azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing | https://github.com/glscopeclient/scopehal-apps | Logs: https://libera.irclog.whitequark.org/scopehal
Degi has quit [Ping timeout: 252 seconds]
Degi_ has joined #scopehal
Degi_ is now known as Degi
Bird|otherbox has quit [Read error: Connection reset by peer]
Bird|otherbox has joined #scopehal
_florent_ has quit [Ping timeout: 252 seconds]
_florent_ has joined #scopehal
Stephie has quit [Read error: Connection reset by peer]
Stephie- has joined #scopehal
tnt has quit [Ping timeout: 252 seconds]
Stephie- has quit [Ping timeout: 245 seconds]
tnt has joined #scopehal
mxshift has quit [Ping timeout: 252 seconds]
mxshift has joined #scopehal
Stephie has joined #scopehal
<d1b2> <johnsel> Cool. I finally had some free time for some dog-fooding ngscopeclient and to look into Rigol MSO8k performance
<azonenberg> And how did that go?
<azonenberg> @johnsel wrt usability/accessibility there's really two fundamental issues we need to work on. the first is making functionality documented/discoverable
<azonenberg> the second is providing alternate input mechanisms where feasible if the primary is difficult for some people to use
<azonenberg> both could use work but they're not the same problem
<_whitenotifier-f> [scopehal-apps] azonenberg closed issue #512: ngscopeclient: file load/save support - https://github.com/glscopeclient/scopehal-apps/issues/512
<_whitenotifier-f> [scopehal-apps] azonenberg commented on issue #512: ngscopeclient: file load/save support - https://github.com/glscopeclient/scopehal-apps/issues/512#issuecomment-1732614548
<d1b2> <johnsel> Well other than the 2 bug reports it went pretty good. I just had my awg and scope connected to each other and did some toying around. W.r.t. to Rigol readouts the Wireshark looks good from our end but the scope keeps sending the same data and a lot of nulls offset over a few seconds when I just request a single (triggered) waveform. I'm reading the programming manual now to try to establish expected behavior vs what we're sending and if that
<d1b2> is expected behavior. Still have to compare it with MSO5k programming manual as well to see if the MSO8k does something different. There are a few ways to read out the waveform, raw, display, certain amount of points. I need to spend a few hours on it still I'm not very familiar with it
<d1b2> <johnsel> I agree that you can separate the issue into 2. I just wrote it quickly to log my finding and how I would generally try to deal with it, but they are obviously application-wide considerations
<d1b2> <johnsel> What I did really like is having control over both instruments with the snapping of windows into the grid system. It's super nice to have full overview of all instruments. I look forward to adding the Thunderscope and then doing some multi-scope toying
<d1b2> <johnsel> I did have issues with the FFT not working correctly but you were working on that and the CI issue prevented me from getting the latest binaries and I didn't want to build from source
<d1b2> <johnsel> So I will revisit that
<d1b2> <johnsel> I do feel like pushing the auto button on the scope sometimes is a lot more attractive than setting up things through ngscopeclient. And then there's some general extra functionality relating to AC/DC and bw limiting that the Rigol driver doesn't know the MSO8k can do
<azonenberg> Yeah there are controls for coupling and BWL in the API and channel properties dialog
<azonenberg> But if the driver doesnt implement them they're no good :p
<d1b2> <johnsel> True true, they are partially implemented though. I guess MSO8k has more options than MSO5k, simple as that
<azonenberg> very possible
<d1b2> <johnsel> I may spend some more time on it today. I had a wedding last night so I don't think I'll be super productive but might make some progress, something something free drinks.
<azonenberg> lol
<d1b2> <johnsel> Are you a drinker? I almost never drink, but I do have a tendency to empty a vat when I do drink
josuah has quit [Quit: josuah]
josuah has joined #scopehal
<_whitenotifier-f> [scopehal] penguin42 commented on issue #803: Rigol: Floating point exception - https://github.com/glscopeclient/scopehal/issues/803#issuecomment-1732626029
<azonenberg> I had a sip of beer in 2013 to satisfy my roommate who told me it tasted better than it smelled
<azonenberg> It tasted exactly like it smelled and confirmed my decision to not drink beer was a wise one
<d1b2> <johnsel> Haha that's is almost verbatim the answer I expected of you. And honestly I agree. I've never seen anyone improve from alcohol consumption. It's strange how ingrained it is, I think I drank my first lemonade with half a beer at 8 or so during our "carnaval" which look something like this https://www.iamexpat.nl/lifestyle/expat-events-festivals/maastricht-carnival with 3-4 days of copious amounts of alcohol and everything is closed
<azonenberg> I did want to honor the tradition of getting pictures of yourself with bottles of booze on your 21st birthday (legal drinking age here)
<azonenberg> So somewhere in my archives there's a photo of me holding a 4" wafer in one hand and a bottle of absolute ethanol in the other, on my 21st bday
<d1b2> <johnsel> Hahahha that's funny
<d1b2> <johnsel> Our legal drinking age for beer and wine used to be 16(18 now), I drank heavily every weekend from 14-18 or so, then a few years not much. Then at 24 or so I had a year or 2 where I drank 1-2 bottles of wine pretending to be a wine snob. At 27 though the hangovers started to become not fun anymore as well it would affect my mood under stress which made me unpleasant to be around so I gave it up entirely except for very special occassions.
<d1b2> Honestly knowing what I know now I'd rather have never started with it. But the Dutch are real alcohol and drug users as a country. Amsterdam is the number 1 city in the EU for cocaine traces in the sewer system
<d1b2> <johnsel> I should have developped a T&M habit earlier in life. So I would not have had the budget to buy fancy wine anymore
<azonenberg> Lol
<azonenberg> Maybe you should have taken up smoking NIST SRM 1082 and 1196
<azonenberg> You won't give yourself lung cancer because you can't afford to buy enough of them to increase your risk significantly :p
<azonenberg> actually they're not *that* bad. SRM 1082 is "only" $419 for a carton of 200 cigs
<azonenberg> and 1196 is $451 for 2 cartons, even cheaper
<azonenberg> I expected significantly higher
<d1b2> <johnsel> Haha that is so funny. I used to smoke, Dunhill International. Which is about the most expensive normal cigarettes you can buy. But yeah if I ever pick up the habit it's reference cigarettes for me only
<d1b2> <johnsel> But honestly smoking cigarettes is even dumber than alcohol. I'm using an e-cigarette now and plan to quit that too with the help of some meds that I know work for me. I used to be smoke free for almost 2 years but met fiancee who did smoke and so I started smoking again.
<d1b2> <johnsel> I prefer meditation nowadays over using alcohol or cigarettes. Far more healthy for you
<d1b2> <david.rysk> that's only what, 4-5x what they normally cost here
<d1b2> <david.rysk> taxes keep going up, probably to discourage smoking
<d1b2> <johnsel> Yup, I think pay 14eu for a pack of Dunhill International here nowadays. But they used to be 6-7 per pack when I smoked them. It's a discouragement tax. They're also forbidding e-cigarette fluid with fruity taste now. Apparently they felt that made it more attractive for young people. which makes some sense. Though ofcourse I can just import it from a few countries over without those rules, as we have free passage of goods in the EU. But if it
<d1b2> prevents young people from starting stupid habits I'm all for it
<d1b2> <johnsel> It's just a mild annoyance for me because my fiancee only likes 1 brand and taste and then I get tasked to find it anyway lol
<d1b2> <johnsel> I'm sure big tobacco sponsored quite some research into how e-cigarettes are bad driving these law changes
<d1b2> <david.rysk> in the US at least, a lot of it is about marketing
<d1b2> <david.rysk> Juul was buying ads on Nickolodeon and Cartoon Network apparently
<d1b2> <johnsel> Yeah that's how it started here. No marketing for it, hide behind counter, then no pictures even for it online
<d1b2> <johnsel> Oh no that was always forbidden here
<d1b2> <johnsel> Juul was never big here anyway, but yeah that makes total sense. Those things are/were also very styled to look like cool tech
<d1b2> <johnsel> I use a big rectangular box with a tip on it to suck on. That has no appeal as a fashion thing
<d1b2> <johnsel> But I know that Juul really targetted children, some people have no morals
<d1b2> <johnsel> Juul even are illegal here it seems, they contain 60mg/ml nicotine in salt form (more addictive, more rapid uptake) where here 18mg/ml is the max
<d1b2> <johnsel> but used to be available anyway given the demand from tourists, says a newspaper 2 years ago
<azonenberg> In other news, i have a (not merged) version of the spectrogram filters with all of the FFTS code gone but am running into some challenges batching vkFFT calls
* azonenberg muffled swearing in general direction of vkFFT
<d1b2> <johnsel> I'm sure the swearing is well deserved
<azonenberg> If i set "specify offset at launch" false, it works perfectly but i can only ever do a FFT at the beginning of a Vulkan buffer
<azonenberg> if i set "specify offset at launch" true, then set the offset to zero - which should be the samething
<azonenberg> i get invalid memory accesses in the shader
Darius has quit [Ping timeout: 240 seconds]
<d1b2> <johnsel> weird
<d1b2> <johnsel> I looked at the code but this too far outside my comfortable zone to make a guess what is happening really without studying it for some time
<d1b2> <johnsel> seems it affects the dispatch plan
<d1b2> <johnsel> probably something somewhere is not being set up properly in the adjust plan
<d1b2> <johnsel> but other than that I can't really tell
Darius has joined #scopehal
<d1b2> <johnsel> FFT looks good now
<d1b2> <johnsel> That's not custom rendering code though for the FFT+popups right?
<d1b2> <johnsel> I found another bug in it
<azonenberg> the fft peak labels are indeed custom