Is the F4PGA / Chipsalliance project still active?

[betocool]

Hey all,

about two or so years ago I was very interested in the whole Open Source / Yosys / Risc-V FPGA news that were coming out about... well... open source tools to synthesize, place and route FPGAs. I also looked into Litex and Migen a bit (that's the Python HDL description language), but of late I've been busy with other things and this kinda went off my mind.

I was looking again today after a longer absence, I can see the GIT repo https://github.com/chipsalliance/f4pga hasn't been updated in 10 months, and there are not many "new" on the https://f4pga.org/ website.

Is this still active? Is it very slow? I can't easily find what chips are supported how exactly either, so, it looks a bit neglected for lack of a better word. Academic interest only?

Cheers,

Alberto

[laugensalm]

That's the former Symbiflow toolchain, right?
I could so far not see the extra value in it apart from the already existing [Design Entry of whatever kind] -> yosys -> nextpnr -> [ bitffile programming of your target] flow that is pretty straightforward already. I don't really understand the motivation of its sales points, hence.

I'd characterize the situation with these toolchain flows like: Everyone seriously working with it sets up their (his/her) configuration that works well with a particular project, but not everyone sets a standard, and often those enthusiastic projects don't gain enough momentum to justify investment into a fancy org and website or attract more developers after those Summer of Code projects have ended. Also, broad support for a large number of targets is often very underestimated in complexity.

Nevertheless, I'd say the yosys-nextpnr-toolchain is far from academic and good enough to handle some industrial applications where the last f_max optimization is irrelevant.

[betocool]

Momentum, that I think is they key word. It feels they ran out of momentum. I remember a bit more talk about 4 or so years ago, but now it seems to have slowed down.

Then again, I am not as active as I would like to, in the end out of simplicity I end up using the same old suspects, Quartus, Vivado, etc.

Cheers,

Alberto

[dolbeau]

FPGA internals are mostly fully closed still, so such open-source toolchains have to do a lot of reverse-engineering and guesswork.
With the multiplications of open-sourced PDKs for ASICs, it seems to me the momentum for open-source toolchains has moved toward ASIC flows. Those PDKs are for non-leading-edge processes that usually have regular MPW slots for academics at somewhat reduced cost, so there is quite a bit of interest there. For instance, The IHP S13G2 is available from europractice.