This is actually something completely different. :)
This is a show case for eFabless that uses SkyWater PDK to build actual ASICs on a 130nm process. They are making a toy 8x8 FPGA, just so that people can play with an FPGA ASIC that was custom manufactured for that group buy.
Working in the area makes this text on Tensil’s page me extremely curious: “This gives you the performance of custom silicon at the cost of commodity hardware. With Tensil, you can roll out your dream chip in weeks instead of years, at prices measured in thousands instead of millions.” So they designed some fuse programmable ASIC for machine learning? Setting the fuses makes this ASIC “custom silicon”? ASIC for thousands? Sounds to good to be true.
Nope, just a small batch test-chip run. And even out of a few dozen chips, I think we only did packaging for ~5 chips.
It boots Linux, but at 6 mm^2, it's not really big enough to contain an on-chip memory controller, so going out to DRAM is through an FPGA controller and thus is painful.
One of the most interesting parts I found was it's not created from a spec, but from scanning and etching an actual chip, useful work for all the chips without complete documentation.
That's awesome. See, it's stuff like this that's why you're among my favorite chip designers despite being new to the game. ;) Look forward to the write-up.
I am impressed by the fabrication technology for these heterogenous systems. It was already complex enough to fabricate a CPU, but chiplets require incredible precision for placing the die on the interposer wafer for soldering. And the coplanarity of the whole assembly is critical, else it will be impossible to effectively cool. It's incredible that any of it works.
From 'scratch' meaning 74 series LSI chips here. Similar in spirit perhaps to the mighty Magic-1 (http://magic-1.org/), albeit on breadboard (lots of breadboards) rather than wire-wrap. Kudos to everyone who build their own ALU ;-)
Really, the best way to solve this is likely with FPGAs, not ASICs. If you're not using an original chip, does it really matter if it's an FPGA or not? Already, most of these reproduction PCB projects, including this one, end up using PLDs to replace hard-to-find specialied ICs (and even those PLDs are at risk of EOL). So it's not like they're using 100% original chips, anyway.
I understand the urge, but unless the cost of ASIC manufacturing drops dramatically, I don't see this as being practical for small runs like hobby reproduction boards.
Edit: I forgot to add: this is impressive work, nonetheless!
reply