Is the "double" datatype 32-bit or 64-bit on SHARC?

The code written with the "double" data type runs in 32-bit single-precision floating point. The double type for SHARC is configurable to be either 32-bits single-precision or 64-bits double-precision. By default it is 32-bits.

If you need 64-bit double-precision, you can either use long double or enable it through the complier option "-double-size-64". In CCES, this setting is available under Project > Properties > C/C++ Build > Settings > Tool Settings > Compiler > Processor > Double size to the required value.