如何在代码中强制Windows为Pygame游戏启用独显及优化字体渲染

一、代码实现强制使用独立GPU #

在Windows环境下，要让程序优先调用独立GPU，可通过以下代码层面的方式实现：

1. 设置SDL环境变量（推荐） #

在初始化pygame前，指定SDL使用Direct3D 11渲染后端——该后端通常由独立GPU提供硬件加速，触发系统分配独立GPU资源：

import os
# 初始化pygame前设置环境变量
os.environ['SDL_RENDER_DRIVER'] = 'd3d11'
import pygame
from pygame._sdl2.video import Window as SdlWindow, Renderer, Texture, Image
# ... 后续代码

2. 创建Renderer时指定硬件加速标志 #

创建Renderer对象时，添加pygame.SDL_RENDERER_ACCELERATED标志，强制启用硬件加速，确保程序优先使用独立GPU：

renderer = Renderer(screen, flags=pygame.SDL_RENDERER_ACCELERATED)

部分笔记本的NVIDIA Optimus或AMD可切换显卡会自动识别硬件加速程序，但上述代码可主动强制程序绑定独立GPU，避免系统默认分配集显。

二、字体渲染优化 #

你当前每帧都通过font.render()生成Surface再转Texture，会产生大量无意义的CPU-GPU数据拷贝和对象创建，优化核心是缓存Texture，仅在内容变化时更新：

优化后的完整代码 #

import pygame
from pygame._sdl2.video import Window as SdlWindow, Renderer, Texture, Image
import random
import math
import os

# 强制使用d3d11渲染后端，触发独立GPU
os.environ['SDL_RENDER_DRIVER'] = 'd3d11'

pygame.init()

screen_size = [900, 900]
screen = SdlWindow(title='Sdl2 Test', size=screen_size)
renderer = Renderer(screen, flags=pygame.SDL_RENDERER_ACCELERATED)

image = pygame.transform.scale(pygame.image.load('sdl2_image.png'), (75, 48))
image = Texture.from_surface(renderer, image)

font = pygame.font.SysFont('Arial', 50)

spriteGroup = pygame.sprite.Group()

clock = pygame.time.Clock()

# 字体渲染缓存变量
last_fps = -1
fps_texture = None
fps_rect = None

last_sprite_count = -1
count_texture = None
count_rect = None

while True:
    clock.tick(60)
    keys = pygame.key.get_pressed()
    mouse_x, mouse_y = pygame.mouse.get_pos()

    for event in pygame.event.get():
        if event.type == pygame.QUIT:
            pygame.quit()
            exit()

    renderer.clear()
    
    if len(spriteGroup) < 5000:
        for i in range(0, 8):
            rect = image.get_rect()
            rect.x = 450
            rect.y = 450
            
            sprite = pygame.sprite.Sprite(spriteGroup)
            sprite.rect = rect
            sprite.image = Image(image)
            sprite.angle = random.uniform(-math.pi, math.pi)
            sprite.x = sprite.rect.x
            sprite.y = sprite.rect.y

    for sprite in spriteGroup:
        sprite.x += math.cos(sprite.angle)
        sprite.y += math.sin(sprite.angle)

        if sprite.x < 0 or sprite.x > screen_size[0] - sprite.rect.size[0]:
            sprite.angle = math.pi - sprite.angle
        if sprite.y < 0 or sprite.y > screen_size[1] - sprite.rect.size[1]:
            sprite.angle = -sprite.angle
        
        sprite.image.angle += 1
        
        sprite.rect.x = int(sprite.x)
        sprite.rect.y = int(sprite.y)
    
    spriteGroup.draw(renderer)

    # 优化FPS渲染：仅数值变化时更新Texture
    current_fps = round(clock.get_fps())
    if current_fps != last_fps:
        last_fps = current_fps
        fps_surface = font.render(str(current_fps), True, (0, 0, 255))
        fps_texture = Texture.from_surface(renderer, fps_surface)
        fps_rect = fps_texture.get_rect()
    if fps_texture:
        fps_image = Image(fps_texture)
        fps_image.draw(fps_rect, fps_rect)

    # 优化精灵数量渲染：仅数值变化时更新Texture
    current_count = len(spriteGroup)
    if current_count != last_sprite_count:
        last_sprite_count = current_count
        count_surface = font.render(str(current_count), True, (0, 0, 255))
        count_texture = Texture.from_surface(renderer, count_surface)
        count_rect = count_texture.get_rect()
    if count_texture:
        count_image = Image(count_texture)
        draw_pos = (0, screen_size[1] - count_rect[3], count_rect[2], count_rect[3])
        count_image.draw(count_rect, draw_pos)

    renderer.present()

优化要点说明 #

• 缓存触发逻辑：记录上一次的FPS和精灵数量，仅当数值发生变化时，才重新执行Surface渲染和Texture转换操作。
• 减少数据传输：Surface转Texture涉及CPU到GPU的数据拷贝，频繁执行会严重消耗性能，缓存后仅在必要时执行该操作。
• 复用对象实例：避免每帧创建新的Texture和Image对象，减少内存碎片和对象销毁的额外开销。

内容的提问来源于stack exchange，提问作者IlIIIlIIlllI