Open jfilhoGN opened 7 years ago
João,
Thanks for your interest in cuda-edu. I can only provide limited support, but I will look into your issue soon (hopefully this evening). What version of ubuntu are you using? What version of libclang?
On Tue, Apr 11, 2017 at 1:00 PM, João Martins Filho < notifications@github.com> wrote:
Hello my name is João Martins, Computer Science student at the Federal Technological University of Parana - Brazil
We are working with Heterogeneous Computing and would like to be using its cuda-edu platform for a mini-course with university students.
Following your step by step on the Github Wiki https://github.com/sean-dougherty/cuda-edu/wiki/Getting-Started-on-Linux, we found an error in the Makefile (sending in attachment), you already Did you have this error?
[image: error_in_makefile] https://cloud.githubusercontent.com/assets/12972663/24928226/1e7e6eba-1ed8-11e7-98ad-b296a2ecce6f.png
Thanks! João Martins
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/sean-dougherty/cuda-edu/issues/8, or mute the thread https://github.com/notifications/unsubscribe-auth/AGJU8cnsF95IqQEgSl0TCeJigRPkcK5Mks5ru9vDgaJpZM4M6k6- .
Thanks for the help @sean-dougherty. The ubuntu version is 16.04 Libclang version 3.8
Can you please tell me which compiler you're using and what version?
I am not seeing your problem, but I am using an Ubuntu 14.04 machine. I can try it later on a newer version of ubuntu.
On Tue, Apr 11, 2017 at 1:15 PM, João Martins Filho < notifications@github.com> wrote:
Thanks for the help @sean-dougherty https://github.com/sean-dougherty. The ubuntu version is 16.04 Libclang version 3.8
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/sean-dougherty/cuda-edu/issues/8#issuecomment-293386973, or mute the thread https://github.com/notifications/unsubscribe-auth/AGJU8f9U98iJdGEhvjqe_P20WkNiLodtks5ru99XgaJpZM4M6k6- .
gcc version 4.7.4 make version 4.1 g++ version 4.7.4 clang version 3.8.0 llvm version 3.8
@sean-dougherty would like an info. We will be doing a mini-course at the GPU-Cuda university, but we only have one machine with a video card (GeForce GTX 780), can we perform multiple runs on this machine through cuda-edu? (We are looking for the same style as WebGPU)
cuda-edu requires a fully c++11 compliant compiler, which means g++ version 4.8 or higher. It's very possible this is why you're getting your compile error. It would be okay for you to force clang to be used. If you edit Makefile.conf, at line 3 you can edit it to be:
CXX=/usr/bin/clang++ Unfortunately, when I run on my 16.04 machine, I get all kinds of include errors when libclang is running in cuda mode (-x cuda). This is when using llvm 3.8. It worked fine for me on my 14.04 machine, which uses llvm 3.4.
can we perform multiple runs on this machine through cuda-edu?
No, cuda-edu does no such thing, since that was the role of WebGPU. How many students will there be?
No, cuda-edu does no such thing, since that was the role of WebGPU. How many students will there be?
There will be between 10 and 20 students
I will force clang ++ to version 4.8 here and I'll tell you if it worked. Have you worked with WebGPU?
It worked @sean-dougherty, really the bug was in the g ++ version, I put it in 4.8 and Make compiled successfully.
Have you worked with WebGPU?
Not as an administrator, no. You'll need to talk to the people who maintain that project if you need help.
On Tue, Apr 11, 2017 at 7:58 PM, João Martins Filho < notifications@github.com> wrote:
It worked @sean-dougherty https://github.com/sean-dougherty, really the bug was in the g ++ version, I put it in 4.8 and Make compiled successfully.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/sean-dougherty/cuda-edu/issues/8#issuecomment-293458536, or mute the thread https://github.com/notifications/unsubscribe-auth/AGJU8b-sdwzc1kKR7uhyoUGeRIkWNgP_ks5rvD3mgaJpZM4M6k6- .
Okay @sean-dougherty. I will be contacting them, thank you very much for the help, I will be presenting to my advisor about cuda-edu.
I know students found cuda-edu to be very helpful, but you should understand before adopting it for a course that I'm the sole author and I'm technically forbidden from even working on it by my employer. That is, you'll get very limited support from me.
On Wed, Apr 12, 2017 at 10:12 AM, João Martins Filho < notifications@github.com> wrote:
Okay @sean-dougherty https://github.com/sean-dougherty. I will be contacting them, thank you very much for the help, I will be presenting to my advisor about cuda-edu.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/sean-dougherty/cuda-edu/issues/8#issuecomment-293646386, or mute the thread https://github.com/notifications/unsubscribe-auth/AGJU8cxvg2RDtk02TTdT0vTWPbskBbVBks5rvQX_gaJpZM4M6k6- .
Thanks a lot for the help @sean-dougherty. I talked to my teacher and at first we will not be using cuda-edu, the idea is to try to use WebGPU, but I want to thank you for your commitment to help me set up cuda-edu, in the future we may be using your project.
Hello my name is João Martins, Computer Science student at the Federal Technological University of Parana - Brazil
We are working with Heterogeneous Computing and would like to be using its cuda-edu platform for a mini-course with university students.
Following your step by step on the Github Wiki, we found an error in the Makefile (sending in attachment), you already Did you have this error?
Thanks! João Martins