diff options
| author | Brecht Van Lommel <brechtvanlommel@gmail.com> | 2018-02-18 02:51:46 +0300 |
|---|---|---|
| committer | Brecht Van Lommel <brechtvanlommel@gmail.com> | 2018-02-18 02:53:03 +0300 |
| commit | 1dcd7db73d13443c59dd824abd9cacbf6bc88997 (patch) | |
| tree | 444b31fa05c90248653eb404a2d506bd4c08b49d /intern/cycles/kernel/svm/svm.h | |
| parent | 9e717c0495a3f9b71d3895d35df1e15518b6ca2c (diff) | |
Code cleanup: remove some more unused code after recent CUDA changes.
Diffstat (limited to 'intern/cycles/kernel/svm/svm.h')
| -rw-r--r-- | intern/cycles/kernel/svm/svm.h | 3 |
1 files changed, 1 insertions, 2 deletions
diff --git a/intern/cycles/kernel/svm/svm.h b/intern/cycles/kernel/svm/svm.h index a8f99d23b7d..fae9f783483 100644 --- a/intern/cycles/kernel/svm/svm.h +++ b/intern/cycles/kernel/svm/svm.h @@ -30,8 +30,7 @@ * in local memory on the GPU, as it would take too many register and indexes in * ways not known at compile time. This seems the only solution even though it * may be slow, with two positive factors. If the same shader is being executed, - * memory access will be coalesced, and on fermi cards, memory will actually be - * cached. + * memory access will be coalesced and cached. * * The result of shader execution will be a single closure. This means the * closure type, associated label, data and weight. Sampling from multiple |