08:57 <nickoe> mithro: See azonenberg with his fancy multi fpga scope project in #kicad https://www.antikernel.net/temp/maxwell-40.png

17:34 <Lofty> Okay, so, the SkyWater readme seems to imply you're meant to use it with OpenROAD

17:34 <Lofty> Except OpenROAD itself seems...very difficult to install/use

17:37 <prpplague> Lofty: there is a fork of openROAD that is planned to be released with specific support for the SKY130 PDK

17:37 <prpplague> Lofty: called OpenLANE

17:37 <Lofty> ...Oh, great

17:37 <prpplague> Lofty: that will have much better installation, support, and functionality for the openPDK

17:38 <Lofty> Frankly just "much better installation, support and functionality" would be all I need

17:38 <Lofty> OpenROAD doesn't even bundle KLayout

17:38 <Lofty> Or Lemon

17:38 <prpplague> Lofty: yea you probably won't see that with openlane either

17:39 <prpplague> Lofty: mainly because people might want to use something else for DRC such as Magic

17:39 <Lofty> Are Google trying to save money on fabbing by making the PDK as difficult to develop with as possible or something?

17:40 <Lofty> So, anyway

17:41 <Lofty> I'm gonna put aside my frustration with OpenROAD

17:41 <Lofty> So, I hit make, downloaded a bunch of stuff

17:41 <Lofty> ...Where are the liberty cell libraries?

17:42 <mithro> Lofty: Type `make timing`to get the liberty files

17:42 <Lofty> That's...just a bunch of AssertionErrors

17:43 <Lofty> https://gist.github.com/Ravenslofty/23b5b7f43f23a0996f9dad2b4c9b1fe8

17:45 <mithro> Lofty: `git submodule update --init`

17:47 <mithro> Lofty: Also, snide comments like "Are Google trying to save money on fabbing by making the PDK as difficult to develop with as possible or something?" are not appreciated. We would like everything to be easier but it requires work and there is only so much time and effort available. FYI -- Google is paying the same amount if there ends up being 1 design for the full 40.

17:48 <mithro> Most of these tools only have a small number of users at the moment and hence are going to have a lot of bugs and issues, hence why we are trying to make it more accessible to everyone.

17:52 <Lofty> Payment for ASICs is per-wafer, right?

17:56 <Lofty> Of course, there's only so much time and effort for things, but there are a lot of little things that can be done to make things easier for beginners

17:58 <mithro> Lofty: per mask set mainly

17:58 <Lofty> In the time you spent telling me off for being snide, you could have ignored the remark and instead taken the time to modify the readme to mention that the Git repository uses submodules, hm?

17:59 <mithro> Lofty: Would love pull requests to help do these things! Just getting these things out the door have been a long hard slog (for many mainly non-technical reasons) which hasn't left much time in improving things for beginners.

18:02 <Lofty> prpplague: yes, I'm being dickish, I realise.

18:04 <Lofty> https://github.com/google/skywater-pdk/issues/76 <-- but I'm not even the first to encounter this

19:10 <prpplague> mithro: karimyaghmour shared this interesting article from 2003 about photomask pricing and tech - https://www.edn.com/sticker-shock-for-photomasks/

19:12 <karimyaghmour> Yeah, 130nm masks back then would've cost $600K :)

19:32 <mithro> I can tell you that it doesn't cost $600k USD for a 130nm mask set today

19:36 <Lofty> Now that I'm a lot less annoyed at the world

19:37 <Lofty> I'd like to apologise for being a dick to people earlier

19:39 <mithro> Lofty: Thank you for the apology! Hopefully we can fix the issues you identified as quickly as possible

19:39 <karimyaghmour> mithro: I bet :) I was reading about EUV lithography a while back and that seemed pretty esoteric: https://spectrum.ieee.org/semiconductors/devices/leading-chipmakers-eye-euv-lithography-to-save-moores-law

19:41 <mithro> karimyaghmour: Mask sets for 7nm are individually actually quite cheap, you just need a huge number of them and they become consumables due to the high energies

19:43 <karimyaghmour> mithro: interesting. I didn't know that. thx for this.

19:43 <mithro> I should also rephrase that -- cheap is relative :-P

19:44 <karimyaghmour> yup, well the NRE for the chips themselves is (in as far as I understand it) the far larger cost. And, fwiw, that's the one thing that I think open source can most impact.

19:44 <mithro> "cheap" compared to cost of an engineers salary in the bay area maybe....

19:47 <karimyaghmour> Former HP exec responsible for HPUX was saying at RISC V summit in December that his budget for his HPUX team was $200-300M whereas his Linux team cost him $2-3M. That's a significant cost saving for a paradigm shift.

20:27 <mithro> karimyaghmour: Yeap