Parallelism and Concurrent Programming

C omputing performance—typically measured in millions of instructions per second (MIPS) or floating-point operations per second (FLOPS)—has been improving at a staggeringly rapid and consistent rate over the past four decades. In the late 1970s, the Intel 8087 floating-point coprocessor could muster only about 50 kFLOPS ( https://www.w3.org/1998/Math/MathML" display="inline"> 5 × 10 4 FLOPS), while at roughly the same time a Cray-1 supercomputer the size of a large refrigerator could operate at a rate of roughly 160 MFLOPS ( https://www.w3.org/1998/Math/MathML" display="inline"> 1.6 × 10 8 FLOPS). Today, the CPU in game consoles like the PlayStation 5 or the Xbox Series X produces roughly 10-12 TFLOPS (10 13 FLOPS) of processing power, and the fastest supercomputer, currently the United States' Frontier supercomputer, has a LINPACK benchmark score of 1.102 EFLOPS (exa-FLOPS, or a staggering https://www.w3.org/1998/Math/MathML" display="inline"> 1.102 × 10 18 floating-point operations per second). This is an improvement of nine orders of magnitude for personal computing devices, and ten orders of magnitude for supercomputers.