<Qboid>
[#173] title: Reduce garbage production by FlightGUI | The current dev-build of FAR creates megabytes of garbage under certain circumstances. The reason seems to be that FAR recreates the output texts for the flight data window every physics frame, for every vessel.... | https://github.com/ferram4/Ferram-Aerospace-Research/issues/173
<rsparkyc>
probably going to try applying that pr locally
<rsparkyc>
ferram4: i'll let you know if i see anything weird
<xShadowx>
rsparkyc: Sarb did some tests a while ago with GC and strings, if you're bored enough you could try finding his results / asking, and improve on that PR too, i seem to remember a couple steps made an improvement over just new stringbuilder
<rsparkyc>
cool, i might do that
<rsparkyc>
thx
<xShadowx>
i mean at some point it prolly crosses into OCD to get 0 GC but just passing on infoz i noticed ^.^
<rsparkyc>
you'll never have 0 GC
<rsparkyc>
at least not in c#
<rsparkyc>
but low hanging fruit is always good
<xShadowx>
you can reach (atleast in many cases) 0 garbage create by code you control
<rsparkyc>
i would think any time use use the `new` keyword, you're creating the potential to create garbage
<rsparkyc>
especially if your using that in the game loop
<rsparkyc>
so pretty much everything would have to be a singleton
<xShadowx>
ive seen a few people strive for 0 GC, or atleasst minimal, makes me wonder how much extra cpu is used with such methods
<xShadowx>
or if cpu saved etc
<rsparkyc>
yeah, that GC stutter in-game is pretty crappy though
<rsparkyc>
i'd give up a few FPS to have that go away
<xShadowx>
same
<xShadowx>
esp with >700 mods
<rsparkyc>
700? wow
<xShadowx>
.......i like a complete game
<rsparkyc>
i guess so
<rsparkyc>
do you kOS everything too?
<xShadowx>
not as much as i should
<xShadowx>
im bad at kos lol
<rsparkyc>
lol
<rsparkyc>
great, just updated TacLS, and now my save says i'm missing a part...
<xShadowx>
thats always fun
<xShadowx>
half reason i switched to a custom install
<xShadowx>
now i just update dll, and diff any cfg / copy changes i like
<rsparkyc>
weird thing is, nothing in the diff looks like it should have messed it up
<rsparkyc>
lol, i'm an idiot
<rsparkyc>
apparently i have a hard time pasting things into my gamedata folder
<rsparkyc>
I'm like "weird…i cant find taclifesupport anywhere in my ModuleManager.ConfigCache…"
* xShadowx
claps slowly
<Probus>
xShadowx, didn't you say you use a ramdisk?
<xShadowx>
ya why?
<Probus>
What program do you use to set it up?
<xShadowx>
ramdisk
<xShadowx>
lol
<xShadowx>
theres a ton of diff options
<Probus>
I saw some you could buy. That's why I asked
<xShadowx>
just google and browse for one you like
<rsparkyc>
oh wow
<rsparkyc>
The launch an Ariane 5 rocket with Brazilian and Korean communications satellites that was set for this week has been postponed indefinitely after protesters blocked access to the French Guiana spaceport.
<xShadowx>
hah
<stratochief>
flat earthers?
<rsparkyc>
who knows
<xShadowx>
you arent one of those silly people who think earth is round are you
<xShadowx>
just because its round in a video game doesnt make it real
<stratochief>
You sure ate your Wheeties and put on the troll cap today, hey xShadowx?
<xShadowx>
stratochief: eh? thats my first troll today ;3
<Starwaster>
just because the Earth is round in real life doesn't make it real.
<Starwaster>
true story: The ISS has never completed an orbit. It fell off the edge of the world years ago and has never been seen again
<Starwaster>
o.O
<stratochief>
Starwaster: that first line sounded like a Jaden Smith line
<Starwaster>
ok so the exploding engines thing has been fixed
<xShadowx>
:)
<Starwaster>
I'm still not really clear on the root cause except... I think it's really in EngineThermodynamics where the ambient temperature seems to spike when decoupling engines on some craft
<Starwaster>
for the time being I just removed that as a variable in SolverRF on non-combusting engines
<Starwaster>
(spiking as in over 1000000 Kelvin)
<stratochief>
Starwaster stomping bugs just for fun :)
rsparkyc has quit [Quit: Leaving.]
<Starwaster>
stupid little squishy bugs
<Starwaster>
I forget how to actually play this game... seems like I'm always doing mod stuff now
<stratochief>
Starwaster: lol. well, NK and I, and plenty of others have youtube series you can watch if you ever want to re-learn :)
<stratochief>
and those series wouldn't be possible without modders and bug squishers like you
<xShadowx>
is it sad that i found NK streaming RP-0 to be better than the worthless crap on tv nowadays?
<xShadowx>
i miss the age of spacey cifi shows :(
<xShadowx>
s/cifi/scifi
<Qboid>
xShadowx meant to say: i miss the age of spacey scifi shows :(
<stratochief>
xShadowx: lots of people get surprised by how enjoyable watching others play video games is :)
<xShadowx>
oh im not one of those people
<stratochief>
I just realized a week or so ago that there was another season of the Expanse out. that was damn fine
<xShadowx>
i hate streamers, id ratherplay ;p
<xShadowx>
NK was just good at it :D and mimic reql life
<stratochief>
I mimicked NASA pretty well up until 1963 or so, now I'm around 1966-1967, got a Gemini Direct onto the Moon and going for MOL/Gemini to get some station experience
<xShadowx>
:)
SirKeplan is now known as SirKeplanzzz
<stratochief>
1967-1969 we'll do a variety of Apollo wetlab Skylab missions, a double launch Saturn V extended lunar stay using a LEM derived lunar shelter, then onto Mars
<stratochief>
although I'll be getting a start on successful Mars landing this year, even a bit of basic ISRU to get out feat wet
<xShadowx>
and a mars colony!
<stratochief>
naw, I'm a Mars Direct man. not sure what I'll be doing post 1980 though, maybe a colony
<stratochief>
I'm just trying to get as much done through the 70s, with the savings derived from skipping Apollo, only using AAP and skipping Shuttle
<xShadowx>
skipping o;
<xShadowx>
i havent seen a good shuttle done
<stratochief>
all the methalox technology NASA develops during the 70s might inspire some private aerospace to build big methalox rockets, who knows what the 1980s hold :)
<stratochief>
xShadowx yeah. in reality or on KSP. RIP affordable, regular trips to LEO
<stratochief>
Kosmonaut Crash has a decent little shuttle
<xShadowx>
ii always see people make symmetrical shit, shuttle gotta work for that balance
<stratochief>
he is still working on landing it well, but it is mostly inexperience and bouncy AF landing gear that hurt him. good shuttle besides that
<stratochief>
I really enjoy his series, more than my own :)
<TheKosmonaut>
stratochief: ok? When did I ask about a shuttle? Lol. But now I'm checking
HypergolicSkunk has quit [Quit: Connection closed for inactivity]
blowfish has joined #RO
<blowfish>
I took a look at RF, I think I know what's causing the NRE but it's not going to help the overheating issue
<blowfish>
oh, looks like Starwaster already figured it out
<blowfish>
no, that's weird. t0 should be ambient temperature
<blowfish>
going to have to do some debugging
<blowfish>
the NRE issue will be good to fix anyway though
<blowfish>
confirmed the NRE is fixed by ignoring the overheat box outside of the active vessel
<blowfish>
Starwaster: I am observing that the vessel ambient temperature is some ridiculous value for the first couple of frames after detaching. Is that what you saw?
<blowfish>
might be a stock issue
<Starwaster>
yeah
<Starwaster>
I took it out of the equation for now
<Starwaster>
did not address the NRE
<blowfish>
I found a simple fix for the NRE, just check for active vessel before trying to update the overheat box
<blowfish>
sounds like it might be time for a new release of SolverEngines and RF
<blowfish>
SolverEngines also has the normalizedOutput issue which was fixed
<Starwaster>
Might be some other RF issues that could be looked into
<Starwaster>
one thing I've been meaning to do soon was look into the analytic functions again for boiloff.
Hohman has quit [Quit: ChatZilla 0.9.93 [Firefox 52.0.1/20170316213829]]
<soundnfury>
hmm, looks like realheat still hasn't been recompiled for 1.2.x? what's holding that one up, anything or just tuits?
Senshi has joined #RO
<Starwaster>
nathan hasn't really been around much and RH is his baby
<soundnfury>
fair enough
<soundnfury>
further hmmmms: my load appears to hang on VSR IVAs, and also produce an NRE in the log. I've baleeted the IVA folder, let's see if it works now
Wetmelon has quit [Ping timeout: 206 seconds]
<soundnfury>
ahh, I was using an old version of VSR
<blowfish>
there are lots of versions of that floating around, since Ven I would say has unofficially abandoned it, but many people are interested in keeping it up to date
<soundnfury>
blowfish: actually, Ven appears to have done a release on 4th Feb
<blowfish>
ah yeah, not so long ago
<blowfish>
no commits to master since then
Senshi has quit [Ping timeout: 180 seconds]
blowfish has quit [Quit: Leaving]
<soundnfury>
well this is weird, KCT is throwing Method not found: 'MagiCore.MathParsi
<soundnfury>
ng.ParseMath'.
<soundnfury>
and yet, I have the latest MagiCore installed
<soundnfury>
has anyone else managed to run KCT on 1.2.2?
<soundnfury>
a-HA! magico did a breaking change, so I need *newer* than newest MagiCore :/
<egg|zzz|egg>
!seen nathankell
<Qboid>
egg|zzz|egg: I last saw NathanKell on [25.03.2017 06:39:45] in #RO saying: "And on that note, night all :)"
Senshi has joined #RO
HypergolicSkunk has joined #RO
SirKeplanzzz is now known as SirKeplan
ferram4 has quit [Ping timeout: 200 seconds]
AlimOncul has joined #RO
Probus has quit [Read error: Connection reset by peer]
AlimOncul has quit [Client Quit]
Probus has joined #RO
Probus has quit [Read error: Connection reset by peer]
Probus has joined #RO
rsparkyc has joined #RO
Probus has quit [Read error: Connection reset by peer]
Probus has joined #RO
<github>
[RealismOverhaul] AlphaMike741 opened pull request #1586: Launchers pack Falcon 1 part names update (master...KK-F1-newnames) https://git.io/vSIyu
Probus has quit [Read error: Connection reset by peer]
Probus has joined #RO
Probus has quit [Read error: Connection reset by peer]
rsparkyc has quit [Quit: Leaving.]
Probus has joined #RO
Starwaster has quit [Quit: Leaving]
<Probus>
Sorry, something is failing in my computer
Probus has quit [Read error: Connection reset by peer]
rsparkyc has joined #RO
rsparkyc has quit [Quit: Leaving.]
Probus has joined #RO
Probus has quit [Read error: Connection reset by peer]
Probus has joined #RO
Regi has joined #RO
<Regi>
Is this a proper place to report bugs/glitches without making a forum post?
Probus has quit [Read error: Connection reset by peer]
<waerloga>
Regi: github depending on which part
Hohman has joined #RO
Probus has joined #RO
<Regi>
I'd probably go to GitHub for something like frozen vessels, correct?
Regi has quit [Ping timeout: 180 seconds]
VanDisaster has quit [Ping timeout: 206 seconds]
rsparkyc has joined #RO
diomedea has joined #RO
leudaimon has joined #RO
<stratochief>
ohh, poor Regi. what does frozen vessels even mean?
<stratochief>
Probus: or are you in Wifi? that is generally the problem BadRocketsCo had when he was on and off all the time
Wetmelon has joined #RO
Starwaster has joined #RO
<zilti>
Well, this is definitely the last time ever I bought any kind of product from ASUS.
rsparkyc has quit [Quit: Leaving.]
<Starwaster>
Been there threatened that
<zilti>
A company that has as standard what I considered a "monday product" six years ago but kept because it worked good enough is just not one I'll ever buy anything again. My new one is basically the same crap they sold six years ago, only with a new chipset and flashy light. Even the software suite is the same heap of crap.
<waerloga>
huh...never had problems with the asus products I've gotten
<waerloga>
they are typically on my shortlist for motherboards and have been for a number of years
<zilti>
Final nail in the coffin was the BIOS update. Of course their tool, which looks still almost identical to what it was like six years ago, screwed up. PC wouldn't boot anymore. "No problem, there's 'ASUS CrashFree BIOS 3 on it!'"
<zilti>
Well guess again, it doesn't work.
<zilti>
And a quick Google search shows that "ASUS CrashFree BIOS 3" is shoddy since 15 years
<zilti>
Fifteen years. I was still in grade school back then!
<Probus>
I had a SATA cable come loose. Very weird. Anything I did on my computer was delayed about 15 seconds.
<Probus>
My MB is ASUS. Had an AdRock before that and I didn't like that one at all.
<zilti>
Well, I guess I'll have to return my mainboard for RMA tomorrow and then live without a PC for a couple weeks... damnit...
<Probus>
Maybe you can get them to send the replacement at the same time you are returning it.
<zilti>
Hmm, yes, iirc the shops replace it if you return it broken in the first five workdays. Then again there's an absolute shortage and I'll have to wait two weeks anyway
<Starwaster>
Ooooooo a flashy light? What more could you ask for?
<zilti>
The four feedback lights my old mainboard had :P
Hohman has quit [Quit: ChatZilla 0.9.93 [Firefox 52.0.1/20170316213829]]
<zilti>
And the built-in "stick a flashstick into the designated port, turn on the power and press here to flash the BIOS" button that'd work even without a CPU and I never used
AlimOncul has joined #RO
Wetmelon has quit [Ping timeout: 206 seconds]
zilti has quit [Quit: WeeChat 1.7]
blowfish has joined #RO
zilti has joined #RO
<xShadowx>
zilti: i'm the oposite, every brand i touch fails me except asus, motherboard wise, about given up and wont touch non asus now
<Starwaster>
Stratochief: Coming soon to Heat Pumps: Proper analytic handling
<stratochief>
Starwaster: cool! at what point does the thermal system go analytic now? 100x? 1000x?
<zilti>
Ooh, looks like I'll install another mod :P
<Starwaster>
Uses whatever is set in PhysicsGlobals. If another mod or the player changes that then that's what gets used
<Starwaster>
at the same time I'm re-enabling analytic handling in Real Fuels so that they work together properly. No idea what that does to stock radiators but then stock radiators still don't handle analytical mode anyway. AFAIK
<xShadowx>
if i wanted stock, i wouldnt install a mod <3
AlimOncul has quit [Quit: Web client closed]
<xShadowx>
stock ksp is like windows, all the included functions are "good enough for basic use" but 3rd party is ment to replace em :)
<zilti>
I thought heat isn't even a thing that matters in stock anyway?
<xShadowx>
pure stock? it makes you blow up now
<zilti>
Up in space?
<xShadowx>
mostly only gain such heat from re entry, or ......wel....if you're that close to the sun you have mental issues more important than the heat :P
<zilti>
"I'm convinced that before the year 2000 is over, the first child will have been born on the moon." - Wernher von Braun
<xShadowx>
he didnt anticipate human greed and laziness
<zilti>
Heh there are lots of neat and also funny quotes by him. I didn't even know that.
<zilti>
I only knew some by Korolev, among them one of my favourite quotes: "The genius of a construction lies in its simplicity. Everyone can build something complicated."
<xShadowx>
i wonder whatd happen if you made a a large chamber to simulate gravity, likely with centrifugal force, but keep say 2+ G inside, then some kid from day hes born spends 20 years in it - come out strong and able to beat up everyone?XD
<zilti>
Hmm, actually I expect the answer to this to be "yes" :P
Thomas is now known as Thomas|AWAY
<Starwaster>
he'll come out shorter but a lot stronger than a 1G person
<Starwaster>
btw I love that bit you quoted from Korolev about simplicity. I haven't heard it before but I have a similar view
<Starwaster>
zilti
egg|laptop|egg has joined #RO
<egg|laptop|egg>
zilti: btw, Iskierka tested the Linux build, it appears to work (as well as the windows build does at least) https://imgur.com/a/sMbcM
<zilti>
Starwaster: weirdly enough, I'm absolutely unable to find that quote in english. I can only find it in german :P
<egg|laptop|egg>
(will release on the new moon as scheduled)
diomedea has quit [Quit: Bye!]
<Starwaster>
now that my feral cat is tame, I'm finding it dangerous to be under the covers in bed. She loves to assault me
<Starwaster>
she's kind of rough actually
<zilti>
egg|laptop|egg: nice! I thought about learning CMake btw, and chose Principia as my learning ground. I'll send you the CMake file in the end, if you like it you can keep it. It also generaltes sln and UNIX Makefiles, so you could get rid of having separate stuff.
<zilti>
egg|laptop|egg: On other news, I bricked my new mainboard with a failed BIOS update
<egg|laptop|egg>
zilti: frankly I'm happy with the makefile as it is, I have some confidence in its correctness; I'll next touch the build system when bazel works with our filenames and in the meantime forget about that lovecraftian madness
<zilti>
bazel?
<egg|laptop|egg>
that makefile certainly isn't idiomatic, but it works
<egg|laptop|egg>
(doesn't support Unicode filenames right now though)
<egg|laptop|egg>
(they're working on it)
<egg|laptop|egg>
also afknomz
egg|laptop|egg is now known as egg|nomz|egg
egg|zzz|egg is now known as egg|cocktail|egg
<zilti>
I gotta admit, I feel like bazel looks like a case of the "Not invented here, we'll build something on our own"-hipster-syndrome.
<Starwaster>
I like standardization
egg|nomz|egg has quit [Ping timeout: 180 seconds]
<Starwaster>
huh.... AnalyticInfo docs say 8 params but it actually passes 9..... wtf is the 9th?
<Starwaster>
ok apparently I just can't count
egg|laptop|egg has joined #RO
<egg|laptop|egg>
zilti: it's good though
<egg|laptop|egg>
(as in, it doesn't drive you insane when building Google-sized codebases with it; whereas make does)
<zilti>
egg|laptop|egg: (That's why there's stuff like Autotools for Unix and CMake for cross-platform ;) )
<egg|laptop|egg>
zilti: it's not like Googlers don't know about those
<egg|laptop|egg>
zilti: but they don't really guarantee correctness in the same way
<egg|laptop|egg>
(bazel copies the stuff to a clean place, uses hashes to check if the files have changed)
<egg|laptop|egg>
(and of course at Google it does massively parallel madness, where when you have a coffee it does 3000 hours of compilation)
<egg|laptop|egg>
none of the Cmake etc. help with correctness
egg|laptop|egg is now known as egg|nomz|egg
<zilti>
Correctness how though? CMake etc. re-compile files and stuff that depends on them when they get changed.
<zilti>
Using hashes instead of timestamps seems like a way to waste more CPU time and in the worst-case screw up a build when two file versions result in the same checksum
egg|nomz|egg has quit [Ping timeout: 180 seconds]
<Starwaster>
Which is executed first, VesselModules or PartModules?
BasharMilesTeg_ has joined #RO
<xShadowx>
i literally tried that last night and now totally no clue ;3
* xShadowx
thinks a post should get made showing the order
BasharMilesTeg has quit [Ping timeout: 201 seconds]
rsparkyc has joined #RO
rsparkyc has quit [Quit: Leaving.]
<soundnfury>
xShadowx: re G chambers - hasn't been done with people, but has been done with chickens
<xShadowx>
humans growing slowly over many more years has more time to adapt to the gravity
KevinStarwaster has joined #RO
Starwaster has quit [Quit: Leaving]
<soundnfury>
xShadowx: just means you gotta keep people in there for longer :P
rsparkyc has joined #RO
Senshi has quit [Quit: Leaving.]
<leudaimon>
xShadowx, all bone and muscle development happens at high G in both cases, I see no reason to have very different results
<egg|cocktail|egg>
zilti: timestamps change all the time when file contents don't (if you use version control)
<egg|cocktail|egg>
zilti: CPU time from hash calculations is dominated by compilation time, false optimization
<zilti>
egg|cocktail|egg: You can have your VCS to keep timestamps though (isn't that even default?)
<egg|cocktail|egg>
zilti: yes, but even then, you can edit a file and undo that
KevinStarwaster has quit [Quit: Leaving]
<egg|cocktail|egg>
this happens a lot
KevinStarwaster has joined #RO
<egg|cocktail|egg>
or you can have colliding timestamps rather more easily than colliding hashes with version control
<egg|cocktail|egg>
zilti: and correctness as in not assuming things about your build system
<egg|cocktail|egg>
isolating the files that get used as input
KevinStarwaster has quit [Client Quit]
Starwaster has joined #RO
Starwaster has quit [Client Quit]
Starwaster has joined #RO
<egg|cocktail|egg>
because builds quickly get nonstandard; not everything is just "feed this to the C compiler and the deps are the includes"
<egg|cocktail|egg>
you have custom code generators etc. in any sufficiently complex project (Principia has that)
Starwaster has quit [Client Quit]
<zilti>
egg|cocktail|egg: Also, you can tell CMake to work with hashes instead (or, more precisely, have it generate the makefiles or visual studio projects so they check for hashes instead of timestamps for builds)
Starwaster has joined #RO
<egg|cocktail|egg>
zilti: anyway, if you know how to make a better build system than what Google switched to from make-based things, I think you can find yourself a very well-paid job :-p
egg|cocktail|egg is now known as egg|zzz|egg
<zilti>
egg|cocktail|egg: I just feel like what they promise to solve all is already solved, and it's another of those Google projects that are maintained by Google so they have control over it and for the sake of it.
<egg|zzz|egg>
makefiles don't isolate the input files
<egg|zzz|egg>
so the compiler can have hidden dependencies that the makefile doesn't know about
<egg|zzz|egg>
and that make the build nondeterministic
<egg|zzz|egg>
blaze was developed internally for the purpose of being an internal tool far before it was released
<egg|zzz|egg>
because make just wasn't up to the taks
<egg|zzz|egg>
s/ks/sk
<Qboid>
egg|zzz|egg meant to say: because make just wasn't up to the task
<zilti>
The compiler can only compile what it knows about, so you can just clear out the standard include paths in the makefile? Yes, I can understand that make itself isn't up to everything
<egg|zzz|egg>
zilti: not everything is the C++ compiler
<zilti>
I mean, that's why e.g. Autotools were created in the first place (with the problem that those are UNIX-only)
<egg|zzz|egg>
complicated builds have many things that take many inputs
<egg|zzz|egg>
in principia we build tools.exe, that generates other code
<egg|zzz|egg>
nothing is going to verify the dependencies of thta
<egg|zzz|egg>
large codebases are full of that
<soundnfury>
egg|zzz|egg: my eggsperience has always been that Makefiles are plenty up to the job
<soundnfury>
and that includes code-generators in the build, I have those in e.g. 'harris'
<egg|zzz|egg>
anyway this discussion can serve no further purpose
<soundnfury>
heh
<egg|zzz|egg>
s/.*/this conversation can serve no purpose anymore
<Qboid>
egg|zzz|egg meant to say: this conversation can serve no purpose anymore
zilti has quit [Quit: WeeChat 1.7]
zilti has joined #RO
rsparkyc has quit [Quit: Leaving.]
VanDisaster has joined #RO
<Starwaster>
crap, forums are down. Anyone know the NEW way to check for resource availability without actually requesting resources? I seem to recall a different method now
ferram4 has joined #RO
<xShadowx>
Starwaster: part.getconnectedresourcetotals i think :|
<xShadowx>
wish it gave connected resource parts list heh
<Starwaster>
that was it
<xShadowx>
:)
Probus-Laptop has joined #RO
Probus-Laptop has left #RO [#RO]
Probus-Laptop has joined #RO
UmbralRaptor is now known as GradingRaptor
rsparkyc has joined #RO
Probus-Laptop has quit [Ping timeout: 206 seconds]
HypergolicSkunk has quit [Quit: Connection closed for inactivity]