configure cuda autodetect #70
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
added 5 commits
January 19, 2013 19:30
…usly report cuda as working on systems that provide the tools to compile and link cuda code but do not provide the dynamically linked libraries to cuda binaries.
…ufft if the cuda checks failed. Properly fenced all the code depending on those libraries. Unfortunately the fences don't work because the precompiler constant CUDA gets set even when the checks fail.
…uda, including the run check, succeed. Since in that case, the cuda libraries are left out and all cuda code is fenced off, this should produce a working ESPResSo binary even on system where the cuda build tools are properly set up but the dynamic libraries are not properly referenced.
…broken cuda setups. Found and fixed two more bugs, one in the code fences and one in the cuda indicator for the makefile.
… files which contain declarations referencing cuda datatypes can also be processed by compilers other than nvcc.
ghost
assigned 
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Zweiter Anlauf. Ich hab den run check eingebaut. Damit das aber das Problem auch loest musste ich abhaengig von dessen Ergebnis die originale library list wiederherstellen. Das gab dann allerdings Probleme mit dem Code, da es einige stellen in der cuda_init.c gab die noch sauber geifdeft werden mussten.
Ich hab dann auch noch alles Plattformunabhaengig gemacht. Das einzig verdaechtige im Cuda Teil der configure.ac ist ein case statement. Allerdings sagen die autotool Seiten, dass es ok ist, so wie's da steht.
Am Ende hab ich seehr ausfuehrlich alle Kombinationen von configure-Einstellungen, Cuda-Toolkit, ld-Einstellungen und LB_GPU durchprobiert und jedesmal configure, make und Espresso getestet. Dabei habe ich herausgefunden, dass es eine Reihenfolge gibt in der man bei configure Cuda ausschalten kann und nachher trotzdem LB_GPU compilieren ohne, dass er merkt, dass Cuda fehlt. Das ging indem man an an der richtigen Stelle nicht make clean macht. Ein weiterer include in Dominiks code hat das aber behoben.