#off-topic-tech
1 messages Β· Page 111 of 1
wifi adapters aren't that expensive are they?
And ARLT seems to offer fully custom too, or might be able to change the MB to one with Wifi etc. when asked.
ARLT system plus this?
www.mindfactory.de/product_info.php/Gigabyte-GC-WBAX200-PCIe-x1-WLAN-Adapter_1335410.html
But again for fully custom, 150e extra.
should be easy enough to install
good to know ill head there soon thx
Mainly online store, but if you live near, warranty work is so much simpler.
QLC SSD, tho good QLC
not a fan of the CPU pick cuz it locks you into AM4 with nowhere to upgrade to
rest is fine tbh, tho idk if 1k EUR is a good price for it
Yeah, problem is that AM5 adds too much cost with proper GPU etc.
shit SSD
that ultimately depends on usecase tbh
also I can't judge Endorfy's quality at all. I checked the Supremo FM5 and Kitguru review looks to be good
Windows file transfer is so painful
I've been waiting over 5 minutes for this to even start copying
Arlt gives zero info on the PSU they used which is a big red flag
yeah, small files suck. compress, copy, extract
I may do that π
And in this case using Alternate.de as source would be better for warranty anyways.
But their preconfigured systems in the right price range are crap.
Yeah, no where near perfect system.
But only one with good set of CPU, RAM and GPU in about right price range.
But yes, PSU can be anything (but as the computer has actual warranty, I wouldn't expect them to use the absolute bottom of the barrel stuff), and the 16GB RAM could be more.
And of course the drive has been skimped on to absolute worst you can basically get.
Same as basically any prebuild that isn't made-to-order.
And that seems to raise the price a lot in Germany compared to that one, by hundreds of euros for same set of components.
made-to-order can't guarantee that everything works together. so they raise the prices to cover the possibly higher number of waranty cases
too lazy to whip up Geizhals or PCPP now to make an actual 1k EUR list rn
Problem is that system completion, that seems to be about 120-150e depending on place.
prebuilt/finished build is a requirement. so a pure part list won't work
And that specific system is somehow cheaper for the parts than getting them separately, even leaving the OS and building costs out of it.
cause the don't charge tax for every part, just the finished product
Also volume etc. which might get some of the parts under normal prices for fixed setup.
if you live near me, it just costs providing me with food/drinks and being a likeable person who's willing to learn
Not me, somewhere north of Frankfurt it seems for the OP
also a factor yes
plenty people that are willing to build either for free or a low extra
@dire igloo final exam on wednesday 
after that only project doc and presentation but that's ez
Viel Erfolg und gutes Gelingen
with luck end of january I'm gonna be officially an IT system engineer
nice
can't wait to never use any of the stuff I learned ever in the real world.
hooray!
unrelated but i'm finlly done with my stalker 2 testing
and i guess the RX6600 is now a 720p card
slight edits
given the devs were in Ukraine when the you-know-what happened I can't really be mad at them for not fully optimising it
Seems mainly that they didn't include any real low resolution textures at all.
So if you don't have at least 8GB, or possibly more, VRAM, you are fucked.
Missing the VRAM numbers for medium and low in that file.
yeah i forgot about them
easy fix
Especially is UE5 in general has that possible "care about whole system VRAM usage" idiocy in it, with current various background programs VRAM usage.
wtf? testing on an unoptimized gfame like stalker 2? π
though when i was testing low/medium i saw 6000mb on medium 55xx on low if i remember corectly
So with the normal UE limitations, even with Low textures, anyone with 6GB card is fucked.
its like testing Cyberpunk when it got released. Madness. Test are garbish.
But that UE5 thing where when it would go over the internal limits, it drops FPS to that 5-15 FPS stable value, where the exact one depends on the system, would explain the stuff around.
For VRAM usage, with those VRAM usage numbers.
Basically lot of GPUs just don't have enough VRAM to work properly even on Low textures.
And it would be even worse with any Electron apps on background etc.
smooth steady garbage
perf
And as UE does that before the VRAM is actually full, not easy to know about.
that is also what i meant when i wrote: does the game break?:
So many people even with Satisfactory 1.0 who just couldn't accept their FPS dropping to that low value after changing biomes is from trying to use too high textures for their VRAM.
As any actual monitoring doesn't show the VRAM as being full, as UE is just doing self-fuckery based on internal stuff.
Before the VRAM is actually full.
its also been pretty anoying that i have to wait for the shaders to compile every time i start the game
im gonna test medium and low while actully paying atention to the VRAM utilization
from the tests i saw, stalker uses about 8gb of vram on average at 1080p high textures iirc
so dropping texture settings should allow for an 8gb card to do fairly well
Testing with high VRAM card isn't really usable.
As properly coded game will use as much VRAM as it can, if there is any available.
in the little settlement yes,but in the wilderness no..
i see
but the game isn't optimized if you read internet π
And for UE5 on 8GB card, it will only use max 6.5GB of VRAM, ever, if the engine isn't modified.
i would test myself but i have 16gb of vram so it isnt that useful
really? i have used UE5 and it has absolutely used more than that
7900GRE right?
yeah
For system total?
Or the game itself?
editor
By default UE4 and UE5 in release build only use max of 81% of dedicated VRAM.
that makes sense, expecting people to have discord and chrome open
Well, ask Windows for a number, that seems to usually be 90% of VRAM, and then multiplies that by 0.9, to get internal limit.
ZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ
And then texture usage max is 70% of that 81%.
this did confuse me as ue5 has support for shader cache
Doesn't make sense, as the engine does very stupid shit to not have GPU driver manage RAM as VRAM, and tanks the performance much worse than letting the GPU driver to do its thing as it likes.
And that limit has been there since first UE4 version.
maybe its a leftover console thing
but the GPU driver puts that stuff into system ram when it has to
maybe
i cant think of a reason why they would do that otherwise
Were just paranoid about going over VRAM in past.
And haven't refactored since, after that has not made any sense anymore.
maybe gpu drivers were bad at managing it back in 2013/2014
And doubled down in UE5.
In UE4, with DX12, the engine would just crash when it hit those internal limits.
In UE5, with DX12, it starts to manually swap stuff every frame.
On DX11, both just downgrade random textures.
And that manual swapping or whatever it is doing, causes that very low but stable FPS state when those internal limits are hit.
better than just crashing
great frametime graph
true
Yeah, but if the self-limiters weren't there, then the GPU driver would just start doing stuff instead, with better performance than the UE5 internal implementation.
That is the stupidity, Epic reimplemented GPU driver functionality, at worse performance.
Yikes, for medium
GPU driver functionality to use RAM as VRAM extensions were a thing BEFORE UE4 was released.
it may have been, but if they werent when it was being developed, that could explain it
medium VRAM usage in the settlement: dGPU 7200MB /process 6000MB
But seems the devs have edited away those UE internal limiters, as none of those are hitting them.
But basically seems from both detail comparison images and those graphs, that the game doesn't actually have low detail mode.
Just bunch of really close together settings.
im getting around 60 epic settings 1440p tsr ultra quality
nanite makes low model settings hard since there arent actually any lower poly LODs
yeah, tsr π
looks fine. I dont notice a difference
When the worst result is 38 vs 21.5 FPS between epic and medium, or 101 vs 86 for 4090 for example.
Author of tom's hardware hits the nail on the head:
It's still frustrating to see Unreal Engine 5 apparently requiring so much help to reach higher levels of performance**. Framegen in particular should not be the baseline recommendation**. It has some clear drawbacks, and tends to be more useful as a marketing feature. We'd rather have slightly less advanced graphics that run well, over interpolated frames to smooth out some of the choppiness. Other games and game engines seem to manage this much better than UE5, but Unreal Engine gets used a lot because it makes life easier on the game developers β often at the cost of performance.
And it is from Epic...
it really isnt UE5 specifically, fortnite is UE5 and it scales pretty well
Ah, not framegen, just upscaling.
medium VRAM usage in the wilderness: dGPU 6200mb /process:5000MB
You can get per process dedicated and shared VRAM numbers from the details tab:
After enabling the columns, not shown by default.
@stray badger You played Stalker 2?
like 30 mins yea
Oh, ok π
isnt that what MSI afterburner shows?
No, that is total system use.
I read some claims the first 2 hours are pretty good. But then game falls apart from bugs, glitches etc.
Wonder if it's true or some nonsense.
Minimized to icon MS Outlook is using 1GB of VRAM, Chromium browser with few windows is using 1.8GB...
and where can they be enabled?
Right-click on any of the headers.
Then Select Columns, find and tick.
As in Task Manager Details tab.
Review builds were broken, day 0 patch fixed most things. Game has been patched several more times as well
Still has some anomalies that i have yet to encounter
Well let me know after you played a few hours π
30 mins is not much to go on π
I will probably wait a year before buying the game, let them put out a few more patches.
i've played the game for 54 mins (before testing) and i havent seen a single bug yet
haha, probably wouldn't take me long to find bugs
i have tried that but it didnt work
also if this doesnt show /process util then what is it that it shows?
Might show then.
nevermind found it
ZZZZZZZZZZZZZZZZZZZZZZZZZzzzzzzzzzzzzzzzzzzz
they are showing the same values
Ok, so then that was with that extra setting really the games usage.
low VRAM usage dGPU6500mb /process 5600MB (in the settlement)
dGPU5800MB /process 4800MB in the wilderness
the amount of ghosting on low is crazy
TAA?
yeah,FSR does a bit better
Or just see what raising just AA quality does?
is that a painting as desktop background?
or is that the game in the back? π
shader compile screen
Ah ok. Because if thats how the game looks. Oof.
also would not recomend playing without sounds
i would love it if the amd overlay didnt take 2 hours to open
which one?
alt + z opens side bar. ctrl + o opens stat overlay. alt + r for fullscreen overlay.
its not the drivers fault though
fullscreen
because thats the only one thats usable
whaaaat?
- They open instantly for me.
- They all work fine for me, no issues.
as i said its not the drivers fault
rather the fault of THE shitiest SSD you can buy
π€
Are you saying because of the SSD, overlays for AMD open slow?
Maybe your system is just overloaded π
no its literaly EVRYTHING that is on my C drive that opens slowly stutters etc...
AMD Software seems to have issue π€£
i also found no enemis here
move around a bit
they can spawn around you / in front of you
it's a an issue a lot of people are talking about.
i already looted the whole place
though FSR looks like straight up garbage in this game
this thing is F*CKING teryfing
it's a horror survival game π
In a radiation zone (nuclear disaster).
Something I really miss in shooters, is having advanced features like leaning etc.
this game has leaning
pigeon VS games journalist
bro...
leaning is so shitty
bro ...
I hate that, just make it automatic when I'm peeking a corner, every ubisoft game has that
and ubisoft is a low bar
I just think leaning is more logical and realistic then turning around a corner like you do in counter strike.
The Dark Souls of tutorials.
stalker 2s leaning system is automatic
What was that xbox game that allowed us to cover behind obstacles?
good
Gears of Wars!
q & e can be used for so much better stuff
Games like Arma 3 and Tarkov have best shooter mechanics imo.
Tarkov lets your select your crouch height. So you can make yourself smaller, without going into full crouch.
You have to press crouch + mousewheel up or down to adjust height. This allows you to peak over obstacles and cover yourself.
tbh that sounds like way too much
It does. But it becomes second nature.
Same for leaning, you could control the leaning angle.
All this, so you could avoid getting one tapped instantly.
And lose all your gear.
all that only to get killed by a hacker
100% π
But that's another story. The mechanics are pretty cool, I whished more shooters evolved.
A lot still feel and play like counter strike in 2000.
But maybe I am comparing tactical shooters with arena shooters.
I wish the moment a cheat was detected they would work around the macro and coding to prevent it rather than banning
bc a DMA card can literally just fix that
bc its kinda hard to be banned when ur running your cheats thru a clean pc
how about detecting the cheat in the first place
COD does it better. Bots hunt down cheaters. And make their life misserable. Or so the developers claim.
and developers need to focus more on writing script in their code making it virtually insane to try to script/hack
Probably all marketing. And doesn't work. But atleast it's a new idea.
CS:GOs overwatch was insane actually
Original System Shocks lean system was pretty clunky :/
CS:GO had* a system called overwatch
π π π π π π π π π π π π π π π π
oh
it was community based review of reports
i was boutta say blizzard pays cheaters
it was actually pretty good
literally you can find an old artical of devs giving community access to certain source codes
I want to play games, not be a judge for 10-15 minuts.
I got so many reports I had to review. It was bullshit.
it was completly optional with a few rewards
guess they learned from that
dota 2 so bad they cant even finish their season 2 of their show while league has alr started π
and then removed it again in CS2 lmao
I quit Dota 2 ages. Was my favorite game.
you kill some 8-year old then go to sleep and then you get bannedπ
overybody loved cs:go overwatch
Because you cannot trust community.
stop saying dota was ur favorite game
People will say innocent people are guilty for fun and giggles.
say it was smite rn
it was actually pretty accurate, you needed like 10 votes to convict someone
It's the internet. You cannot hope for common sense.
just like for ppl in the whole politcal misinformation scheme
hell we are even being brainwashed now into beleiving their was no seahorse emoji and yet i have a picture on my phone of me having it
so idk bruh this whole gaslighting thing aint gon work
I think they added it again.
And best of all, community is in the dark.
Testing STALKER 2 on the RX 7900 XTX 24GB graphics card! I tested 1080p, 1440p and 4K resolutions, with and without FSR and frame generation, at Epic settings!
β± Timestamps β±
Intro, specs, stuff - 0:00
4K Epic - 0:48
4K Epic / FSR Quality - 3:01
4K Epic / FSR Performance - 5:27
4K Epic / FSR Performance / FG - 7:32
4K Epic / FSR Quality / FG -...
Valve removed overwatch. Without mentioning it in patchnotes it seems.
Then they added it back a few months later. Without any further information.
But who wants kernel level anti cheat?
"Huh, what? Common. Oh no. The game is bugged..." π
Already done for comparison images?
https://www.pcgamer.com/hardware/stalker-2-heart-of-chornobyl-best-graphics-settings-guide/
i wonder if textures could be downscaled at runtime to create lower texture levels
low doesn't look that bad ....
Which was the point of why I said the game doesn't have really low settings.
or maybe during/after shader compilation
As even the lowest settings are very close to the Epic ones in quick glances.
Ok, I am not crazy.
Because I was thinking the same Baldur.
I really have to stare at both images to find the differences. Which are there obvious.
The full resolution versions:
but the images on pcgamer are downscaled weirdly.
object details doesnt do much, the only assets it affects are foliage, characters, and guns
Need to first get the fullscreen button on bottom right, then use the "view original" even after that.
The devs have been drinking paint if they think that's "low"
π
look at the concrete
a lot of detail in stalker is actually nanite geometry
the concrete it's visible, but low isn't that bad.
I may just be a stupid simple gamer
look at the roof of the house in distance.
But the concrete is one of my least concerns
trees seem to be where the most detail is lost since they are not nanite
it looks almost the same as epic. Just the lightning seems different.
may be weather
Yeah, or just the sun moving.
just the difference in timing in actually making the screenshots
i more so meant comparisons between upscalers
Yup. But I can't be sure. Don't know the game and if it has dynamic sun / weather etc.
wrong comment
ground textures look pretty gnarly at low
that i pinned
it does. thats a pretty central part of stalker
low textures actually really does lower texture quality, and it is noticeable, but a lot of details are actually real geometry, which is why it still looks good
im convinced they are at the same settings
I see some lightning change.
foliage and NPCs are the only things that change
all static meshes are nanite
nanite cannot lower quality
You have a humvee in the back, the window color changes.
probably just lighting changes
and it seems the crate gets a little bit more detail.
object detail may be a bit bugged tho, since the things that should change dont
but really, I have to peer into the screenshots.
is likely destructable and therefore not nanite
its the exact same picture just 2 seconds apart
https://www.youtube.com/watch?v=lEc-py9zqKs go to 17:20
Testing STALKER 2 on the RX 7900 XTX 24GB graphics card! I tested 1080p, 1440p and 4K resolutions, with and without FSR and frame generation, at Epic settings!
β± Timestamps β±
Intro, specs, stuff - 0:00
4K Epic - 0:48
4K Epic / FSR Quality - 3:01
4K Epic / FSR Performance - 5:27
4K Epic / FSR Performance / FG - 7:32
4K Epic / FSR Quality / FG -...
Some of the plants get tiny bit less detailed.
When I put both on screen zoomed and switched quickly between them.
You sure Baldur? Sure you not imagining things. It's not the lightnign changing? π
One part was just plant moving in "wind", some others had actual tiny changes.
it may have actually just changed LOD bias, not actually lowered which LODs are used
But very tiny.
foliage looks the same to me, even zoomed in.
so it switches LODs at less distance
im adamant that they are the exact same settings just 2 second apart
anyways time to actually play the game lol
Nah, just lot of UE5 games use too high base detail levels for everything as "engine handles it", and then just use the base multiplier based settings, without any tuning.
Causing everything to basically always be at max, as the too high base * 0.2 is same as too high base * 1.0.
One is epic. One is low 
In the end, after all the steps.
nanite does not have any way to lower detail, so static meshes will not have any change
When the Epic instruction is to use the highest level LOD you would normally use in older engines.
And not the full resolution ones, before any lowering.
but if your whole game is static meshes
it aint
Then why does it all look the same on the screenshot?
skeletal meshes cant use nanite
read baldur's thing
NPCs (skeletal meshes) and foliage do not use nanite in stalker (tho UE5.3 iirc added nanite for foliage)
anything breakable wont either
and the guns are also skeletal meshes
But there are multiple previous UE5 games that basically used 10-100x too high detail meshes.
The highest detail the artists had created.
And expected the engine "just handle the LODs".
When they should have created the highest detail LODs from those full detail models for the actual release version.
I know another game that has terrible LOD with mountains 
(insert cough) ark ascended
And then the engine would have handled the lower LODs from those, for the objects it could do that.
They look like triangles in the distance. Or pyramids. But when you get closer, it turns into mountains.
That is DX11 bug as Nanite doesn't work on DX11 in UE5.
"bug"
None of the new graphical features in UE5 over UE4 work on DX11.
All need DX12 or Vulkan.
That was DX12.
Normal on DX11, not normal on DX12.
realy?
So I would really expect you to have DX11 in use for some reason.
i thought only dx 11 does that
Or very weirdly broken game install.
Couple of other cases where it was in the end DX11, when the person thought they had been running DX12.
I play on DX12.
Only game I can't is MHW, it crashes with cutscenes on DX12 and switches to DX11.
MHW?
Tried it twice, always crashes with cutscenes and reverts to DX11. So I think that game isn't optimized for DX12. And also the reason why it automatically switches after a crash.
Make #1038092680493801533 thread?
Include screenshot of the ingame video options with that renderer showing?
Monster Hunter World
I just assumed that was LOD issue π
Or that they put in triangles as texture for far away mountains
optimizationπ
I would expect that there is still single extra level of far terrain LODs in the game, but they are buggy and contain those or something similar.
As when Nanite is in use, there aren't fixed terrain LODs anymore.
But if the game actually is on DX12, you would be the only known case to me with that issue even then.
But I cannot try to debug it here.
I am not sure. It's a screenshot from around release.
I just logged into the game quickly, DX12, and mountains seemed fine on first glance. Didn't test extensive.
Don't bother. It's not an issue. I am playing other games atm anyway π
i love how stalker 2s shader compile bar starts off full but then after a second it goes back down to 5%
such a troll....
the new monster hunter rise will have it too π¦
Or that is again your drive fucking stuff up.
As the shader cache file is by default saved under C:\Users\
So it sees that there is full cache, tries to read it, fails, and start to compile again.
but like, it only needs to compile sharders once no?
devs gotta first make it so you dont HAVE to use FG
or do games have to recompile everytime on boot?
Normally not, but as he has cursed model SSD as OS drive.
lightning, you really seem into tech, why not buy better SSD? π
black friday is around the corner.
but im pretty sure the shader compilation is like this for everyone
capitalism at its peek.
People convinced me to get Lexar and I don't regret it.
People also convinced me to get AMD. 
no DRAM cache and a tiny SLC cache
Shouldn't affect reading, only writing, but there is something fucky going on, and it being QLC makes reads harder to do, but even when it is cursed.
There is one thing on my mind, something I cannot explain. So after I replaced my CPU. (I think almost 6-7 months ago), my PC crashed only 1 time (near the beginning of the switch). Same exact crash BUT ever since then, I hope I don't curse myself now, it's been stable.
Critical 14/08/2024 21:51:23 Kernel-Power 41 (63)
So I am super confused. Because it seems switching the CPU fixed the problem.
this has happened before just not with upscaling.. and not at high
And maybe that one time was a fluke? But ... what can caused it? And how comes it didn't happen before.
Or maybe... the problem isn't fixed. And in a year or two my system will start crashing regularly.
time to restart the game
holy redit thread... https://www.reddit.com/r/stalker/comments/1gvt4j2/man_these_shaders_sure_are_compiling/
I am going to reset my eventviewer for now.
ππππΆπππβΉοΈπ£π’ππβ οΈ
Was listening to HUB, apparently Stalker 2 really pushes distant LOD to minimize pop in and visible loading. I'd love that to be more of a trend.
And that might be part of the performance problem?
Oh, almost certainly
That kind of thing is always balance between noticeability and performance.
But we can't know what it'll take to get it done until we're given the option to try.
If the 5090 and Zen 5 X3D of any sort can't do it well, that's an ominous sign.
Normal UE console commands would probably help.
Problem would be to find the right commands.
i'd rather have a bit of poppin than have
FSR AA
first settlement
8 avg
smooth 60 at first settlement
If UE tools can open the data files, then that gets lot easier, by getting access to the INI files inside that contain all the settings the game is setting for graphical details.
looks great
i can get there too with:
HIGH preset 1080p FSR Q FG:ON
wilderness
93 avg
80 1%
57 0.1%
first settlement
72 avg
5 1%
4 0.1%
if i had a little more VRAM i wouldnt be stut-tering
Omg, eventviewer is a mess.
it IS
I cleared all windows logs. But apparantly, there is a shit ton under applicationsd and service logs.
And I have to click on them manually, and then clear log.
And I have to do that for the whole tree to the right.
Clicking on them one by one and then clear log.
Or just allow Windows to keep recycling the logs as time goes by?
Or execture a powershell / command prompt script that I don't want to investigate now.
I just assumed there was a button: purge logs.
It only keeps certain amount in physical disk space.
But I guess I am asking too much from Microsoft.
I just wanted a clean slate. Because half of the logging stuff was when my system was unstable.
Baldur, tell me Linux does it this way π
AFAIK SystemD JournalD doesn't have any easy way to clear stuff up.
You just again just set time and file size limits on how long to keep stuff.
Linux it's just purge this folder and you are good.
Not really.
Lot of software don't like that, including that JournalD.
Some need to be restarted to resume logging, some cannot handle it at all.
Wait. Since when is what software more important than what user wants? And why is software brittle enough to break when logging is removed.
waow the weather fx look nice
If software breaks because logging folder dissapears, that worries me.
the game does look great
the loading screen doesnt...
Yikes. π₯Ά
There is reason why logrotate has per program configs how to do log rotation so it doesn't break the specific software.
it is a painting
Some need that restart, some need you to do weird stuff etc.
reminds me of old stalker games
I understand that.
But I don't understand why software is dependent on logs.
the painting looks great
this part doesnt
I did say they would not LOG more until stuff was done.
Its more of an opinion that I am stating; removing a log folder, shouldn't break the software.
ah i see
But there are some that do crash if they cannot log.
i feel for those on 8 core or less rn
But very small minority.
You can write some code that if an exception is thrown you do x and y thigns to see what's going on and then write code to fix it.
For example if there is an exception that folder "logs" cannot be found, recreate it.
If there is an exception a log file can't be found, create a new one.
i have 12 threads and it *used to do this in 2 minutes
And if restarting software fixes it, it mostly because on startup, program checks for folders and recreates them as needed.
But doesn't do these checks during runtime.
Sorry missed that. Fair point.
and the ssd is also not maxing out
Yeah, but the point was, even on Linux, the stuff isn't designed to work with clearing of logs willy nilly.
And even JournalD doesn't give any command line option to do it on command.

hmmm maybe i should delete the shader cache next time
NEVERMIND I LIED
Ok, journalctl does have -vacuum-size -time -files options, that allow to truncate the journal, with some limits.
Since SystemD 218.
no matter the jank stalker has, at least it works 100% on linux
I have hopes for Stalker 2 still. Developers just need some time. π
From what I've seen, it has everything it needs to become a good game.
But none of those work on current journal file.
Only older ones.
But then there is second option to archive all the current data first and start new file.
now where are those shader files...................
It seems WAY better than Cyberpunk at release for example.
thats a VERY low bar
I mean, it was a tripple A game with shitton of money and development power behind it
the bar has been low for a long time now for games.
You know atleast give us an option to clear all logs on restart.
So that way all software gets the chance to create new logs on startup etc.
And nothing should break during runtime.
I dunno, my guess is programs break when they can't log is because they're coded to write their output to a file. if it can't do that then depenidng on out it fails it causes an exception and the cpu has to handle that, normally by breaking the program
I know.
But in that exception, you can catch and create new folder.
Which means, they clearly not doing that in a catch.
The exception just escalates and depending on what code they wrote, program might stop working.
assuming it's a file not found error.. I think stdrr will automatically create a file and append it if the file doesn't exist.
But that's not clean code practices.
In many setups, the programs will not have rights to create a file in the log directory.
^ that too.. esp if the log is in /var/logs/programlog
And need outside help in that, as security measure.
linux is funny with perms.
... well they have to be created somehow.
Which the log rotation program does for example.
and the fact that some programs can handle it and some not, smells like fishy code.
Yeah, but with logs, the base idea is to keep certain settable amount of time of them.
With possibly also file size limit as sanity check.
at least if the program crashes, you have an error in systemd.
But there are situations I imagine where a user wants a clean slate.
spring cleaning and stuff π
if it doesn't, and continues to run without logging, that could cause issues later on if something goes wonky and you can't go back and check.
Pretty rare?
You can just filter the logs you view to be from specific point?
logrotate will purge old logs
changing hardware and you don't want clean install of windows.
I usually just add -b (since last boot) when checking JournalD for example.
Is that a linux feature?
yeah.. we were talking linux so.. xD
No, we are talking both π
That is the program that handles changing log files for most of the old text file type logs, with set schedule and keeping X amount of them.
But that doesn't matter. I just disliked the fact I cannot purge logs on Windows 11 eventviewer and said: I think Linux does this better.
pretty much everything I do that's a service, like servers... file sharing etc... I use linux systems for.. my windows pc is just a pretty ssh terminal lol
channel got busy
Not really.
Both keep X amount of time of them.
And with both there are stuff you can do to purge the logs badly.
I am getting at the point to ditch Windows. Might need a few more years.
Thinking about it more and more.
i started the game with the EXE thats located in the save files and i SWEAR it took tenth the time
especially with the recent "recall" shenanigans, miucrosoft is still pushing out that feature.
But even those Windows Event log entries are just files.
That could be nuked.
My point was.
And if you looked at my screenshot.
I can't help you with windows stuff. I know how to do things with the gui on it.. but as far as the backend I know nothing of... with my linux systems, aside from the steam deck that's arch and I know fuck all about... I do everything in terminal with no gui, except for pve and a few other services that have a web interface
I have to click on the folders one by one, for the WHOLE tree, and purge them.
I cannot do it from parent.
we pretty much just have to hope for steam to continue to kill it with their deck and wait for a desktop OS lol
What you see in the tree, is A-B-C... it has a whole bunch more.
Yes, they cannot be nuked easily in GUI.
And neither can Linux log files traditionally.
JournalCtl finaly has a way to do it, even when it needs mixing several options to work properly.
Point was that you talked about deleting the log files as files in Linux.
You could do the same in Windows.
As each of those is just a file somewhere under C:\Windows\ or C:\ProgramData\Microsoft\
Yeah.
for /F "tokens=*" %1 in ('wevtutil.exe el') DO wevtutil.exe cl "%1"
That tells wevtutil to clear each tree, not file deletion.
this was fun to deal with... audio is a mounted smb share
the game looks VERY weird with AA tuned off
Gets list of all the options, then clears each one.
Almost like the painting. Keep tweaking. You getting there π
That should also deleted the files no?
Empties them, doesn't delete.
Oh empties them is fine for me π
Basically automated way to go through all the entries and clicking the clear.
Yup.
That's what I want. But still ... I expected better of Microsoft.
Maybe I am just too demanding. But these things seem so trivial to me I don't understand why nobody realized it at Microsoft that this is bad UX.
you should've known better
I have this very old PC I got from my grandfather once for a birthday. I was thinking, if I can get it to work, back-up all the files on my NAS, nuke the OS and put Linux on it. Should be a fun project.
but I haven't turned the thing on for 20 years π
Seems what a lot of people do with old hardware, you put linux on it
then you start buying hardware to install linux on it.. it's a diesease
MS is the disease. Searching for the cure gets addictive because MS keeps getting worse and worse.
Oh, Linux gaming definitely isn't there. There are people IRL who want it more than I do, when they start dumping Windows, that's the sign I'm looking for.
unrelated but Why does TSR keep halucinating??????????????
sharpening the sky.
i actually think that TSR looks the best in this still frame
probably because i LOVE high sharpness
but the sky is weird, the grain
well i dont think people play games to look at a still frame
so lets look at them in motion
Yeah I agree with that. But you cannot deny it's an anomaly.
im not denying it
but herg said it too, at what point are we pixel staring
i think i named the wrong one FSR and
how things perform we can measure. how we perceive graphics as "good" is subjective and is harder to measure.
true
i only care if i can run the game
(im an igpu user)
intel iGPU
fair, but we were discussing different upscalling methods and how they look π
they all sound like the result of gpt
in the end, we all just want to game, I played games on a macbook pro for a while.
not the best gaming experience either lol.
until I got enough money to buy a PC.
i used to play on a VEGA 3
with a 3200u
that had 8gb of 2400MT RAM
upgraded my pc to 32 gigs, running modded games like ksp or cities skylines on 12 was hell
"Worst" graphics I can remember is:
If you start with that, everything is a upgrade now π
likely
the problem with 12 gigs of ram is that its not a "normal" ammount of ram
so games like cities skylines would suck on it
Nah, the problem is that some of it isn't dual-channel.
And so you have two different speeds of RAM at same time.
8GB of full-speed, and then 4GB of half-speed.
4GB of both sticks at full speed, then the extra without counterpart as half-speed.
do 6GB sticks even excist?
back to upscaling,i think TSR and FSR try to keep detail while XESS just blurs everything
DDR5 24Gbit dies were the first not direct power of two.
So that is 8+4GB.
Yeah, I got that. Just wondered if 6GB stick excisted π
yea
but then TSR and FSR do this weird ghostines with these trees
i took the ram from a broken mini pc (ssd transplant gone wrong)
Aah!
holy moly.
dear god TAA
FSR
fsr leaves the same thing but bigger and more not togheter
here?
i cant realy see the ghosting on FSR here
its just a little pixelated
(keep in mind TAA is at 100% while everything else is at 66%)
@safe trench Bro WHAT?
i just accidentaly clicked on your name
erm
Never click on a profile lol
FIRST ADEMENT!
adement
adement? you mean amendment?
i forgot that i was listening to music LOL
i intetionally miss spell words when its jokingly
Bro is listening to youtuber intros in his spare time
real
Streamable doesn't compress as heavily
i going to mute the audio this time:)
spying π«£
OR i wont because if i do that it looks like youtube compression all over again
i actually can
best song ever has to be thick of it
||for the safety of me and my family this is a joke||
says literally everyone about their own music π
its my Playlist not my music
nah my musci is hot garb 50% of the time
you probably listen to Gzuz
Listen to peak playlist by A. Buster_123 #np on #SoundCloud
screamo rap, tearout, metal
i mean its defedently all over the place
nahnah trust me this is 1005% peak
it's German gangster rap
yeah i dont fuck with german rap
WTF IS THIS??????
peak music
TRUE
certified peak music clasic
https://www.youtube.com/watch?v=xYBoaJIwQmM
this is the shit I listen to
hello and thank you
been super busy for a while but i got tons of stuff in the vault just waiting to be released. Hope you can enjoy this lil lyric video in the mean time.
lyrics
Straight line acting
Vector tactics
Amnesia
Blackened retrospective
An eye for an eye
The debt corrected
Your life for a lie like contraceptive
Go separate me o...
it is kinda ur right
yeah, track is pretty hype
nowhere near mine frfrfr
I don't trust you. 
my playlist feels so normal and boring compared to you guys lol
he also does his own album art in blender which has been my desktop wallpaper for a while now
Spice it up π
add the 2nd one in my playist FRFR
sick.
don't know any song I think ...
not even september?
Kendrick Lamar did drop a new album yesterday.
I am listening to that since it dropped.
(back to upscaling) i finally get to do justice for/to image stability https://streamable.com/v7o3vv https://streamable.com/guwk9a https://streamable.com/lpdj21
Oh, now that I hear it. Yeah!
I know the original.
Damn, 774 mln views on YT over 11 years for the original song.
teardown theme is peak for me tho
yea its a clasic
I like it π
been listening to a lot of normal techno too
but nothing can keep up with good old tearout dubstep
with the ear damage its meant to do
or old skool dub: https://www.youtube.com/watch?v=Ye5QXjqTDqE
BACKBONE with Stormzy live from Ushuaia π₯΅
Listen here: https://ChaseStatus.lnk.to/BACKBONEID
Join the Massive & Crew email updates to never miss out π©
https://bio.to/ChaseStatusID
Stream the music of Chase & Status here: https://chasestatus.lnk.to/musicID
Explore the Chase & Status Store: https://chasestatus.lnk.to/officialstoreID
Lyrics:
Che...
why do i feel like game soundtracks sound better and have so much more effort than regular music?
chase and status are still going strong after all these years. Song is 3 months old.
this feels straight out of beat saber lmao
Stormzy is just a beast on the mic and that beat is sick.
I though the music genre would be dead by now. But nop.
very uk drum&bassy
will someone shoot me after this?
ah fuck you are right! It's drum n bass not dubstep.
my bad π
i will agree that sometimes its true i really liked "Why cant love be true" from PCBS2
Was listening to the Witcher 3 soundtrack again a few days ago, I agree
i really like the teardown maintheme tho
if you want somewhat modern dubstep that's still leaning towards the (good era) 2012-2015 dubstep you gotta go black tiger sex machine or crankdat
those names ...
so happy YT is not flagging it for anything
Yeah crankdat nails it
watching a live set atm "Lost lands 2024"
lmao
corpo needs u to compare the difference with this picture and this picture
it gives the same energy ngl
i just learned that yoshi means adopted kid π
god damn since when does the 1gb raspberry pi 4 cost 45 euros
Just testing ping/latency basically.
First without load, then with max download, then with max upload.
And comparing what the latencies were.
What asre you on about?
How that buffer bloat test works?
Very simple thing.
As long as that "speed test" part gets to your connections max speed.
If it doesn't, then it will show whatever.
Both images have the bottom important "Upload active" graphs cut off, so no way to see where the difference might have been between the runs.
A test should give the same results to be reliable imo.
Also that upload is lower in the second one, so there could easily be current limiting factor between you and the test server, that doesn't allow the upload to actually get into full speed status.
Nothing over Internet can be reliable.
There is always the caveat that the interconnects between your connection and their connection can cause results to vary.
And that site doesn't have anywhere near as many test servers as speedtest.net for example has.
Then it's useless to me.
So you need to run both that, and then speedtest, to see if the speeds are lower on the bufferbloat one.
If they are, then the buffer bloat results aren't valid.
I am not going to buy a new router, based on a F-C grade, from a tool, that is flaky.
You can do the same thing manually.
Set up ping to your routers default gateway, then load the connection to max in one direction, and see how the result changes.
So I am between A and F. 
The IP your routers WAN page says is the default gateway for it.
Nah, not worth the trouble. Was thinking what I need to look out for on Black Friday. But this router is going to be a nightmare. Barely support from ISP. And I am not even sure my router is that bad.
I just keep rolling the dice untill I get an A and leave it at that π
Back to F. Fuck this.
But very stable from test to test and day to day for me:
Because of either ISP or router issues.
Who knows π€·ββοΈ But if it gave consistent a F that would be less worrysome then C-D-F.
your internet is never going to be reliable like that
No, that kind of jumping is expected in some situations.
And as that error bar is huge in your test...
What a crap statement, baldur just proofed it can be stable.
if you're not getting consistent results, then your internet isn't consistent
Basically the quality of your connection is all over the place even during the test.
So we hav Baldur, who has consistent an A.
i'm having weird behavior with files. and i suspect there's some limitation to max number of files in folder
because it's stuck at 260908, which is kinda close to 256k
but i'm not getting any errors related to write, so... idk
We have me, with fluctuating grades on bufferbloat within minutes.
For the download, upload is much more consistent, but still bad compared to working connection.
And as you compare to that unloaded error bar in your own test.
Basically either your router is shit, your ISPs connection to the selected test server is crap, or both.
And running that manual ping to gateway during the test would let you see which.
As that will just test the latency to nearest ISP router.
So only your internal LAN, your router, and your link to that ISP router.
Mhhh... I have a few ideas.
You get that ip, then just run "ping -t ip" and then run that test or some other speedtest.
Which OS and file system?
Ubuntu 24, ext4
Then a no to max file number but a yes to Software handling that amount of files
With 256K files in one directory you should probably start thinking about organising them in sub dirs
maybe, maybe...
What kind of internet do you have? Cable?
I already figured it out π
Your isp is shit?
How can that be?
aren't they all?
Or It's your router? Lol
Oh right, didn't post screenshot from my macbook on wifi.
Might just be that the MacOS is doing self-limiting when it notices the response time goes way out.
No, PLC:
Oh... yeah those are bad
Macbook got 88.9 Mbps on wifi 5, that's 3-4x what I get on my PC with cable.
And got a B score for bufferbloat.
Yeah, power line networking is usually just bad.
Only really usable when nothing else can be used.
Normally next summer, I am in my new house without PLC.
That's my case.
But you see? I am not crazy in taking these tests with a grain of salt.
Because I don't even think it's more router.
How's this [:
Which is the point to then confirm with manual pinging to various points on the way.
Alright
With that as the reason, pinging your own routers internal IP while running bandwidth test would have revealed that the problem is inside your own LAN.
Or using my macbook to connect to router and skipping cable + PLC π
Ping -t ip... I'll try that later
Goddammit, if WIFI get's better result than cable, you know you done and fd up
but functionally that won't do much for me. except if somewhere it acutally removes limit
Our pull a ethernet:p
Not really, Power line networking is dead tech exactly because of Wifi getting so much better.
Eh....what?
People all go cable because of stability and overall better experience.
Hasn't been updated in long time.
Still being sold, but with old tech.
It is..
Who in here has cable? β
Moca adapters are the new thing
Pretty sure 90% in here uses cable
I did say "Power line networking", using houses AC wiring for networking.
Not that cabled networks are dead.
That MOCA (using houses TV antenna cables) is the newer replacement.
And like you saw, even WiFi is usually better today than PLN you have.
But funny how a buffer bloat test picks it up, nothing to do with buffer rofl.
Yes it has, buffers in the PLN adapters.
You could try eeros or whatever to extend your coverage. Mesh with wireless backhaul is the new thing as well..
Weird thing is, I tested this PLC setup a year ago.
And it was the other way around, WIFI was worse back then.
Just shows you how unreliable the electric grid is.
Yup, shared with everyone behind the same transformer.
Shit changes. Pln can work perfect until you plug in a new tv.. or phone charger
Exalctly, so tired of this shit ... I am always fixing problems half assed because of irl reasons.
And someone else behind the same transformer plugging in something that causes "interference" in that phase will fuck up it.
Noisy electronics etc.
Well, I get that you're moving. There's always running a cat5 cable out the window and into the window of the room where you want it lol
MOCA in turn is basically just shorter range cable modem tech.
So the cabling is much higher quality than AC wiring.
But moca as a wired backhaul and a mesh system would probably give very good results
How do you know my PLC adapter has buffer? 
Otherwise it couldn't cause buffer bloat etc.
And would just cause lot of packet loss instead as the quality of the connection between the adapters varied.
How do you know it's not packet loss?
can't latency be caused by packet loss?
No, lost packet isn't counted for latency.
Oh right, because of how we test without load first.
Only the ones that get through.
Ok, good to know. Well. I continue my black friday hunt lol.
Next: robot cleaner π
btw for anyone in benelux: tweakers.net/deals/black-friday/echte-techdeals
shows real price drops etc for hardware
My roomba is a lifesaver since I hate constantly vacuuming dust that's collecting under my bed
The only problem I have with roomba is that their spare parts tend to be expensive
well, while you're router shopping, have a look at ubquiti's offerings if you're looking for higher-end stuff.. once you get out of the $100-200 price range.. consumer hardware offerings from netgear/tplink whoever really arn't worth it
any experts in Python's generators?
Does it require phone app?
Roomba can work without the app afaik (although app offers some benefits)
Oh, and correction to previous talk about TP-Link consumer routers cloud accounts etc.
The account isn't mandatory, but even if you don't create one, it will connect to the "cloud" anyways.
For the tplink deco system... if you don't use their app you can't pair the decos, set your ssid, or configure them for ap mode... the web interface has zero config options..
You can get a 3 toiletpaper deco kit for the same price as a single U6+ and your still need the network controller, although that could be self hosted..
The u6+'s will do something like 900mbps over wifi with a client that can handle it... I could never get the decos to go over ~500mbps
Usbc
lat's time i checked latest iPhone was called "X"
Good point.
My galaxy S9 unfortunately started having issues with a second sim slot and my sister gave me a spare iPhone 12 Pro Max. I don't really like iOS tbh....
Or maybe I need to use it for a few more months in order to notice the advantages compared to Android like my coworker said although I doubt it will change my opinion
none
They both call, text, and browse the internet... android keys you install Firefox and extensions, ios blocks them.. android let's you bluetooth files, ios only keys your do it with other apples.. ios has a better messenger for texting by miles. Things I've noticed by using both
It's more like pick your poison. Both OS have upsides and downsides.
I am used to iPhone, Android had in the beginning a really ugly UI.
Agreed. I hated Android in the past but it is much better now
I remember the old days when you didn't have enough storage space and you had to move apps to sd card
For sure, but it's choosing between Google and Apple in the end.
At least with Android, I have a lot of brands/price ranges to pick and normal prices for accessories like adapters and cables. I still use my S9 wireless charger for iPhone because an adapter for iPhone is expensive imo
EU forced apple to go USB C, so charging is a non issue with apple now.
For iphone, you also have a wide rrange of prices available.
It just depends, do you want the newest of the newest iphone, or can you wait a few years?
Cheapest I can find:
iphone 6 and 7 (refurbished).
Highest price:
When I was a kid, I always wanted the "best smartphone" π . Now that I am older, galaxy A series is enough for me lol
Just need a phone for basic stuff
of course samsung is half of them
Point being, a lot of time I would agree with Apple being overpriced bullshit. But for iPhones, they seems to have more competitive pricing than greedy prices like they do with other products.
I would avoid foldable phones like plague, my coworker has galaxy Z flip (I think) and there are already screen problems present
But when you look into the hardware, iPhone are probably still overpriced. So hardware/price.
its still not the like 399 of like 6 years ago last time i stepped in a place that actually sold phones
The screenshots I posted are from a website that monitors multiple stores and aggregates them. Forgot the english word for it.
Price watching website?
scalpers paradise
Good thing I work for an ISP so I can get phones through company for cheap π
π§
Nah, happy for you.
Freudenfreude π
I noticed most of my coworkers have recent galaxy S models
I am so proud of my iphone. It's a relic. But it's holding on π
I think it's 10 years old by now.
My galaxy S9 is still technically working but there is something wrong with the second sim slot. Incoming phone calls get dropped after few seconds and I made sure it's not related to settings
Also did factory reset but It didn't help
Btw, Apple gives software support to around 10 years on average (since model introduction to the market)?
It looks like this:
So to unlock my phone, I have to double tap it. But you can place the button wherever you want.
anyway, I am out for a bit. Will be back later π
Keep in mind theyβre all mostly folding too
you got a flip a fold a fold a fold a pile of trash and then apple so yea i guess fair enough
I wonder, what would be the good amount?
I can split by year, and even by month. but then it will require some glob searches to be recursive...
Depends on what's inside the files
No idea then
True but still 3k
yeah, i need like... 1250k of them to save π
and for some reason i noticed that on laptop they simply wont do more than tha weird number, even though there should have been like 100k more by now...
and on PC i'm doing funny thing - i saved those files and basically re-parsing them to keep everything clean and consistent. and that takes some time... I could have probably just split the workload into processes, but that eneds a bit of configuration. so 300 files / minutes is decent too
I wonder if Apple wil ever make foldables.
but i'm starting to fear SSD durability for such load
it's probably comparable to downloading and unpacking the game
maybe but for $12K at this rate
sometimes i think that Kaggle is a joke with completely unreliable UI (plots not showing, buttons disappearing, disconnects from kernel, messy files tree, cunbersome navigation within notebook, broken navigation of saved notebook...)
Apparantly there are rumors on a foldable iphone, arriving in 2026.
are we going to get one before mission to Mars starts?
Who knows, they might cancel it π€·ββοΈ
ok so this is totally weird - i changed structure to split by month
no changes to contents of files. but more deterministic search for them - no glob, straight check for filename
and speed from 300 gone past 2000 / m
crazy
i finally remembered that my expectation of 750k files was from a larger search range... so perhaps it's ok
they decrease in value fast also
I wonder what percentage of foldable phones survive warranty period without any screen or hinge issues
Well actually over to 4200 /minute
So now it really is fast to reitertae parsing of insividual events
Logging to FS becomes problemat this pace.
Scrapy generates a lot.
128 rotating files x2MB is like nothing for it
but ipads do
@soft bloom have you considered using a database instead?
Because this sounds a lot like you're moving into that territory
Yeah, but what kind?
For this project i am trying to keep tech simply, do that it is as easy to setup as it could be. At least that should be an option.
The most db used so far is sqlite and i doubt it can work with dozens of gigabytes
SQLite isn't really a database, more of object store.
So basically whole file as single entry in this case.
For example.
Ah, sorry, mixed with NoSQL.
Databases should have no issues with that normally
I mean, individual size of compressed QuakeML is couple kilobytes
i know that SQL can do large relational queries,
but this is just fetching large volume of bytes by single key
so clearly more of a job for object storage, but still idk how well it would perform in comparison
And system will loose simplicity of being able to just open a folder and read the file like normal person
So... I am fine with idea of having index in SQL to speedup search (like i did with http cache),
but even an option to have something like Mongo fully replace the storage raises doubts. I would have to benchmark
i mentioned both MongoDB and object storage because both could solve this, but am aware they are different (though this feels more like object storage task)
also, considering ease of finding the right file, i worry that any of these will make no difference due to operational overhead
although... i have no idea if there's one with FS
Funny, a year ago, with any other project, i would use some fancy tech head on - no questions asked...
ok indeed, that was even more that i should have had. so all good - just by chance it was close to 256k...
perhaps the weirdest search query in a while π
i had a bit of a pickle because I am saving parsed data to JSONlines, coupled by year
So to save a bit of space i then join and convert it to CSV
but this created problem i didn't expect because that turned internal lists and dicts (which there's a couple of them) into just stringified python objects
so when reading resulting CSV back to pandas it was just a string
and this is where i think LLMs shine, because I asked this Claude 3 and it revealed to me existance of ast module
which i haven't heard of before, and it has fitting literal_eval method that won't do many nasty things
...
as i was writing this, i tested what would i get just by inserting the same prompt into DuckDuckGo - it return answer from StackOverflow, lol
so, i guess instead of saying this wouldn't be as easy i should say we should try natural langauge queries in traditional search engines
exploding pandas are truly dangerous:
not sure if it was mentioned already, but one of the reasons to split a large number of files into subdirectories is also the limitations of the underlying file system. For example ext3/4 becomes super slow north of a few tens of thousands of files, where it can take multiple seconds to just create a new file, because assumingly hash-collisions in the dictionary or the dictionary of the node becoming too large etc. different file systems, different limits/problems, though.
having to share presentations on discord because teams sucks is peak school π€£
you mean cheap school? 
i guess yea im using my laptop as a hotspot for my classmates because none of there wifi cards can reach and then were having to present to eachother on discord because it uses less data
we can all agree on "Peach school"
Is windows supposed to become unresponsive mess when Steam moves gamefiles to system drive?
The only indictaion of load is busy system drive
Hitting system drive really hard always slows things down
Depends on the drive. So yes, it can happen
Yeah, Teams sucks
@twin dew I've been struggling setting up a static IP, can't understand shit.
Ordered a static IP from my ISP for my router which has a sim card in it. (Huawei 818-263)
Also changed the APN as instructed by the ISP.
But now I can't figure out what to do next so my laptop could have a static IP.
I need it for good remote desktop software(moonlight or parsec) to work so I could stream my games while in service.
And as we all know, you're smart so I could use your help.
Because that stuff doesn't make your computer have static IP?
But the router.
You can then set up static DHCP lease on the router for the computer, or static IP on the computer inside LAN.
And do port forwarding on the router, in theory.
And as Huawei doesn't have anything more than quick start guide for that router, no idea what it allows to do inside the management to accomplish those steps.
Yes.
That sets the computer inside LAN with that specific MAC address to have always the same IP.
And then you set that port forwarding so that any traffic the router gets to specific ports, is forwarded to that IP.
static IP on the computer inside LAN
This? Couldn't get it working.
No.
Aight
That is for the WAN
So now port forwarding...
This? support.parsec.app/hc/en-us/articles/360045297592-How-to-port-forward-Parsec
This makes that whenever DHCP client connects to the router to get internal LAN IP, if it has that MAC, it gets that same one every time.
Basic idea, yes.
Aight, will mess with it. Not sure if I got the time/motivation to do it today.
Thank you!
My devices have all static IP based on MAC adress.
My router assigns these local IP's based on MAC if not mistaken.
IPv6 SLAAC? π
If IPv4 and you did not set then manually its probably just long DHCP lease times.
Or they just extend the lease often enough to keep their IP
That and lazy-reallocation
Lazy fucking DHCp server smh
I set them manually.
nerver experienced any changes/slowdowns while doing so, but then again im nerver generally a good point of reference
hdd just killed itself was luckily able to return it, had 15mins left on the return window
there is something sketchy about 6.12.1 and virtualization
Argh, still didn't work
what's the issue?
support.parsec.app/hc/en-us/articles/115002601011-Error-Codes-6023-and-6024-Unable-To-Negotiate-A-Successful-Connection
hmm can you ping the machine?
especially the port. You can use the "telnet (ip) (port)" without the brackets
Asked my mom to restart the PC, closed rustdesk with task manager to see if it might've caused a conflict, now can't remote in again.
I need to go soon, will fuck around with this later today.
Will have the time to try the ping thing today, got to check the host IP.
Can also recheck if firewall could be the issue but I doubt it. Literally every IT problem guide mentions firewall while it has never been the issue.
are you using a VPN or are you going in raw?