wafer.space Community

Does anyone else have a problem where this Discord server keeps being shown as unread even when there are no new messages? I can normally clear it for a while by looking at the announcements channel, but it happens repeatedly.

Leo Moser (mole99) started a thread. 2026-04-02 11:39 a.m.

Is it just me, or has gitter.im been down? When I try to log in with web, it says "no login methods available" (edited)

6:56 a.m.

I've never had a good experience with Matrix in general :< (edited)

https://www.tomshardware.com/tech-industry/artificial-intelligence/nvidia-says-ai-cuts-10-month-eight-engineer-gpu-design-task-to-overnight-job-company-is-still-a-long-way-from-ai-designing-chips-without-human-input

Nvidia says AI cuts 10-month, eight-engineer GPU design task to ove...

Nvidia reveals how AI speeds up its chip design process.

i think I saw it posted somewhere here, what was the rough cost of a GF180 reticle mask set? Seem to recall it's like 100k something?

Hi! Is anyone of you going to be at Hackaday Europe in May? Apart from Mr. Venn, of course

Pretty much all the TT crew will be there not just Matt

2

BreakingTaps

i think I saw it posted somewhere here, what was the rough cost of a GF180 reticle mask set? Seem to recall it's like 100k something?

Somewhere between $100k and $200k depending on feature set and such.

got it, thanks @Tim 'mithro' Ansell!

https://youtu.be/h6GWikWlAQA

Dr.Semiconductor

Making RAM at Home

Got OSHWA certification on Aegis

@needsadrink|woke I was told you would be interested in my power brick

1

Nice lol

Do people here have opinions on the best way to get started with FPGA/ASIC design? My background is that I have a Master's degree in computer science, but it contained almost no courses in this direction (only one course where we created a simple RISC-V core in VHDL that ran on an FPGA, which was a lot of fun, but I still feel like I know nothing and should start learning from scratch). I love reading about TinyTapeout and wafer.space and what people do there, and would love to learn about this, but every time I try to get started I get lost trying to figure out the best way to start, and lose motivation. I would like to create some designs that I can send to TinyTapeout (and maybe even wafer.space, but that's probably a too large investment for a hobby for me), and they don't need to have a practical purpose. I know the Zero to ASIC course exists, but it's fairly expensive at $700, and I think I'd rather learn from a book/text than videos. But I could afford it and it seems to be the most popular resource aimed at beginners that don't want to study for years before getting something done, does anyone have any opinions if it's worth the price? And are there any other recommended resources? Should I simply find a book about Verilog and read that, and if so, which? (edited)

i've heard good things about this book https://nandland.com/book-getting-started-with-fpga/ but i haven't personally read it

Leon

Do people here have opinions on the best way to get started with FPGA/ASIC design? My background is that I have a Master's degree in computer science, but it contained almost no courses in this direction (only one course where we created a simple RISC-V core in VHDL that ran on an FPGA, which was a lot of fun, but I still feel like I know nothing and should start learning from scratch). I love reading about TinyTapeout and wafer.space and what people do there, and would love to learn about this, but every time I try to get started I get lost trying to figure out the best way to start, and lose motivation. I would like to create some designs that I can send to TinyTapeout (and maybe even wafer.space, but that's probably a too large investment for a hobby for me), and they don't need to have a practical purpose. I know the Zero to ASIC course exists, but it's fairly expensive at $700, and I think I'd rather learn from a book/text than videos. But I could afford it and it seems to be the most popular resource aimed at beginners that don't want to study for years before getting something done, does anyone have any opinions if it's worth the price? And are there any other recommended resources? Should I simply find a book about Verilog and read that, and if so, which? (edited)

Do you have a particular goal? I don't think $700 is particularly expensive for what all is included

there's a verilog/vhdl introduction on that website too

A master's education that would give you the tools for a novel desisgn and tapeout starts around 40k here for reference

Christopher

Do you have a particular goal? I don't think $700 is particularly expensive for what all is included

Not really, other than making something (hopefully interesting/unique) and sending it to TinyTapeout. I'm hoping that as I learn, I come up with ideas.

kris

i've heard good things about this book https://nandland.com/book-getting-started-with-fpga/ but i haven't personally read it

Cool, thanks for the link.

I feel like I might have a strong contender for the weirdest w.s submission :p

I'd start with fpgas personally then, but its up to you. There's a big "missing middle" problem in EE, where there's lots of stuff to start and almost nothing to bridge the gap between the cutting edge and the start

1

8:44 p.m.

(usually because you're taking a tiny chunk of any given problem and spending a huge amount of time drilling down into a narrower and narrower spec)

For me the problem right now is

there's lots of stuff to start

because I don't know what to choose

So I hoped I could get some opinionated recommendations here to help me with this.

Leon

For me the problem right now is

there's lots of stuff to start

because I don't know what to choose

So I hoped I could get some opinionated recommendations here to help me with this.

the classic example is a CPU. I think that's boring, so instead I made my chip play chess.

Lofty

the classic example is a CPU. I think that's boring, so instead I made my chip play chess.

Yeah, I don't really want to do that (though I don't mind following some tutorial/book that uses this as an example for teaching stuff).

Leon

For me the problem right now is

there's lots of stuff to start

because I don't know what to choose

So I hoped I could get some opinionated recommendations here to help me with this.

I learned on AMD's university program. You can get free access to fpgas at fpgas.online

At least a "normal" general purpose CPU

Christopher

I learned on AMD's university program. You can get free access to fpgas at fpgas.online

I assume I need to be enrolled at a university that offers that? Also that sounds too serious for me, I don't want to make this my carreer or anything like that

Leon

I assume I need to be enrolled at a university that offers that? Also that sounds too serious for me, I don't want to make this my carreer or anything like that

No its all free

8:48 p.m.

Vivado is probably the best environment other than the FOSS tools. EDA is pretty brutal

I have a kind of hot take that you do not need an FPGA to develop for hardware

Vivado is what we used in the course I mentioned. I hated it

8:50 p.m.

I'd like to stick with open source tools, assuming that's viable.

that in practice you spend so much time in sim and such that "synthesising the hardware" is kind of the very last step in the process

1

Lofty

that in practice you spend so much time in sim and such that "synthesising the hardware" is kind of the very last step in the process

Yeah but having a blinky is nice for students

Leon

I'd like to stick with open source tools, assuming that's viable.

LibreLane uses FOSS tooling

8:50 p.m.

(including a program I maintain :p)

Leon

I'd like to stick with open source tools, assuming that's viable.

FOSS tools are plenty viable until you hit certain parts where they arent

8:51 p.m.

They'll work fine for a cpu

I think there's good value in using FPGAs for development, to get a feel for running designs on some kind of hardware with quicker turnaround, but definitely agree that getting used to simulation and testing in simulation early is very worthwhile.

8:52 p.m.

one thing you tend to find out quickly is debugging on-the-metal is a lot more involved than debugging software

1

8:53 p.m.

harder to instrument things after the fact / instrument things without impacting timing, etc

Christopher

They'll work fine for a cpu

(I got the impression Leon didn't want to make a CPU...)

make a gpu instead :P

Not a terrible suggestion, but I personally kinda suffer from scope creep when given such an idea

I feel like the open toolchain stuff was feeling very good on the supported lattice parts, and it's been a couple years since I've really poked at it, so I'd assume there are probably some other decently supported families. And of course all the work done for going to actual silicon.

We're a first party toolchain for a vendor now :3

2

it really is fun seeing history repeat itself

8:58 p.m.

I remember development in the mid-90s where every workstation vendor, etc, had a proprietary C compiler you had to wrangle with and wrote gcc off as a toy.

I think we're still being written off as a toy

8:59 p.m.

^^;

oh they won't stop that even after you win

Lofty

(I got the impression Leon didn't want to make a CPU...)

Substitute CPU for CPU-like (or simpler) abstraction then

(edited)

gah. now thinking about arguing with management in the early 2000s because they were convinced we'd be better off paying $7500/seat for the ARM compiler than using gcc

1

9:02 p.m.

"but look at how much better it optimizes!" "you do understand that all the critical path stuff (the vm core, the blitters, etc) are already hand-tuned assembly, right?"

Lofty

We're a first party toolchain for a vendor now :3

CCGM1A1?

namibj

CCGM1A1?

the CCGM1 chips in general

afaik arguably it's just one chip

10:28 p.m.

just multi-chip packaging

eeeeh

but yes

Christopher

A master's education that would give you the tools for a novel desisgn and tapeout starts around 40k here for reference

wafer.space is offering $4k USD and $7k USD slots (which get you 1,000 parts), Tiny Tapeout also offers tiles at ~$200 USD.

Lofty

I have a kind of hot take that you do not need an FPGA to develop for hardware

Plenty of RTL developers don't use FPGAs (except by proxy of their very expensive emulation hardware from proprietary vendors having FPGAs inside them). There are plenty of options around using simulation and such too. I personally think FPGAs are cool because of their speed of iteration. Plus you can always get access to them via https://fpgas.online

Brian Swetland

harder to instrument things after the fact / instrument things without impacting timing, etc

You can get suprisingly quite far with the hardware equivalent of printf debugging (IE blink an LED) --

Brian Swetland

I feel like the open toolchain stuff was feeling very good on the supported lattice parts, and it's been a couple years since I've really poked at it, so I'd assume there are probably some other decently supported families. And of course all the work done for going to actual silicon.

Sadly, the Xilinx FPGA support can still be pretty rough in areas and I don't know anything about the state of Altera parts.

The future might perhaps be Gowin parts

Asking on the proper channel now, I assume there must be some voltnuts around here right?