sb0 changed the topic of #m-labs to: ARTIQ, Migen, MiSoC, Mixxeo & other M-Labs projects :: fka #milkymist :: Logs http://irclog.whitequark.org/m-labs
mithro has quit [Ping timeout: 240 seconds]
cyrozap has quit [Ping timeout: 240 seconds]
shenki has quit [Ping timeout: 240 seconds]
shenki has joined #m-labs
cyrozap has joined #m-labs
mithro has joined #m-labs
<travis-ci> [rust-managed] whitequark tagged v0.2.1 at master: https://github.com/m-labs/rust-managed/commits/v0.2.1
<travis-ci> m-labs/rust-managed#14 (v0.2.1 - 3e3dacf : whitequark): The build passed.
<travis-ci> m-labs/rust-managed#15 (master - 3e3dacf : whitequark): The build passed.
sb0 has quit [Quit: Leaving]
sandeepkr has quit [Write error: Connection reset by peer]
sandeepkr has joined #m-labs
<GitHub86> [smoltcp] whitequark pushed 3 new commits to master: https://github.com/m-labs/smoltcp/compare/92e9fcb307a3...019d03d758df
<GitHub86> smoltcp/master 019d03d whitequark: Implement socket sets.
<GitHub86> smoltcp/master d8c460e whitequark: Update docs.
<GitHub86> smoltcp/master 7e7c0d1 whitequark: Add the use_collections feature.
<GitHub197> [smoltcp] whitequark pushed 1 new commit to master: https://github.com/m-labs/smoltcp/commit/70a09735c99acece4685e27805deaf638a095a3b
<GitHub197> smoltcp/master 70a0973 whitequark: Make binding the UDP socket an explicit operation.
<travis-ci> m-labs/smoltcp#19 (master - 70a0973 : whitequark): The build passed.
<GitHub142> [smoltcp] whitequark pushed 1 new commit to master: https://github.com/m-labs/smoltcp/commit/8f28e99a0cae516df8a4c62b4cb7f0de5a3c1e56
<GitHub142> smoltcp/master 8f28e99 whitequark: Make interfaces not own the sockets.
<travis-ci> m-labs/smoltcp#20 (master - 8f28e99 : whitequark): The build passed.
<mithro> Where does libbase come from?
<whitequark> misoc?
<mithro> whitequark: I mean where did the code originate? It seems to be a mismash of stuff?
<whitequark> I think nowhere in particular
<GitHub30> [smoltcp] whitequark pushed 1 new commit to master: https://git.io/vM0IC
<GitHub30> smoltcp/master 0ccd620 whitequark: Swap the data and endpoint in UdpSocket methods....
<travis-ci> m-labs/smoltcp#21 (master - 0ccd620 : whitequark): The build passed.
<GitHub90> [smoltcp] whitequark pushed 2 new commits to master: https://github.com/m-labs/smoltcp/compare/0ccd6205f790...d5638f469fb2
<GitHub90> smoltcp/master e2b4753 whitequark: Add functions to check if the UDP socket buffers are empty or not.
<GitHub90> smoltcp/master d5638f4 whitequark: Return the amount of bytes sent from UdpSocket::send_slice.
<travis-ci> m-labs/smoltcp#22 (master - d5638f4 : whitequark): The build passed.
FabM has joined #m-labs
<rjo> whitequark: the Cargo.lock's should either be added to the repository or gitignored.
<rjo> whitequark: more accurately: carg should either always update the file and then it should be ignored. or it should not automatically update the file and then it should be in the repo.
Gurty has quit [Ping timeout: 240 seconds]
Gurty has joined #m-labs
sandeepkr has quit [Read error: Connection reset by peer]
sandeepkr has joined #m-labs
rohitksingh has joined #m-labs
rohitksingh has quit [Quit: Leaving.]
sb0 has joined #m-labs
mumptai has joined #m-labs
sandeepkr_ has joined #m-labs
kuldeep_ has joined #m-labs
sandeepkr has quit [Ping timeout: 255 seconds]
kuldeep has quit [Ping timeout: 260 seconds]
jbqubit has joined #m-labs
<jbqubit> rjo: "I don't know which of the changes made the difference. But it looks ok now from my end. You should see a signal on channel 1." I see 4 LEDs on KC705 pulsing at about 1 Hz. No output on DAC.
<rjo> where are you measuring?
<rjo> jbqubit: ^
<jbqubit> I checked all four outputs. Now connected to DAC0 and DAC1
<jbqubit> I rebooted kc705.
<rjo> jbqubit: J17?
<jbqubit> Yes
<rjo> jbqubit: LED state (all eight, be precise), and what do you see on ttl_sma?
<rjo> (user_sma_gpio_n)
<jbqubit> user_sma_gpio_n is 0V (1 us/div on scope)
<jbqubit> 012345678 is 0011fff where f is ~1 Hz flash
<jbqubit> 12345678 is 0011fff where f is ~1 Hz flash
<rjo> you said you see four leds flashing. what now?
<jbqubit> 12345678 is 0011ffff where f is ~1 Hz flash
<jbqubit> They flash on and off in unison.
<GitHub137> [artiq] sbourdeauducq pushed 2 new commits to master: https://git.io/vME1L
<GitHub137> artiq/master 7c699e2 Sebastien Bourdeauducq: drtio: add FIFO space request count debug API
<GitHub137> artiq/master c25186f Sebastien Bourdeauducq: drtio: print packet error descriptions in log
<rjo> is the led state the same now?
<jbqubit> 0001ffff
<rjo> jbqubit: did you trigger on user_sma_gpio_n?
<rjo> and the signal on the dac channels is not huge. use the settings from the screenshot i had sent around.
<jbqubit> Now I am triggering on user_sma_gpio_n. For a few seconds I saw a waveform on the scope but its now gone.
<jbqubit> The trigger pulse persists. 20 ms period
<jbqubit> What program are you running?
<rjo> it "persists"?
<jbqubit> output on user_sma_gpio_n remains periodic with period 20 ms
<jbqubit> No synchronous output on DAC0 and DAC1.
<rjo> yeah. the ttl pulses where there before, you just didn't look close enough.
<rjo> at what scale?
<jbqubit> 100 mV/div,
<rjo> H?
<jbqubit> 2 us/div
<jbqubit> LED pattern is back to 0011FFFF
<rjo> please provide a photo of the setup
<jbqubit> sent
<jbqubit> signal
<rjo> the pulse is at the falling edge of the ttl pulse and the setting is 200ns/div.
<jbqubit> confirmed
<rjo> jbqubit: awesome joe. what a waste of time. that pulse has always been there. i didn't change a thing.
<jbqubit> OK. Are you running demo_2tone.py?
<rjo> jbqubit: yes. always. i just added ttl edges to be able to debug the operator.
<rjo> and i guess you know how you have to run kernels given the broken networking stack, right?
<jbqubit> OK. Please update demo_2tone.py to include the output on user_sma_gpio_n. It's helpful.
<jbqubit> I'm aware that lwip requires sending kernels to flash. I will try that out now.
<GitHub95> [artiq] jordens pushed 1 new commit to master: https://git.io/vMESV
<GitHub95> artiq/master 394ffd8 Robert Jordens: phaser/demo_2tone: add ttl and led pulse
<sb0> bb-m-labs_, stop build artiq-board need-cpu
<bb-m-labs_> build 308 interrupted
<bb-m-labs_> build #308 of artiq-board is complete: Exception [exception interrupted] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/308 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
<sb0> bb-m-labs_, stop build artiq need-cpu
<bb-m-labs_> build 1208 interrupted
<bb-m-labs_> build #1208 of artiq is complete: Exception [exception interrupted] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/1208 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
<rjo> jbqubit: demo_2tone is the idle kernel right now.
<sb0> bb-m-labs_, stop build artiq need-cpu
<bb-m-labs_> build 1209 interrupted
<jbqubit> ok
<bb-m-labs_> build #1209 of artiq is complete: Exception [exception interrupted] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/1209 blamelist: Robert Jordens <rj@m-labs.hk>
<sb0> whitequark, you are 4 flterms on /dev/ttyUSB0 that look stale, should I kill them?
<sb0> *have
<rjo> jbqubit: consider the ~400ns delay between the ttl and the pulse a confirmation of the DAC side of my PID latency estimate.
<jbqubit> (artiq3) britton@britton1:~/artiq3/artiq/artiq/examples/phaser$ artiq_compile repository/demo_2tone.py
<jbqubit> (artiq3) britton@britton1:~/artiq3/artiq/artiq/examples/phaser$ artiq_mkfs phaser_config.bin -s mac 00:0a:35:03:1e:53 -s ip 192.168.1.71 -s startup_clock i -f idle_kernel repository/demo_2tone.elf
<jbqubit> (artiq3) britton@britton1:~/artiq3/artiq/artiq/examples/phaser$ openocd -f board/kc705.cfg -s ~/anaconda3/envs/phaser2/share/openocd/scripts -c "init; jtagspi_init 0 /home/britton/.migen/bscan_spi_xc7k325t.bit; jtagspi_program phaser_config.bin 0xb80000; xc7_program xc7.tap; exit"
<jbqubit> Excerpt from corelog....
<jbqubit> [ 211322417us] TRACE(runtime::session): comm<-kern RTIOInitRequest [ 211322526us] INFO(runtime::session): resetting RTIO [ 211336358us] TRACE(runtime::session): comm<-kern NowSave(252337300856) [ 211336693us] TRACE(runtime::session): comm<-kern RunException { exception: Exception { name: 0x40819030, file: 0x40818fc0, line: 61, column: 8, function: 0x40818ff6, message: 0x40819060, param: [252337300856, 0, 0], phantom: Phantom
<jbqubit> I've added self.ttl_sma.pulse(1*us) to while True loop
<jbqubit> I see no pulses on user_sma_gpio_n
<rjo> well there is an exception.
<rjo> jbqubit: check package versions. misoc, rustc, llvm, the whole thing.
<rjo> jbqubit: check whether you messed up things by installing things with setup.py.
<rjo> jbqubit: oh. please always check with the original script first before you edit things.
<jbqubit> I did check with original script. And saw nothing triggering on DAC0.
<sb0> rjo, are you using any of the HKG KC705s?
<rjo> "triggering"?
<rjo> sb0: no.
<rjo> jbqubit: does the original script give you an exception?
<jbqubit> both version compile without exception
<jbqubit> This might be the problem...
<jbqubit> $ artiq_corelog WARNING:artiq.coredevice.comm_generic:Mismatch between gateware (3.0.dev+684.g7af152e) and software (3.0.dev+687.g394ffd8.dirty) versions
<rjo> dude no. you posted a log with an exception.
<jbqubit> I've rolled back to version of artiq that matches my installed gateware.
<jbqubit> and ran setup.py develop --user
<jbqubit> As regards "jbqubit: does the original script give you an exception?" I thought you were asking about the artiq_compile step.
<jbqubit> I now see that there is a RTIOCollision.
<jbqubit> I've resolved RTIOCollision. Now I see waveform.
<sb0> the DAC waveform?
<jbqubit> Yes. I see the DAC waveform.
<sb0> great!
<sb0> out of curiousity, what was causing the JESD ready timeout?
<jbqubit> I don't yet know what was causing the JESD ready timeout. Changes vs when SB and Joe were debugging things are a) lower amplitude clock b) different prototype board. I'm checking now to see if I can revert.
<sb0> so, a random selection of 0.0213% of the drtio packets are truncated somehow, apparently at the transmitter. this is what is causing most of the symptoms it seems...
<jbqubit> sb0: Just confirmed problem was too high clock power on J1. +11 dBm produces behavior you and I saw. Today I reduced power to +6 dBm and it works.
<jbqubit> I see the too high clock power reproduce.
<sb0> jbqubit, good
hozer has joined #m-labs
mumptai has quit [Quit: Verlassend]
<GitHub31> [artiq] jbqubit opened issue #653: cannot unify numpy.int? with float https://git.io/vMuq0
<GitHub40> [artiq] jordens closed issue #653: cannot unify numpy.int? with float https://git.io/vMuq0
<GitHub120> [artiq] jordens commented on issue #653: The argument is a float. Give it a float.... https://git.io/vMumC
<whitequark> rjo: fwiw the problem with that issue is that adding this coercion would require annotations elsewhere
<whitequark> it's basically a zero-sum game
<GitHub60> [artiq] jbqubit commented on issue #653: - This is inconsistent. I can set a frequency 50*MHz.... https://git.io/vMuYj
<GitHub169> [artiq] jordens commented on issue #653: No. It's not inconsistent 50*MHz is a float. It is 50*1e6.... https://git.io/vMu3Y
<GitHub161> [artiq] jordens commented on issue #653: No. It's not inconsistent 50*MHz is a float. It is 50*1e6.... https://git.io/vMu3Y
<GitHub58> [artiq] jordens commented on issue #653: No. It's not inconsistent `50*MHz` is a float. It is `50*1e6`.... https://git.io/vMu3Y
<GitHub191> [artiq] jbqubit commented on issue #653: > No. It's not inconsistent 50MHz is a float. It is 50e6.... https://git.io/vMusX