How to set individual bits in 64-bit double-precision floating number

Hi everyone,

I am wondering if it is possible to set individual bits in 64-bit double-precision floating number in MATLAB.

I know that I can use "format hex" to read the bits. But how can I set it?

Thanks Kevin

Answer 1

[Walter Roberson]

Example:

typecast(bitset(typecast(pi,'uint64'),53,1),'double')

Bit numbering is 64 for the most significant bit, down to 1 for the least significant bit.

As the IEEE 754 floating point representation is being altered, the change I made here of effectively adding 1 to the exponent resulting in doubling the result (because IEEE 754 has binary representation, not decimal representation.)

[Kevin]

Hi Walter,

Excellent. This is exactly what I am looking for.

Side question. Were you working for CIBC investment? Long time ago I think there was a Walter (if my memory is right) very active in the MATLAB newsgroup and working for CIBC investment.

Thanks Kevin

[Walter Roberson]

I have never worked for CIBC or any finance company.

I find a record of Michael Robbins in comp.soft-sys.matlab who worked for CIBC. Looks like he had about 75-ish posts. My active time on comp.soft-sys.matlab did not overlap with him.

[Kevin]

You are correct. I was thinking about Michael Robbins.

Thanks again for your help.