[Rd] Floating point issue

[Dirk Eddelbuettel]

On 10 July 2022 at 17:43, Antoine Fabri wrote:
| He is my session info :
| 
| R version 4.1.3 (2022-03-10)
| Platform: aarch64-apple-darwin20 (64-bit)
| Running under: macOS Monterey 12.0.1

Along with the different session infos posted by numerous people who are
unable to replicate your issue (myself included) and a bit more off-line
thinking about I started to suspect you may have an M1 chip. Which this
confirms.

Tomas and Simon had a deep dive into possible floating point concerns in

  https://blog.r-project.org/2020/11/02/will-r-work-on-apple-silicon/index.html

I also seem to recall that it was also said somewhere (though apparently not
in that post) that the M1 does not have long double. Post such as this one

  https://developer.apple.com/forums/thread/673482

seem to suggest that 80 bit accuracy is Intel specific -- so this may in fact
be an issue of sailing too close to the hardware boundary.

Dirk

-- 
dirk.eddelbuettel.com | @eddelbuettel | edd using debian.org