Hi all, I have a problem with camera performance and I’ve searched for a long time but I haven’t found anything so I hope you can help me.
I have an environment and a panel 2d for controlling camera movements. In this panel there is a 2d button to move vertically the camera with a LerpFunc.
That LerpFunc uses an “easeOut” effect to get the movement much more elegant.
When camera is framing a few objects (low number of polygons) it flows very fast and when it is framing a lot of polygons it’s not the same… is flows a little bit slowly.
Is there any solution about this trouble?
Thanks a lot.
Not really so much the number of polygons, but is to the number of objects the cpu/gpu has to spend time drawing them usally. Read this page should help =)
Thanks a lot for your reply… I’ve tried your suggest but it doesn’t change a lot…
I think it’s my grpahic card, but i’d like to write a program a little bit flexible to have in every computer a discrete level of performance… ( I’d like the same) I don’t want a super speed in a good computer and a super slow speed in a “normal” computer.
The frame rate remain constant during program execution… So I think it’s well done…If you had another suggest I’d thank you very much.
My graphic card is a Intel GMA 4500MHD… very common card. I’m programming it on my laptop to have an idea of performance in a very common pc.
ORIGINAL OUTPUT
Known pipe types:
wglGraphicsPipe
(all display modules loaded.)
3 total nodes (including 0 instances); 0 LODNodes.
0 transforms; 0% of nodes have some render attribute.
68 Geoms, with 35 GeomVertexDatas and 2 GeomVertexFormats, appear on 1 GeomNodes
.
38358 vertices, 38358 normals, 0 colors, 9532 texture coordinates.
GeomVertexData arrays occupy 1052K memory.
GeomPrimitive arrays occupy 351K memory.
42 GeomVertexArrayDatas are redundant, wasting 309K.
160 GeomPrimitive arrays are redundant, wasting 272K.
61483 triangles:
4156 of these are on 804 tristrips (5.16915 average tris per strip).
57327 of these are independent triangles.
7 textures, estimated minimum 2079K texture memory required.
OPTIMIZED OUTPUT
Known pipe types:
wglGraphicsPipe
(all display modules loaded.)
3 total nodes (including 0 instances); 0 LODNodes.
0 transforms; 0% of nodes have some render attribute.
43 Geoms, with 35 GeomVertexDatas and 2 GeomVertexFormats, appear on 1 GeomNodes
.
38358 vertices, 38358 normals, 0 colors, 11622 texture coordinates.
GeomVertexData arrays occupy 1066K memory.
GeomPrimitive arrays occupy 361K memory.
28 GeomVertexArrayDatas are redundant, wasting 308K.
154 GeomPrimitive arrays are redundant, wasting 278K.
61483 triangles:
472 of these are on 231 tristrips (2.04329 average tris per strip).
61011 of these are independent triangles.
7 textures, estimated minimum 2079K texture memory required.
the intel gma chip has a few major drawbacks.
one of them is, all vertex data is processed in software.
the more vertices in your scene, the more work has to be done by your cpu. 50k vertices still sound somewhat reasonable.
if you can, try to reduce the ammount of vertices.
another thing would be the fillrate. that’s rather easy to test. if you reduce your window size and the framerate increases a lot. you propably want to tacle it.
i’d guess on the vertex-part simply because that always turned out to be my bottleneck for GMA chips