sorear changed the topic of #riscv to: RISC-V instruction set architecture | https://riscv.org | Logs: https://libera.irclog.whitequark.org/riscv
chripo has quit [Quit: ""]
chripo has joined #riscv
lagash has quit [Quit: ZNC - https://znc.in]
lagash has joined #riscv
chripo has quit [Client Quit]
chripo has joined #riscv
jack_lsw has joined #riscv
jack_lsw has quit [Client Quit]
mahmutov has quit [Ping timeout: 246 seconds]
Gravis_ has joined #riscv
Gravis has quit [Ping timeout: 256 seconds]
jacklsw has joined #riscv
mahmutov has joined #riscv
mahmutov has quit [Ping timeout: 268 seconds]
pabs3 has quit [Remote host closed the connection]
pabs3 has joined #riscv
k355l3r5yndr0m3 has joined #riscv
k355l3r5yndr0m3 has left #riscv [WeeChat 3.5]
prabhakarlad has quit [Quit: Client closed]
prabhakarlad has joined #riscv
kaph_ has joined #riscv
kaph has quit [Ping timeout: 246 seconds]
KombuchaKip has joined #riscv
BOKALDO has joined #riscv
riff-IRC has quit [Remote host closed the connection]
riff-IRC has joined #riscv
motherfsck has quit [Ping timeout: 246 seconds]
ZipCPU has quit [Ping timeout: 260 seconds]
ZipCPU has joined #riscv
solrize has joined #riscv
<solrize> does anyone have a higher end riscv linux box running?
<drmpeg> Higher end than a HiFive Unmatched?
winterflaw has joined #riscv
<solrize> drmpeg, hi5 unmatched sounds fine, i wonder if you could run "openssl speed rsa2048"
<solrize> thx
<drmpeg> Sure.
<solrize> this is same hw class as a current raspberry pi more or less?
<drmpeg> It's slower than a Pi4.
<solrize> ok maybe pi3 ?
<drmpeg> 1006 2048 bits private RSA's in 10.01s
<drmpeg> 41147 2048 bits public RSA's in 9.99s
<solrize> not bad
<solrize> thanks
<drmpeg> Faster than a Pi3.
<solrize> trying to find my file with pi measurements
<solrize> wow pretty good
<solrize> ah can you also do rsa1024? that's what these old measurements are
<solrize> sorry
davidlt has joined #riscv
<drmpeg> 5897 1024 bits private RSA's in 10.00s
<drmpeg> 131284 1024 bits public RSA's in 10.00s
<drmpeg> Second run.
<drmpeg> 5922 1024 bits private RSA's in 10.00s
<drmpeg> 131726 1024 bits public RSA's in 10.00s
<solrize> close enough. wow that is really pretty fast, i wonder if the setup i tested earlier was misconfigured. also there are no rpi numbers in this file but i think i do have them somewhere
<pabs3> what does pi3/pi4 get?
<drmpeg> I have my Unmatched running at 1.5 GHz.
<solrize> drmpeg, can you try: time echo "scale = 1000; 4*a(1)" | bc -l
<solrize> that computes pi to 1000 decimals
<solrize> pabs i have some x86 and arm servers but no pis in this file for the rsa bench
<solrize> this is a pretty old file and i forgot what was in it
<solrize> sec
<drmpeg> real0m1.266s
<drmpeg> real0m1.262s
<solrize> thanks, can you also do 2000 places?
<solrize> yeah you are beating pi3 slightly
<solrize> and crushing pi0
<drmpeg> real0m7.098s
<solrize> thanks
<solrize> beating pi3 by 2x, but 2x slower than pi4, wonder what was going on
<davidlt> solrize, what's the issue?
<solrize> davidlt, i just see that the pi4 is 4x faster than the pi3 at that test, which is kind of odd
<davidlt> ah
<solrize> the unmatched is about halfway between the two
<solrize> https://i.solrize.net/scaleway.html i updated the file
<drmpeg> We'll see how this Intel/SiFive "Horse Creek" thing pans out.
<drmpeg> Could be a long wait though.
<solrize> i hadn't heard of that
<solrize> i'd like to see a simulation of taniwha's chip
<drmpeg> I guess it's a SiFive P550 on an Intel fab.
<solrize> is that much different than the unmatched?
<drmpeg> It's triple issue and out of order.
<solrize> wow ok
<solrize> are there some non-intel ones out there now?
aburgess_ has quit [Ping timeout: 248 seconds]
<drmpeg> Nope.
pecastro has joined #riscv
Andre_H has joined #riscv
Andre_H has quit [Client Quit]
littlebobeep has quit [Ping timeout: 240 seconds]
littlebobeep has joined #riscv
\dev\ice has quit [Quit: leaving]
ivii has joined #riscv
Sofia has quit [Ping timeout: 240 seconds]
avoidr has quit [Ping timeout: 260 seconds]
avoidr has joined #riscv
Sofia has joined #riscv
drmpeg has quit [Ping timeout: 260 seconds]
jacklsw has quit [Ping timeout: 260 seconds]
jmdaemon has quit [Ping timeout: 260 seconds]
drmpeg has joined #riscv
mps has joined #riscv
kaph_ has quit [Ping timeout: 272 seconds]
<mps> trying to build u-boot 2022.04 for alpine linux we got errors like this `/home/buildozer/aports/main/u-boot/src/u-boot-2022.04/arch/riscv/cpu/start.S:43: Error: unrecognized opcode `csrr a0,0xf14'`
<mps> build machine is x86_64 host with qemu-risc
<drmpeg> mps: I'm having the same issue here.
<mps> drmpeg: also with qemu or native
<drmpeg> With riscv-gnu-toolchain on x86.
<drmpeg> Native on HiFive Unmatched is okay.
seninha has joined #riscv
<mps> drmpeg: aha, thanks for info (I don't have any riscv hardware to check)
<drmpeg> I think it's a specific commit. Lemme see if I can find it.
wingsorc has quit [Quit: Leaving]
<drmpeg> mps: It's the toolchain.
<drmpeg> I have another x86 computer with an older version of riscv-gnu-toolchain, and it works fine.
<mps> drmpeg: thanks for info, very useful to know
<drmpeg> If you wait around for the US guys to wake up, they'll know more about it.
<mps> btw, could you post your gcc build params, especially what -march is
<mps> sure, I'm not in hurry
<drmpeg> Yeah, something changed, but I can't quite remember exactly what it is.
<drmpeg> In fact, I think it may have to do with -march.
<drmpeg> Oh wait, I think it has to do with binutils-2.38.
<mps> yes, also I'm building with this version
<mps> ah, not this bug
<drmpeg> I think it's about this.
<davidlt> GCC and binutils now default a new ISA version
<davidlt> csr stuff is now a separate extension
<mps> I wonder why it is not merged in u-boot
littlebobeep has quit [Quit: leaving]
<davidlt> you need to configure your toolchain to an older (i.e. previous) ISA standard version, or make sure that new extension are enabled by default (remember that RV64GC definition changed)
<davidlt> It was IIRC
littlebobeep has joined #riscv
<davidlt> This only affects things like OpenSBI, kernel, u-boot, grub, etc. that manually set a minimal ISA (i.e. not rv64gc)
<drmpeg> I don't see that patch in u-boot.
<davidlt> it looks like patch missed 2022.04 due to CI failures
<davidlt> seems to fail on rv32 stuff
<mps> also I think so
<davidlt> just pull it in :) it will be in the next release for sure
prabhakarlad has quit [Quit: Client closed]
<drmpeg> This is what you call "tribal knowledge". :)
<mps> I'm trying to build u-boot 2022.04 for next alpine stable release so have to do fix it soon
<davidlt> mps, so for now the easiest thing is to change your toolchain build recipe to go back to the old default
<davidlt> in that case you don't need to patch anything until everything is fixed (again, not too many packages would be affected and patches for them were posted)
<mps> davidlt: not sure changing toolchain will be accepted now that the release is practically frozen
<davidlt> mps, in that case just pull in the patch
GenTooMan has joined #riscv
<davidlt> mps, you might want to add this toolchain ISA change into release notes
<mps> davidlt: yes, I agree it is better solution
<davidlt> mps, do you plan to ship with GCC 12?
<mps> no, with 11 but some patches are pulled from 12
<davidlt> I think, a combo GCC 11 (old ISA) + binutils 2.38 (new ISA) was tested
<mps> davidlt: lets see, building it now (slow with qemu)
<drmpeg> patch works good here.
<davidlt> mps, well QEMU (with the right hardware) is faster compared to Unmatched :)
<drmpeg> yes
<drmpeg> on x86
<mps> davidlt: I think qemu on aarch64 is faster than on x86, but our riscv dev machine is lxc on x86
<mps> i tested riscv with qemu on apple M1 and it is (well) fine
<davidlt> You can use more threads on QEMU which is beneficial on large projects
<davidlt> Also not to mentioned RAM, I think I tested 32G or 64G max on QEMU
<davidlt> and newer QEMU (not sure if already merged) will allow more than 8 vCPUs, but anything above that might not be that beneficial
<mps> RAM is not big problem for most of software in my tests
<dlan> davidlt: hey, I've switched to vanilla kernel(5.17.1) for unmatched board which running great, but do you see any missing bits still?
<davidlt> dlan, define missing bits? like DT nodes, or fixes, or something else?
<drmpeg> I'm only adding one patch for 5.17.
<drmpeg> The LED stuff.
<drmpeg> Gotta have blinky lights!
<mps> I'm waiting for first riscv64 board nearly four months
<drmpeg> Also, you need a recent OpenSBI for the reset stuff to work.
<davidlt> drmpeg, yeah, just pull the latest OpenSBI, there was an issue with WDT (if you have DA9063 enabled and WDT on your kernel)
<drmpeg> On Ubuntu, it's black listed.
<davidlt> drmpeg, you might want to backport 2-3 patches related to PCIe (I believe 2 of them are schedule for v5.18; at least I saw a PR sent out)
<davidlt> drmpeg, for riscv64 or in general?
<mps> drmpeg: yes, I updated openSBI for alpine linux already
<drmpeg> Just for riscv64.
<davidlt> drmpeg, that's because Ubuntu is running at older kernel. v5.17 and most likely v5.16 is all OK.
<mps> davidlt: v5.18-rc1 is released, are these PCIe patches there already
<drmpeg> Ahh, I can try it.
<davidlt> Pull:
<davidlt> Ben Dooks (2):
<davidlt> PCI: fu740: Drop redundant '-gpios' from DT GPIO lookup
<davidlt> PCI: fu740: Force 2.5GT/s for initial device probe
<davidlt> it's already merged on U-Boot side (in 2022.04), but on Linux side it will take ages to be reviewed
<davidlt> basically there is an issue with PCIe, and you might have not noticed it ever depending on your setup.
<davidlt> commit messages explains it better
<mps> davidlt: I see
<drmpeg> That's pretty specific though.
Gravis_ is now known as Gravis
Andre_H has joined #riscv
aerkiaga has joined #riscv
Andre_H has quit [Ping timeout: 248 seconds]
winterflaw has quit [Remote host closed the connection]
davidlt has quit [Ping timeout: 256 seconds]
<mps> drmpeg: thank you for help (and davidlt, who left channel), now u-boot is upgraded for next alpine linux on all arches
Forty-Bot has quit [Ping timeout: 252 seconds]
motherfsck has joined #riscv
jacklsw has joined #riscv
sobkas has joined #riscv
jacklsw has quit [Quit: Back to the real world]
davidlt has joined #riscv
kaph_ has joined #riscv
___nick___ has joined #riscv
winterflaw has joined #riscv
aerkiaga has quit [Remote host closed the connection]
JanC has quit [Read error: Connection reset by peer]
JanC has joined #riscv
sobkas has quit [Remote host closed the connection]
sobkas has joined #riscv
vagrantc has joined #riscv
___nick___ has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
___nick___ has joined #riscv
___nick___ has quit [Client Quit]
___nick___ has joined #riscv
raym has quit [Ping timeout: 268 seconds]
mahmutov has joined #riscv
raym has joined #riscv
adjtm has quit [Read error: Connection reset by peer]
adjtm has joined #riscv
geranim0 has joined #riscv
bob has joined #riscv
bob has quit [Read error: Connection reset by peer]
BOKALDO has quit [Quit: Leaving]
freakazoid333 has quit [Read error: Connection reset by peer]
freakazoid333 has joined #riscv
davidlt has quit [Ping timeout: 256 seconds]
sobkas has quit [Quit: sobkas]
theruran has quit [Quit: Connection closed for inactivity]
___nick___ has quit [Ping timeout: 272 seconds]
mthall_ has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
jmdaemon has joined #riscv
Sofia has quit [Ping timeout: 240 seconds]
EchelonX has quit [Quit: Leaving]
Sofia has joined #riscv
ivii has quit [Remote host closed the connection]
Sofia has quit [Ping timeout: 240 seconds]
wingsorc has joined #riscv
h2t has quit [Quit: ZNC - https://znc.in]
Sofia has joined #riscv
Slide-O-Mix has quit [Ping timeout: 260 seconds]
h2t has joined #riscv
Slide-O-Mix has joined #riscv
wingsorc__ has joined #riscv
wingsorc has quit [Read error: Connection reset by peer]
theruran has joined #riscv
geranim0 has quit [Remote host closed the connection]
winterflaw has quit [Ping timeout: 240 seconds]