Introduction to Blue Gene Seminar Topic:
It is a type of supercomputer developed by IBM and Lawrence Livermore National Laboratory. It has the speed of 360 teraflop peak of power computing. They are embedded with extremely powerful processors (100 GFLOPS). They contain two 32 bit POWER PC cores. This systems contains SDRAM-DDR memory, in which node has the maximum memory of 2 GB. The complete system contains 64 cabinets. It means the total memory of the systems has 16 TB of memory. This system support five different types of networks: Torus, Tree, Ethernet, JTAG, and global interrupts.
They basically use Linux Kernel. Blue gene doesn’t have any bios in their nodes. They have POSIX interface created by BLRTS. Messages are passed by point-to=point packet addressing mode. This functionally is provided by daemon called CIOD. They are control I/O daemon. They provide the function of job control. CIOD are used in two scenarios:
Console shell: it is mostly used for simulation and testing purpose. User has limited number of commands.
Job scheduler: it has a special interface that is used to implement the same protocol that is defined in CIOMAN and CIOD.
Blue gene is familiar in programming and they only use standards tools. They have two modes: Linux for I/O nodes and BLRTS for computer nodes. They use parallel programming languages like HPF, because they require regular computation.
For designing the blue gene hardware many types of applications are used like DOE’s NNSA laboratories. This program is the application for blue gene science program, they provide prototyping platform that used to research for application that is suitable for cellular architectures. It provide simulation environment so we don’t require a hardware for demonstrating the complete and functional software environment.
Download Seminar Topic and Report on Blue Gene .