Difference between revisions of "WG211/M23Scholz"
From WG 2.11
(Created page with "== Stream it if you can - not just when your GPU runs out of memory == This talk presents our work on enabling out-of-core computations on GPUs, i.e., computations on arrays...") |
(→Stream it if you can - not just when your GPU runs out of memory) |
||
| Line 1: | Line 1: | ||
| − | == Stream it if you can - not just when your GPU runs out of memory == | + | == Stream it if you can - not just when your GPU runs out of memory (Sven-Bodo Scholz) == |
This talk presents our work on enabling out-of-core computations on GPUs, i.e., computations | This talk presents our work on enabling out-of-core computations on GPUs, i.e., computations | ||
Latest revision as of 16:35, 24 March 2024
Stream it if you can - not just when your GPU runs out of memory (Sven-Bodo Scholz)
This talk presents our work on enabling out-of-core computations on GPUs, i.e., computations on arrays that are bigger than what can be held within the device memory of a given GPU. We propose code generation schemes for streaming map-like operations as well as stencil computations across GPUs. We also provide a performance evaluation of these two schemes. As it turns out, computations of all sizes benefit from the streaming approach.
