Xilinx Virtex-4 APU Controller Questions

Hello all,

I am in the process of creating a floating point FCM that interfaces with the APU controller on a ML403 development board. After studying the documentation, I have a few questions:

  1. In the PowerPC405 Processor block reference (UG018), on page 207, a generic FCM Load Instruction example is presented. Is this an example of a floating point register load instruction? i.e. if the instruction "lfd 1,rX(0)" (Load floating point double) is issued, will the APU follow this example and automatically fetch the data indexed at rX and then send the instruction plus data to the FPU and let it decode that the data should go in floating point register 1? Or is this load example for a special type of UDI?

  1. How do you enable hardware floating point support in the EDK software compiler?

  2. On page 188 of the PPC Block Reference, it is stated that three groups of FP instructions can be selectively disabled: complex arith, conversion, and estimates. How do you inform the compiler, either the EDK or a ppc gnu cross compiler that these instructions are not available? If that isn't possible, what happens if they try to execute that instruction? Does the APU or FCM simply issue an exception and hope that the software handles it?

Any advice or suggestions would be greatly appreciated. Thank you.

Sincerely, Jarrod Wood

Reply to
Jarrod Wood
Loading thread data ...

Jarrod,

I am having the same problem. I am trying to put in an FPU core from coregen into EDK and have it interface to the APU. So far, I haven't been able to figure it out, and Xilinx support isn't very helpful with this. I am wondering if you got any further. If you have any more infomation it would be very helpful if you could post it.

Thanks

Dmitriy Bekker Jet Propulsion Laboratory

Reply to
Dmitriy Bekker

Ditto I am looking at using 4 of the fpu cores provided with ISE and writing the instruction decode and glue logic myself. This seems like a generic enough problem that there must be an easier way. I have been having the same problems -mhard-float does not make GCC target an FPU and the documentation on the APU is slim. Let me know if you hear anything/ have a great breakthrough, ill do the same.

Reply to
Justin

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.