How to improve the performance of the following code?

Question

glBegin(GL_TRIANGLES)
for face in faces:
   for vertex_i in face:
      glVertex3f(*vertices[vertex_i])
glEnd()

Is there any way to replace the double for loops with a more efficient function/method?

Answer 1

Yes, you can use the fixed function attributes and define an array of vertex data glVertexPointer / glEnableClientState. In this case you can draw the triangles by glDrawElements. indices has to be an array of the triangle indices and vertexarray and array of the coordinate components. For Instance:

Once at initilization:

vertexarray = [c for v in vertices for c in v]
indices = [i for f in faces for i in f]

Per frame:

glVertexPointer(3, GL_FLOAT, 0, vertexarray)
glEnableClientState(GL_VERTEX_ARRAY)
glDrawElements(GL_TRIANGLES, len(indices), GL_UNSIGNED_INT, indices)
glDisableClientState(GL_VERTEX_ARRAY)

---

Alternatively you can use numpy.array

import numpy

vertexarray = numpy.array(vertices, dtype=numpy.float)
indices = numpy.array(faces, dtype=numpy.uint32)

glVertexPointer(3, GL_FLOAT, 0, vertexarray)
glEnableClientState(GL_VERTEX_ARRAY)
glDrawElements(GL_TRIANGLES, indices.size, GL_UNSIGNED_INT, indices)
glDisableClientState(GL_VERTEX_ARRAY)