Potentially awesome open-source idea
by Unknown on Jan 31, 2004 |
Not available! | ||
Tom Hawkins wrote:
On Wednesday 28 January 2004 10:49 am, Bill Cox wrote:
[large snip]
Another idea is that the author of a core should be able to offer
users configuration options, and then have the core customized. For example, did you want a 32 bit MAC, or a 16 bit MAC (probably a poor example). If we have a module-generation capability that's easy to work with, I think we can pull this kind of flexibility off. I think I can see at least three software projects here: - A Web based GUI - A tool to glue cores together, initially using wishbone - A module generation system (Confluence and/or OpenLPM and/or JavaHDL) A while back I offered up Confluence to the OpenCores community for this very concept: flexible IP and web-based netlist generation. The offer still stands. On a similar note, a Confluence user recently ported OR1200 to the language. (Ken, are you going to post ORCF anytime soon?) Bill, it's a great idea. If there's anything I can do to assist, let me know. Regards, Tom Hi, Tom. I am very impressed with confluence. You've done something pretty cool there. My main strike against it is that it's not open-source. Of course, the other choices all seem to have draw-backs. Perlilog is Verilog-centric, and I recently found out just how much the VHDL users around here insist backing their language. JavaHDL requires the user to install the JVM. My old OpenLPM project is still buggy, and not as cool as Confluence, but it could probably be offered as open-source. I'm not against charging for software, and I think you deserve to make money on Confluence. However, I feel an SoC Builder will get more interest if it's an open-source front-to-back. Is there any way to have an open-source implementation of Confluence without diminishing your ability to commercialize it? I know that's a tough nut to crack... Bill |
Potentially awesome open-source idea
by Unknown on Jan 31, 2004 |
Not available! | ||
On Saturday 31 January 2004 05:33 am, Bill Cox wrote:
Tom Hawkins wrote:
I agree. It would be ideal if OpenCores only depended on open-source
software. But the truth is every OC project requires commercial EDA.
(If I'm wrong, please show me the open-source path from RTL to GDSII.
;-))
If it eases one's mind, Confluence is delivered without license
protection. Once you have the executable, nothing's going to cut you
off; unless of course Linux falls out of existence. For legality, I
would have no problem signing agreements to ensure OpenCores keeps a
perpetual license.
Personally, I think CF fits well with OC; especially for design reuse.
If projects can be unified under a single CVS tree, cross project
instantiation is trivial. Need a processor? Just do:
OpenRisc = import "/processors/OpenRisc1000/OR1200.cf"
Regards,
Tom
--
Tom Hawkins
Launchbird Design Systems, Inc.
952-200-3790
http://www.launchbird.com/
>On Wednesday 28 January 2004 10:49 am, Bill Cox wrote:
> >[large snip] >
>Another idea is that the author of a core should be able to offer
>users configuration options, and then have the core customized. >For example, did you want a 32 bit MAC, or a 16 bit MAC (probably > a poor example). If we have a module-generation capability > that's easy to work with, I think we can pull this kind of > flexibility off. > >I think I can see at least three software projects here: > >- A Web based GUI >- A tool to glue cores together, initially using wishbone >- A module generation system (Confluence and/or OpenLPM and/or >JavaHDL) > >A while back I offered up Confluence to the OpenCores community > for this very concept: flexible IP and web-based netlist > generation. The offer still stands. > >On a similar note, a Confluence user recently ported OR1200 to the >language. (Ken, are you going to post ORCF anytime soon?) > >Bill, it's a great idea. If there's anything I can do to assist, > let me know. > >Regards, >Tom Hi, Tom. I am very impressed with confluence. You've done something pretty cool there. My main strike against it is that it's not open-source. Of course, the other choices all seem to have draw-backs. Perlilog is Verilog-centric, and I recently found out just how much the VHDL users around here insist backing their language. JavaHDL requires the user to install the JVM. My old OpenLPM project is still buggy, and not as cool as Confluence, but it could probably be offered as open-source. I'm not against charging for software, and I think you deserve to make money on Confluence. However, I feel an SoC Builder will get more interest if it's an open-source front-to-back. |
Potentially awesome open-source idea
by Unknown on Feb 2, 2004 |
Not available! | ||
Aloha!
Quoting Tom Hawkins tom at launchbird.com>:
I agree. It would be ideal if OpenCores only depended on open-source
Sure there is. You could use Icarus Verilog to go from RTL to GTL and then use
Magic (albeit with some scripts) to generate GDSII. The result will not be great
and not production quality, but it's possible. Magic have and still is used to
build real chips.
--
Med vänlig hälsning, Yours
Joachim Strömbergson - Alltid i harmonisk svängning.
VP, Research & Development
----------------------------------------------------------------------
InformAsic AB / Hugo Grauers gata 5B / SE-411 33 GÖTEBORG / Sweden
Tel: +46 31 68 54 90 Fax: +46 31 68 54 91 Mobile: +46 733 75 97 02
E-mail: joachim.strombergson at informasic.com Home: www.informasic.com
----------------------------------------------------------------------
software. But the truth is every OC project requires commercial EDA. (If I'm wrong, please show me the open-source path from RTL to GDSII. ;-)) |
Potentially awesome open-source idea
by Unknown on Feb 2, 2004 |
Not available! | ||
Aloha!
Quoting bporcella bporcella at sbcglobal.net>:
For some inspiration I recommend looking at the SOPC builder by Altera. It is in
effect a Perl based tool that allows you to select cores. Connect them using the
Avalon fabris, set addresses, interrupts etc. Each core has it's own tab (i.e. a
specified customization ability that the SOPC frameork just needs to be able to
launch.
After setting all features needed for the given SoPC-design, you press Generate
and it will produce a system with corresponding RTL, SW-interface, memory map
etc. Vey neat. I belive you could download a trial version of SOPC-builder fromn
Alteras webpage. It's IMHO definetly worth a testrun.
And there is no magic trick behind this either. It's base configuration files
for the Avalon fabric, address maps and symbol lists. The Perl scrips copies and
modiies the files as needed. We (InformAsic) have a similar setup for our own
internal SoC-platform. Not as fancy with GUI etc, but still it generates the
complete SoC for us in very short time.
There is another benefit with these systems: You lower the risk of changing
parts that shouldn't be changed. When adding cores to a system you just add
THOSE files and configurations, the rest of the system is automatically included
and "off limits" from mistakes that tend to happen when done manually.
--
Med vänlig hälsning, Yours
Joachim Strömbergson - Alltid i harmonisk svängning.
VP, Research & Development
----------------------------------------------------------------------
InformAsic AB / Hugo Grauers gata 5B / SE-411 33 GÖTEBORG / Sweden
Tel: +46 31 68 54 90 Fax: +46 31 68 54 91 Mobile: +46 733 75 97 02
E-mail: joachim.strombergson at informasic.com Home: www.informasic.com
----------------------------------------------------------------------
> -- Users post pre-synthesized cores to OpenCores (as well as RTL
> source), in a generic netlist format (this is something I know how to do). > -- We write open-source tools that can combine those cores > automatically, taking core-specific options, and running the author's > customization scripts. > -- This tool then spits out gate level Verilog/VHDL to the desired > target (Xilinx, Altera, LSI, etc). > > The impact on design productivity is potentially huge. > > Any interest in discussing the idea? > > Bill I don't have any basic problem with the idea - it seems relatively straightforward to implement. but don't see how impact on productivity could be potentially huge (or even significant) -- unless of course the tool also puts together the composite testbench and writes the total verification suite. |
Potentially awesome open-source idea
by Unknown on Feb 2, 2004 |
Not available! | ||
Hi, Joachim.
Joachim Strombergson wrote:
Aloha!
... For some inspiration I recommend looking at the SOPC builder by Altera. It is in effect a Perl based tool that allows you to select cores. Connect them using the Avalon fabris, set addresses, interrupts etc. Each core has it's own tab (i.e. a specified customization ability that the SOPC frameork just needs to be able to launch. Thanks for the reference. I'll definately take a look. It sounds much like what I had in mind.
After setting all features needed for the given SoPC-design, you press Generate
and it will produce a system with corresponding RTL, SW-interface, memory map etc. Vey neat. I belive you could download a trial version of SOPC-builder fromn Alteras webpage. It's IMHO definetly worth a testrun. And there is no magic trick behind this either. It's base configuration files for the Avalon fabric, address maps and symbol lists. The Perl scrips copies and modiies the files as needed. We (InformAsic) have a similar setup for our own internal SoC-platform. Not as fancy with GUI etc, but still it generates the complete SoC for us in very short time. There is another benefit with these systems: You lower the risk of changing parts that shouldn't be changed. When adding cores to a system you just add THOSE files and configurations, the rest of the system is automatically included and "off limits" from mistakes that tend to happen when done manually. I guess this shows that an SoC Builder can be done and has some merit. I'm quite excited about working on it. Bill |