01:27 <miek> i'm hitting a segfault while working on the usb decoding: https://gist.githubusercontent.com/miek/acebcf419c94d94131a3b9b80fc31801/raw/69921365da7c43ffa4ebb2db5a912eaf1c1e4c05/gistfile1.txt

01:28 <azonenberg> miek: will investigate shortly. Reviewing your trigger PR right now

01:28 <miek> i don't really know enough about the different layers to debug it properly though, so i'm not sure if it's just the decder outputting garbage or if it's a bug downstream

01:28 <azonenberg> (just got up, lol)

01:28 <miek> mornin' :)

01:29 <azonenberg> Yes i woke up at 17:00 local time :p

01:29 <miek> lol

01:30 <azonenberg> BTW how are you liking the trigger APIs? Everything seem good or need work?

01:30 <azonenberg> (I think you're the first dev to add a new trigger type other than me)

01:32 <miek> no issues so far, seems good to me :)

01:37 <azonenberg> Great. And yeah the initial implementation was rather limited and only worked for level and certain kinds of pattern triggers

01:37 <azonenberg> this is much more generic

02:07 <_whitenotifier> [scopehal] azonenberg closed pull request #389: AgilentOscilloscope: support Nth edge burst trigger - https://git.io/JL2ii

02:07 <_whitenotifier> [scopehal] azonenberg pushed 2 commits to master [+4/-0/±8] https://git.io/JLVTo

02:07 <_whitenotifier> [scopehal] miek 0f4600a - AgilentOscilloscope: support Nth edge burst trigger

02:07 <_whitenotifier> [scopehal] azonenberg fc1e983 - Merge pull request #389 from miek/nth_edge_burst AgilentOscilloscope: support Nth edge burst trigger

02:56 <_whitenotifier> [scopehal] azonenberg pushed 1 commit to master [+1/-0/±4] https://git.io/JLVY4

02:56 <_whitenotifier> [scopehal] azonenberg bec752e - De-embed / CTLE filters now GPU-accelerate all computations. Reused several buffers to avoid wasted allocations and copies.

02:56 <_whitenotifier> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/JLVYu

02:56 <_whitenotifier> [scopehal-apps] azonenberg e701d04 - Updated submodules

02:59 <azonenberg> 16 -> 4.95 ms for de-embedding on 800K points

03:01 <azonenberg> This is still with probably unnecessary copies from gpu to cpu memory and back. Not a problem with only one gpu-accelerated filter in the graph but long term I would like to be able to push stuff to GPU and keep it there for as long as possible

03:01 <azonenberg> But guess what, it's still a 3x speedup even with sloppy buffer handling and lots of wasted pcie bandwidth :p

16:25 <d1b2> <mubes> A very small user report; Build instructions have suffered a slight bit of bit-rot. For debian/ubuntu opencl-dev and libglew-dev appear to be needed to be added to the dependencies.

16:40 <_whitenotifier> [scopehal] miek opened pull request #390: USB high-speed decoding - https://git.io/JLrKr

16:56 <miek> https://i.imgur.com/bSW5dl7.png :D

16:58 <monochroma> yay

17:02 <_whitenotifier> [scopehal] miek opened pull request #391: AgilentOscillscope: request max points - https://git.io/JLrXh

17:03 <_whitenotifier> [scopehal] miek synchronize pull request #391: AgilentOscillscope: request max points - https://git.io/JLrXh

17:03 <_whitenotifier> [scopehal] miek edited pull request #391: AgilentOscilloscope: request max points - https://git.io/JLrXh

19:50 <bgamari_> azonenberg, I'm not particularly familiar with cmake but it seems a bit odd the installation directory here is hard-coded: https://github.com/azonenberg/graphwidget/blob/master/CMakeLists.txt#L15

19:50 <bgamari_> this will break on some distributions (e.g. NixOS, in my own case)

19:54 <d1b2> <david.lenfesty> yeah, unless there's a specific reason that should be ${CMAKE_INSTALL_PREFIX}

19:55 <d1b2> <david.lenfesty> woops, ${CMAKE_INSTALL_PREFIX}/lib

19:55 <bgamari_> I think it can even just be a relative path

19:55 <bgamari_> according to https://cmake.org/cmake/help/latest/command/install.html