Welcome, Guest. Please login or register.

Author Topic: FPGA for dummies  (Read 59536 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: FPGA for dummies
« Reply #44 from previous page: December 13, 2011, 02:51:12 AM »
Quote from: psxphill;671231
It's not 1:1 because the cell in the FPGA is nothing like the gates in an ASIC.


If you make a scan and put that into a custom ASIC, then yes it is an exact 1:1 copy. I was under the impression that the definition of simulation/emulation in this thread included
ASICs that did not come from the original VHDL/schematics.

Quote
undocumented behaviour is not the same because alot of them rely on the analogue behaviour of digital gates and the cells in the FPGA aren't going to behave like that at all.


Analog behavior of the digital gates?! Eh? How does an AND gate behave any differently than an AND gate due to this magical Analog nonsense?
 
Quote
when you build an asic from VHDL you would just put the gates in, but electronically they are still going to be different than the original. So any analogue effects (resistance/capacitance etc of the circuit) may differ just because of different lengths and the gate chemistry.
 
If you take a look at resid then you'll see how difficult it is to simulate a sid chip using digital concepts.
The colour generation in vic is analogue based, because it operates in NTSC/PAL colour space.


You can determine what is a resistor, what is a capacitor, etc. in the die scan. To a little probing to associate resistance/capacitance/etc. values to feature sizes, and implement those values in your custom ASIC remake. Analog clone. Direct from the original die itself.
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: FPGA for dummies
« Reply #45 on: December 13, 2011, 06:51:09 PM »
Quote from: Thorham;671294
What if it's a robotic duck?


Which side of the uncanny valley is it on?
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: FPGA for dummies
« Reply #46 on: December 13, 2011, 07:02:51 PM »
Quote from: Thorham;671294
No, they can't. Same goes for the brain, because it's the neural connections that make a brain able to do anything in the first place. No neural connections=no functionality=no intelligence.


Step 1: Synapse.
http://www.gizmag.com/researchers-create-artificial-synapse/18482/
http://www.mit.edu/newsoffice/2011/brain-chip-1115.html

Step 2: Neuron....
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: FPGA for dummies
« Reply #47 on: December 14, 2011, 02:41:49 AM »
Quote from: Thorham;671335
And without step 3 it will all do nothing.


Step 3 is profit.
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: FPGA for dummies
« Reply #48 on: December 14, 2011, 02:45:06 AM »
Quote from: amigadave;671297
realize that there is no ONE definition of what a "Real" Amiga is anymore.

Just a note, I haven't been trying to argue about what is real or fake Amiga. My part in the debate is about what is a real or imaginary circuit implementation. OK, I guess I have asked questions about what or who makes something real instead of a simulation or emulation in a custom chip, and what or who makes the same thing not a simualtion/emulation, but that all links back to trying to understand why one custom chip is a real circuit and another custom chip is a figment of everyone's imagination, which is how some people here feel about it.
« Last Edit: December 14, 2011, 02:47:34 AM by billt »
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: FPGA for dummies
« Reply #49 on: December 14, 2011, 02:52:12 AM »
Quote from: Thorham;671335
Perhaps it's a partial replica? Doesn't matter, because it was designed and made by the makers of Amiga.


And the "legitimate" makers of Amiga also made the Walker, which has an FPGA in it. I understand this implements some new "custom chip" because they couldn't afford the megabucks to get an ASIC (custom chip) manufactured. Is Walker a partial replica, a simulator, an emulator, an actual computer, or something else? Or is Walker not legitimate enough for people here to discuss this way?
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: FPGA for dummies
« Reply #50 on: December 18, 2011, 03:02:50 AM »
Quote from: psxphill;671772
Generally chips from different sources are made from the same mask and are just second source suppliers.
Quote


I don't know if that's true at all, but I do know that it's not generally true.

Some companies get bit by part changes. Supposedly identical parts from different suppliers can have enough difference to cause a product defect (probably a new timing or crosstalk issue due to faster/stronger outputs). Sometimes a single supplier changes their chip and this can cause a defect, without any notice or updates to documentation. One can be left scratching his head for quite a while until a chip change becomes known.

Also understand that even the same mask set in a different factory can give a different result and break timing or crosstalk.
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: FPGA for dummies
« Reply #51 on: December 19, 2011, 04:39:22 AM »
Quote from: psxphill;671836
It's very well known.
 
http://en.wikipedia.org/wiki/Am286

If AMD's had a higher possible clock speed, then I really don't believe it was the same masks as suggested elsewhere. Being pin-compatible and based on Intel's microcode does not a mask duplicate. Though it does seem your wikipedia article does suggest the 386 situation was a result of Intel not sharing masks, it still seems that AMD did some changes to the 286 design to get higher frequencies. And the wikipedia article says that it's common for those agreements to allow tthe first-source company to make designs from the second-sourcers. So there must be some design going on to have that in there.
 
Quote
The 68000 was second sourced by alot of companies, Mostek, Rockwell, Signetics, Thomson & Toshiba.

I remember something about some of them having quirks in compatibility, I think Signetics in particular. Can't be same masks.
 
Quote
The fools thing was aimed at the "my room lamp emulates light". It's even a rubbish analogy, because the lamp emits light.

If it was not made by Edison himself, then some in this thread seem to believe it should not really be a light bulb... And CFL and LED replacements are not light bulbs, so they must only be simulators...
« Last Edit: December 19, 2011, 04:44:40 AM by billt »
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: Question by an FPGA dummy.
« Reply #52 on: January 03, 2012, 02:46:32 AM »
Quote from: Thorham;674031
Seeing how I've become interested in FPGA computing, I would like to play with this, and seeing how a switch (emulated relay using bits as signals) emulator on a computer shouldn't be too hard to write, I would like to know how easy or difficult such virtual circuitry would be to translate to actual circuitry that can be implemented using an FPGA (or with wires and transistors, or an ASIC, or...)?

Any pointers are appreciated :)


Your relay sounds like a passgate. (also called a transmission gate in some circles) In an ASIC you might just put a passgate there, but this implies tristate wires which are frowned upon in modern design practices in terms of design using gates (logic gates I mean, not transistor gates). (be that verilog, vhdl, schematics, whatever, it's all a poor habit to be in unless there is particularly good reason for them) It's frowned upon as it implies the possibility of a floating input somewhere when all drivers on the net are "off"/disconnected. Floating inputs are bad, in that they can do weird things, which can be bad. So tristate busses, like you do see on PCB boards (PCI, Zorro, ISA, etc) are frowned upon inside a chip. Using a passgate can be troublesome in timing analysis. In my FPGA silicon days, we could not plan for timing because of inability to time through passgates. So we just built them, and thigns went as fast as things turned out to work right. It's also difficult to design an FPGA with timing/speed in mind, as what exactly needs to go how fast? You don't know what it will be doing in a customer design, so how do you optimize for that or have a particular goal?

In an FPGA, to do a passgate, you probably controlling a mux to pass the input value out or not. But in not passing the input value out, in an FPGA you probably need to pick an "off value", in that you can't pass a Z (complete disconnect), you have to pass (select) either a 1 or 0 (still connected to something). Proper design methods/habits/whatever need to be used, so rather than design using "relays" as you call them, you need to design using more appropriate elements. And that applies to all digital things inside the chip. (Analog chip design is a whole other world with whole other rules. I've only scratched the surface there as a layout guy, and know nothing of circuit design that we layout guys are given to implement)
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: Question by an FPGA dummy.
« Reply #53 on: January 03, 2012, 05:20:11 AM »
Quote from: Fats;674049
BTW, a transistor is actually a switch and in CMOS technology you basically have two types: a nMOS one that is open when input is high and a pMOS one that is closed when input is high.


In CMOS, the N and P are used together. pmos are used on the high voltage side, and nmos on the ground side. nmos give a weak high voltage connection, and pmos give a weak ground connection, which is why we don't use one alone anymore. using pmos on high side gives a good high connection, and nmos gives a good ground connection, and together you have good connections to both voltages.
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: FPGA for dummies
« Reply #54 on: December 17, 2012, 08:01:43 PM »
To revive a year old thread, there is now an online simple class about FPGA chips that seems relevant to this. It does compare and contrast software running on a processor and hardware configuration of an FPGA, and talks some about how things work inside. I've finished 2 of 5 sessions now, and am enjoying it. (On another note, I've just finished a university course in Computer Architecture, also very enlightening and interesting stuff about the control logic of a processor)

http://university.eetimes.com/lecture-calendar.asp?cid=901#lecture_track_cgid_3
Bill T
All Glory to the Hypnotoad!
 

Offline billt

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 910
    • Show all replies
    • http://www.billtoner.net
Re: FPGA for dummies
« Reply #55 on: December 23, 2012, 07:10:46 PM »
Hi all, i didn't intend to rekindle an argument about what is or is not emulation. the links i posted go to a technical education series of webinars introducing fpga technology to newbies. I thought the op of this thread and some others would be interested in learning.


People designing with fpgas design circuits that are placed into the chip by configuration.

Whether or not you feel the result is emulation or not i don't feel like participating in again, though i do have my choice on that debate.
Bill T
All Glory to the Hypnotoad!