CUDA Highlighting settings

1, will c:/documents and Settings/all users/application data/nvidia corporation/nvidia CUDA sdk/doc/syntax_highlighting/ Copy the Usertype.dat file from the visual_studio_8 to the C:/Program files/microsoft Visual Studio 8/common7/ide directory (appended to the original if it already exists).

2. Open the Vs2005-> tools (T), Options (O), Projects and Solutions->vc++ Project Settings->c/c++ file extension (add *.cu later), include the extension (after which *.CU is added). Other default on the line.

3, or under Tools (T), open the text editor, file extension, in the right-hand extension column, fill in the. Cu, select Microsoft Visual C + + in the editor, and select the following (map files without extensions to (M)), The corresponding follow-up is Microsoft Visual C + +.

4. Open the registry and add sub-keys under Hkey_local_machine/software/microsoft/visualstudio/8.0/languages/file extensions/. cu then copy. The key value of the CPP to. cu. This will indicate that CU is also a project file for VC under vs.

5. Open the Registration form

Hkey_current_user/software/whole tomato/visual Assist x/vanet8 Add a key value in the Extsource key. Cu;.cuh.

6. Open the Visual Assist property, add Cuda's header file directory under projects C + + directories custom, in order to be in visual Assist When generating a rule, find the specific definition of CUDA itself to generate a visual assist keyword, such as __global__.

In this step, platform select Custom,show directiories for select stable include files

7, for this highlight, I do not know the above configuration is right, I am also tiger. Previously configured the environment, with the Open-Yong Wizard, version is cuda_vs_wizard1.2, highlighting is not very good, this time to use Cuda_vs_wizard_w32.2.0.zip, do not know whether the two have any relationship, not quite understand. Thank you for your courage here.

8, then there is in the fifth step, when adding a key value more than before. Cuh, this does not know and the success of the highlight has no relationship, did not try.

9, because the previous highlight failed, so this time again the same as the previous reset, changed the place in the 7th and 8 steps have been explained. Steps 2nd and 3rd have also been reset. The success of this set, let me very fruitful, so I put these feelings written down, in order to give themselves later set up a blueprint.

http://blog.ipattern.org

http://blog.ipattern.org/archives/281-C + + simulation of CUDA structure alignment

http://blog.ipattern.org/archives/283 "Cuda template class do not overload operator="

http://blog.ipattern.org/archives/303 "MSVC solution to the problem of not using STL vectors for aligned structures"

Notebook updated, finally the Cuda development environment to match up
Long wanted to study cuda, unfortunately, the graphics card is not good, this time the notebook video card replaced by the Nvidia GeForce 8400M GS, finally can play cuda.
Installation is still quite smooth, first install Cuda Toolkit 2.0, then install Cuda SDK 2.02, and then install Nvidia graphics driver 178.08, of course, visual Studio2005 has been installed, so the basic development environment is done.
From the SDK to find a cuda example project, compile and run through.
Then try the first example of "Memory copy" on the Cuda Chinese station, but how to create a Cuda project in Visual Studio2005. Google also help, found the Cuda VS2005 Wizard This tool, can automatically create a Cuda project, is really convenient, the example is immediately compiled through, thanks to Zhao Kaiyong. I don't know when to let NVCC run under the command line.
Cuda source code file name defaults to Cu,visual studio in the absence of syntax highlighting, you can in the "tools → options → text editor → file extension", the CU will be added.

Http://www.nvmod.com/forum-14-1.html