map point to 1 (or inc or whatever)
now just segmentate map and you will have find your objects
after segmentation you have position and size of all objects
so finding of bounding boxes should be easy
You can start with a variant of the rectangle bin-packing algorithm and add rotation. There is a method "Guillotine bin packer" and you can download a paper and a library at github.

transform a path along an arc

Im trying to transform a path along an arc.
My project is running on osX 10.8.2 and the painting is done via CoreAnimation in CALayers.
There is a waveform in my project which will be painted by a path. There are about 200 sample points which are mirrored to the bottom side. These are painted 60 times per second and updated to a song postion.
Please ignore the white line, it is just a rotation indicator.
What i am trying to achieve is drawing a waveform along an arc. "Up" should point to the middle. It does not need to go all the way around. The waveform should be painted along the green circle. Please take a look at the sketch provided below.
Im not sure how to achieve this in a performant manner. There are many points per second that need coordinate correction.
I tried coming up with some ideas of my own:
1) There is the possibility to add linear transformations to paths, which, i think, will not help me here. The only thing i can think of is adding a point, rotating the path with a transformation, adding another point, rotating and so on. But this would be very slow i think
2) Drawing the path into an image and bending it would surely lead to image-artifacts.
3) Maybe the best idea would be to precompute sample points on an arc, then save save a vector to the center. Taking the y-coordinates of the waveform, placing them on the sample points and moving them along the vector to the center.
But maybe i am just not seeing some kind of easy solution to this problem. Help is really appreciated and fresh ideas very welcome. Thank you in advance!
IMHO, the most efficient way to go (in terms of CPU usage) would be to use some form of pre-computed approach that would take into account the resolution of the display.
Cleverly precomputed values
I would go for the mathematical transformation (from linear to polar) and combine two facts:
There is no need to perform expansive mathematical computation
There is no need to render two points that are too close from each other
I have no ready-made algorithm for you, but you could use a pre-computed sin or cos table, and match the data range to the display size in order to work with integers.
For instance imagine we have some data ranging from 0 to 1E6 and we need to display the sin value of each point in a 100 pix height rectangle. We can use a pre-computed sin table and work with integers. This way displaying the sin value of a point would be much quicker. This concept can be refined to get a nicer result.
Also, there are some ways to retain only significant points of a curve so that the displayed curve actually looks like the original (see the Ramer–Douglas–Peucker algorithm on wikipedia). But I found it to be inefficient for quickly displaying ever-changing data.
Using multicore rendering
You could compute different areas of the curve using multiple cores (can be tricky)
Or you could use pre-computing using several cores, and one core to do finish the job.

Efficient way to handle large runtime-generated tile maps?

I am coding a 2 dimensional, tile based (orthogonal tiles) iPhone game. All levels are procedurally generated when the app is first played, and then persist until the user wants a new map. Maps are rather large, being 1000 tiles in both width and height, and the terrain is destructible. At the moment it is rather similar to Terraria, but that will change.
To hold map/tile information I am currently using several 2 dimensional c style arrays. This works well, but I am concerned as to the amount of memory this takes up, as the arrays are all defined as short array[1000][1000], which takes up (1000 * 1000 * sizeof(short)) bytes of space.
This is not particularly desirable when the iPhone doesn't have an incredibly large amount of memory to work with, especially when the user is multitasking. The main problem is that there is no way that I can use a specific tile map format such as .tmx, because all the levels are procedurally generated. Performance could also be an issue, because if a tile is destroyed at index(x, y), then I need to change the data in that index. I have also thought about writing tile map data to a text file, but I think there would be difficulties or performance issues when accessing or changing data.
Keeping all this in mind, what would be an efficient and fast way to handle my tile data?
My gut feeling on this is Core Data structured such that each tile element has relationships to the tiles around it. There's some non-trivial overhead here, but the advantage is that you can release tiles that aren't onscreen from memory and fault them back when you need them. As you move in a direction, you can query for the tiles in that direction, and you can fairly cheaply dump memory when you're in the background. This would get rid of the "several" 2D arrays and move all the data into a single object. In principle, the grid could be infinite in size this way, since everything is by relationship rather than coordinate.
You could similarly approach the problem using SQLite, querying for rows and columns in a given range. You might mark the objects as NSDiscardableContent and put them in an NSCache, which could dramatically improve memory performance. You could still generate an effectively-infinite grid as long as you allow coordinates to be both positive and negative.