Load store action in vulkan & ogles 的解决方案
metal的带宽之前的blog有讲
这篇主要是vulkan 和ogles的解决方案
https://www.khronos.org/registry/vulkan/specs/1.1-extensions/man/html/VkAttachmentDescription.html
typedef struct VkAttachmentDescription {
VkAttachmentDescriptionFlags flags;
VkFormat format;
VkSampleCountFlagBits samples;
VkAttachmentLoadOp loadOp;
VkAttachmentStoreOp storeOp;
VkAttachmentLoadOp stencilLoadOp;
VkAttachmentStoreOp stencilStoreOp;
VkImageLayout initialLayout;
VkImageLayout finalLayout;
} VkAttachmentDescription;
这样 load store action和programmable blending
vulkan上都有对应实现了
(vulkan还能单独操作stencil的load store。。。)
这样 这两步在unity上应该 都是跨平台支持的 不需要分平台支持了
下面就是metal2那些 feature 在vulkan上的实现。。。
这里八成是要分平台写 因为metal里面的那些关键字别的平台又没有对应。。
=======================================================================================
https://developer.arm.com/solutions/graphics/developer-guides/understanding-render-passes/how-render-passes-work
opengl下 上述方案对应的解决方案
ogles是隐式控制pass切换的
- The application calls
glBindFramebuffer()
to change theGL_FRAMEBUFFER
orGL_DRAW_FRAMEBUFFER
target. - The application calls
glFramebufferTexture*()
orglFramebufferRenderbuffer()
to change the attachments of the currently bound draw framebuffer object when the drawing is queued. - The application calls
eglSwapBuffers()
to signal the end of a frame. - The application calls
glFlush()
orglFinish()
to explicitly flush any queued rendering. - The application creates a
glFenceSync()
for some rendering in the current render pass and then callsglClientWaitSync()
to wait on the completion of that work, or an equivalent behavior with a query object
vulkan和metal是显式控制切换的
this means avoiding:
- Reading in older framebuffer values at the start of a render pass if they are going to be overdrawn.
- Writing out values at the end of each render pass which are transient and are only needed for the duration of that render pass.
这两步就是load store
mali gpu
避免tile memory和ddr之间的traffic
这部分的traffic都是带宽 带宽除此之外还包括texture sample
https://developer.arm.com/solutions/graphics/developer-guides/understanding-render-passes/efficient-render-passes
1.避免切出去再切回来
2.尽量合并 共用renderpass
OGLES:
Load
glClear()
glClearBuffer*()
glInvalidateFramebuffer()
Caution: Only the start of tile clear is free. Calling glClear()
or glClearBuffer*()
after the first draw call in a render pass is not free, and this results in a per-fragment clear shader.
For Vulkan, set the loadOp
for each attachment to either of:
VK_ATTACHMENT_LOAD_OP_CLEAR
VK_ATTACHMENT_LOAD_OP_DONT_CARE
Caution: If you call VkCmdClear*()
commands to clear an attachment, or manually use a shader to write a constant color, it results in a per-fragment clear shader. To benefit from the fast fixed-function tile initialization, it is much more efficient to use the render pass loadOp
operations.
不影响最终结果的情况下 mali里面invalidate operation 优于 a clear operation.
Store
用这个 glInvalidateFramebuffer
OpenGL ES, you can notify the driver that an attachment is transient by marking the content as invalid using a call to glInvalidateFramebuffer()
as the last draw call in the render pass.
Note: If you write applications using OpenGL ES 2.0, you must use glDiscardFramebufferExt()
from the [EXT_discard_framebuffer][EXT_dfb] extension.
depth能用这个么。。。GL_COLOR_ATTACHMENTi
, GL_DEPTH_ATTACHMENT
, GL_STENCIL_ATTACHMENT
, and/or GL_DEPTH_STENCIL_ATTACHMENT
....可以 处理depth
这样解决方案就ok了
For Vulkan, set the storeOp
for each transient attachment to VK_ATTACHMENT_STORE_OP_DONT_CARE
. For more efficiency, the application can even avoid allocating physical backing memory for transient attachments by allocating the backing memory using VK_MEMORY_PROPERTY_LAZILY_ALLOCATED_BIT
and constructing the VkImage
with VK_IMAGE_USAGE_TRANSIENT_ATTACHMENT_BIT
.
这块是metal里 memoryless在vulkan上的解决方案
depthstencil在vulkan metal上的建议是 别包一起 我们确实遇到这个问题了 就一个mask却把depth一起开了 这种状况是要避免的
-----------------
EXT_discard_framebuffer
这个unity有支持
===================
看了下unity代码 load store action它在ogles里也是支持的
那之后测下msaa下 是否生效就可以了
metal下msaa是不生效的 有bug要修
https://www.cnblogs.com/minggoddess/p/11447389.html
Load store action in vulkan & ogles 的解决方案的更多相关文章
- msaa mrt load store action unity
unity buildin renderpipeline 和lightweight rp 对于开了msaa的rt 的load store action设置失效 buildin的时候set render ...
- unity里blit的load store action设置
做blit的 load store action时 用 setrendertarget做 之后blit 参数用 BuiltinRenderTextureType.CurrentActive https ...
- TBDR下msaa 在metal vulkan和ogles的解决方案
https://developer.arm.com/solutions/graphics/developer-guides/understanding-render-passes/multi-samp ...
- vulkan load store and memoryless
https://www.jendrikillner.com/article_database/ https://community.arm.com/developer/tools-software/g ...
- Unordered load/store queue
A method and processor for providing full load/store queue functionality to an unordered load/store ...
- Load store and memoryless
metal https://developer.apple.com/library/archive/documentation/3DDrawing/Conceptual/MTLBestPractice ...
- 报错:Unable to load configuration. - action - file:/E:/apache-tomcat-8.0.37/webapps/20161102-struts2-3/WEB-INF/classes/struts.xml:11:73
第一种报错: 严重: Exception starting filter struts2Unable to load configuration. - action - file:/E:/apache ...
- 批量Load/Store指令的寻址方式
批量Load/Store指令用于实现在一组寄存器和一块连续的内存单元之间传输数据.也称为多寄存器寻址方式,即一条指令可以完成多个寄存器值的传送.这种寻址方式可以用一条指令最多完成传送16个通用寄存器的 ...
- 出现错误:Unable to load configuration. - action - file:/E:/Java/Tomcat7.0/apache-tomcat-7.0.68-windows-x64/apache-tomcat-7.0.68/webapps/SSH2Integrate/WEB-INF/classes/struts.xml:8:43
严重: Exception starting filter struts2 Unable to load configuration. - action - file:/E:/Java/Tomcat7 ...
随机推荐
- poj 2362:square
题目大意:给你T组数据,每组数据有n个棍子,问你能不能用这些棍子拼成一个正方形(所有都要用上,而且不能截断棍子). Sample Input 34 1 1 1 15 10 20 30 40 508 1 ...
- Java代码是怎么运行的
前言.... 作为一名 Java 程序员,你应该知道,Java 代码有很多种不同的运行方式.比如说可以在开发工具中运行,可以双击执行 jar 文件运行,也可以在命令行中运行,甚至可以在网页中运行.当然 ...
- CentOS7 卸载mysql(YUM源方式)
防止重装 yum方式 查看yum是否安装过mysql yum list installed mysql* 如或显示了列表,说明系统中有MySQL yum卸载 根据列表上的名字 yum remove ...
- nginx+uwsgi02---django部署(推荐)
参考 https://blog.csdn.net/weixin_39198406/article/details/79277580 https://www.cnblogs.com/alex3714/ ...
- ubuntu下java的安装与执行
一.安装java sudo add-apt-repository ppa:linuxuprising/java sudo apt-get update sudo apt-get install ora ...
- hdu 1068 最大子序列和变形,,,
#include<iostream> #include<cstdio> #include<cstring> #include<algorithm> #d ...
- python selenium5 模拟点击+拖动+按照指定相对坐标拖动 58同城验证码
#!/usr/bin/python # -*- coding: UTF-8 -*- # @Time : 2019年12月9日11:41:08 # @Author : shenghao/10347899 ...
- 使用cublas 矩阵库函数实现矩阵相乘
2014-08-10 cublas中执行矩阵乘法运算的函数 首先要注意的是cublas使用的是以列为主的存储方式,和c/c++中的以行为主的方式是不一样的.处理方法可参考下面的注释代码 // SOME ...
- Socket的神秘面纱
Tcp/IP协议是目前世界上使用最为广泛的协议,是以Tcp/IP为基础多个层次上的协议的集合.也称Tcp/IP协议族或Tcp/IP协议栈. TCP: Transmission Control Prot ...
- 使用Vue CLI脚手架搭建vue项目
本次是使用@vue/cli 3.11.0版本搭建的vue项目 1. 首先确保自己的电脑上的Node.js的版本是8.9版本或者以上 2. 全局安装vue/cli npm install @vue/cl ...