Home    Bloggers    Messages    Webinars    Resources   
Tw  |  Fb  |  In  |  Rss
Warren Miller

IP for FPGAs: What Does the Future Hold?

Warren Miller
Page 1 / 2 Next >
Marc Perron
Marc Perron
11/19/2012 5:27:55 PM
User Rank
Beginner
Re: Test & Verification
Hi Warren and congrats for this excellent series of post.

I completely agree with the idea of having IP blocks that include not only the desired function, but also a bundle of verification functions that help designers to reach their ultimate goal: having a fully working system on time and within budget.

In my area of expertise - motor control software IP - this is particularly important since the application is the management of energy and if an error occurs, this can lead to important system damage (i.e. burn a motor / power stage). Not just a simple "system reboot".

However, the design of such verification function is a field of expertise in itself that's entirely related to the expertise domain (i.e. motor control, this is true also for other complex applications such as image processing). Also, the translation of those functions into a form that's usable by a third-party (which can be internal or external) has also a cost (testing, documentation, etc.).

This is true for DSP/MCU-based design which are only SW configurable devices. Hence, for FPGA-based designs, which have an order of complexity higher than DSP/MCU based design (because of the programmability of the HW), it is obvious that it is also true.

I think the fullfillement of those needs belong to 3rd party system-level IP providers (such as Alizem in the field of motor control), i.e. who package their expertise in the form of a licensable IP product.

This is exactly the topic I have presented last month at the IEEE Industrial Electronics Society (IES) Annual meeting for which I have been invited speaker of the Industry Forum. You can access my presentation on my blog:

"FPGA-based Custom Motor Drives Design: The Role of 3rd-party System-Level IP"

The conclusion of this presentation is: the role of 3rd-party system-level IP providers is to provide products that makes it so easy to the system designers that they can bring their own system to the next level (i.e. focus on their true product differentiation). Just like Google did with its "self-driving" car.

Best Regards, Marc.

50%
50%
Myplanet
Myplanet
11/16/2012 2:38:48 AM
User Rank
Guru
Test & Verification
"How about IP blocks that help us develop the necessary verification and test "platforms" our newly complex designs will require?"

Warren, idea is excellent because testing and verification can be done at the root level. In certain design tools itself; they had integrated such facilities at minimal level, I mean the basic do's and don'ts for simple designs.



100%
0%
William Murray
William Murray
11/14/2012 4:35:57 PM
User Rank
Blogger
Re: where will the next development bottlenecks be?
Think it has been bandied about -- using the FPGA array to accellerate the P&R --

50%
50%
Aser
Aser
11/14/2012 4:09:31 PM
User Rank
Expert
where will the next development bottlenecks be?
The configurable computer is very attractive idea. It can utilize the high end FPGAs very well. But the main brake is the compiling time, more precisely, the P&R time. For such a computer the P&R time must be decreased dramatically. The feature is that the compilation result can be much less efficient than as usual.

The IP core library must be cheap or even free. Such cores must be generated automatically for the needed throughput, dimensions, etc.

And IP core interface (both spatial and sequential)  must be standardized. 

50%
50%
JezmoSSL
JezmoSSL
11/14/2012 4:03:31 PM
User Rank
Blogger
Re: Agreed on needing a verification IP framework -- prefereably vendor independent
validation any FPGA design let alone IP cores has always been a thorny issue,all too often you see the attitude of'we'll test in on the actual hardware' which is very difficult to do.

I think the main issues which causes this is masivley complex designs which take  even more complex testbenches,  so yes we do need a tool set which will generate a testbench which gives all the coverage you require and that ir something which seems to have been neglected.

At the moment if you want a testbench which gives full coverage for a large design you can at least double your design time and associated costs.

50%
50%
William Murray
William Murray
11/14/2012 3:18:22 PM
User Rank
Blogger
Agreed on needing a verification IP framework -- prefereably vendor independent
Agreed on needing a verification IP framework -- prefereably vendor independent --

 

50%
50%
More Blogs from Warren Miller
Warren has finally started to write some HDL code to implement his chess-playing FPGA, but he's not a professional coder, so he needs our help and advice.
What might we see in new Ultra Low Density (ULD) CPLD families three-to-five years down the road? Are there new technologies or programmable structures that will find their way into ULD devices?
Following our evaluations, the resources required by a chess-playing FPGA implementation would seem reasonable, even for a small or midsized device.
A number of challenges are faced by the users and manufacturers of ultra-low-density devices (ULDs).
We are ready to consider how to use our Move Generator to traverse the tree of possible moves efficiently and find the sequence that produces the best board position.
flash poll
follow us on twitter
follow Xilinx on twitter
like us on facebook
like Xilinx on facebook
All Programmable Planet     About Us     Contact Us     Help     Register     Twitter     Facebook     RSS