TUFLOW Message 3018: Difference between revisions

Content deleted Content added

Revision as of 15:42, 2 March 2023 view source Emilie Nielsen (talk \| contribs) Administrators 6,249 edits No edit summary ← Older edit		Latest revision as of 14:32, 5 September 2023 view source Pavlina Monhartova (talk \| contribs) Bureaucrats, Administrators 5,384 edits No edit summary
(3 intermediate revisions by 2 users not shown)
Line 4: \|type=[[ERROR]] \|message_desc=A GPU memory request has failed. Although the initial memory ~~esitmate~~estimate was passed, as HPC requests ~~contguous~~continuously blocks of memory for fields it is possible that a request may fail as the model size approaches the available free memory on the GPU. \|suggestions= * Check that no other models are currently running on the selected GPU. * <u>[[TUFLOW_Simulation_Speed \| Reduce RAM requirements]]</u> for the model. * Reduce cell size. * Change to single precision if using double precision. * Run the model on the CPU. ~~\|prelink=[[TUFLOW_Message_3017\|Message 3017]]~~ \|uplink=[[3xxx_TUFLOW_Messages\|3xxx Messages]] ~~\|nextlink=[[TUFLOW_Message_3019\|Message 3019]]~~ }}