Getting started with FPGA

On Aug 19, 1:17=A0am, Symon <symon_bre...@hotmail.com> wrote:
> For every nice reply, it's important to have a dissenting one. The OP
> post's IP (NNTP-Posting-Host: 213.15.253.202) resolves to something
> called IBM regus UK. Whatever that is. Seems to be some kind of
> training. I would suggest that 'Rupert' could be best served by asking
> his trainers the answer to his query. Especially if the prices listed
> match the course 'Rupert' is on. Anyway, IBM regus has gotta be better
> than Doulos, right?
>
> Cheers, Syms.

I work in an office rented in a building owned by Regus (in Manchester
in the UK), no idea what IBM has to do with Regus, but there you go.

I've nothing to hide. I work as a Java programmer for a company that
produces software for stock markets and traders, an area where FPGAs
are gaining ground if and when they can provide a speed advantage. Its
a bit of an arms race really; why all this need to go faster? to beat
the competition. As I said in my OP, I'm not looking into FPGAs with a
specific application in mind. Maybe somewhere down the line it might
become a work project, if it could be proved to give a speed
advantage, if there's the time and funding to do it and so on. In that
case, I would likely still be on the software side and someone who
knows what they are doing would do the hardware.

For now, I just want to take it up as a hobby and learn something
about it and have some fun doing it.

On Aug 18, 7:21=A0pm, Tim Wescott <t...@seemywebsite.com> wrote:
...
> There are open-source simulators out there, but the FPGA companies are
> pretty tight with their synthesis algorithms (and even the mapping
> between the bit files and the FPGA functions), so you won't find any
> open-source synthesizers.
...
> The FPGA is soldered down to the board. =A0During early development you
> fire things up and write bit files directly to the chip (at least in the
> Xilinx workflow) with a JTAG programmer. =A0FPGA chips (at least those
> that don't have on-board PROM) are capable of loading configurations
> from EEPROM, and most development boards will have a suitable EEPROM
> that you can load your bit file onto so the board will come up by itself.

Hi Tim,

Thanks for this introductory overview of the process, very helpful
indeed.

One question about the specifics on programming the synthesized data
onto the chip: This is generally achieved through a JTAG port, with a
cable that goes from the USB port (serial/parallel ports on older
models), through a box of tricks, and out onto a JTAG connector. Is
this cable specific to each vendor or is the USB -> JTAG conversion
standard accross all? If I buy a cable from Xilinx, will it work on
all boards with Xilinx chips, no matter where I buy the board? I
notice that some of the sites offering prototyping boards also sell a
cable. The cables are a significant cost, so I'd like to make sure I
get the right cable with the maximum usability across different
boards.

Rupert

On 8/19/2010 10:03 AM, rupertlssmith@googlemail.com wrote:
> On Aug 19, 1:17 am, Symon<symon_bre...@hotmail.com>  wrote:
>> For every nice reply, it's important to have a dissenting one. The OP
>> post's IP (NNTP-Posting-Host: 213.15.253.202) resolves to something
>> called IBM regus UK. Whatever that is. Seems to be some kind of
>> training. I would suggest that 'Rupert' could be best served by asking
>> his trainers the answer to his query. Especially if the prices listed
>> match the course 'Rupert' is on. Anyway, IBM regus has gotta be better
>> than Doulos, right?
>>
>> Cheers, Syms.
>
> I work in an office rented in a building owned by Regus (in Manchester
> in the UK), no idea what IBM has to do with Regus, but there you go.
>
> I've nothing to hide. I work as a Java programmer for a company that
> produces software for stock markets and traders, an area where FPGAs
> are gaining ground if and when they can provide a speed advantage. Its
> a bit of an arms race really; why all this need to go faster? to beat
> the competition. As I said in my OP, I'm not looking into FPGAs with a
> specific application in mind. Maybe somewhere down the line it might
> become a work project, if it could be proved to give a speed
> advantage, if there's the time and funding to do it and so on. In that
> case, I would likely still be on the software side and someone who
> knows what they are doing would do the hardware.
>
> For now, I just want to take it up as a hobby and learn something
> about it and have some fun doing it.
>

I take it all back!

> I take it all back!

All of the sarcy comments, ever?

:-)


Nial. 

On Aug 18, 8:07=A0am, "rupertlssm...@googlemail.com"
<rupertlssm...@googlemail.com> wrote:
> What about books? Can anyone recommend a good book to get started
> with?

This is not an HDL book but provides an overview of the Xilinx design
process.  Free to download at:
  http://www.xilinx.com/support/documentation/boards_and_kits/ug500.pdf

Xilinx also has an in-depth tutorial which is good to start with:
  http://www.xilinx.com/support/documentation/dt_ise12-2_tutorials.htm

There should also be introductory tutorials for whatever board that
you buy.  Check out the board documentation first before you buy.

My favorite book is HDL Chip Design by Douglas J. Smith.  It has both
Verilog and VHDL code side-by-side and often shows a circuit
representation of how the code synthesizes.  It's more of an example
book then a reference book, but I like it.

Bryan

On 8/19/2010 1:36 PM, Nial Stewart wrote:
>> I take it all back!
>
> All of the sarcy comments, ever?
>
> :-)
>
>
> Nial.
>
>
No, only last night's drunken paranoia!