Hacker News Reader

Comment by askvictor | Hacker News Reader

Yes, it has an arm core. The manufacturer pays the royalty to arm. Very different ball game in my opinion to proprietary codecs. There are no non-arm derivatives. The closest non-arm MCU you'll find is the esp32 but that's completely different, and lacking probably the most useful thing about the rp2040: the PIO.

View on HN

I guess I will avoid that chip and favor those without an ARM block then. Luckily there are many.

But indeed, would be good if those ARM free SOC would also be free-ed from HDMI and MPEG blocks.

Getting rid of ARM cores is only the middle of the journey for better hardware freedom, HDMI and MPEG codecs must be a target too.

Why do you think ARM shouldn’t get paid for the work they did?

Work once, get paid once is the model that most drives efficiency and innovation. If an entity is getting paid without working or innovating (work once, get paid forever), they have no reason to further work or innovate. This hampers the progress of science and the useful arts instead of furthering it.

ARM continues to innovate (“work”). They release new core IPs every year. It’s gotten to the point that companies stopped making their own implementations and just license ARM’s versions.

I suspect it's not that ARM shouldn't get paid, it's that companies producing open source hardware should get paid instead.

If you do all the real work of implementing ARM ISA, you will have to pay royalties anyway (where those are legal though).

That's why it is critical to favor worldwide royalty free ISA like RISC-V. This is not a silver bullet but does a good enough job.

Not to favour but to have an alternative.

And the geopolitical issue of totalitarianism may bite one day by open source nuclear arm design.

It depends. May be paying ok. Not that apple cannot enjoy fruitful result even if somehow it must pay a bit on some core.

Everyone knows what GP is getting at, those opposing GP knows it leads to fragmentation and vendor lock-in, not standardization and open sourcing.

That was my assumption, indeed.

I said to favor chips without ARM ISA cores, for instance with RISC-V ISA cores instead.

there is unfortunately absolutely nothing that get nearly close to the rp2040 price/performance/features...

ESP32-C3 and ESP32-C6 are pretty close. Many products could be made with either. The RMT unit is not as flexible as PIO, but will handle basic I/O cases at least.

Fortunately, we have many RISC-V SOC alternative without ARM cores able to do a more than good enough job?

Name one which offers something similar to the RP2040 state machines (PIO) which I can actually buy (not to speak of the RP2040's very competitive prizing) outside of China?

You mean there is zero RISC-V SOC able to do a good enough job, and that without ARM cores?

You're asking them to prove non-existence? Come on, just drop some references already.

From your GP:

> I guess I will avoid that chip and favor those without an ARM block then. Luckily there are many.

What would your personal top picks be for a use-case like the OP?

Believing that HN people are not able to search online for RISC-C SOCs which do not contain ARM cores, but able to do a good enough job, is ...

I believe it when I see it. Have fun LARPing with your "RISC-C SOCs" (if you even run any...).

We should be able to buy a RP2350 with a hardware fuse burned to disable the ARM cores for good. Ofc, those RP2350 should be cheaper since ARM royalties would not be paid.

Can't wait to the day, when I finally can get a RP2350 for $0.78 instead of the current $0.80. Think of the possibilities of what to do with those savings! I gladly forget everything I ever learned about ARM for such.

Can you expand on why you are so much against ARM?

If I am a chip designer, I don't want to implement an ISA which requires to pay royalties (where those are legal), I would prefer to be free and go RISC-V which does a more than good enough job.

Reply on news.ycombinator.com