What are FPGAs used for today and with what tools?

[asmi]

One of my all-time favourite ICs is the smallest Cyclone IV part from Altera, the EP4CE6E22. It has a quantity of logic and memory that's a good fit for a number of projects, and way more I/O than I often need. I wish it came in 64 or 100 pin packages.

That's stage 1. Stage 2 will begin when you come across the project that will require external memories (as they tend to consume a lot of pins, and often has very specific requirements as to which pins you can or can not use, which often leads to underutilized IO banks due to IO voltage conflicts, or layout reasons), at that point you will realize that you need more pins, not less. It will also be about time when you also realize that you need more FPGA resources than these simple old chips can provide. Stage 3 will be when you want to use serial transceivers for hi-speed communications - usually these starts as wanting to implement something like DisplayPort, which is among the simplest Gbps-class interfaces to implement, and cover pretty much any modern serial standard out there - and this will further restrict which devices you can use, as not all of them have required number of transceivers and/or they are running fast enough for your needs. I've been there a couple of years ago, and since that I've seen few others follow similar path.
FPGAs are fun. But unfortunately not cheap, and the price goes up rather quickly as you climb the functionality ladder

[james_s]

what has not changed is that the xilinx tools still suck ... Altera is much better.

I've spent quite a lot of time in Xilinx ISE and Altera Quartus and my conclusion is that they both suck in different ways and both do a few things quite well. Overall I have found I prefer the Altera workflow but Xilinx has a much easier to use simulator and a few other advantages. I see no clear winner.

[james_s]

That's stage 1. Stage 2 will begin when you come across the project that will require external memories (as they tend to consume a lot of pins, and often has very specific requirements as to which pins you can or can not use, which often leads to underutilized IO banks due to IO voltage conflicts, or layout reasons), at that point you will realize that you need more pins, not less. It will also be about time when you also realize that you need more FPGA resources than these simple old chips can provide. Stage 3 will be when you want to use serial transceivers for hi-speed communications - usually these starts as wanting to implement something like DisplayPort, which is among the simplest Gbps-class interfaces to implement, and cover pretty much any modern serial standard out there - and this will further restrict which devices you can use, as not all of them have required number of transceivers and/or they are running fast enough for your needs. I've been there a couple of years ago, and since that I've seen few others follow similar path.
FPGAs are fun. But unfortunately not cheap, and the price goes up rather quickly as you climb the functionality ladder

One of my favorite parts is an ancient Cyclone II, complete dev boards using it are under $15 and I have managed to fit all kinds of interesting projects into it. Complete 8 bit computers, several early arcade games and other interesting vintage hardware. I have a DE2 dev board with a much larger FPGA and have found it handy for prototyping due to the onboard displays and switches and stuff but I have never come anywhere close to filling it. Most of my projects end up being memory constrained and I run out of block RAM long before I run out of logic. External memory can solve this but as you say, it consumes a lot of pins, especially when you have something like an arcade game that has video RAM and ROM and system RAM and ROM operating in different clock domains.