Talk:Bit-level parallelism

This article is rated Stub-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects:

Computing This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing ??? This article has not yet received a rating on the project's importance scale.

It would be good if someone who understands this topic well -- and I do not -- would explain why bit-level parallelism has run out of steam. Why not 128-bit or 256-bit processors? Is it just that the benefits of bit-level parallelism apply mainly to arithmetic, and 32- or 64-bit word sizes comfortably accommodate numbers within the range of most calculations? Or is there another reason?

Eodell (talk) 03:48, 18 March 2009 (UTC)Reply

For general purpose computing, perhaps the biggest driver to 64-bit address/integer sizes is RAM & disk sizes breaking the 4 GiB barrier.

A 32 bit address can only go up to 2^32, 2 to the 32nd power or 4,294,967,296, so when our memory/storage approached that scale (or less for esoteric reasons), we needed larger numbers. Since this was a broad issue, the most effective way to solve this was to go to 64 bit address/integers in the processors' cores.

Now 64 bits addresses 2^64 integers (or bytes in memory)--18,446,744,073,709,551,616. As things are going that should be enough for many decades. Assuming we can even figure out what to do with that (we probably will).

Outside of values in computing, there are already relatively few places where we need to deal with with integers larger than 4 billion (2^32) so for most purposes 32 bits was enough and 64 bits is already overkill. We won't need to go beyond 18 quintillion to count things.

There are also very real costs to doubling word size: all of the places using them in memory double as do others, the processors become more complex, and certain very frequently used operations/algorithms take longer.

Hmmm. Maybe some of this belongs in the article. src (talk) 00:52, 13 January 2025 (UTC)Reply

@Eodell There are 128-bit and 256, even 512-bit processors, but those are mathmatical registers. General purpose registers which can be used to point to memory addresses are still not exceeding 64-bit because we hardly have one tarabyte of memory for CPUs. Dannyniu (talk) 11:39, 15 June 2016 (UTC)Reply