Zobrazeno 1 - 2
of 2
pro vyhledávání: '"Brett Neuman"'
Publikováno v:
The International Journal of High Performance Computing Applications. 34:519-531
Enhanced-precision global sums are key to reproducibility in exascale applications. We examine two classic summation algorithms and show that vectorized versions are fast, good and reproducible at exascale. Both 256-bit and 512-bit implementations sp
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::e14adc954beea84ad54382eaebc9398b
https://doi.org/10.1177/1094342020938425
https://doi.org/10.1177/1094342020938425
Autor:
Laura Monroe, Robert W. Robey, Christopher M. Mauney, Shane Fogerty, Brett Neuman, Vanessa Job, Terry Grove
Publikováno v:
CORRECTNESS@SC
Due to accumulated round-off error, mathematically equivalent floating-point summations can yield different computational results. Errors propagated across timesteps can be substantial and can lead to significant inaccuracy in the final results. We f
Externí odkaz:
https://explore.openaire.eu/search/publication?articleId=doi_________::72d41b562f6201fa0aba40e4e9de1e33
https://doi.org/10.1109/correctness51934.2020.00007
https://doi.org/10.1109/correctness51934.2020.00007