OpenGL 核心配置文件在 OS X 上令人难以置信的减速 [英] OpenGL core profile incredible slowdown on OS X
问题描述
我在我的引擎中添加了一个新的 GL 渲染器,它使用了核心配置文件.虽然它在 Windows 和/或 nvidia 卡 上运行良好,但在 OS X 上却慢了 10 倍(3 fps 而不是 30).奇怪的是,我的兼容性配置文件渲染器运行良好.
I added a new GL renderer to my engine, which uses the core profile. While it runs fine on Windows and/or nvidia cards, it is like 10 times slower on OS X (3 fps instead of 30). The weird thing is, that my compatibility profile renderer runs fine.
我使用 Instruments 和 GL 分析器收集了一些跟踪:
I collected some traces with Instruments and the GL profiler:
https://www.dropbox.com/sh/311fg9wu0zrarzm/31CGvUcf2q
它表明应用程序将其时间花在 glDrawRangeElements 上.我尝试了以下几点:
It shows that the application spends its time in glDrawRangeElements. I tried the following things:
- 改用 glDrawElements(无效)
- 翻转剔除(对速度没有影响)
- 禁用一些 GL_DYNAMIC_DRAW 缓冲区(无效)
- 绘制时在 VAO 之后绑定索引缓冲区(无效)
- 将索引转换为 4 字节(无效)
- 使用 GL_BGRA 纹理(无效果)
我没有尝试将我的顶点与 16 字节边界对齐 和/或将索引转换为 4 字节,但说真的,如果那会是问题,那为什么他妈的会执行标准允许吗?
What I didn't try is to align my vertices to 16 byte boundary and/or convert indices to 4 byte, but seriously, if that would be the issue then why the hell does the standard allow it?
我正在创建这样的上下文:
I'm creating the context like this:
NSOpenGLPixelFormatAttribute attributes[] =
{
NSOpenGLPFAColorSize, 24,
NSOpenGLPFAAlphaSize, 8,
NSOpenGLPFADepthSize, 24,
NSOpenGLPFAStencilSize, 8,
NSOpenGLPFADoubleBuffer,
NSOpenGLPFAAccelerated,
NSOpenGLPFANoRecovery,
NSOpenGLPFAOpenGLProfile, NSOpenGLProfileVersion3_2Core,
0
};
NSOpenGLPixelFormat* format = [[NSOpenGLPixelFormat alloc] initWithAttributes:attributes];
NSOpenGLContext* context = [[NSOpenGLContext alloc] initWithFormat:format shareContext:nil];
[self.view setOpenGLContext:context];
[context makeCurrentContext];
尝试了以下规格:
- radeon 6630M,OS X 10.7.5
- radeon 6750M,OS X 10.7.5
- geforce GT 330M,OS X 10.8.3
你知道我可能做错了什么吗?同样,它适用于兼容性配置文件(尽管不使用 VAO).
Do you have any ideas what I might do wrong? Again, it works fine with the compatibility profile (not using VAOs though).
更新:向 Apple 报告.
UPDATE: reported to Apple.
更新:Apple 并不在乎这个问题……无论如何,我创建了一个实际上很好的小型测试程序.现在我用Instruments对比了调用栈,发现在使用引擎时,glDrawRangeElements做了两次调用:
UPDATE: Apple doesn't give a damn to the problem...anyway I created a small test program which is actually good. Now I compared the call stack with Instruments, and found out that when using the engine, glDrawRangeElements does two calls:
- gleDrawArraysOrElements_ExecCore
- gleDrawArraysOrElements_Entries_Body
而在测试程序中它只调用第二个.现在,第一个调用执行类似于立即模式渲染的操作(gleFlushPrimitivesTCLFunc、gleRunVertexSubmitterImmediate),因此很明显会导致速度变慢.
while in the test program it calls only the second. Now the first call does something like an immediate mode render (gleFlushPrimitivesTCLFunc, gleRunVertexSubmitterImmediate), so obviously casues the slowdown.
推荐答案
最后,我能够重现减速.这太疯狂了……这显然是由在my_Position"属性上调用 glBindAttribLocation 引起的.现在我做了一些测试:
Finally, I was able to reproduce the slowdown. This is just crazy... It is clearly caused by glBindAttribLocation being called on the "my_Position" attribute. Now I did some testing:
- 1 是默认值(由 glGetAttribLocation 返回)
- 如果我将它设置为零,那就没问题
- 如果我将它设置为 1,渲染会变慢
- 如果我将它设置为更大的数字,它又会变慢
显然我重新链接了程序(检查代码).这在实现中不是问题,我也用正常"值对其进行了测试.
Obviously I relink the program (check code). It is not a problem in the implementation, I tested it with "normal" values too.
测试程序:
https://www.dropbox.com/s/dgg48g1fwgyc5h0/SLOWDOWN_REPRO.zip
如何复制:
- 使用 XCode 打开
- 打开 common/gleext.h(不要被名字打扰)
- 将 GLDECLUSAGE_POSITION 常量从 0 修改为 1
- 编译并运行 => 慢
- 变回零 => 好
这篇关于OpenGL 核心配置文件在 OS X 上令人难以置信的减速的文章就介绍到这了,希望我们推荐的答案对大家有所帮助,也希望大家多多支持IT屋!