VCS & FPGAs - Benifit to owning an FPGA

[forkconfig]

I was thinking about buying an FPGA but I want to find out if there is any benefit over just using VCS.
Thus far I have been developing hardware and verifying using a SV TB I develop.
When would I need an FPGA? Is it only when I actually want to control something like a robot or an LED? Or is there added benefit to owning an FPGA?

 

[TrickyDicky]

You can write all the code you want and run as many simulations as you want, but unless it actually does something in the real world, whats the point? (other than to learn HDL)
Using a real FPGA will show you about timing closure etc.

 

[forkconfig]

So the added benefit is the ability to interface with things in the real world as well as learn about timing because of issues I may run into.
Is that correct?

 

[TrickyDicky]

it really depends what your aims are. Always doing simulations is all well and good, but when something falls over on the bench you need to be able to get a scope or logic analyser out.

But most time should be spent in simulation. Verification up front can save integration time later. Bug fixing in integration always takes longer than it would have taken in verification.

 

[BlackHelicopter]

What kind of hardware are you developing and where does your 'hardware' currently reside, on your computer? I'm not quite sure what you're missing about the whole concept of an FPGA. Don't you have any interest on interfacing with it on a physical level in the real world?

Most of the time people create HDL code for the sole purpose of putting it on an FPGA and using it in the real world. Take a car for instance. You don't drive around in CAD models and mechanical simulations do you? When you buy a TV - in place of the digital electronics would you rather get a bunch of VCS simulations and binary files, OR.. the (application-specific) FPGAs?

So I would say the benefit to having an FPGA is that you actually get to use what you've made in the real world, unless you're content with creating things in a vortex somewhere.

Edited: I see you said "added" benefit. If you owned an FPGA, you would find out that designs sometimes don't work in the real world as they do simulations and for extremely "interesting" reasons. There are multiple issues that can arise like: problems communicating with a peripheral IC, metastability, voltage levels, timing, propagation delays, reset anomalies ..etc., all resulting in tons of head scratching. Hey, think you're finished with your design!? Nope! Someone found a way to break it! Or better yet.. they got the 1 (out of 1000) that uncovered an intermittent issue!

 

[TrickyDicky]

Oh yeah, thats another point.
FPGAs will run 1000s of times faster than a simulation can.

 

[forkconfig]

Great answer, thanks!

Right now I'm just starting to learn Verilog.
So I was considering just using VCS till I felt comfortable designing hardware.
Figured in 1 year I would be better prepared and then I could buy a newer FPGA.

What kind of hardware are you developing and where does your 'hardware' currently reside, on your computer? I'm not quite sure what you're missing about the whole concept of an FPGA. Don't you have any interest on interfacing with it on a physical level in the real world?

Most of the time people create HDL code for the sole purpose of putting it on an FPGA and using it in the real world. Take a car for instance. You don't drive around in CAD models and mechanical simulations do you? When you buy a TV - in place of the digital electronics would you rather get a bunch of VCS simulations and binary files, OR.. the (application-specific) FPGAs?

So I would say the benefit to having an FPGA is that you actually get to use what you've made in the real world, unless you're content with creating things in a vortex somewhere.

Edited: I see you said "added" benefit. If you owned an FPGA, you would find out that designs sometimes don't work in the real world as they do simulations and for extremely "interesting" reasons. There are multiple issues that can arise like: problems communicating with a peripheral IC, metastability, voltage levels, timing, propagation delays, reset anomalies ..etc., all resulting in tons of head scratching. Hey, think you're finished with your design!? Nope! Someone found a way to break it! Or better yet.. they got the 1 (out of 1000) that uncovered an intermittent issue!