There are currently no mainstream general-purpose processors built to operate on 256-bit integers or addresses, though a number of processors do operate on 256-bit data. CPUs feature SIMD instruction sets (Advanced Vector Extensions and the FMA instruction set etc.) where 256-bit vector registers are used to store several smaller numbers, such as eight 32-bit floating-point numbers, and a single instruction can operate on all these values in parallel. However, these processors do not operate on individual numbers that are 256 binary digits in length, only their registers have the size of 256-bits. Binary digits are found together in 128-bit collections.
Modern GPU chips move data across a 256-bit memory bus.
256-bit processors could be used for addressing directly up to 2256 bytes. Already 2128 (128-bit) would greatly exceed the total data stored on Earth as of 2010, which has been estimated to be around 1.2 zettabytes (over 270 bytes).
The Efficeon processor was Transmeta's second-generation 256-bit VLIW design which employed a software engine to convert code written for x86 processors to the native instruction set of the chip.
Researchers at the University of Cambridge use a 256-bit capability pointer, which includes capability and addressing information, on their CHERI capability system.
The DARPA funded Data-Intensive Architecture (DIVA) system incorporated processor-in-memory (PIM) 5-stage pipelined 256-bit datapath, complete with register file and ALU blocks in a "WideWord" processor in 2002.
This page is based on a Wikipedia article written by authors
Text is available under the CC BY-SA 3.0 license; additional terms may apply.
Images, videos and audio are available under their respective licenses.