unums as new numerical computation unit vs. floats

[Robert Herman]

Has anyone looked into, or familiar with unums?

Supposedly a new number format for computation to replace floats and their problems.

CRC book on Amazon: http://www.amazon.com/The-End-Error-Computing-Computational/dp/1482239868#

Slidecast:  http://insidehpc.com/2015/03/slidecast-john-gustafson-explains-energy-efficient-unum-computing/ 

Rob

[Robert Virding]

No I haven't seen them before. Question is if they will take off and make it into the processor. Maybe we could actually match against "floats" in a useful manner.

Robert

[Robert Herman]

They are in the exploratory phase with people all over writing implementations with small bits as proof of concept. I have no doubt that if they find no big hurdles, they will be implemented. I believe I already found mention that somebody has implemented a software unum processor in an FPGA.

I think his point of even if the software implementation of unums, akin to software FPU libs, is the same or even a bit slower, it at least puts you in a bounded region of your answer, unlike floats that can be way off when rounding errors and other ills occur.

Rob

[H Durer]

Sounds like they'll also want to use interval arithmetic.
https://en.m.wikipedia.org/wiki/Unum_(number_format)

--
You received this message because you are subscribed to the Google Groups "Lisp Flavoured Erlang" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lisp-flavoured-e...@googlegroups.com.
To post to this group, send email to lisp-flavo...@googlegroups.com.
Visit this group at https://groups.google.com/group/lisp-flavoured-erlang.
For more options, visit https://groups.google.com/d/optout.

[TR NS]

I read a good bit about it. It seems to have some merit but I think it is a bit more complex than it really needs to be, especially the 2.0 version.

[Robert Virding]

They seem very interesting but as LFE runs on top of the BEAM, the Erlang VM, it is not something we can really do anything about. It is out of our direct control.

Robert