Ninth INFN International School on: "Architectures, tools and methodologies for developing efficient large scale scientific computing applications" ESC17 - Bertinoro (Forlì-Cesena) Italy 22-28 October 2017

Session

Session 4

26 Oct 2017, 08:30

Bertinoro

32. Efficient floating-point computation and vectorization

Vincenzo Innocente (CERN)

26/10/2017, 08:30

33. Efficient floating-point computation and vectorization

Vincenzo Innocente (CERN)

26/10/2017, 09:15

34. Efficient floating-point computation and vectorization

Vincenzo Innocente (CERN)

26/10/2017, 10:30

35. Consolidation

26/10/2017, 11:15

36. GPUs and the Heterogeneous programming problem

Dr Tim Mattson (Intel)

26/10/2017, 12:15

37. GPU programming with OpenCL: Ideas and the host program

Dr Tim Mattson (Intel)

26/10/2017, 14:30

38. Consolidation

26/10/2017, 15:15

39. Consolidation

26/10/2017, 16:45

57. From HPC to Deep Learning, a short journey on modern techniques

Piero Altoè (NVIDIA)

26/10/2017, 18:00

Software design of legacy high-performance computing (HPC) applications is heavily influenced by the frequent changes on the architectures. The classic CPU-based approach with communication libraries is not sufficient to scale beyond the limit of Amdahl law and Dennard scaling. Silicon manufacturers design alternative architectures and programming models to overcome this limitation, leading to...