sorear changed the topic of #riscv to: RISC-V instruction set architecture | https://riscv.org | Logs: https://libera.irclog.whitequark.org/riscv | Matrix: #riscv:catircservices.org
sevan has quit [Ping timeout: 272 seconds]
sevan has joined #riscv
prabhakalad has quit [Quit: Konversation terminated!]
prabhakalad has joined #riscv
sevan has quit [Ping timeout: 268 seconds]
fuel has quit [Remote host closed the connection]
fuel has joined #riscv
sevan has joined #riscv
n_crm has quit [Ping timeout: 246 seconds]
naoki has quit [Ping timeout: 264 seconds]
mengzhuo has joined #riscv
test924 has joined #riscv
jacklsw has joined #riscv
naoki has joined #riscv
naoki has quit [Client Quit]
naoki has joined #riscv
BootLayer has joined #riscv
pabs3 has quit [Read error: Connection reset by peer]
[exa] has quit [Ping timeout: 260 seconds]
pabs3 has joined #riscv
heat has quit [Ping timeout: 264 seconds]
fuwei has joined #riscv
eshep_ has joined #riscv
eshep has quit [Ping timeout: 256 seconds]
eshep has joined #riscv
eshep_ has quit [Ping timeout: 264 seconds]
mengzhuo has quit [Ping timeout: 268 seconds]
BootLayer has quit [Quit: Leaving]
q66 has quit [Ping timeout: 268 seconds]
eshep has quit [Ping timeout: 268 seconds]
eshep_ has joined #riscv
coldfeet has joined #riscv
stefanct has quit [Ping timeout: 252 seconds]
stefanct has joined #riscv
coldfeet has quit [Remote host closed the connection]
seasharp has quit [Ping timeout: 255 seconds]
seasharp has joined #riscv
Revy[m] has quit [Quit: Idle timeout reached: 172800s]
mengzhuo has joined #riscv
atka has joined #riscv
atka has quit [Client Quit]
atka has joined #riscv
alperak has joined #riscv
ldevulder has joined #riscv
eshep has joined #riscv
eshep_ has quit [Ping timeout: 246 seconds]
prabhakalad has quit [Quit: Konversation terminated!]
Noisytoot has quit [Remote host closed the connection]
Noisytoot has joined #riscv
TMM has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
TMM has joined #riscv
prabhakalad has joined #riscv
ldevulder has quit [Quit: Leaving]
ldevulder has joined #riscv
cleger has joined #riscv
danilogondolfo has joined #riscv
ncopa has quit [Quit: Alpine Linux, the security-oriented, lightweight Linux distribution]
clandmeter has quit [Quit: Alpine Linux, the security-oriented, lightweight Linux distribution]
clandmeter has joined #riscv
ncopa has joined #riscv
danielinux has quit [Ping timeout: 272 seconds]
danielinux has joined #riscv
q66 has joined #riscv
[exa] has joined #riscv
mengzhuo has quit [Ping timeout: 255 seconds]
ema has quit [Quit: leaving]
sevan has quit [Changing host]
sevan has joined #riscv
ema has joined #riscv
eshep has quit [Read error: Connection reset by peer]
eshep has joined #riscv
eshep_ has joined #riscv
Starfoxxes has joined #riscv
eshep has quit [Ping timeout: 264 seconds]
cronos has quit [Quit: ZNC - https://znc.in]
cronos has joined #riscv
cronos has quit [Client Quit]
cronos has joined #riscv
eshep_ has quit [Ping timeout: 240 seconds]
bjoto has joined #riscv
bjoto has quit [Remote host closed the connection]
eshep has joined #riscv
bjoto has joined #riscv
coldfeet has joined #riscv
coldfeet has quit [Remote host closed the connection]
jacklsw has quit [Quit: Back to the real world]
heat has joined #riscv
la_mettrie has quit [Ping timeout: 240 seconds]
<conchuod> courmisch: Did you ever run linux on the small core on the k230?
naoki has quit [Quit: naoki]
Stat_headcrabed has joined #riscv
ncopa has quit [Changing host]
ncopa has joined #riscv
unlord has quit [Ping timeout: 268 seconds]
unlord has joined #riscv
Nixkernal has joined #riscv
Stat_headcrabed has quit [Quit: Stat_headcrabed]
<conchuod> Esmil: have you changed your mind on this? https://lore.kernel.org/all/20240627115236.618-1-naoki@milkv.io/
<conchuod> Hmm either the fastest response to a thread ever, or we both looked at it at the same time...
<Esmil> conchuod: yeah, sorry. catching up after the summit
* bjdooks is happy to have gotten back to UK
<conchuod> At least you have a reason for being behind Esmil :)
<conchuod> Mines just being lazy
felixonmars has quit [Remote host closed the connection]
felixonmars has joined #riscv
<courmisch> conchuod: yes
<courmisch> conchuod: that's what the vendor SDK does , FWIW
<conchuod> courmisch: if it's not too much trouble, what's proc cpuinfo say about that CPU?
<courmisch> conchuod: it lies about supporting V, so it's useless
<conchuod> I'm less interested in lies about V and more about the marchid/mimpid values. Do you recall if they were 0?
<courmisch> haven't looked at them since the day I got the device, 9 months ago
<conchuod> Fair enough!
fuwei has quit [Quit: Konversation terminated!]
fuwei has joined #riscv
greaser|q has quit [Quit: HYDRA IRC LOL]
greaser|q has joined #riscv
Finde has quit [Read error: Connection reset by peer]
Finde has joined #riscv
Amanieu has quit []
Amanieu has joined #riscv
la_mettrie has joined #riscv
<palmer> IIUC it's the same as the K1, just clocked faster
<courmisch> and more RAM
<ja_02> courmisch so ill drop [PATCH v3 4/8] RISC-V: Check Zicclsm to set unaligned access speed
<courmisch> uh?
<ja_02> conchuod sry
<ja_02> i was hopeing that there could be a way to skip speed probing as its quite slow
<conchuod> Ye, I guess. It was only really useful before you added the speed probing.
<ja_02> wdym?
<bjdooks> I finally got around to updating ivan's opensbi build updates and sedning them
crest has quit [Changing host]
crest has joined #riscv
another| has quit [Remote host closed the connection]
la_mettrie has quit [Quit: WeeChat 4.2.1]
<conchuod> ja_02: That the Zicclsm test was only a useful idea before the speed probing was added, when all that was being done was testing for a trap.
<conchuod> Because it does tell if you misaligned accesses are supported
another has joined #riscv
<ja_02> ah i understand now. the supported part is fast though so it doesn't really matter
eshep has quit [Ping timeout: 252 seconds]
eshep has joined #riscv
psydroid2 has joined #riscv
n_crm has joined #riscv
n_crm has quit [Changing host]
n_crm has joined #riscv
BootLayer has joined #riscv
psydroid2 has quit [Quit: KVIrc 5.0.0 Aria http://www.kvirc.net/]
psydroid2 has joined #riscv
frkzoid has joined #riscv
frkazoid333 has quit [Ping timeout: 268 seconds]
<courmisch> using indirect branches to call VSETVLI zero, zero with different immediate vector types... seems about as slow as using VSETVL. Somewhat unsurprisingly
<courmisch> so only way to optimise functions with small fixed-size inputs will be to replicate them entirely for each hardware VLEN :(
<courmisch> (unless the function only uses a single VTYPE, but that's not so commonly possible)
<courmisch> unlord, haasn ^ but my benchmarks were not too thorough
<unlord> courmisch: hmm, I was seeing benefits on the dav1d stuff I thought
<unlord> I was setting vsetvl only once though
TMM has quit [Quit: https://quassel-irc.org - Chat comfortably. Anywhere.]
TMM has joined #riscv
<courmisch> you can amortise VSETVL if you loop
khem has quit [Quit: WeeChat 4.3.3]
<courmisch> but that will barely work if you don't loop[3~
<courmisch> and it just doesn't work if you don't loop and you need to change element size
<courmisch> well, it depends on the ratio of instructions per vsetvl, obviously
khem has joined #riscv
<sorear> there's always "wait a year and see if performance characteristics of common cores have changed", I very much doubt c908 will have staying power
<courmisch> this is much more of an x60 than c908 problem
<courmisch> c908 has the issue, but it's kind of irrelevant because it's got the smallest hardware vector size that we care about
<courmisch> thus code that *works* at all on Zv128l should be optimal for C908 anyway.
<courmisch> But it won't be optimal for Zvl256b or larger. That's why I call it an x60 problem
<courmisch> sorear: and I agree that we should wait and see, but I seem to be alone in voicing that opinion in FFmpeg, dav1d, etc
<sorear> do we no longer care about dynamic LMUL for variable-length vector stuff on c908?
Stat_headcrabed has joined #riscv
Stat_headcrabed has quit [Client Quit]
_catircservices has quit [Quit: Bridge terminating on SIGTERM]
_catircservices has joined #riscv
_catircservices has quit [Client Quit]
<courmisch> if the variable length is iterated a lot, then computing VTYPE dynamically and using VSETVL is almost free, that's not much of an issue
_catircservices has joined #riscv
<courmisch> the problem now is small fixed-size inputs
<courmisch> for that we cannot use VSETVL, it's too slow
<sorear> vsetvl is actually slower than vsetvli on existing hardware? i didn't expect that to be a problem until next year and was worried the optimizations using it would all have to be reverted
<courmisch> it seems to be, but maybe I'm measuring wrong
<unlord> courmisch: most of the DSP functions I've been looking at in dav1d have loops after the vsetvl*
<courmisch> unlord: but don't they need to change SEW?
<unlord> some don't
<courmisch> for instance, if you need to clip s16 to u8, you have to change SEW (and therefore VTYPE)
<unlord> the blend ones for example
<courmisch> yeah in that case it's easy
<courmisch> but it's not always possible
<sorear> I wonder if any of the "clip s16 to u8" cases could be profitably converted into "clip s16 to s8" with a -128 offset transform
<courmisch> yes that works in some cases
<courmisch> but in the general case, it's adding one extra instructions (not counting vsetvl's)
<sorear> I'm thinking of cases where the "extra instruction" can be folded into the next pass at zero cost
<courmisch> converting the multimedia world to signed pixel samples, yeah, good luck
<courmisch> I don't know case where it could readily be done in isolation. Maybe there are, but I don't have the wherewithal
dobson` has quit [Quit: Leaving]
psydroid2 has quit [Ping timeout: 264 seconds]
fossdd__ has joined #riscv
fossdd_ has quit [Ping timeout: 255 seconds]
BootLayer has quit [Quit: Leaving]
fossdd_ has joined #riscv
psydroid2 has joined #riscv
dobson has joined #riscv
fossdd__ has quit [Ping timeout: 268 seconds]
fossdd__ has joined #riscv
fossdd_ has quit [Ping timeout: 256 seconds]
fossdd___ has joined #riscv
fossdd__ has quit [Ping timeout: 256 seconds]
fossdd_ has joined #riscv
fossdd___ has quit [Ping timeout: 256 seconds]
fossdd__ has joined #riscv
fossdd_ has quit [Ping timeout: 246 seconds]
fossdd_ has joined #riscv
fossdd__ has quit [Ping timeout: 246 seconds]
fossdd_ is now known as fossdd
fossdd__ has joined #riscv
fossdd__ has quit [Remote host closed the connection]
fossdd has quit [Ping timeout: 246 seconds]
fossdd_ has joined #riscv
fossdd__ has joined #riscv
fossdd_ has quit [Ping timeout: 256 seconds]
fossdd___ has joined #riscv
psydroid2 has quit [Quit: KVIrc 5.0.0 Aria http://www.kvirc.net/]
psydroid2 has joined #riscv
fossdd_ has joined #riscv
fossdd__ has quit [Ping timeout: 256 seconds]
fossdd__ has joined #riscv
fossdd___ has quit [Ping timeout: 256 seconds]
fossdd_ has quit [Ping timeout: 256 seconds]
fossdd___ has joined #riscv
fossdd__ has quit [Ping timeout: 255 seconds]
fossdd_ has joined #riscv
fossdd__ has joined #riscv
fossdd___ has quit [Ping timeout: 264 seconds]
fossdd_ has quit [Ping timeout: 268 seconds]
atka has quit [Quit: WeeChat 4.3.3]
eshep has quit [Ping timeout: 246 seconds]
eshep has joined #riscv
Starfoxxes has quit [Remote host closed the connection]
eshep has quit [Ping timeout: 264 seconds]
eshep has joined #riscv
Andre_Z has joined #riscv
ja_02 has quit [Remote host closed the connection]
ldevulder has quit [Quit: Leaving]
ja_02 has joined #riscv
cleger has quit [Ping timeout: 256 seconds]
eshep has quit [Ping timeout: 252 seconds]
SpaceCoaster has quit [Ping timeout: 252 seconds]
SpaceCoaster has joined #riscv
eshep has joined #riscv
another is now known as another|
Andre_Z has quit [Quit: Leaving.]
Andre_Z has joined #riscv
psydroid2 has quit [Quit: KVIrc 5.0.0 Aria http://www.kvirc.net/]
alperak has quit [Quit: Connection closed for inactivity]
DesRoin has quit [Ping timeout: 268 seconds]
DesRoin has joined #riscv
eshep has quit [Ping timeout: 264 seconds]
eshep has joined #riscv