04:40 <cyrozap> Heh, so apparently my RV32I-on-8051 emulator isn't as fast as I thought, and I was just running it on a _very_ fast 8051 (ASMedia xHCI USB host controller). I ported my RISC-V code over to an old RTD2660/RTD2662 board I had laying around (LCD controller chip), and wow, it's _slow_. Like, the serial port's running at 115200 8N1, but I can see each individual character being printed on the console.

04:40 <cyrozap> Looking at it with a logic analyzer, the "effective" bitrate is closer to 2400 bps. But of course, this RTD266x chip is running at 27 MHz, and it's using Synopsys's DW8051 which runs at a minimum of 4 clock cycles per instruction. In contrast, the ASMedia ASM1142 runs at 78 MHz and most likely executes single-byte instructions in a single clock cycle.

07:04 <TD-Linux> obviously your next goal should be to hunt down the fastest superscalar 8051 implementation you can find

07:47 <cyrozap> Yeah, I was half-jokingly considering designing one a while ago. Maybe I should actually do it, and then submit it to that Efabless/Skywater/Google free ASIC thing. Though it'd have to be designed, laid out, and tested within a week, since the deadline's the 30th, lol :P

07:55 <Ultrasauce> sounds like a plan to me

08:09 <cyrozap> Just in case it wasn't clear, that was a joke--there's no way I'd even finish the design in a week, let alone write all the testbenches for it and figure out the ASIC toolchain on top of that.

08:14 <cyrozap> I'm half tempted to submit something stuipd, like a PWM + state machine that just continuously plays a meme song.

08:19 <implr> iirc they wanted to do ~2 runs per year? or something like that

08:19 <implr> so you can always shoot for the next one

08:27 <cyrozap> But by then most people will have their projects ready, and so there will be a lot more competition for space. :(

08:32 <cyrozap> And I think I'm sort of burnt out on hardware design. The actual design part is fun, but writing all the testbenches to make sure things work is such a drag and sucks all the joy out of the process for me, especially when I can't for the life of me figure out what I did wrong.

08:39 <TD-Linux> I know that feel, that's why so many of my projects take forever. it's not so bad once I can get myself in the zone

08:40 <TD-Linux> doing a superscalar cpu would be fun, but also I really don't know anything about the design. I'd like to design a fast 68k core

08:40 <TD-Linux> I'm currently clocking up ao68000 on a ecp5. it's a pretty cute design, with a java microcode generator

08:41 <TD-Linux> the microcode features branches and a call stack

09:37 <TD-Linux> love the new gtkwave that makes my eyes bleed at hidpi

09:40 <TD-Linux> oh it's because my distro rebuilt it with gtk3

12:27 <TD-Linux> well I fixed it https://github.com/gtkwave/gtkwave/pull/36

13:56 <azonenberg> cyrozap: yeah i would have loved to get something on this shuttle

13:56 <azonenberg> but i just dont have the time

13:56 <azonenberg> i have my hands in too many places already

21:59 <pie_> im not saying it isnt, but is riscv even a good architecture?

22:00 <sorear> that sounds like a value judgement

22:04 <sorear> like, what are you trying to do with it

22:04 <awygle> do Yosys and nextpnr work on Mac?

22:04 <awygle> (not the new weird ones)

22:04 <daveshah> Yes

22:05 <awygle> Cool, thank you