COST IC0805

COST IC0805

European Cooperation in Science and Technology

The aim of the project is the adjustment of particular applications and taking full advantage of the computing power of hardware accelerators, GPU in particular, more often installed in supercomputing hybrid systems. Contemporary graphic processors are multicore architectures installed along with multicore systems of up to a few Tflops per computing node.

 

Advance graphic processors facilitate creation of a few thousands threads, which significantly increases performance of parallel codes. The available programming standards, like OpenCL and CUDA, allow using GPU for general-purpose computation (GPGPU). Even though graphic cards offer relatively cheap solutions which expand computational power, the use of GPU is difficult due to the necessity to adjust algorithms to new computational architectures.

Scheduled start date: 2009-05-07
Scheduled finish date: 2013-06-05
Status: completed
 

Contact:
Krzysztof Kurowski
+48618582072
Calendar
<< November 2017 >>
Mo
Tu
We
Th
Fr
Sa
Su
  
  
  1
  2
  3
  4
  5
  6
  9
11
18
19
20
21
22
23
24
25
26
27
28
29
30
  
  
  
Tenders