|
Message
From: h.larsen@r...
Date: Thu, 16 Jan 2003 10:33:13 +0100
Subject: RE: [oc] Final Year University Projects
Headline: High throughput FFT core, More FIR achitectures
Hello friends,
For signal processing, FFT and FIR cores are ofcourse central blocks. There
are however often different requirements to throughput/point/tap/resource
ratios. Some applications require fast, high throughput, some compact with
minimal hardware, and other need many points/taps. I.e. one core does not
fit them all.
The current "Radix 4 Complex FFT" OpenCore is made using multi-pass
processing, in-place ram data operation, one stage at a time, saving
hardware, at the cost of speed. This fits many applications, but not all.
Similar arguments goes for FIR filters. Traditionally addressed with
tradeoffs between parallel / serial /bit-wise processing schemes.
So I could see interests for
higher throughput FFT core, i.e. one made with pipelining and hardware for
each stage
and for additional FIR architectures.
All are challenging and interesting projects.
regards
henning engelbrecht larsen
> -----Original Message-----
> From: John Dalton [mailto:john.dalton@b...]
> Sent: 16. januar 2003 00:46
> To: cores@o...
> Subject: [oc] Final Year University Projects
>
>
> Hi all,
>
> One of my University contacts has asked me to
> come up with some ideas for final year Electrical
> Engineering projects. Each student must complete
> one of these projects to get their degree. Typically
> they will spend 2-3 days per week on the project
> over a 10 month period. The work may be involve any
> combination of software, firmware or hardware design
> and construction.
>
> This is an opportunity us to get some more cores
> produced. Does anyone have any suggestions for
> useful cores or tasks, which may be suitable for
> a student?
>
> Apart from producing new cores, I'm thinking this
> might be a good opportunity for a student to build
> a complete system from opencores, and in doing so
> iron out some bugs.
>
> There are no guarantees of useful results as the
> student may flunk, or no student may chose the project.
> Also in the case of writing a new core there is the
> risk the student may chose not to release their work
> to opencores (though I will be suggesting that they do).
>
> If you are going to reply, please do so quickly as the
> students will be choosing projects in the next week or two.
>
> Regards
> John
>
|
 |
Browse
