Bug hunt #78
Bug hunt #78
Conversation
… through Jenkins, but I'm not sure that it will also perform as expected. The code passes the testcases I've tried, but whether the physics of the LB_BOUNDARIES(_GPU) is correct I cannot vouch for. I've also changed a line of code that was erronously commented out by Dominic.
|
Thank you for doing this! I'm not sure, however, whether I want to actually merge it. As long as Jenkins shows red lights, Stefan knows that something is wrong, and apparently, there is more wrong than just compilation and test cases. Just ask Axel about that... Thanks, anyway! |
|
Sorry, there still appear to some problems. I don't quite understand why it On 21 March 2013 16:26, Olaf Lenz notifications@github.com wrote:
|
… lb_stokes_sphere_gpu.tcl test case still fails. I think there is a mysterious factor of 2 floating around somewhere in the code. I've added several comments to where I was uncertain about a piece of code, always with // ***** (lb_boundaries.c and lbgpu.cu). For the rest I think I've removed pretty much all the mess that was there.
…r for the random_tcl.c file, for which their implementation lead to unexpected and undesired behavior.
I've also put in the changes to the random_tcl.c file, where atol and atoi were not used correctly, and several other files where similar use was made of atol and atoi.
I have repaired almost all the errors that were plaguing the master over the past few days. There is still something wrong with the LB boundary forces for the GPU, but I can't figure out what it is. There are some other important bugs removed from the code as well. This includes a line in the lbgpu.cu that was falsely commented out.