<courmisch>
sorear: lpc is effectively a series of integer dot-products, each depending on all previous ones (see C code in libavcodexc/flacdsp.c)
<sorear>
i forsee two classes of implementations where this could be unproductive: in-order implementations that suppress microops for vector chunks containing only tail elements (tu/ta exists specifically to enable this)
<sorear>
and out of order implementations with a single rename stage where vsetvl stalls/flushes the entire pipeline until LMUL is known (unlikely to show up on benchmarks, I think you're the first user besides preemptive context switches)
psydroid has quit [Read error: Connection reset by peer]
psydroid has joined #riscv
<courmisch>
sorear: this only sets vtype once at the beginning, and never during the loop, though
Stat_headcrabed has joined #riscv
psydroid has quit [Remote host closed the connection]
psydroid has joined #riscv
mahk has quit [Ping timeout: 255 seconds]
Stat_headcrabed has quit [Quit: Stat_headcrabed]
wingsorc has quit [Ping timeout: 272 seconds]
bjoto has quit [Remote host closed the connection]
bjoto has joined #riscv
prabhakalad has quit [Ping timeout: 256 seconds]
prabhakalad has joined #riscv
<conchuod>
palmer: Have you seen any reports of failing builds in next? There's an interaction between smaeul's https://lore.kernel.org/r/20240327045035.368512-4-samuel.holland@sifive.com and irqchip driver changes that will need to be handled in Linus' merge commit cos riscv_ipi_set_virq_range() lost an argument.
billchenchina- has joined #riscv
billchenchina has quit [Remote host closed the connection]
billchenchina- has quit [Max SendQ exceeded]
billchenchina- has joined #riscv
mahk has joined #riscv
prabhakalad has quit [Ping timeout: 252 seconds]
damian101 has joined #riscv
damian101 has quit [Remote host closed the connection]
damian101 has joined #riscv
mahk has quit [Changing host]
mahk has joined #riscv
knolle has quit [Ping timeout: 256 seconds]
knolle has joined #riscv
dilfridge has quit [Ping timeout: 240 seconds]
dilfridge has joined #riscv
mlw has quit [Ping timeout: 252 seconds]
Noisytoot has quit [Excess Flood]
Noisytoot has joined #riscv
Noisytoot has quit [Ping timeout: 272 seconds]
seasharp has quit [Ping timeout: 252 seconds]
seasharp has joined #riscv
luca_ has joined #riscv
luca_ is now known as OwlWizard
Noisytoot has joined #riscv
pavelow has quit [Quit: No Ping reply in 180 seconds.]
pavelow has joined #riscv
mlw has joined #riscv
OwlWizard has quit [Quit: OwlWizard]
Tenkawa has joined #riscv
<Esmil>
f
knolle has quit [Ping timeout: 246 seconds]
knolle has joined #riscv
Andre_Z has joined #riscv
Stat_headcrabed has joined #riscv
heat has joined #riscv
ntwk has quit [Read error: Connection reset by peer]
<kelvium>
hi. i've got a mangopi mq-pro on allwinner d1. it semi-works: the pwr led is lit but it boots like 1 in 50 times i try to power it on. like it doesn't even write logs to serial output, usually. is this a known issue or is my board just doomed? i've got it to boot and it even worked for some time. sorry if this is off topic, i don't know where to ask :)
BootLayer has joined #riscv
<Tenkawa>
Esmil: is there a way to completely clear out the spi on the VF2 boards?
<Tenkawa>
Esmil: without beingable to access the cli
<Tenkawa>
Esmil: I've been able to get different OpenSBI/u-boot combinations flashed via xmodem on it (with different dates) but they all still have the same hang in OpenSBI in the SPI
<Tenkawa>
(well what we found is it was after OpenSBI)
<Tenkawa>
It just has no more output
vagrantc has joined #riscv
Andre_Z has quit [Quit: Leaving.]
<courmisch>
so Alibaba says the Post has my BPi-F3 and the Post says Alibaba has it
<courmisch>
:shrug:
<bjdooks>
the old quantum package condition
<courmisch>
OTOH, the Post says it arrives tomorrow and Alibaba says it arrives in 3 weeks
<Tenkawa>
heh
<courmisch>
enterprise IT at its best
<Tenkawa>
courmisch: just hopw it wasn't like my last Argon40 order... took a month
<Tenkawa>
er hope
<Tenkawa>
courmisch: '
<Tenkawa>
customs "outgoing" rejected it 2 weeks in
<Tenkawa>
so they had to resend the whole order to me
<courmisch>
somehow Alibaba has managed the customs without my intervention
<courmisch>
I'm sure they totally didn't dodge any duties
<Tenkawa>
courmisch: this is the first time I've had "this" much delay
<courmisch>
is "this" the first time your shipment went through Cainiao?
<courmisch>
which I reckon is Alibaba's logistic subsidiary
<bjdooks>
it's amusing to see how fast couriers find something when you point out the value it is insured for
<Tenkawa>
courmisch: this was Argon40 shipping.. not Alibaba
<bjdooks>
you've lose £25k of experimental hardware with hard to source ics...
<bjdooks>
oh sorry sir, turns out it accidentally got lost in our warehouse and we found it
<Tenkawa>
bjdooks: had that happen in town with it falling in a sorter
<Tenkawa>
after they shipped me a replacement already
<Tenkawa>
the carrier services here can be "ifft"
<Tenkawa>
er iffy
<courmisch>
bjdooks: BPi-F3 is under 100€ so that strategy is unlikely to work
<courmisch>
but it's funny how I was almost first outside China to receive my K230. Didn't work out so well this time
<Tenkawa>
courmisch: over on Discord there's been questions about its cores/performance already with the lack of info
<courmisch>
I don't mean to give any... real... names but
<Tenkawa>
Someone got theirs today and has been testing it
<courmisch>
cyyself already posted benchmarks
<courmisch>
and I think unlord already got his
<Tenkawa>
Doesn't sound like it would stand up perf wise even the JH7110
<Tenkawa>
er to the
<courmisch>
it's got double the cores, though
<Tenkawa>
I just wish I knew why mine is stuck booting
<courmisch>
and they're both in-order
Stat_headcrabed has joined #riscv
coldfeet has joined #riscv
damian101 has quit [Remote host closed the connection]