Update Picoblaze Code in Bitstream

Hello,

in my actual design im using a few picoblazes. Now I wonder if it is possible to update the code in the bitstream without a new implementation run like it is possible with the microcblaze. I checked data2bram but it allows only an update of 16 Bit wide Brams, not the necessary 18 Bit.

Thanks, Michael

Reply to
Michael Dreschmann
Loading thread data ...

checked

Check out Ken Chapmans reply to this in the PicoBlaze forum:

formatting link
snipped-for-privacy@.eea991

Reply to
Gabor

is

the

Darn... the link is broken...

Here's the post:

Ken Chapman - 03:54am Dec 14, 2004 PST (#1 of 1) Reply

A new version of DATA2MEM is in development and will support the update of the

1024x18 memory used for a KCPSM3 program. I have carried out some initial tests last week and it worked for me.

Please contact me directly ( snipped-for-privacy@xilinx.com) if you would like to help me verify it further.

Thanks to the Xilinx Software Development team for adding support of BRAM in this aspect ratio.

Regards,

Ken

Reply to
Gabor

Hello,

Here it works, thanks for your help

Michael

Reply to
Michael Dreschmann

Michael,

As of ISE 7.1 Data2MEM does support 18-bit datawidths.

You need to add a BMM and MEM file to the top level of your ISE project.

The BMM should have syntax to describe the instantiation of the BlockRAM you instantiated. An example BMM file is:

//////////////////////////////////////////////////////////////////////// ADDRESS_SPACE picoblaze1 RAMB18 INDEX_ADDRESSING [0x00000000:0x000003FF] BUS_BLOCK top/picoblaze1/myram [17:0]; END_BUS_BLOCK; END_ADDRESS_SPACE; ////////////////////////////////////////////////////////////////////////

In the above BMM file example the name 'picoblaze1' could be anything. Further the label 'top/picoblaze1/myram' is the instantiation name of the blockRAM with complete hierarchy. To obtain the instantiation name of the blockRAM you could run the tools one and then open up FPGA Editor to see the complete instantiation name of the blockRAM.

The second file you have to add to the Project Navigator project is a MEM file. A MEM file is essentially a HEX file (output from KCPSM assembler) with '@0' as the first line. Currently, you must create the MEM file by hand from the HEX file.

Every time the time stamp of the MEM file changes then the 'Generate Bitstream' task becomes unchecked in Project Navigator. All you have to do is is Run 'Generate Bitsteam' and the a bitstream will be generated with the updated values without running through Synthesis/Translate/MAP/PAR.

Cheers, Shal> Hello,

--
------------------------------
Shalin Sheth
Embedded Applications Engineer
General Products Division
Spartan-3 Generation FPGAs
http://www.xilinx.com/spartan3
http://www.xilinx.com/spartan3e
------------------------------
Reply to
Shalin Sheth

The following may be of interest to you (full-text PDF available from the link):

formatting link

Programmable Parallel Coprocessor Architectures for Reconfigurable System-on-Chip.

Williams, John A. and Bergmann, Neil W. (2004)

We propose a hybrid rSoC parallel processing architecture consisting of a central 32-bit RISC microprocessor interconnected to an array of 8-bit microcontrollers as coprocessing nodes. The central processor runs an embedded Linux operating system, with the coprocessor nodes mapped into a virtual file system, by which they can be controlled and reprogrammed. The hardware and software architectures are detailed, and several useful application contexts are proposed. Supporting theoretical analysis is also presented.

Regards,

John

Reply to
John Williams

Impressive. Did you look at data links between the PicoBlaze units ? - that would further off-load the main CPU, and ease situations where a single task proved too much for one Picoblaze ?

-jg

Reply to
Jim Granville

No I didn't, however PicoBlaze interfacing is so simple that adding either direct-wired or buffered (FIFO) links between multiple Picos should be pretty straight forward.

We since have expanded the concepts in that paper to multi-microblaze systems - their native FSL interfaces make it trivial to connect them up to each other in arbitrary topologies, and also makes it easy to drop a bit of custom hardware in place of a CPU if you decide you need more performance or smaller logic utilisation.

BTW, the device driver developed for that work is in the public uClinux source tree, it's a generic uClinux driver for FSL peripherals. So, you can interact with pretty much any FSL core from within uClinux, using intuitive open/read/write system calls, just like any other file/device.

Regards,

John

Reply to
John Williams

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.