PureBasic Forums - English

Allow scientific notation numbers when 64, 80, 128 bit floats are allowed in PB.
Example:

var.f=0.436024071944526E-17

I hope that requests for 80 bit and 128 bit floats does not delay the introduction of 64 bit IEEE double floats.

I have done a lot of scientific computing over the years and have never found the need for more than 64 bits.

Don't forget the FPU only has 80 bit precision.

Can anyone explain why they need 80 or 128 bit floats?

geoff wrote:Can anyone explain why they need 80 or 128 bit floats?

Well, that's pretty simple. With 80 or 128bit floats you can map every long (and with that I mean the real longs, the one with 64bits ) to a float number. With only 64bit floats you can't do that.

Certainly a good feature to have. But the 64-bit floats first. Please. .