<Hypercube3D_>
Hoping to open sauce the boards soon after building it :)
<jrtc27>
you know it's perfectly possible to like riscv without saying things like "arm sux"
<jrtc27>
arm objectively does not suck, you don't need act like a child with this us vs them attitude
<Hypercube3D_>
I make living with arm :) :)
jjido has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
<la_mettrie>
so we need a bit more mature topic for a flame war here
<jrtc27>
technical discussions and debates are fine, childish flamewars are not
<jimwilson>
qwestion, contributing to GNU projects requires a signed paper copyright assignment, contributing to google source requires a click through copyright assignment, contributing to the linux kernel requires a developer certificate of origin, the major FOSS players are all careful about copyrights and patents
<jimwilson>
qwestion, also, there is a lot more money in hardware than FOSS software, so hardware groups are more likely to be sued for infringement, and need to be more careful
<Hypercube3D_>
Average RISC-V microcontrollers about 1/3rd price compared to ARM now.
<Hypercube3D_>
There is no shortage of China made chips in China
<Hypercube3D_>
I think patents are for holding back others now
<Hypercube3D_>
Let them try and fail :)
<Hypercube3D_>
China projected to go into slight oversupply for RISC-V by 2023 before stabilizing chip production.
<Hypercube3D_>
Prices will be under $1
<Hypercube3D_>
Hopefully they sell dice as well - in which case its cheaper still.
<Hypercube3D_>
Several thousand fabless companies caught in a rush to make new chips
<Hypercube3D_>
Can't wait to savour all their chips when it comes out :)
vagrantc has joined #riscv
<Hypercube3D_>
Its about $90,000 and 9 months to get a new chip into production at blazing speed using Taiwan route
<Hypercube3D_>
China route I don't fully know as yet. but China is self sufficient in 28nm tech which is great for making 3mmx3mm
<Hypercube3D_>
3mm x 3mm dice typical of microcontrollers of today with about 64K RAM
<Hypercube3D_>
China will be self sufficient to 14nm end of this year.
<Hypercube3D_>
That would imply the RISC-V will all have to go a up a not in specs otherwise the dice becomes a little too small
<Hypercube3D_>
(up a notch in specs)
<Hypercube3D_>
As geometry shrinks, the number of IO pins shrinking - not enough space for IO pins around a small dice.
<Hypercube3D_>
Making more IO pins cost money and leaves empty silicon to be filled.
<Hypercube3D_>
Filling all that silicon with complex circuits not the answer - e.g. adding PWM controllers with a gazillion features no one will ever want to use
<Hypercube3D_>
Adding 10MB flash doable - but again not clear who might want to use so much of it on a regular basis
Andre_H has joined #riscv
<Hypercube3D_>
Raspberry Pi 2040 (ARM) put two cores in the same chip and a programmable IO controller.
<Hypercube3D_>
Not sure multi-cores are answer if no one is going to use where memory contention is the bottleneck.
<Hypercube3D_>
The programmable IO controller and its relatives like FPGA may be some kind of solution - so you program up extra functions as needed at the hardware level - still not sure its something everyone wants
mahmutov has joined #riscv
aburgess has quit [Ping timeout: 256 seconds]
peeps[zen] has joined #riscv
peepsalot has quit [Ping timeout: 250 seconds]
cwebber has quit [Remote host closed the connection]
<qwestion>
> qwestion, contributing to GNU projects requires a signed paper copyright assignment, contributing to google source requires a click through copyright assignment, contributing to the linux kernel requires a developer certificate of origin, the major FOSS players are all careful about copyrights and patents
<qwestion>
Linux has some members only Ml's iirc( spectre discussion happened hidden for sec reasons ) but i *think* it was made public after embargo?, how many of these discussions stay closed for years+? Hw money and difficulty/patents, explains diff, i think ur right
cwebber has joined #riscv
<Hypercube3D_>
The only thing I can snow flake out right now is that China closes today :(
<Hypercube3D_>
Happy Chinese New Year!! :)
<gordonDrogon>
not until tuesday? Tiger year...
<gordonDrogon>
although I'd not blame them for closing early...
<Hypercube3D_>
Gives time for travel to regions in time for New Year - some factories closed Tuesday.
<gordonDrogon>
well - more time for me to think & ponder about a chip that fits my needs before ordering...
<gordonDrogon>
although I doubt I'll ever find something I like without going down the fpga route, but who knows...
<Hypercube3D_>
I got some Nezha / D1 boards to play with :)
<Hypercube3D_>
I couldn't find any D1s yet
mahmutov has quit [Ping timeout: 250 seconds]
<gordonDrogon>
my needs are simple - probably too simple - RV32IM (F is a bonus) but the key part is RAM - 512KB to 1MB is ideal - and a little bit of flash but anything > 32KB is a bonus.
<gordonDrogon>
there are a few out there that fit, but all have extra "stuff" like the ESP32-C3 which is fine as I could ignore it, but it would be nice to have a "clean" system.
EchelonX has joined #riscv
aburgess has joined #riscv
j0hnny has quit [Ping timeout: 256 seconds]
unmanbearpig has quit [Ping timeout: 256 seconds]
zkrx has quit [Ping timeout: 256 seconds]
ZipCPU has quit [Ping timeout: 256 seconds]
muurkha has quit [Ping timeout: 256 seconds]
muurkha has joined #riscv
ZipCPU has joined #riscv
zkrx has joined #riscv
unmanbearpig has joined #riscv
jjido has joined #riscv
edf0 has quit [Ping timeout: 256 seconds]
dilfridge has quit [Remote host closed the connection]
edf0 has joined #riscv
j0hnny- has joined #riscv
dilfridge has joined #riscv
jjido has quit [Client Quit]
<gordonDrogon>
and tonights task - write RV code to execute the bytecode output by the BCPL compiler for the switch instruction. the bytecode has the most CISCY thing I've personally seen - literally the 1-byte opcode then a binary tree of values + labels to jump to. I think the compiler writer got lazy at that point and left it as a mere implementation detail...
mahmutov has joined #riscv
alMalsamo has quit [Ping timeout: 276 seconds]
alMalsamo has joined #riscv
mahmutov has quit [Ping timeout: 250 seconds]
mahmutov has joined #riscv
<sorear>
sometimes I wonder if we should have an enforceable policy against that sort of 'childish flamewar'
<jrtc27>
I mean we don't have any official policy, right?
<jrtc27>
other than network rules
<la_mettrie>
channels like this seem to work pretty smoothly in general. i wouldn't be worried.
<sorear>
i don't think we needed one until recently
vagrantc has quit [Ping timeout: 240 seconds]
vagrantc has joined #riscv
cousteau has joined #riscv
BOKALDO has quit [Quit: Leaving]
<Hypercube3D_>
A snow flake charter will see ban of 6 char post such as 'Arm :)'
<jrtc27>
if you're trying to piss everyone off then you're doing a great job of it, otherwise I suggest you grow up and stop posting nonsense like that
<Hypercube3D_>
Arm compilers are nasty - because there are so many ARMs around, many don't get attention for declaring registers and bit fields
<Hypercube3D_>
Unless of course its Atmel Studio which is fantastic
<Hypercube3D_>
I'm looking into RISC and I see same mistakes as in ARM
<Hypercube3D_>
CHV307x chips I have compiler working
<Hypercube3D_>
Registers are declared, but not with a union of their bitfields.
<Hypercube3D_>
Is this common? I hope it can stop.
<Hypercube3D_>
I don't fancy programming in binary bitfields :(
EchelonX has quit [Read error: Connection reset by peer]
EchelonX has joined #riscv
<Hypercube3D_>
A random example is this: NVIC->IPRR[((uint32_t)(IRQn) >> 5)] = (1 << ((uint32_t)(IRQn) & 0x1F));
<Hypercube3D_>
What I would like to be able to do is NVIC-IPRR.ClearIRQ = 1;
<Hypercube3D_>
(I mean NVIC->IPRR.ClearIRQ = 1;)
<Hypercube3D_>
Another example: NVIC->SCTLR &= ~(1<<3); //meaningless binary bitfield operation for wait for interrupt
<Hypercube3D_>
It would be more elegant if it can be written like this: NVIC->SCTLR.WFI = 0;
jmdaemon has joined #riscv
mahmutov has quit [Ping timeout: 256 seconds]
gdd has quit [Ping timeout: 250 seconds]
gdd has joined #riscv
jjido has joined #riscv
EchelonX has quit [Read error: Connection reset by peer]
EchelonX has joined #riscv
nvmd has joined #riscv
nvmd has quit [Client Quit]
nvmd has joined #riscv
jmdaemon has quit [Ping timeout: 240 seconds]
jmdaemon has joined #riscv
aerkiaga has joined #riscv
aerkiaga has quit [Remote host closed the connection]
elastic_1 has quit [Quit: elastic_1]
elastic_dog has joined #riscv
<vagrantc>
hrm. the microUSB -> serial out the back of my hifive unmatched seems to have stopped recognizing ... i see pins on the board for uart0 and uart1 ... what voltage do those pins operate at? are there any jumpers to use to get the pins on the header to work?
<vagrantc>
board otherwise appears to be working
<vagrantc>
and does tty/4
<vagrantc>
er, is ttySIF0 uart0 on the board? ... and what voltage do these operate at? ... i see a 1v8 pin on the header ...