From: Jim Phillips (jim_at_ks.uiuc.edu)
Date: Sat Dec 05 2015 - 08:50:57 CST
Sorry, no change on the Phi port in 2.11. We're going to focus on
self-hosted KNL for upcoming machines at NERSC, Argonne, and TACC.
Jim
On Sat, 5 Dec 2015, Francesco Pietra wrote:
> Hello:
> May I ask whether the much increased speed of 2.11b1 CUDA with respect to
> namd 2.10 CUDA (I could enjoy of it on my personal GPUs computer) finds (or
> will find in a next beta version) a parallel with namd Phi? This is to plan
> our work because at our supercomputing center GPUs have been abandoned for
> computing nodes in favor of Phi (2 mic for a node of 16 cores)
>
> thanks
> francesco pietra
>
This archive was generated by hypermail 2.1.6 : Tue Dec 27 2016 - 23:21:37 CST