jacklsw has quit [Read error: Connection reset by peer]
vagrantc has quit [Quit: leaving]
BOKALDO has joined #riscv
jacklsw has joined #riscv
riff-IRC has quit [Remote host closed the connection]
riff-IRC has joined #riscv
huseyinkozan has joined #riscv
winterflaw has joined #riscv
jjido has joined #riscv
nun has quit [Ping timeout: 252 seconds]
nun has joined #riscv
alMalsamo has quit [Ping timeout: 276 seconds]
_inky has quit [Ping timeout: 260 seconds]
ivii has joined #riscv
alMalsamo has joined #riscv
NishanthMenon has quit [Ping timeout: 268 seconds]
llamp[m] has quit [Ping timeout: 268 seconds]
gruetzkopf has quit [Ping timeout: 268 seconds]
Starfoxxes has quit [Ping timeout: 268 seconds]
rsalveti has quit [Ping timeout: 268 seconds]
NishanthMenon has joined #riscv
gruetzkopf has joined #riscv
rsalveti has joined #riscv
<la_mettrie>
it seems that StarFive's VisionFive SoC I ordered from allnetchina.cn on 5th of January is still in a "confirmed" state (instead of "shipped"). i wonder if i should try to cancel it and think about something else.
psydroid has joined #riscv
TianruiWei[m] has joined #riscv
Andre_H has joined #riscv
llamp[m] has joined #riscv
Starfoxxes has joined #riscv
Andre_H has quit [Client Quit]
Andre_H has joined #riscv
_inky has joined #riscv
Andre_H has quit [Client Quit]
KombuchaKip has quit [Ping timeout: 256 seconds]
Hypercube3D has joined #riscv
Hypercube3D_ has joined #riscv
<Hypercube3D_>
qwestion: I'm working through what is available in China for RISC-V searching shopping sites and then checking out their developer IDEs
<Hypercube3D_>
windoze only IDEs I give up immediately. those with linux I try out. The more interesting ones are those with IDE's for linux on ARM and RISC-V
<josuah>
Hypercube3D_: I missed the question in that, but I guess you are after RISC-V that can be driven nicely from Linux?
<josuah>
I am currently wrapping my head around GD32VF103's USART register to get it working
<josuah>
I got it blinking though :)
<josuah>
I am on my way for getting some code for a few boards so I can run "make" and "make flash" and get the board to blink
<Hypercube3D_>
Definitely driving anything from Linux is nice because you can pull in scripts and other IDEs (like GAMBAS) to make extremely powerful PC integrated systems
<josuah>
there is also an ESP32-C3, never played with it
<Hypercube3D_>
I got 100pcs of ESP32-C3 to make new boards and some dev boards to try out the dev environment - I will be playing with those in a few days.
<josuah>
there is also the FE310 by SiFive, but not a lot of peripheral on it
mithro has quit [Ping timeout: 245 seconds]
<Hypercube3D_>
I can't seem to order any SiFive chips - not sure why they don't sell to stockists
<josuah>
I bought mine from digikey
<josuah>
but maybe digikey bought the few they had as a regular custommers, not as distributors
mithro has joined #riscv
<josuah>
still has RTC, PWM, I2C, SPI, UART, GPIO though
NishanthMenon has quit [Ping timeout: 252 seconds]
<Hypercube3D_>
i never could order from anywhere SiFive chips - now the shortages kill that idea 100%
<Hypercube3D_>
They are mad not to order chips and start selling.
<Hypercube3D_>
I'll order a few boards (expensive for what they are) - but I also need at least 100pcs chips to build boards - otherwise it gets binned eventually :(
<josuah>
the price is still high for an MCU, maybe because it is not massively produced yet
jimwilson has quit [Remote host closed the connection]
jimwilson has joined #riscv
<Hypercube3D_>
doesn't sound right - about $90,000 to produce a chip and 9 months if done flat out speed
<Hypercube3D_>
Each wafer may cost $3000 (depending on process/wafer size) and likely yield 1000 to 10,000 chips costing $0.2 per chip
<Hypercube3D_>
After packaging it cost $0.4
<Hypercube3D_>
A lot of instant profit if done right - no need to sell at high price and kill your market
<josuah>
are you planning to build a Digital Quantum Computer (learned the word today, from your project) prototype out of MCUs?
<Hypercube3D_>
Sadly it is my duty to dominate the world :)
jjido has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
<josuah>
that sounds like a weird strategy indeed
<josuah>
maybe it did put a lot of debt onto the authors
<gordonDrogon>
do these boards/chips just look like USB storage to the host?
<josuah>
gordonDrogon: maybe it is possible to do it like what the RP2040 does, but at least for the GD32VF103, the easiest way to flash it was through USB, yes, but with DFU protocol
<gordonDrogon>
I'd like something I can preferably just drop a binary image into and have it load and run it ...
<josuah>
with dfu-util
<gordonDrogon>
ah, ok. I'll look at that (linux friendly I hope)
<josuah>
cp $file $target is barely any harder than ./flash-chip.sh (or in my case "make flash")
<josuah>
gordonDrogon: I am flashing it from OpenBSD without trouble :D
<gordonDrogon>
once I find a board with enough RAM that is (which for me is 512KB to 1MB)
<gordonDrogon>
I'm not after Linux though - RV32IM although I can do without M at a pinch.
<gordonDrogon>
I have my own retro style OS written in a very retro language; BCPL... (just call me weird if you like! :)
<gordonDrogon>
I currently have most of it running in an emaulator written in BCPL running on a retro 8/16 bit CPU. it's very very slow, but makes it ok for testing/bringup.
<gordonDrogon>
just a fun/personal project.
<gordonDrogon>
ESP32-C3 is about the best so-far, but I've started to tentatively look at my own FPGA solution for it.
<josuah>
they ported their virtual machine (in the java sense, but a nimble and simple one) to various platforms
<gordonDrogon>
that's quite interesting - sort of in the same idea as pico-8 - ish...
<gordonDrogon>
BCPL runs on-top of a bytecode VM too, so my task of getting BCPL going on RV is essentially porting/re-writing my existing bytecodeVM from the 65816 world to the RV world.
<gordonDrogon>
My aims are for 'self hosting' - which I have to a good degree with my existing system - it's just a bit slow.
<josuah>
but building a VM for a language sounds an extremely good use of an FPGA!
<josuah>
also a fun project
<gordonDrogon>
ultimately it would be good if I could build a CPU (in an FPGA) that executed the byecode directly but that may be beyond my ken (for now, anyway)
<gordonDrogon>
however RV ASM is making it remarkably easy - far (far!) easier than writing it in '816 assembler anyway.
<gordonDrogon>
although arguably I did the 'hard work' in '816 code, so translating it to RV is not that hard.
<gordonDrogon>
and despite being 32-bit, the RV code is more code-dense than the 65826 code is, and clock tick for clock tick much faster.
<josuah>
gordonDrogon: and you would have built a compiler!
<gordonDrogon>
no - that was already written
<gordonDrogon>
and I had a binary for it too, so the bootstrap paradox for that had been 'solved' many years ago. BCPL dates for about 1966..
<gordonDrogon>
BCPL was used to write B which was used to write C ...
<gordonDrogon>
I suspect the very first BCPL compiler was written in Algo.
<gordonDrogon>
SpaceCoaster, yes - BCPL has traditionally compiled into an intermediate code there are several variants and the latest (for a couple of decades or more) is a bytecode
<gordonDrogon>
I think it would be relativly easy to "JIT" compile the bytecode into native on the fly too - although from a niave point of view just copy in the snippets I've written for each byte instruction...
<gordonDrogon>
it would save a few cycles dispatch per opcode.
<gordonDrogon>
there is another code the compiler can output which is supposedly easier to translate to a native format but I've not looked at it.
winterflaw has quit [Ping timeout: 276 seconds]
jeancf has quit [Ping timeout: 256 seconds]
<SpaceCoaster>
Is this your first foray into FPGAs?
aerkiaga has joined #riscv
<gordonDrogon>
yes
<gordonDrogon>
possibly jumping in at the deep end, but there are cores that seem to be ready to go, according to the somewhat scant documentation for it.
<gordonDrogon>
and I have quie enjoyed writing RV asm code. it's fasr easier than the old 65'816..
<SpaceCoaster>
It was a long time since I read an chip manual and probably the first time I paid for one!
<SpaceCoaster>
I think the draft or most of the contents are online. It explains the rational for the choices.
<gordonDrogon>
I'd done a lot of sparc & i860 many years ago so it felt familiar - my plan to get to grips with it was to write my own RISC-V emulator - which I did although even running C on it, I found I only needed 22 base instructions to run most of the simple code I could use to test it with.
<gordonDrogon>
I've not counted but I don't think I'm using much more than dozen base instructions in my VM interpreter....
<gordonDrogon>
the plethora of registers is what's good though - I can keep all state in them. no need for ram to run the bytecode interpreter at all.
<SpaceCoaster>
Looking forward to hearing about it. Might take a while to arrive with the Chinese new year etc.
jjido has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]