改善精灵渲染性能？

Question

我正在寻找一种方法来提高PySDL2的性能，从而渲染出大量的精灵。贝娄是我的尝试，试图尽可能多地使用硬件渲染：

import sdl2.ext as sdl2ext, sdl2
import sys
from random import randint

def run():
    sdl2ext.init()
    disp_mode = sdl2.video.SDL_DisplayMode()
    sdl2.SDL_GetDesktopDisplayMode(0, disp_mode)
    window = sdl2ext.Window('pysdl2 benchmark', size=(disp_mode.w, disp_mode.h),
            flags=sdl2.SDL_WINDOW_SHOWN | sdl2.SDL_WINDOW_FULLSCREEN_DESKTOP)
    window_w, window_h = window.size
    window.show()

    renderer = sdl2ext.Renderer(window)
    factory = sdl2ext.SpriteFactory(sprite_type=sdl2ext.TEXTURE,
            renderer=renderer)
    spr_image = sdl2ext.image.load_image('big_sprite.png')
    sprites = [factory.from_surface(spr_image) for _ in xrange(100)]
    for spr in sprites:
        spr.position = randint(0,160), randint(0,90)
        spr.w, spr.h = spr.size
        spr.dx = randint(1,16); spr.dy = randint(1,16)
    renderer_spr = sdl2.ext.TextureSpriteRenderSystem(renderer)

    running = True
    while running:
        events = sdl2ext.get_events()
        for event in events:
            if event.type == sdl2.SDL_QUIT:
                running = False
                break
        renderer.clear(COLOR_CARMINE)    # instance of sdl2ext.Color
        renderer_spr.render(sprites)
        renderer.present()
        # Updating sprites positions
        for spr in sprites:
            if spr.x + spr.w + spr.dx > window_w or \
                    spr.x + spr.dx < 0:
                        spr.dx = -spr.dx
            if spr.y + spr.h + spr.dy > window_h or \
                    spr.y + spr.dy < 0:
                        spr.dy = -spr.dy
            spr.x += spr.dx
            spr.y += spr.dy
        #window.refresh()
    sdl2ext.quit()
    return 0

由于PySDL2的文档相当稀疏，我很好奇，如果我可能错过了一些重要的东西，而且编写的代码在某种程度上是不理想的。