<d1b2>
<kule_57901> Hi @azonenberg it was very generous of you to introduce and provide the background of the projects. Yes @aleksorsist did lead me to this channel.. The hardware we are currently working on is a production monitoring system (i.e having 2 to 4 ADC channels and encoders ) which is used to capture process signals and parameters for analysis. We are currently trying to investigate different UI libraries to plot the live data acquired. I
<d1b2>
have been looking into the cool Thunderscope for quite some time , After sharing about the scope client our small team really liked the capabilities of it. The hardware in general is having an Intel CPU with FPGA connected over PCIe . Our application running on the CPU would specifically configure the AFE and retrieve data over DMA using the C2H interface of the XDMA core , yet to be tested :). However, we are currently at a stage where we can
<d1b2>
configure the ADC as required and aquire data from it. So, as mentioned I am trying to evaluate the requirements and posibility of gluing it to our application. Most of our drivers to configure and read are based on KMDF as its a windows based device.
<d1b2>
<azonenberg> Well, ngscopeclient certainly sounds possible to integrate for that purpose
Degi_ has joined #scopehal
Degi has quit [Ping timeout: 246 seconds]
Degi_ is now known as Degi
<d1b2>
<fpga_zealot> I got a question for windows install, I ran cmake, but there is no Makefile in build...
<d1b2>
<azonenberg> @david.rysk @bvernoux ^
<d1b2>
<fpga_zealot> Oh I think i need a UCRT64 shell
<d1b2>
<fpga_zealot> that didn't change anything...
<d1b2>
<azonenberg> i know that mingw / msys2 has a bunch of different build environments that are confusing
<d1b2>
<azonenberg> did you double check you followed the build steps in the manual exactly, to the letter?
<d1b2>
<fpga_zealot> I think all the pacman stuff in the wrong shell env
<d1b2>
<azonenberg> So it looks like your CMake is trying to generate Ninja build scripts
<d1b2>
<azonenberg> not Make scripts
<d1b2>
<azonenberg> I think you probably need to override it on the CMake command line to specify the correct one. I'm just not sure if it's Unix Makefiles, MSYS Makefiles, or MinGW Makefiles
<d1b2>
<fpga_zealot> I assume MSYS
<d1b2>
<azonenberg> I wonder if this was an upstream CMake change? we probably will need to update the docs to reflect this
<d1b2>
<azonenberg> That seems the most logical to me, try it first
<d1b2>
<azonenberg> but yeah either way, file a github issue for this because we should either be telling people to compile with ninja or to explicitly target make
<d1b2>
<fpga_zealot> What is the CMAKE flag I should be using?
<d1b2>
<azonenberg> (I'm not involved in windows building/packaging whatsoever)
<d1b2>
<azonenberg> Check cmake documentation, i've never had to override
<d1b2>
<david.rysk> The instructions should be saying to use ninja not make, I’ll check today
<d1b2>
<20goto10> On the latest Windows build from GitHub Actions, I2C (and many others) are disabled in Create Filter > Bus. Is this related to AVX being disabled on Windows?
bvernoux has joined #scopehal
<d1b2>
<azonenberg> No. That's probably due to incorrect typing of the input
<d1b2>
<azonenberg> Most of those filters expect a digital input
<d1b2>
<azonenberg> If you try to feed them an analog input they don't know what to do with it
<d1b2>
<azonenberg> Try thresholding it and then decoding the resulting digital signal
<d1b2>
<azonenberg> (we have a long-standing ticket for automatically inferring a threshold block in this use case)
<d1b2>
<azonenberg> but the UX around it is a little tricky
<d1b2>
<20goto10> thank you!
<d1b2>
<azonenberg> The Windows AVX issue is purely performance related (and becoming less important as most stuff we used AVX for is getting pushed to GPU)
<d1b2>
<azonenberg> no functionality should be disabled
bvernoux has quit [Read error: Connection reset by peer]
bvernoux has joined #scopehal
<d1b2>
<fpga_zealot> Trying out ninja build, seems to be working, using cmake --build . -j4 then ninja -j4