tusko has quit [Remote host closed the connection]
tusko has joined #riscv
tusko_ has joined #riscv
tusko has quit [Remote host closed the connection]
tusko_ is now known as tusko
tusko has quit [Remote host closed the connection]
loggervicky has joined #riscv
tusko has joined #riscv
loggervicky has quit [Remote host closed the connection]
loggervicky has joined #riscv
davidlt_ has quit [Ping timeout: 272 seconds]
ntwk has quit [Ping timeout: 250 seconds]
loggervicky has quit [Ping timeout: 244 seconds]
loggervicky has joined #riscv
epony has quit [Ping timeout: 252 seconds]
epony has joined #riscv
loggervicky has quit [Ping timeout: 244 seconds]
ntwk has joined #riscv
tusko has quit [Remote host closed the connection]
tusko has joined #riscv
mps has quit [Quit: Lost terminal]
davidlt_ has joined #riscv
epony has quit [Ping timeout: 252 seconds]
dor has joined #riscv
epony has joined #riscv
BootLayer has joined #riscv
m5zs7k has quit [Ping timeout: 240 seconds]
m5zs7k has joined #riscv
prabhakarlad has joined #riscv
epony has quit [Remote host closed the connection]
epony has joined #riscv
jacklsw has quit [Quit: Back to the real world]
jacklsw has joined #riscv
jacklsw has joined #riscv
dor has quit [Ping timeout: 244 seconds]
jacklsw has quit [Quit: Back to the real world]
jacklsw has joined #riscv
prabhakarlad has quit [Ping timeout: 244 seconds]
ZipCPU_ has joined #riscv
ZipCPU has quit [Ping timeout: 250 seconds]
ZipCPU_ is now known as ZipCPU
qwer has quit [Read error: Connection reset by peer]
qwer has joined #riscv
bauruine has joined #riscv
Trifton2 has joined #riscv
Trifton_ has quit [Ping timeout: 260 seconds]
prabhakarlad has joined #riscv
Andre_H has joined #riscv
epony has quit [Ping timeout: 252 seconds]
epony has joined #riscv
jacklsw has quit [Quit: Back to the real world]
wgrant has quit [Ping timeout: 260 seconds]
qwer has quit [Ping timeout: 240 seconds]
wgrant has joined #riscv
stefanct has quit [Quit: quit]
stefanct has joined #riscv
elastic_dog has quit [Ping timeout: 250 seconds]
elastic_dog has joined #riscv
mps has joined #riscv
jjido has joined #riscv
midnight has quit [Ping timeout: 268 seconds]
dor has joined #riscv
midnight has joined #riscv
jjido has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
epony has quit [Ping timeout: 252 seconds]
epony has joined #riscv
frkzoid has joined #riscv
freakazoid332 has quit [Ping timeout: 255 seconds]
freakazoid332 has joined #riscv
frkzoid has quit [Ping timeout: 244 seconds]
jjido has joined #riscv
jjido has quit [Client Quit]
Trifton_ has joined #riscv
Trifton2 has quit [Ping timeout: 250 seconds]
loggervicky has joined #riscv
eroux has quit [Ping timeout: 250 seconds]
enthusi_ is now known as enthusi
eroux has joined #riscv
rneese has joined #riscv
rneese has quit [Client Quit]
awita has joined #riscv
Trifton_ has quit [Quit: Error: no route to host]
Raito_Bezarius has quit [Ping timeout: 250 seconds]
epony has quit [Ping timeout: 252 seconds]
jobol has joined #riscv
epony has joined #riscv
Raito_Bezarius has joined #riscv
Raito_Bezarius has quit [Max SendQ exceeded]
awita has quit [Ping timeout: 244 seconds]
Raito_Bezarius has joined #riscv
awita has joined #riscv
Raito_Bezarius has quit [Max SendQ exceeded]
Raito_Bezarius has joined #riscv
Raito_Bezarius has quit [Max SendQ exceeded]
rvalles has quit [Ping timeout: 255 seconds]
jmdaemon has quit [Ping timeout: 255 seconds]
Raito_Bezarius has joined #riscv
bauruine has quit [Remote host closed the connection]
pedja has joined #riscv
prabhakarlad has quit [Quit: Client closed]
jacklsw has joined #riscv
Andre_H has quit [Ping timeout: 244 seconds]
Andre_H has joined #riscv
epony has quit [Ping timeout: 252 seconds]
epony has joined #riscv
bjoto has quit [Quit: WeeChat 3.6]
bjoto has joined #riscv
loggervicky has quit [Ping timeout: 244 seconds]
GenTooMan has quit [Ping timeout: 250 seconds]
crabbedhaloablut has quit [Remote host closed the connection]
crabbedhaloablut has joined #riscv
loggervicky has joined #riscv
foton has quit [Quit: %Bye, bye, ...%]
foton has joined #riscv
aburgess_ has joined #riscv
GenTooMan has joined #riscv
aburgess has quit [Ping timeout: 250 seconds]
loggervicky has quit [Ping timeout: 244 seconds]
loggervicky has joined #riscv
foton has quit [Quit: %Bye, bye, ...%]
vagrantc has joined #riscv
foton has joined #riscv
___nick___ has joined #riscv
___nick___ has quit [Client Quit]
___nick___ has joined #riscv
loggervicky has quit [Ping timeout: 244 seconds]
loggervicky has joined #riscv
jacklsw has quit [Read error: Connection reset by peer]
loggervicky has quit [Ping timeout: 244 seconds]
GenTooMan has quit [Ping timeout: 244 seconds]
<muurkha>
from solrize's link: "Now Canonical has announced the launch of Ubuntu builds for the SiPeed LicheeRV, a tiny single-board computer with a RISC-V chip and a $17 starting price." super exciting!
<muurkha>
lychees are much tastier than raspberries anyway
zjason` has quit [Read error: Connection reset by peer]
josuah has quit [Ping timeout: 260 seconds]
zjason` has joined #riscv
<conchuod>
palmer: it's easier to reply here I think since I'd image there'll be back and forth.. try gcc-12 & binutils 2.35
josuah has joined #riscv
<palmer>
post-2.38
jjido has joined #riscv
<palmer>
sorry that wasn't specufic
<bjdooks>
muurkha: anything not runnign slack is a good idea
<palmer>
my argument was that 2.38 should be the newest required binutils version for any of the ISA string ELF note stuff, that and after we just stopped emitting errors for that field
<conchuod>
palmer: the problem is that the old bintuils doesnt understand what zicbom is
<palmer>
ya, that makes sense
<conchuod>
so we need to disable zicbom if <2.38
<conchuod>
which is what my patch was doing
<palmer>
ya, that makes sense. but IIRC the code was checking for 2.39
<palmer>
did I just misunderstand something?
<conchuod>
I picked the versions that the support is added in
<conchuod>
if we are going to set a version, we may as well pick the right ones imo
<palmer>
I guess "support" is kind of vague here: the linker doesn't do anything meaningful for those fields, even when it understands that Zicbom is an extension
<palmer>
that happened in 2.38, I don't think 2.39 really does anything different
<palmer>
(the assembler will understand Zicbom, but that's a bit different)
<conchuod>
so do you want me to resubmit it w/ 2.38 as the min versions (for binutils, lld is different)
<conchuod>
that way the rationale behind the version number would be different for binutils and lld
<palmer>
I just wanted to make sure that was actually the minimum version, not that I'd missed something when disabling the bits
<palmer>
it's different for LD and GAS ;)
<conchuod>
its all greek to me
<palmer>
OK, well, IMO the minimum supported LD for Zicbom is 2.38 (as opposed to 2.39)
<conchuod>
you're the boss!
<palmer>
so if you're OK with me squashing that in I'll do so, I'm just not sure anyone's tested it
<conchuod>
idk I don't have 2.38 here, lemme go get it
<conchuod>
Andrews got a better solution for this stuff but it depends on a bunch of refactoring so hopefully we can just backport this to fix the build and replace it with his solution once that lands
<palmer>
ya, I think having the simple fix for now is best
<palmer>
no sense in having a bunch of random build failures
<solrize>
muurkha, that board sounded comparable to a rpi zero original version?
<solrize>
i mean it's progress
<conchuod>
i'll try 2.38 then Palmer & then reply to the mail
<palmer>
thanks
<muurkha>
solrize: I was thinking a RPi original version
<solrize>
ok similar i guess
crabbedhaloablut has quit [Remote host closed the connection]
jjido has joined #riscv
<solrize>
idr whether it has usb host ports and ethernet
<muurkha>
I think the former at least
<solrize>
cool
crabbedhaloablut has joined #riscv
<conchuod>
palmer: ye 23800 is grand
<solrize>
i guess that is like the rpi model a
<muurkha>
yeah
<muurkha>
which is pretty impressive
vagrantc has quit [Quit: leaving]
bjoto has quit [Ping timeout: 244 seconds]
<muurkha>
I wonder if Raspberry Pi will come out with a new model next year with a custom SoC with one of the RP2040 peripheral processors bolted onto the side of a XuanTie
<muurkha>
if they don't die next month they have enough volume to justify a full-custom chip, and Honey Badger don't care, they open-sourced the design on purpose
<solrize>
what who? what xuantie and has rpi ever done anything with one? honey badger ??? rpi (or xuantie) might die next month? i don't understand any of that
<solrize>
do you mean for a pi 5? yeah they might include those peripheral processors but they also seem hot on neural processors gak
<solrize>
they don't even have a published api for the gpu in the existing soc right?
<muurkha>
Raspberry Pi is in a serious conflict with Broadcom right now, and they are having a very difficult time with fulfillment as well
<solrize>
oh interesting i didn't know that there was a bcm conflict
<conchuod>
im going to make a soc in my shed and itll be the next rpi
<muurkha>
that's the rumor anyway
<muurkha>
wrt GPU, the VideoCore IV instruction set is fully documented, but the firmware blob they load onto it is not
<solrize>
i thought they were deep in bed with bcm since forever and that was why their product always had cringy bits
<solrize>
who is honey badger?
<muurkha>
PingTouGe is Chinese for "honey badger"
<solrize>
um ok who is pingtouge ?
<muurkha>
that's the AliBaba subset that developed the XuanTie core in the AllWinner D1 in the Lichee board you linked
<muurkha>
yeah. I mean I think it's more likely that they'll use ARM if they do a full-custom chip
<muurkha>
*AliBaba subsidiary
<solrize>
it would be interesting though doesn't sound like them
<muurkha>
well, apparently the pioasm thing was original to them, though of course "channel processors" or "peripheral processors" have been developed many times before
<solrize>
it is kind of neat but stupidly limited
<muurkha>
how would you make it less limited?
<solrize>
redesign it so it is can have more than 32 instructions
<muurkha>
solrize: what kinds of things have you done with it where that was a limitation?
jmdaemon has joined #riscv
<solrize>
i haven't done anything with it but there are obvs things that it would be good for with more than 32 instructions, that's why the beaglebone had much more powerful pru's. there wasn't a chip area constraint either. the rp2040 die is almost all memory, even the 2nd cortex m0 core is tiny and was probably an afterthought since they could easily afford the space
<muurkha>
what kind of things?
<solrize>
cryptography comes to mind right away
<solrize>
all the random weird protocols for things like video cameras
<solrize>
what is nicehash? another crypto currency?
<muurkha>
hmm, you'd like to run cryptography on a glorified PWM driver because it has constant time execution?
<muurkha>
I'd think it would work fine for video camera protocols in its current form but I haven't actually tried it
<muurkha>
nicehash?
tusko has quit [Remote host closed the connection]
tusko has joined #riscv
<solrize>
idk what nicehash is either. yeah, crypto, also things like error correction codes, and i think motor control came up here yesterday. idr if that hdmi hack used all 32 insn slots. your description of it as a glorified pwm driver is pretty accurate in its current form
<muurkha>
why would you need ECC to run in constant time and have ultra-low latency though?
<solrize>
bluetooth audio or wifi video for example
<muurkha>
but you can do the ECC on the main processor on the other end of a FIFO
<solrize>
more latency
<muurkha>
the main processor can totally handle doing ECC at submillisecond levels I'd think?
<muurkha>
I mean the zen of pioasm (as I understand it, I haven't actually used it either) is that it allows you to program state machines for things like CAN and I²C that need guaranteed submicrosecond response times that are usually obtained with custom circuitry
<muurkha>
or precisely timed pulses for things like NeoPixels
<solrize>
well you'd need some fancy buffering between the main proc but maybe it is doable. idr whether dv25 has a large block ecc layer like cd-rom has
<muurkha>
not that it has a lot of computing horsepower, it doesn't, it just makes it easy to get very deterministic timing
<solrize>
neopixels are fairly relaxed, CAN idk about but if it can do CAN that is encouraging
<muurkha>
and it has DMA to the main processor memory to make fancy buffering easy
<solrize>
the fancy buffering costs instruction space
<muurkha>
I wonder if anyone has done CAN with it or if there's some processing
<solrize>
i had the idea of using it as a proximity token but it is slower than i would like for that
<muurkha>
uh, I meant "or if there's some problem"
<muurkha>
it does seem like it would be hard to drive the recessive and dominant voltages directly with GPIOs I guess but couldn't you use some resistors instead of a chip?
<solrize>
realtime?
<solrize>
oh hmm
<solrize>
drive current
<solrize>
you'd need amplifiers
<solrize>
but also idk what is in the protocol
tusko has quit [Remote host closed the connection]
<solrize>
keyboard scanning is another thing but i guess the main cpu is fast enough for that
<muurkha>
at the point that you're using amplifiers it makes sense to use a transceiver chip
<muurkha>
yeah, keyboard scanning doesn't actually have to be fast at all, you can do it as a background task that pauses for up to a few milliseconds under load
<solrize>
i have a keyboard with crappy firmware that i can't change, but maybe i'd like to intercept the scan lines which are being lit by the existing processor