sb0 changed the topic of #m-labs to: https://m-labs.hk :: Logs http://irclog.whitequark.org/m-labs :: Due to spam bots, only registered users can talk. See: https://freenode.net/kb/answer/registration
_whitelogger has joined #m-labs
rohitksingh has joined #m-labs
rohitksingh has quit [Ping timeout: 272 seconds]
rohitksingh_ has joined #m-labs
rohitksingh_ has quit [Client Quit]
rohitksingh has joined #m-labs
rohitksingh has quit [Quit: No Ping reply in 180 seconds.]
rohitksingh has joined #m-labs
rohitksingh has quit [Ping timeout: 244 seconds]
rohitksingh has joined #m-labs
rohitksingh has quit [Quit: No Ping reply in 180 seconds.]
rohitksingh has joined #m-labs
rohitksingh has quit [Ping timeout: 245 seconds]
rohitksingh has joined #m-labs
<sb0> whitequark: but what is the purpose of tbaa_now?
<GitHub-m-labs> [artiq] sbourdeauducq pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/2e66788c6c9c310de50914ee7b47d36c69324b74
<GitHub-m-labs> artiq/master 2e66788 Sebastien Bourdeauducq: compiler: support little endian target when storing now
<whitequark> sb0: it was an attempt to communicate to llvm that now doesn't ever alias anything
<whitequark> i'm not sure if it worked in real life, i remember running artifical tests on it
<whitequark> which seemed to do the thing
<whitequark> alias analysis in llvm is tailored to things like spec2006
<whitequark> which do not really represent real-life and moreso useful workloads
<bb-m-labs> build #2740 of artiq is complete: Failure [failed lit_test] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2740 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
rohitksingh has quit [Read error: Connection reset by peer]
rohitksingh has joined #m-labs
<sb0> whitequark: why is module.data_layout an empty string? is llvm ever supposed to return that?
<whitequark> sb0: data layout is set by artiq compiler
<whitequark> in compiler/target iirc
<whitequark> llvm does not set it anymore, for a long time
<sb0> whitequark: yes. so the artiq compiler sets it to empty string (when run on the host), but then afaict this is passed to LLVM, and then when reading it back it also comes from LLVM
<sb0> it's a python property
<whitequark> hm, it should use the proper API to ask the JIT for the data layout, I think
<whitequark> however, I'm not sure why it started failing all of a sudden
<whitequark> did you upgrade LLVM?
<sb0> no
<whitequark> maybe the JIT is broken? do llvmlite unittests pass?
<GitHub-m-labs> [artiq] sbourdeauducq pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/8940009e1a65a2d5b80e5c3b5e0ab91ddca18ce1
<GitHub-m-labs> artiq/master 8940009 Sebastien Bourdeauducq: compiler: pass data_layout string to llvm.create_target_data before determining endianness
<sb0> bb-m-labs: stop build artiq typo
<bb-m-labs> build 2741 interrupted
<bb-m-labs> build #2741 of artiq is complete: Exception [exception conda_build_output] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2741 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
<GitHub-m-labs> [artiq] sbourdeauducq pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/dd03fdfd1a53bf58fe135280828be6a93a401bcb
<GitHub-m-labs> artiq/master dd03fdf Sebastien Bourdeauducq: typo
rohitksingh has quit [Ping timeout: 272 seconds]
<sb0> bb-m-labs: stop build artiq ff
<bb-m-labs> build 2742 interrupted
<bb-m-labs> build #2742 of artiq is complete: Exception [exception interrupted conda_remove] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2742 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
<GitHub-m-labs> [artiq] sbourdeauducq pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/d931967e5c3bb34592c5e7652df7c85e2e9fe90e
<GitHub-m-labs> artiq/master d931967 Sebastien Bourdeauducq: fix previous commits
<bb-m-labs> build #2743 of artiq is complete: Failure [failed lit_test] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2743 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
<sb0> ffs
<sb0> whitequark: can you fix this please
<whitequark> so, why did it break?
<sb0> see the commits, I need to detect endianness in LLVMIRGenerator, which I am doing by looking at the data_layout string, which is inexplicably empty on the host
<whitequark> just add self.target.little_endian instead
<whitequark> ok, sec, there's a slightly more proper way
<sb0> LLVM already knows the endinanness, no?
<whitequark> sb0: str(llvmlite.binding.targets.Target.from_default_triple().create_target_machine().target_data)
<whitequark> this is how you get the data laout for the host with llvmlite
<whitequark> llvm returns you an empty string back because you feed it an empty string
<whitequark> it doesn't know or care that you are building for the host
<whitequark> it just uses the built-in default values it has
<whitequark> the JIT later inserts the correct data layout but it's far too late for you to interrogate it. the module isn't associated with JIT.
<GitHub-m-labs> [artiq] sbourdeauducq pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/981a77834a2d0af95751c1043b80122c45045bf4
<GitHub-m-labs> artiq/master 981a778 Sebastien Bourdeauducq: compiler: use default triple to determine data_layout for JIT
<sb0> whitequark: thanks
<GitHub-m-labs> [artiq] sbourdeauducq pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/421834fa3e191759388e1073803a6fcde7fc3022
<GitHub-m-labs> artiq/master 421834f Sebastien Bourdeauducq: compiler: document Target.little_endian
<bb-m-labs> build #2107 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/2107
<whitequark> sb0: I would personally set the data layout for the native target and don't bother with little_endian as a field
<whitequark> but this also works
<sb0> whitequark: aren't other things that use the data layout and expect it to be empty for JIT?
<whitequark> nope
<whitequark> JIT isn't really special, it's just compiling code for the host CPU and then a few hooks into the OS to actually run it
<whitequark> compiler side is exactly the same
<bb-m-labs> build #2108 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/2108
<bb-m-labs> build #963 of artiq-win64-test is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-win64-test/builds/963
<bb-m-labs> build #2744 of artiq is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2744
<bb-m-labs> build #2109 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/2109
<bb-m-labs> build #2110 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/2110
<bb-m-labs> build #2745 of artiq is complete: Failure [failed python_unittest_2] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2745 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
<GitHub-m-labs> [artiq] sbourdeauducq pushed 1 new commit to master: https://github.com/m-labs/artiq/commit/6aa341bc4470faa92b66da3e9e46d21e896b9a69
<GitHub-m-labs> artiq/master 6aa341b Sebastien Bourdeauducq: test_loopback_gate_timing: fix lat_offset
<GitHub-m-labs> [artiq] sbourdeauducq pushed 1 new commit to release-4: https://github.com/m-labs/artiq/commit/7358262ab3dc2f0a0034f7d0e7d9dfeac68be306
<GitHub-m-labs> artiq/release-4 7358262 Sebastien Bourdeauducq: test_loopback_gate_timing: fix lat_offset
<bb-m-labs> build #2111 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/2111
<bb-m-labs> build #2112 of artiq-board is complete: Failure [failed conda_build] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/2112 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
<bb-m-labs> build #2746 of artiq is complete: Failure [failed] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2746 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
<bb-m-labs> build #2113 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/2113
<bb-m-labs> build #2114 of artiq-board is complete: Success [build successful] Build details are at http://buildbot.m-labs.hk/builders/artiq-board/builds/2114
<bb-m-labs> build #2747 of artiq is complete: Failure [failed python_unittest_2] Build details are at http://buildbot.m-labs.hk/builders/artiq/builds/2747 blamelist: Sebastien Bourdeauducq <sb@m-labs.hk>
rohitksingh has joined #m-labs
<GitHub-m-labs> [artiq] KaifengC opened pull request #1204: kasli: add support for windows platform. (master...bug_fix) https://github.com/m-labs/artiq/pull/1204
<GitHub-m-labs> [artiq] sbourdeauducq commented on pull request #1204 c355430: Elsewhere in ARTIQ, this is tested with ``os.name == "nt"``. https://github.com/m-labs/artiq/pull/1204#discussion_r238104466
<GitHub-m-labs> [artiq] sbourdeauducq commented on pull request #1204 c355430: The function will break if ``sys.platform`` is neither ``linux2`` or ``win32``.... https://github.com/m-labs/artiq/pull/1204#discussion_r238104530
<GitHub-m-labs> [artiq] sbourdeauducq commented on pull request #1204 c355430: This code can be factored:... https://github.com/m-labs/artiq/pull/1204#discussion_r238104575
<GitHub-m-labs> [artiq] sbourdeauducq commented on pull request #1204 c355430: This code can be factored:... https://github.com/m-labs/artiq/pull/1204#discussion_r238104575
<GitHub-m-labs> [artiq] sbourdeauducq commented on pull request #1204 c355430: This code can be factored:... https://github.com/m-labs/artiq/pull/1204#discussion_r238104575
<GitHub-m-labs> [artiq] sbourdeauducq commented on pull request #1204 c355430: Doesn't this block (and break) if the key pressed is not ENTER? https://github.com/m-labs/artiq/pull/1204#discussion_r238104609
<GitHub-m-labs> [artiq] sbourdeauducq commented on pull request #1204 c355430: On Linux, the ``select`` call waits for ENTER, and ``sys.stdin.read`` also waits for ENTER.... https://github.com/m-labs/artiq/pull/1204#discussion_r238104640
<GitHub-m-labs> [artiq] sbourdeauducq commented on pull request #1204 c355430: This code can be factored:... https://github.com/m-labs/artiq/pull/1204#discussion_r238104575
_whitelogger has joined #m-labs
<GitHub-m-labs> [artiq] KaifengC commented on pull request #1204 c355430: Yes, on windows `kbhit()` will immditely return `0`, which is not `True`. But if you press any button, a none-zero value will return, which is `True`.... https://github.com/m-labs/artiq/pull/1204#discussion_r238105924
rohitksingh has quit [Ping timeout: 268 seconds]
<GitHub-m-labs> [artiq] KaifengC commented on pull request #1204 f23536c: Yes, on windows `kbhit()` will immediately return `0`, which is not `True`. But if you press any button, a none-zero value will return, which is `True`.... https://github.com/m-labs/artiq/pull/1204#discussion_r238105924
<GitHub-m-labs> [artiq] KaifengC commented on pull request #1204 f23536c: Yes, on windows `kbhit()` will immediately return `0`, which is not `True`. But if you press any button, a none-zero value will return, which is `True`.... https://github.com/m-labs/artiq/pull/1204#discussion_r238105924
rohitksingh has joined #m-labs
lkcl has quit [Ping timeout: 250 seconds]
lkcl has joined #m-labs
rohitksingh has quit [Ping timeout: 250 seconds]