锋芒拍摄关于风景、人像、纪实、建筑、星野、深空天体的摄影作品都会在这里展示给大家欣赏!

PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST

24 08月
作者: DARKARCHON |分类:天文资料 | 转载时间:2020年08月24日
【特别声明】 本文摘自网络,原文 作者/网站: DARKARCHON

PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST


This guide has been updated to work with the PixInsight 1.8.8-6 version, which has StarNet++ included. It will not work with the old StarNet++ version anymore. If you did this guide before you will need to follow it thoroughly again because all dependencies are updated.

(本指南已经过更新,可以与包含StarNet++的PixInsight 1.8.8-6版本配合使用。它将不再适用于旧的StarNet++版本。如果您在此之前执行了此指南,则需要再次彻底遵循它,因为所有依赖项都已更新。)


If you are reading this you have found this post most likely by a Google search or have been linked to it. What is it for? Well if you are here, you are looking for a way to speed up Starnet++. And there is one, by utilizing the GPU in your system to bring the speed up. We’re talking a factor of at least 5x as fast as before. On my system (Ryzen 7 2700x, GeForce 2080Ti) it went from 3 minutes 45 seconds on my CPU to 25 seconds via CUDA!

(如果你正在阅读这篇文章,你很可能在谷歌搜索中找到了这篇文章,或者已经链接到了这篇文章。它是用来干什么的?好吧,如果你在这里,你正在寻找一种方法来加速Starnet++。还有一种,利用你系统中的GPU来提高速度。我们说的因素至少是以前的5倍。在我的系统(Ryzen 7 2700x,GeForce 2080Ti)上,它从CPU上的3分45秒通过CUDA变成了25秒!)


There is a caveat: this method only works on 64bit Windows and only with nVidia GPUs. It does not require PixInsight but the tutorial will focus on getting it to run within the application. Got those prerequisites? Then let’s continue.

(有一个警告:这个方法只适用于64位Windows,并且只适用于nVidia GPUs。它不需要PixInsight,但本教程的重点是让它在应用程序中运行。有那些先决条件吗?那我们继续吧。)


Note: this only works with a NVIDIA GPU card with CUDA architectures 3.5, 3.7, 5.2, 6.0, 6.1, 7.0 and higher than 7.0. See the list of CUDA-enabled GPU cards.

(注意:这只适用于具有CUDA架构3.5、3.7、5.2、6.0、6.1、7.0和高于7.0的NVIDIA GPU卡。请参阅启用CUDA的GPU卡列表。)


Gathering supplies (收集补给)


(Prerequisites and downloads)(必须条件和下载)

You will need to download several things to get this setup running.

(你需要下载下面三个软件来运行这个安装程序。)


  1、nVidia CUDA 10.1 (Update 2)

  2、nVidia cuDNN v7.6.5, for CUDA 10.1(需要注册后才能下载)

  3、libtensorflow-gpu 2.3.0




Bringing the system online(使系统联机状态)

(Setup)(设置)


Step 1: Replace the tensorflow.dll(步骤1:替换 tensorflow.dll 文件)


  1、Open the downloaded libtensorflow-gpu-windows-x86_64-2.3.0.zip

     (打开下载的libtensorflow-gpu-windows-x86_64-2.3.0.zip)

  2、Extract the tensorflow.dll from ‘lib’ folder to ‘C:\Program Files\PixInsight\bin’, overwriting existing files

     (在压缩文件里找到“lib”文件夹,提取 tensorflow.dll 到“C:\Program Files\PixInsight\bin”目录中,覆盖现有文件)


PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST


Step 2: Install CUDA步骤2:安装CUDA)


  1、Run the cuda_10.1.243_win10_network.exe

       (运行 cuda_10.1.243_win10_network.exe 文件)

  2、Select ‘Custom (Advanced)’

       (选择“自定义(高级)”)

PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST


 

 3、Deselect everything but CUDA -> Runtime -> Libraries

       (取消选择 除 CUDA -> Runtime -> Libraries 之外 的所有内容)

  4、Make sure to deselect the Demo Suite in Libraries too

       (请确保 取消选择 Libraries 库中的 Demo Suite 演示套件



PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST



  5、Press next until installation is done

       (按“下一步”直到安装完成)



Step 3: Install cuDNN(步骤3:安装 cuDNN)


  1、Open the downloaded cudnn-10.1-windows10-x64-v7.6.5.32.zip

       (打开下载的 cudnn-10.1-windows10-x64-v7.6.5.32.zip 文件)

  2、Extract the folder ‘bin’ from the included folder ‘cuda’ to ‘C:\Program Files\NVIDIA GPU Computing Toolit\CUDA\V10.1’

       (将压缩包中“cuda”文件夹中的“bin”文件夹 和 "lib" 文件夹,提取到“C:\Program Files\NVIDIA GPU Computing Toolit\cuda\V10.1”目录下)


PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST



Step 4: Edit Environment Variables(步骤4:编辑环境变量)


  1、You will need to set 2 environment variables in Windows so everything runs flawlessly

       (你需要在Windows中设置2个环境变量,这样一切都能完美运行)

  2、Open the start menu and search for ‘environment’, select ‘Edit the system environment variables’

       (打开“开始”菜单,搜索“环境”,选择“编辑系统环境变量”)

PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST


  3、In the window open ‘Environment Variables’, click on ‘New’ and enter “TF_FORCE_GPU_ALLOW_GROWTH” as name and “true” as value, Press OK to confirm

       (在窗口中打开“环境变量”,单击“新建”并输入“TF_FORCE_GPU_ALLOW_GROWTH”作为名称,输入“true”作为值,按OK确认)

PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST


  4、Look for the variable called ‘Path’, select it and click on ‘Edit’, if not present the folder “C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin” needs to be added by pressing ‘New’ and adding it

       (查找名为“Path”的变量,选择它并单击“编辑”,如果不存在文件夹“C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin”需要通过按“New”新建并添加信息)

PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST



  5、Press OK to close everything

       (按“确定”关闭所有内容)


Step 5: Verify everything works(验证一切正常)


  1、Open the Windows Task Manager, go to the ‘Performance’ tab and select GPU 0

       (打开Windows任务管理器,转到“性能”选项卡并选择GPU 0)

  2、In one of the Drop Downs select ‘Cuda’ as performance metric

       (在其中一个下拉列表中,选择“Cuda”作为性能指)

  3、Note: you might not have ‘Cuda’ as performance metric. This is ok. However then you will just see the GPU usage rise.

       (注意:您可能没有使用“Cuda”作为性能度量。没关系。然而,你会看到GPU使用率的上升。)

PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST


  4、Open PixInsight and load in one of your images, ideally one that is non-linear (or stretch it quickly)

       (打开PixInsight并加载其中一个图像,最好是非线性的(或快速拉伸))

  5、Run the StarNet process from PROCESS -> <All Processes> -> StarNet

       (从进程-><All Processes>->StarNet运行StarNet进程)

  6、Edit it to include the path to the downloaded weights in C:\Program Files\PixInsight\library\

       (编辑它以将下载的权重的路径包含在C:\Program Files\PixInsight\library中\)

  7、If you do not have the files there, make sure to update your PixInsight through Resources -> Updates -> Check for Updates

       (如果没有文件,请确保通过“资源”->“更新”->“检查更新”更新PixInsight)

PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST


  8、Apply the default Process on the image

       (对图像应用默认进程)

  9、Watch the Task Manager GPU or CUDA utilization, it should spike up shortly after StarNet process has begun

      (观察任务管理器的GPU或CUDA利用率,它应该在StarNet进程开始后很快上升)

PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST


  10、If it does not and your CPU spikes instead, you did something wrong. Verify that you did all steps (especially the Environment Variables)

        (如果它没有,而你的CPU峰值反而是你做错了什么。验证您是否完成了所有步骤(尤其是环境变量))



Gotta Go Fast(得快点)

(Benchmarks)(基准)


Not much I can say here, I only tested it on one image as of the time of writing. Nonetheless, those are the results for a 8.3MP image drizzled to 2x scale, StarNet running with Stride 128.

(在这里我不能说太多,我只在一张图片上测试了它。尽管如此,这是一张830万像素的图像的结果,这张图片的比例是2倍,星网以128步的速度运行。)


PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST

StarNet on Ryzen 7 2700x



PIXINSIGHT, STARNET++ AND CUDA – GOTTA GO FAST

StarNet on GeForce 2080Ti



Thanks for reading, enjoy the new sped up StarNet and have clear skies!

(感谢您的阅读,享受新的加速星网和晴朗的天空!)

浏览32 评论0
返回
目录
返回
首页
标签:StarNet PixInsight CUDA nVidia GPUs
NEOWISE Comet Processing Notes(NEOWISE彗星的处理说明) 小米手机任意版本MIUI安装Google Play服务

发表评论

◎欢迎参与讨论,请在这里发表您的看法、交流您的观点。