giovanni-rosotti
commented
7 years ago iorig is an int. Is there any machine where an int is 64 bit?
On Tue, Jun 13, 2017 at 2:01 PM, Richard Booth notifications@github.com wrote:
Currently the code to read iorig in python assumes we are using a 32bit float, but this can now be double precision depending on the compilation options. I think this needs to be fixed at both ends.
In c++ we probably need to cast to an int of the right size before reinterpretting it as a float, and in python we need to detect the itemsize of the float and use that to generate the int. Otherwise we could implement an actual int type data transfer...
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/gandalfcode/gandalf/issues/153, or mute the thread https://github.com/notifications/unsubscribe-auth/ABdSguFIe-i6twLwc-WTEjyO1E-6OSJOks5sDnoKgaJpZM4N4Yzp .
rbooth200
Currently the code to read iorig in python assumes we are using a 32bit float, but this can now be double precision depending on the compilation options. I think this needs to be fixed at both ends.
In c++ we probably need to cast to an int of the right size before reinterpretting it as a float, and in python we need to detect the itemsize of the float and use that to generate the int. Otherwise we could implement an actual int type data transfer...