12-10-2010, 09:41 AM
[attachment=5614]
Blue Gene Architecture
B.Tech Seminar report
by
Deepak M R
2K6711
Abstract
The Blue Gene project was originally aimed at exploring newer ideas in the
eld of supercomputing ,or more more speci
cally ,Massively Parallel Processing. The major motivations include making more e
cient use of computing clusters which tra- ditionally used SMP (Symmetric MultiProcessing) which involved multiple processors accessing the same shared memory areas ,which confronted with many problems in- herent with using shared resources like wasted processor power and synchronization problems. The main motive however was achieving operating speeds high enough to apply to computation intensive applications such as protein folding .This report pro- vides details about the hardware architecture of the blue gene system along with what has transpired in the last 10 years since its inception and its various classi
cations Introduction Blue Gene is an IBM Research project dedicated to exploring newer ideas in su- percomputing striving to provide a satndard in computer architecture and in the software required to program and control such massively parallel systems.The name "Blue Gene" comes from Blue ,which is te corporate colour of IBM and Gene from the original application for which the system was proposed to be built.The intentions of the original project was in developing our knowledge in applying high speed com- putation to such applications that require enormous processing powers or processing speeds that were unattained in those times.The project was initiated as an e
ort to- wards improving the maximum attainable processing powers, more speci
cally, speed in FLOPS(Floating Point Operations per Second).The early Blue Gene system was designed to provide a speed of 360 TFLOPS(Terra-FLOPS). But the latest Blue Gene projects claims to attain speeds in the PFLOPS range(Peta-FLOPS). Blue Gene has always been a major contributor in the High Performance Computing(HPC)
eld. Though originally initiated by IBM, Blue Gene now is a collabora- tive project among LLNL(Lawrence Livermore National Laboratory), United States Department of energy, Industrial Business Management group(Thomson J Watson Research centre and IBM Rochester), Argonne National Laboratory and academia.It won the national medal for technology and innovation from US president Barrack Obama in 2009.
Blue Gene Architecture
B.Tech Seminar report
by
Deepak M R
2K6711
Abstract
The Blue Gene project was originally aimed at exploring newer ideas in the
eld of supercomputing ,or more more speci
cally ,Massively Parallel Processing. The major motivations include making more e
cient use of computing clusters which tra- ditionally used SMP (Symmetric MultiProcessing) which involved multiple processors accessing the same shared memory areas ,which confronted with many problems in- herent with using shared resources like wasted processor power and synchronization problems. The main motive however was achieving operating speeds high enough to apply to computation intensive applications such as protein folding .This report pro- vides details about the hardware architecture of the blue gene system along with what has transpired in the last 10 years since its inception and its various classi
cations Introduction Blue Gene is an IBM Research project dedicated to exploring newer ideas in su- percomputing striving to provide a satndard in computer architecture and in the software required to program and control such massively parallel systems.The name "Blue Gene" comes from Blue ,which is te corporate colour of IBM and Gene from the original application for which the system was proposed to be built.The intentions of the original project was in developing our knowledge in applying high speed com- putation to such applications that require enormous processing powers or processing speeds that were unattained in those times.The project was initiated as an e
ort to- wards improving the maximum attainable processing powers, more speci
cally, speed in FLOPS(Floating Point Operations per Second).The early Blue Gene system was designed to provide a speed of 360 TFLOPS(Terra-FLOPS). But the latest Blue Gene projects claims to attain speeds in the PFLOPS range(Peta-FLOPS). Blue Gene has always been a major contributor in the High Performance Computing(HPC)
eld. Though originally initiated by IBM, Blue Gene now is a collabora- tive project among LLNL(Lawrence Livermore National Laboratory), United States Department of energy, Industrial Business Management group(Thomson J Watson Research centre and IBM Rochester), Argonne National Laboratory and academia.It won the national medal for technology and innovation from US president Barrack Obama in 2009.