Skip to Main Content U.S. Department of Energy
PNNL: High Performance Computing

Performance

Gordon Bell Finalist at SC09 - GA Crosses the Petaflop Barrier

  • NWChem petascale calculation
  • GA-based parallel implementation of coupled cluster calculation performed at 1.39 petaflops using over 223,000 processes on ORNL's Jaguar petaflop system. Apra et al. "Liquid water: obtaining the right answer for the right reasons"
  • Global Arrays is one of two programming models that have achieved this level of performance

Floating-Point
Performance
Courtesy: Edo Apra

Parallel Matrix Multiply

  • Cray XT4
  • Matrix size is 65536x65536

Parallel Matrix
Multiply

Stomp Speedup

Stomp Speedup

Site Links

Our Research

Past Research

Global Arrays Links

News and Events

Feedback

Contacts