Overcomplicated formula in LDR/STR with immediates

[alexrp]

[timpart]

[alexrp]

[ysapir]

But, this means that you now have to explain to the reader what size is, and refer him/her to the decode table?? Thanks, I'd rather stick to the current description...

Plus, you are tying two similar-but-not-identical quantities, as the opcode version should really be read as "size code" rather than "size". Furthermore - what if the "code" changes in future revision?

[timpart]

I agree that the existing formula should be retained. One possibility is to add a little quick reference table giving the result for various data sizes.

Tim

[alexrp]

[ysapir]

[alexrp]

[aolofsson]

To answer the tangent question.

1.) We would of course like the Epiphany to be a general purpose architecture but that is not likely to happen any time soon...
Right now the Epiphany serves as a coprocessor to ARM/MPIS/x86. It's definitely not limited to embedded applications, now or in the future!

2.) It would take something absolutely monumental for us to decide to break binary compatibility. Basically, we will only change things if the feature is completely broken or practically unusable. The only exception is chip level register like IO. For example, note that the Epiphany-III and Epiphany-IV are binary compatible.(minus the excause register which is broken on Epiphany-III anyway)

Andreas