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
<ohsix> whitequark: was smoltcp what you were fuzzing?
<whitequark> ohsix: yes
<GitHub114> [migen] whitequark pushed 1 new commit to master: https://git.io/vQOit
<GitHub114> migen/master 8c7472b Lukas Lao Beyer: Add 'depth' attribute to _FIFOInterface...
<sb0> whitequark, TBytes isn't in artiq.experiment
<bb-m-labs> build #151 of migen is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/migen/builds/151
<sb0> should it be added like TInt32 etc.?
<sb0> whitequark, artiq.compiler.builtins.TBytes doesn't work when used as type annotation though ...
<sb0> I have 3.0.dev+1174.g12357d88
<sb0> whitequark, TBytes()
<sb0> 1763210.1822707946 bps
<sb0> whitequark, by the way, when you remove the type annotation, it produces a huge traceback that causes the core device to run out of memory and crash
<sb0> well it varies. 648921.2758402787 837302.7063632996 1118088.6401711144
<whitequark> sb0: onto it
<whitequark> (all tree)
<whitequark> three*
_rht has joined #m-labs
<whitequark> sb0: this is weird
<whitequark> how is TBytes not in artiq.experiment?
<whitequark> in module language: __all__.extend(types.__all__)
<whitequark> in module experiment: __all__.extend(language.__all__)
<whitequark> and it works over here...
<whitequark> oh I didn't push
<GitHub117> [artiq] whitequark pushed 2 new commits to master: https://github.com/m-labs/artiq/compare/12357d884e7e...f17a6616b2ce
<GitHub117> artiq/master f17a661 whitequark: runtime: ensure management interface buffer can hold log buffer....
<GitHub117> artiq/master d0f7263 whitequark: language: export TBytes and TByteArray.
<sb0> whitequark, throughput looks acceptable now though. shall we tell the oxford guys to have a look already?
<whitequark> sb0: well I want to fix a few more things
<whitequark> but you can tell them, yes
<whitequark> for example rjo suggested implementing delayed ACK and that should play very nicely with our scheduler and is also easy
<whitequark> oh and we should send a duplicate ACK when we drop an out-of-order segment
<whitequark> again easy and high-payof
<whitequark> payoff*
<whitequark> if I'm correct then with those two implemented throughput should not deviate anymore
<GitHub97> [artiq] sbourdeauducq commented on issue #685: @klickverbot @cjbe @dhslichter should be better now. https://github.com/m-labs/artiq/issues/685#issuecomment-310881699
<whitequark> sb0: regarding "huge traceback", that's a heap dump
<whitequark> oh
<whitequark> are you saying the *host* produces a huge traceback?
<whitequark> this is very odd, let's see
<whitequark> ok I see what you mean
<sb0> well. it produces some large crash output because it dumps the value of the bytes array. and simultaneously the core device runs out of memory and crashes. I have not investigated further
<whitequark> you mean with artiq_run -vvv?
<whitequark> or something else
<whitequark> oh, the *exception message* is long
<whitequark> this is silly
<sb0> and the "connection reset by peer" corresponds to a out of memory error in the core device log
<whitequark> yes, I figured it out already
<whitequark> sb0: I found a fairly serious bug
<whitequark> the host may try to reply to an async RPC
<whitequark> looks like this triggers it somehow, I'm not fully sure yet how
<whitequark> figured it out, fixing now
<bb-m-labs> build #663 of artiq-board is complete: Failure [failed conda_build] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/663 blamelist: whitequark <whitequark@whitequark.org>
<bb-m-labs> build #1580 of artiq is complete: Failure [failed] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/1580 blamelist: whitequark <whitequark@whitequark.org>
<whitequark> sb0: timing not met
<GitHub30> [artiq] whitequark pushed 3 new commits to master: https://github.com/m-labs/artiq/compare/f17a6616b2ce...31b52ff1b3fc
<GitHub30> artiq/master 31b52ff whitequark: coredevice: do not desync kernel TCP stream on RPCReturnValueError....
<GitHub30> artiq/master fdb24ef whitequark: coredevice: truncate overlong exception messages....
<GitHub30> artiq/master 1fa8be3 whitequark: artiq_devtool: don't loop forever if core device dies.
<bb-m-labs> build #664 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/664
<bb-m-labs> build #1581 of artiq is complete: Failure [failed python_unittest_2] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/1581 blamelist: whitequark <whitequark@whitequark.org>
<GitHub3> [artiq] whitequark pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/69fa9b38e0d6d6a582f8e0dbe759fd3ec780534c
<GitHub3> artiq/master 69fa9b3 whitequark: test: actually run test_embedding.AsyncTest.
<bb-m-labs> build #665 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/665
mumptai has joined #m-labs
<bb-m-labs> build #501 of artiq-win64-test is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-win64-test/builds/501
<bb-m-labs> build #1582 of artiq is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/1582
<GitHub14> [artiq] whitequark pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/816ec6c52f75bac07693f5c02903bf4ed6a1c6c9
<GitHub14> artiq/master 816ec6c whitequark: artiq_pcap: implement.
<bb-m-labs> build #1583 of artiq is complete: Failure [failed python_unittest] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/1583 blamelist: whitequark <whitequark@whitequark.org>
<GitHub34> [artiq] whitequark pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/6061393bd0576f3d7376c7b74a6b4f34f581cc32
<GitHub34> artiq/master 6061393 whitequark: tools: lazily import paramiko.
_whitelogger has joined #m-labs
mumptai has quit [Ping timeout: 240 seconds]
_rht has quit [Quit: Connection closed for inactivity]
mumptai has joined #m-labs
<rjo> sb0: http://termbin.com/5c9t that's usually the current pattern of timing failure throughout our bitstreams. 15 logic levels on the "writable" logic outputmanager-to-DMA.
<GitHub88> [migen] jordens pushed 3 new commits to master: https://git.io/vQObV
<GitHub88> migen/master 5ccee6c Robert Jordens: vivado: run opt_design before place...
<GitHub88> migen/master 9d19ccb Robert Jordens: vivado: write out checkpoints and synth timing report
<GitHub88> migen/master 46feda9 Robert Jordens: vivado: use non-project workflow
<bb-m-labs> build #152 of migen is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/migen/builds/152
<rjo> bb-m-labs: force build --props=package=artiq-kc705-phaser artiq-board
<bb-m-labs> build forced [ETA 15m34s]
<bb-m-labs> I'll give a shout when the build finishes
<bb-m-labs> build #667 of artiq-board is complete: Failure [failed conda_build] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/667
mumptai has quit [Ping timeout: 240 seconds]
mumptai has joined #m-labs
mumptai has quit [Read error: Connection reset by peer]
mumptai has joined #m-labs
mumptai has quit [Read error: Connection reset by peer]
<travis-ci> batonius/smoltcp#10 (poll_refactor - c00f431 : Egor Karavaev): The build passed.
<travis-ci> batonius/smoltcp#11 (poll_refactor - 81cff0f : Egor Karavaev): The build passed.
<travis-ci> batonius/smoltcp#12 (poll_refactor - 77c64f3 : Egor Karavaev): The build passed.
<GitHub137> [artiq] jordens pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/f4624e086af587bf875cb93b300321c5fa4d82f3
<GitHub137> artiq/master f4624e0 Robert Jordens: conda: bump migen
<rjo> bb-m-labs: force build --props=package=artiq-kc705-phaser artiq-board
<bb-m-labs> build forced [ETA 15m34s]
<bb-m-labs> I'll give a shout when the build finishes
<GitHub173> [smoltcp] batonius commented on issue #19: Here's my attempt at refactoring `EthernetInterface::poll`: https://github.com/m-labs/smoltcp/compare/master...batonius:poll_refactor . I've tried to keep the logic intact, but I've made two changes I think are reasonable: ... https://git.io/vQ3sp
<bb-m-labs> build #668 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/668
<bb-m-labs> build #669 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/669
<bb-m-labs> build #502 of artiq-win64-test is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-win64-test/builds/502
<bb-m-labs> build #1585 of artiq is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/1585
mumptai has joined #m-labs
mumptai has quit [Quit: Verlassend]
_whitelogger has joined #m-labs
<GitHub99> [smoltcp] whitequark pushed 3 new commits to master: https://git.io/vQ3us
<GitHub99> smoltcp/master 8ace8cd whitequark: Set TCP PSH flag when sending the last buffered data.
<GitHub99> smoltcp/master a14c476 whitequark: Expose PSH flag in TcpRepr.
<GitHub99> smoltcp/master 6a09672 whitequark: Fix a few typos in TCP socket code (NFC).
<whitequark> sb0: rjo: I wrote artiq_pcap that lets you grab a packet trace to/from core device
<travis-ci> m-labs/smoltcp#129 (master - 8ace8cd : whitequark): The build passed.