Sorry, you need to enable JavaScript to visit this website.

Some figures and interest for Sound Synthesis

3 posts / 0 new
Some figures and interest for Sound Synthesis

Hi all,

I've been working at times on sound synthesis with FPGAs, and also a bit on the Eco32 processor design and use, and to prevent the style of making the zedboard users into stamp collectors with no new, challenging purposes (or hdmi signal copycats or something), I thought I'd memorize some characteristics and quote some experience from Open Source projects I know of.

A little comparison table:

Spartan 3e-500 (from the well known Xilinx/Digilent devboard):
Sysgat 500K (10kELC), Rambits 73k/360k, Mults 20, DCM 4, UserIO 232

XC7Z020 (ZedBoard)
Sysgat 1300K (85kELC), Rambits -/4480k, MACCs 220, DCM 8(?), UserIO 200(+?)

Virtex-7 (about the biggest and fastest available ?)
Sysgat -(2000kELC), Rambits -/68000k, MACCs 3600, DCM (?), UserIO 1200

Of course heat production and required energy per operation is also a factor...

I think for general "progress" (I'm uni EE, not an conservative IT-er) instead of "just" nice drivers for a ARM based computer systems, there are good uses to be made of the strong connection between the ARM cores and the FPGA fabric on the chip, for fast interface interactions (like a fast working high sampling frequency sound interface), but also for fast kernel response hardware acceleration and of course making the FPGA fabric run-time programmable.
So I'm curious if the whole zedboard can get of the ground good!

Will the Linux made available here be compilable with the gnu-arm compiler, or is there something like a Xilinx Embedded Controller compiler connected with the zedboard ? I know that it is often in the details where the compatibility and smooth working parameters appear to hide...

Oh and if I compile the to-be test projects, will I need to upgrade to ISE 14 (from 13, which I have working on two Fedora 64bit installations) with extras installed, or base installation ?


Ir. Theo Verelst

Welcome to the forums!

I'm so glad you have shown intersted in Zed, and thank you for your breakdown of basic resources between previous development kits and the Zedboard.

Do remember that the Zynq-7000 EPP is based on 28nm Virtex technology, making it significantly faster, lower power, and more efficient at using routing and clocking resources than any previous family of Xilinx FPGA's.

The ARM core portion and many of it's peripherals are straight from ARM, and this can take advantage of any part of the vast ARM echo system available today. We support our tools which are called from SDK, as well as the ARM compiler and gcc-arm.

My understanding is that all of the projects will be in 14.x tools for the Zedboard, so yes, you should upgrade you tools. You can download the webpack from here:

I don't think this whole site

I don't think this whole site and the projects are going to work all too good in the near future, so I herewith suspend confidence in joining the endeavor.

I think the Zinq machine is a great idea and I have all confidence in Xilinx, but I prefer the Open Source character of let's say a left-oriented hippie from Harvard over some jacks of all trades led by the AVNET department with golden watch-on-chain and suit and tie. Don't promise stuff and be late, that's bad business.

I like the board idea, the acedemic price seems more than worth it, and I hope in the future interesting deep projects with it are possible, but for the moment I have more confidence in lets say a brick-ed android EEE device with a nice 7030 in it, and prefer the road of a say a 10 large array of 7040 as a PCIe board to do fun stuff with at the price of a nice graphics parallel card.

The various voices here do not make me like the project, sorry.