
Elusien
Ex IT manager and programmer [mainly C, PERL(/Tk), Javascript and PHP, HTML and CSS also Fortran, Algol & PL/1] and systems analyst [mainly Unix and some ancient O/Ss - COS, NOS/BE, OS/360], keen on scuba-diving and golf.
I’ve mainly worked as a systems analyst on supercomputers: Cray-1A to C90 from Cray Research inc, Vector Parallel Processors from Fujitsu and SMT microprocessor systems linked via Infiniband from IBM.
The last system I was responsible for comprised 2 identical supercomputers, each with around 8,700 2-way symmetric multi-tasking processors, each one probably more powerful than the PC you are using to read this. The whole system was capable of a sustained performance of about 20 Teraflops (320 Teraflops peak performance - 320 million Megaflops [Megaflops = one million floating-point operations (calculations) per second] - the first true supercomputer (the Cray-1) had a peak performance of 80 Megaflops).