Text archives Help

---

Cool, it looks like in a few places folks have just written double when they meant (or didn't mean) to write Real...

I'll fix when I come across it if necessary.... maybe if there is a computation that has to be in double for some reason, it could be tested in both cases and commented so no one else changes it..

Abe

Steven G. Parker wrote:

On May 23, 2005, at 3:28 PM, Abe Stephens wrote:

I thought that the idea of having a Real typedef was that it could be a global switch to change computation from being double precision to single precision etc..

Yes, this was the plan.

This seems to work in most parts of the code, however there is a lot of code, especially computation involving scalars and vectors where this isn't the case (i.e. the code only compiles if Real==double). Is this the desired behavior?

It is not the desired behavior, but the typedef Real is relatively new, so it hasn't been propogated everywhere.


Steve

---

• [MANTA] double & float & Real, Abe Stephens, 05/23/2005
  • Re: [MANTA] double & float & Real, Steven G. Parker, 05/23/2005
    • Re: [MANTA] double & float & Real, Abe Stephens, 05/23/2005

---

Archive powered by MHonArc 2.6.16.