| commit | a65fe6e73c0ff06fa1bcfa01b218222136fff240 | |
| author | Mark Abraham <mark.j.abraham@gmail.com> | |
| Sun, 29 Dec 2019 23:08:15 +0000 (30 00:08 +0100) | ||
| committer | Artem Zhmurov <zhmurov@gmail.com> | |
| Mon, 30 Dec 2019 11:06:33 +0000 (30 12:06 +0100) | ||
| tree | 397acb74738f4b31f848649fbb7a5f0ff775eb9c | treesnapshot (tar.gz zip) |
| parent | 517554ee2bc6d5d746898d26c620388f7554c074 | commitdiff |
Remove use of tuple in LINCS code for GPU implementation
GCC 6.4 for POWER9 couldn't compile any form of increasing the size of
a vector of tuples that I could think of, so got rid of the tuple.
Used some range-based for, rather than explicit indexing plus the (slow)
at() function. There's no need to check indices are in range if there
are no indices!
Change-Id: Ie36a7a6fafe8f6613ec449778cfe2fbf97e02007
GCC 6.4 for POWER9 couldn't compile any form of increasing the size of
a vector of tuples that I could think of, so got rid of the tuple.
Used some range-based for, rather than explicit indexing plus the (slow)
at() function. There's no need to check indices are in range if there
are no indices!
Change-Id: Ie36a7a6fafe8f6613ec449778cfe2fbf97e02007
| src/gromacs/mdlib/lincs_cuda.cu | diffblobblamehistory |