#off-topic-tech
1 messages ยท Page 103 of 1
And maybe use PCIe as the data connection ๐
i love the custom VRM fan
it's mainly for RAM, coz overclocked
ah
SAS is not PCIe. Buuuut NVMe 2.0 has support for HDDs, and NVMe is PCIe ^^
everything is color matched and build for looks as you can clearly see
Usb? Man that's gotta be bad
(if you close your eyes)
i can't believe nobody noticed my noctua grey fans
I know SAS isn't pcie. But just use something faster than sata 3 for data on 60tb drives
ah, misunderstanding then. but yeah. RAID rebuild with >50TB drives on SATA? sounds like hell
to get into working temperature faster
their v1 of it was clearly not designed with AM4 in mind, their V2 they turned the heat pipes 90 degrees so that it'd be front to back airflow
and then i replaced it with the assassin 120SE
last time I installed an air cooler (Dark Rock Pro Vsomthing) I had bleeding hands
original: 1700X/1660
transition: 5600X/6600XT
current: 5600X/6800XT
AIOs are just more convenient if you disassemble your CPU often
Oh yeah even the 22tb drives it's still bad
Missed out on the 6900xt
i bled when i had to take my gpu out to install my 2nd nvme drive. that stupid PCIe lock tab was about 3mm below the cooling fins
oh and i think i'm the only person i've seen with my little red AIC
yeah what's that little red card?
7x USB 3.0, 2 of them internal
some DMA device to cheat in Escape from Tarkov I bet ๐ค
USB hub was not an option?
didn't want yet another thing with external power needed
plus it's a great spot for that stupid gigantic xbox dongle
i think it's supposed to be used for mining, looks kinda wild
perfect for routing my VR stuff
ah, that's where the bend over came from
lol nah
that was about when your PSU has those 2 sata power connectors really close together, and you have to bend the wiring between them to be able to get each one to connect to drives right on top of each other
You bought one of those?
Is it any better than just normal Bluetooth?
eh, didn't buy it, was a freebie from buying the VR stuff from a friend ๐คฃ he was upgrading
Free is good. I'll take free hardware anytime
this stupid shit when it comes to wiring sata power. i changed it when i replaced a drive because of how nervous i was having that bend in the cabling
are there 3 more fans in front of the rad?
yeah push pull self made
i know it's just the angle, but the front fans kinda look like 92mm
EK Nucleus 360 AIO + 3x Arctic P12 Max curve matched with FanControl in push/pull
RIP EK btw
i appreciate the "fuck it, leaving all the AIC tabs out"
"AIC tab"?
ah those little buggers just like to fall out ๐
idk how they haven't come up with a better way for those
just about everything else has been improved
I think I've put them in again
I like how I have like one of the biggest cases on the market and GPU + push pull radiator setup just barely fits
shoutout to nvidia where you actually get a lot of weight for your money
the lian li case?
or was it lancool
duno, saw a review recently from them, case looked nice and also performed well
yea the 207
Sponsor: Get 10% off Squarespace purchases (https://geni.us/BqEpf)
We're continuing to add case benchmarks to our review charts! The Lian Li Lancool 207 is up for review today, a new ATX case for under $100. It's compact, but still ATX. The Lancool 207 is fully focused on thermal performance at the price, taking a number of key steps to improve ...
I'll stick with my Fractal Torrent big tower, swapped out all the fans and put custom fan curves everywhere. Can't beat that case in terms of airflow
i wanna see someone come up with a dedicated back to front airflow case ๐คฃ
so you got yourself a heater in the winter months?
just to see what they could come up with
or if it could maybe solve some of the cooling issues
I was thinking of going full custom water loop for the next build... but if I switch to AMD because of reasons, that would be much work for no gain given that I'd get a X3D CPU
just doesn't scale with more cooling, unfortunately
yea, i'd only spend money on something like that if it was impossible to cool without it
i'm still shocked at how efficient AMD chips are
i'm kinda waiting for a triple tower air cooler ๐คฃ
everything seems more efficient than Intel at this point ๐
my M2 Mac Mini draws around 4W at idle and 7W watching a video. Meanwhile Intel Raptor Lake...
Only watching a video, Intel should be comparable on Kaby Lake. Have they really regressed?
probably
so much of intel has been "well, nobody cares about how much power it sucks"
The problem is that people tend to use PCs as more than TV, so the moment you're doing other stuff, it's way worse. And Intel hasn't figured out sleep.
i am glad that AMD figured out that they'd probably get more sales if they stuffed in an iGPU
that's the one thing i absolutely hate about my system
https://github.com/WerIstLuka/int
i think im finished with my program
it can accept input in every way i could think of
has all of the features i could think of
no bugs (to my knowledge)
it takes one or more integers in any base and can convert them to any other base
Lancool IS Lian Li
that's like saying "idk if it was Asus or Strix" or "I think it was Gigabyte, might've been Aorus tho"
yea had a brain fart
MLID claiming 377mm~ for Blackwell 203 die, basically the same as 379mm for Ada 103. If the 5080 achieves meaningful gains over the 4080, that's going to be impressive as hell for TSMC. If the 377mm thing is true.
I think this is one of the first die size leaks that have come out.
Now I want some case modder to make a PC inside a satisfactory computer
#satisfactory-memes message
๐

Love this
Those angles are not 90ยฐ...
They are at the final infinitely short section.
But the missing requirement is that the lines are straight.
Because that isn't in the normal "Square" requirements directly, but in previous shape family requirements.
Square > multiple steps > Polygon, and polygons can only have straight lines.
In geometry, a polygon is a plane figure made up of line segments connected to form a closed polygonal chain.
In geometry, a line segment is a part of a straight line that is bounded by two distinct end points, and contains every point on the line that is between its endpoints. It is a special case of an arc, with zero curvature.
This horror contains arcs that aren't zero curvature, so it isn't polygon, so it isn't square, as square is subtype of polygon.
But it is a shape with four equal length sides and four right angles.
๐ค
And proud of it! ๐ค
Only if you skip those inherited requirements from the shape family.
And just use the simplified "this specifies the shape from any other in the family".
Yeah, seems they have actually printed that simplified version...
Huehuehue
Actual definition differs.
Should contain that "Is polygon with..."
Or one of those midpoints between square and generic polygon.
Can't you say something dumb for once?
Yes?
I'm still going to be stupid and nit-pick, because I cannot help myself ๐ญ
And I didn't start that before kit asked about the 90 degree angles.
Because it is good take on the simplified requirements.
influence of people who don't care...
i recently noticed that on wikipedia page of my university there's not a single technical person: poets, journalists and politicians. WTF
I never asked, made a statement.
Sorry, wrong phrasing from me. Statement that was wrong.
The corner between arc and line can be 90 degrees, it is just infinitely short section of the arc.
should string alhpa-numberic check be a joint of .isnumeric() and .isalpha() or a subset?
(atm in python it's a joint, so string must have both numbers and letters... but i think it shouldb e renamed to .issigma() and .isalnum() return true for when it's numeric or alpha as well...)
yeah, i am just not hapy that isexepcted different behaviour by that lowercase obsolete name
Usually you have one chech that only accepts numbers, another that only accepts letters, and one that accepts both?
So in this case I would expect that that isnumeris is just numbers (no letters etc. allowed), isalpha is only letters (no numbers etc. allowd), isalnum would be both allowed (but not special characters etc).
wait, it's not true - the other way around... this is weird "morning" already
So the alnum would allow everything in either numeric or alpha.
And any combination of those characters.
i think there's something in my tea...
but where's the 0 equivalent of the letter?
?
looks like isalnum() returns True on pure alpha strings just fine? Maybe I'm misunderstanding here
It does return False when you start adding spaces though
Those just check if all characters in the string are in allowed list or not
like, 1 is natural. 0 isn't
so we often use 1 irl, we start counting with it
with letters, we count from A
with machines, we count from 0, but there's no 0 of letters world
you are correct, i just messed up before writing it without proper check
0-9 for numeric etc.
They don't care what of the allowed set is used.
Just that there are no characters from outside the set.
Depends.
Both 1-starting and 0-starting are a thing, depending on programming language, convention etc.
The actual binary numbers start from 0.
Numbers starting with 1 is normal decimal numbers convention.
Alphabet (in latin alphabet) starts from A.
Etc.
But what is the length of an int? :^)
Well, with most binary encodings, all 0-bits is 0.
I'm not trying to get a silly argument to the death going, why do you ask that?
So with programming languages etc. 0-index starting was logical originally.
As doing 1-index starting needs extra work in the language.
I don't mean "32-bit" or MAXINT btw, I mean "what should happen when you do int.length()?
source of my madness is that for second day straight i can't figure out why my new import script only puts 70k items into DB, while old one put 120k
i see that there are folders missing
But 0-starting has higher chance of people making mistakes and trying to use one slot more than there is.
"But I created 8 slot array! Why is Array[8] invalid!"
I mean, 0-based indexing is something you run into a few times when you start, and then "get" it, right?
Usually, but off-by-one errors with it will happen anyways when you forgot that you need to use array.length-1 etc.
Yeah, that's fair
Of course the programming languages could also give variable/function that does that substraction already with different name.
Where one would answer how many spots there are, and the other the index of the last one.
oh lmao. i was doping .islower() check that fails on numberic strings
Yeah, because that is just subset of the alphabet one, with just the lower case letters, but not the upper case ones.
Usually there is then some version where you can specify the allowed characters freely.
well, that would work.
honestly, not sure if i should hsve checked that in the first place. it's more of defence against my own shenanigans that involved creating files inside predefined fodler structure in previous version of that tool.
and i tend to write strict code...
could have done a check if string is hex, which actually would be more strict and in right places ๐ having lowercase is not a requirements - more of a trait in this case
But .islower in python doesn't seem to care about anything but the letters, everything else is ignored and doesn't affect the result:
The islower() method returns True if all the characters are in lower case, otherwise False.
Numbers, symbols and spaces are not checked, only alphabet characters.
"hello 123".islower() = True
yes. but i needed "00" to also pass the check. and .islower() doesn't like full numeric strings
it's probably because "".islower() si also False
Then you need to do multiple ones?
Or find that more generic one, the quick string functions table I found didn't have one.
Probably needs at least one lower case letter then.
That wasn't in that crap "doc" in found.
Didn't go looking for the proper Python documentation
atm i branch our if string is numeric.
but i should probably just int(hex_string, 16) under supress
oh, i again invented something that doesn't exist... try-except will do
Ok, so while timeit. times which is faster...
storing metadata in sqlite reduced filesystem footprint down to 7GB, down from 100GB
everything is neatly stored in 300MB sqlite file, instead of couple files per cache item
so i may move it back on nvme ssd
there was a question of at want size i should compress, (and also what level of compression to use...), but i didn't benchmark anything and just eyeballed at 16KB
plan:
iterate over thousands of key params for URL generation, convert them to URLs, check if each and every one of them has been cached
reality: you have 30k such params, and querying DB takes forever
The search column isn't indexed?
the key problem i am trying to solve is that i have to make smaller requests to an API, that is unreliable (in a sense that it amy take multiple retries to finally get response, if it even can be retrieved), and where i can split requests by params.
So there's a question of when to branch out. but also, if at some point i have parts of "initial" requests saved/cached - should I pursue trying to get it in full or just continue with smaller parts?
F
i actually forgot
I indexed a different one, but this code rn uses different column
weird, PyCharm's databse tooltip does not support setting column indexing? or is just for SQLite?..
declarative models in Python spoiled me...
I forgot that indexes are a different kind of thing from column properties ๐
i did CREATE INDEX idx_users_name ON users (name);
and noticed small blue thingy near the column in PyCharm's UI, and then noticed separate list of indexes ๐
considering that I had 30k params to go through, while having only 1/10 of date range, i need some better solution
perhaps its time to reverse the process, and dump those URLs into memory (at least it already exists and is tested)
or a separate table that matches cached items with those params
That is the point of an index.
That it simplifies and caches the specific column or column combo for searches in efficient way.
So the search only goes against the index, and only the matching wanted full rows are then accessed.
But this index only works when the match is only on that name column.
And has much less impact if the WHERE contains other columns too.
Depending on database engine it might not be used at all if it isn't full match.
for this purpose, it's a single column exact match
SQLite, unfortunately
And point was that I have no idea what SQLites limitations are.
sure, but... I still think that querying full content of column (all rows) and doing set.intersection is going to be quicker that individual requests to any SQL engine
Point is to make the SQL engine do the intersecting with properly done one query?
I have much more faith in it recently, but still realise that it's not as good as postgres
And just get the end result in one whole dump.
hmmm, ok. i didn't think of delegating mathing of full set to db engine. should try
also, just putting index on that column did wonders i got to say
(looks like engine doesn't use indexing if it was applied outside of its process)
You know... I have gotten a few ads on the webbed site Tumblr.com in my lifetime
But this one is the first one to leave me genuinely puzzled
What the fuck does that mean.
Drugs!
Embrace those Madagascan relations!
lambda s: s.isnumeric() or (s.isalnum() and s.islower())
is much (~6 times) faster than
try:
int(string, 16)
except ValueError:
return False
return True
on random 20-char strings
what would you do on it being a number ?
it's much faster to convert it right away if it is possible
print(chr(sum(range(ord(min(str(not())))))))
Guys don't run this in Python. Worst mistake of my life
I just got invalid syntax
Now try
it must be hex
number as in?..
i don't need numeric value from it
my brain crashed when i tried to remember what a weird expression is not
so you want us to not repeat your mistake by telling us how to repeat it?
actually good way to crash your memory is to build enormous list, which can be done with powers of 2
["-" for _ in range(0xff**0xff)] should be enough (advice to not run it)
I initially thought it was some memory bomb or smth, reality is just boring
Nah its not a memory bomb. I don't wanna be banned
If bad software is getting you banned, then Discord should implode
the whole earth should implode
Arrow Lake pricing coming in below the street price of the 7800 X3D is probably sensible, but 12700KF still available for $166 according to PCPartPicker.
5700 X3D also looking really good at $182.
feel like with 9000 series out that the 5800X3D should be that price now
It's held at around 300 for years now
thanks Steve
Thanks intel
I mean if 5800x3d is much more expensive than 5700x3d
it's probably because people got the hype about it
lol. $15 more than a 5700G
that has to be one of the biggest jumps between price & performance on any platform
wow, 5950X is $330 now
Worth buying new at those prices and skipping each of the current sockets.
that was my thought
if you need something with a lot of cores but not current top-end performance, that sounds like a bargain
So i have this dell latitude 5490
Brought it to shop to fix not starting up and it was fixed by unplugging and plugging battery it started without problem
Now tried to start it it was on got to login screen put password been in for 2 min and bam black screen
Turn it on again just to see amber amber white..
Your fault for getting a Dell tbh
Hold the powerbutton for like 40 seconds.
Otherwise tough luck. Dell is for businesses that also get the pro support because those laptops are so hilariously cheaply made its insane
7 year old laptop, well past its prime
most laptops are designed to last just past warranty expiration
hey guys, i know alot of ppl here have some nice knowledge about debian etc. I need help.
My root server is kinda locked , i cant use my root passwort , i cant ssh into it. I tried mounting it in the rescue system and chroot into it and passwd the root but it wont work. I made a new user with full sudo rights but i cant ssh into it also. ssh conf looks right and i never changed anything since my last server maintenance.
Please i need some urgent help. Any ideas ?
Why didn't the rescue & change password not work?
i have no idea thats my problem ... it says user password changed but after unmounting and reboot i still cant root in
Locally or SSH?
SSH root login is disabled by default in SSHD config.
no root login is enabled
i never had a problem with it in 5 years , never changed my ssh logins
What does sshd log output say when you view those entries in the rescue system?
Probably either in daemon.log or journalctl depending if it is SystemD based or before that.
wait nvm, was wrong , was old log
pam_unix(sshd:auth): authentication failure; logname= uid=0 euid=0 tty=ssh ruser= rhost=
Failed password for bin from
but its 100% not false password, it now so simple
i dont see anything else in the logs
i have the feeling passwd did not change it but my old root pw also dont work anymore
What is the shell for root set in /etc/passwd ?
And you are sure the password change actually affected the right /etc/shadow ?
wait, ill check and make sure
yes it the hash or whatever it is ^^ its changed after passwd in /etc/shadow
Check that /etc/ssh/sshd_config for the PermitRootLogin entry.
And for PasswordAuthentication.
Both at yes?
Authentication:
#LoginGraceTime 2m
PermitRootLogin yes
#StrictModes yes
#MaxAuthTries 6
#MaxSessions 10
To disable tunneled clear text passwords, change to no here!
#PasswordAuthentication yes
#PermitEmptyPasswords no
PasswordAuthentication is outcommend. weird never did that or is it right ?
Should default to yes even now.
so should i delete the # on PasswordAuthentication ?
never changed anything there
Shouldn't matter.
alr
And does local login in the actual system work?
dont have local access , only via rescure system of my provider. kvm access was a thing but they diasbled it sadly a little while ago.
Only have SSH or their rescure boot
And that was for user "bin", not root...
Not your attempt but just random brute force attempt.
Point was to find your attempts entry in the log(s) by time etc.
There will be lot of completely random failed attempts in there for root, bin etc.
so you mean someone bruteforced me and randomly did some weird shit that even a chroot passwd wont work ?
No, but point is that most of the "root" login failures are probably not you.
oh wait nvm . i get what you are saysing
yea thats true.
hmm do you have any idea where to look further ?
By finding the log entries for your login attempt.
In that daemon.log and possibly the auth.log
Or journalctl.
By time and IP.
i think i found something ...
"User root not allowed because shell /bin/zsh does not exist"
zsh might have gotten corrupted on disk and removed?
Or package management things.
Or something changing the shell entry.
weird , in passwd it says /bin/bash but after the reboot the journal still says /bin/zsh not found but the /etc/passwd points to /bin/bash
ill install zsh , lets look
hmm no logs since 11o clock ... not in auth.log not in journal
Something really wrong with the file system probably.
Or is it logging in UTC?
hmm "date" is showing correct time , so can it be logged in a different utc ?
ill start rescure now and try "chsh -s /bin/zsh root - maybe setting it new works
UTC is specific timezone.
Replacement for GMT.
yea but the minutes dont match either
Then.
shells: valid login shells
/bin/sh
/bin/bash
/usr/bin/bash
/bin/rbash
/usr/bin/rbash
/bin/dash
/usr/bin/dash
/usr/bin/tmux
/usr/bin/screen
/usr/bin/sh
/bin/zsh
/usr/bin/zsh
/usr/bin/zsh
could it be that the path of zsh is wrong ? but bin/bash dont work either.
I think something really fucked
Symlinks broken?
Filesystem in general broken?
Those /bin/ entries are mostly symlinks to /usr/bin/
can you help me , how can i check ? Dont looks like its broken but im not sure
Is there / and separate /usr/?
Or on same file system?
What file system format(s) are they using?
When unmounted or mounted only RO in the recovery, what does running fsck on the partition(s) say?
root@rescue ~ # fsck /dev/md2
fsck from util-linux 2.38.1
e2fsck 1.47.0 (5-Feb-2023)
/dev/md2: clean, 996238/121012224 files, 88573752/484019984 blocks
wtf ... tried their vKVM and i managed to login !!! but ssh wont work
WTF , ssh now works too
i dont understand anything now . need a whisky
But thank you so much @twin dew !!!!
Dont know what it was for sure but all it counts that im back in my system!
Nah it started working again
Most likely nearly dead small battery to kickstart laptop
- this laptop isn't my main thing i use
It's more or less to do university
Putting notes there and that is all
Maybe small gaming but with games which are more of older one or are just very simple stuff
Epic has fucked up with UE5.3 even more...
Now the TOTAL system VRAM usage needs to stay below 90% of card VRAM, or they will start doing stupid shit...
enshitification continues...
could it be sign of increased corporate espionage & sabotage?
I like how in recent years there is a direction on YT to improve understanding (and popularize) of intrinsic physics phenomena, like speed of light, speed of sound etc
most recent vidoes from Action Lab about enthropy
and most recent video on holograms from 3blue1brown
And that is real limiting thing, as each Electron software etc. used 300MB VRAM when HW acceleration is on.
as my professor said on lecture on optics - holograms simply didn't get studied enough
when did Windows started adding RAM to total VRAM?
No?
It is just that the electron apps are using the GPU directly, and using VRAM too, at that pretty excessive amount.
alphaphoenix about electronics (and some computer science)
yes, Yes and YES
i mean, when did Windows started allowing GPU to use RAM
Currently Chromium browser is using 500MB of RAM, but 850MB of VRAM.
Microsoft Office Outlook is using 350MB of VRAM, even when has been minimized to tray...
That has been basic thing GPU drivers can do for 15+ years.
the university in my city had (maybe still has) a team working exclusively on stuff like optical tweezers and holograms (nonlinear photonics)
super cool shit
but then why do we have problems like having low VRAM at all?.. aka games reporting that, and not just slowing down but refusing to work
the president of Germany's Physikalisch-Technische Bundesanstalt (federal metrology institute) actually worked in optics quite a lot
i know mine has lab for producing holograms, like those sticky ones (i hope that's not classified, haha)
Why does Epic code their engine to not allow GPU driver to do its thing and just get some performance degradation, and instead does stupid shit that causes even worse problems?
How would I know?
and that makes me question: are there sound holograms?
like, i remember years ago i was wondering if digital sound can recreate the scene from concert, for example
yeah, kinda like that.
but am pretty sure it's a problem in many engines and apps.
and it feels like until recently Windows task manager didn't report total (shared) VRAM
Because it reporting anything GPU related is pretty new thing.
And that was in it once any GPU support was in.
IIRC Win 10 was first one with that.
And not sure if it was there in from beginning or not.
Using RAM as VRAM extension was thing at least from Vista, probably before.
i just realized that US seismic network actually got smaller
and Antartica as well
question is "why"
and last year there was a spike of seismic activity in Romania, very concentrated
and Volcano Discovery doesn't show any volcanoes in there (small area)
and I don't see any tectonic ridges there
and it's too deep for any mining
i hope i will find some simple explanation
I watched a documentary about Yellow Stone while back, scared the hell out of me ๐
Freaking super volcanoes.
yellow stone is literally hell on earth sleeping
waiting for doing servicing again in some years
how convenient to have them distributed around the globe: Yellowstone, Campi Flegrei, Aira
Hello satisfactory community!,
I'm running my own server, using AWS eu-west-3 as I'm located in Spain - Madrid.
Now, I wonder if there's something to configure or, is there some known stability network issues regarding eu-west-3 Paris (can't find much about it), since without changing any configuration, I get a normal 50ms, but when doing something it will become rapidly into 150>, then 300> and even with a friend, into 2000ms. This is not normal behaviour of course, and makes it unplayable.
I really wonder if this is more about the CPU, which I see on 'top' to go around 60 to 100 to 110 % and so on. There's RAM available. But I guess the CPU fluctuating is normal of course.
So I wonder if this power spikes are because of network stability, and I should change my region to a more stable one (I wonder which one), or if I should upgrade de CPU (instance-type).
Thanks!
this channel isent for this or satisfactory at all go to https://discord.com/channels/370472939054956546/902621736602861588
so...
For example, i have some number of non-intersecting starting requests
and let's say i can split each of them in half, and it is done so during downloads when certain number of retries is reached.
Due to the nature of downloads, i will have to start them over and over... (they are long, and sometimes doing retries doesn't help)
So, at the beginning of another download, there should be choice of how to treat those requests
basically, idea is that for each starting request, instead of just checking if they have been saved/cached, i can also check if their 'children' were.
But, there's limited recursion. And i am not sure how to handle it.
given that some request got 1/2 cached, and the other one not, but that other half has it's 1/2 cached (1/4 of original)... i chould only request that 1/4 that is missing, but I don't get a good way to program this. and it feels like there could be some nice solution already.
I mean...
Just recursive func that does splitting, checks if they are cached, calls itself on everything what wasn't cached - kinda simple. But what about cases when nothing is cached on first split, but there are some instances on 1/4ths and 1/8ths? Technically, if they cover full range - no need to do extra downloads. So it makes sense to check in full depth... HMMM
why do you need to check
either you downloaded either you did not
if you downloaded you should know about it (how where when how much)
this is either bad design or you are afraid of running a basic log tree
Today, the National Security Agency (NSA) released a copy of an internal lecture delivered by Rear Admiral Grace Hopper from August 19, 1982. Known as one of the most influential figures in the development of early computing technologies, Hopperโs contributions have left an indelible mark on the field of computer science, particularly in the rea...
she'd still be smarter than most people in any room
this is an incredible lecture
Catch is the main event IMO
yup, they've done everything else
from the KSP discord ๐คฃ
the catch will be wild, but so will starship reentry, i think it might be on the daylight side this time
ok Ars Technica wins the headline writing for it
in ideal world, i wouldn't even need retries: generate, get responses, parse and save - ez lemon squeze.
But, so far, I cannot be certain that parsing wont change (extend) in future, so for now I am focusing on just downloading data, so primary property for deciding if request should be made is check if it was cached.
what I mean is whatever you're using for requests should report it failed or completed or partially completed which you can save
no need to to "check*
it's same as checking if it was cached (cause i tweaked caching to only work on successful responses)
there's no problem on getting to know if individual request was sent/cached/saved, but to understand next time what requests to send, because there's splitting of requests.
aka some starting requests have failed but all it's split ones where successful - so it can't decide just by starting request to resend it - it needs go deeper
it does not need to go deeper because you just want to retry the split request that's failed
"it's the same as checking that, the only thing is I'm missing information now" no way
"I'm struggling to get detailed information on each request results" then process these details at each requests
instead of trying to recover the info after having handling it
Huh, see, that's the trick - if request failed, it could be because it's impossible to get retrieve. There's no guarantee!
That's why splitting is there in the first place - it wouldn't exist if i could just retry initial requests until i get response.
And that's the decision i am not sure how to make - retry failed or split and try smaller requests (current implementation is that there is limit on retries and limit on how deep splitting goes). And similar decision has to be made at the start of new run when potentially some requests where already processed or cached.
Oh SHIT
Some extra fire, but otherwise good.
ok, that was epic
one chine got obliterated
something's still burning, i bet it's hydraulic fluid
i'm still kinda thinking it'll go boom like that one starship test that landed kinda hard & sat there for a few minutes before exploding
now i can't wait to see reentry in daylight
looks like that little fire is out finally
still kinda worried about the molten stuff falling off
SpaceX has intentionally gone with minimum protection as baseline, and has been increasing as needed as shown by tests.
reentry coming up
oh plasma already
damn, had hoped the reentry was going to be sooner than it was so it could be in daylight
Doesn't really matter when there is no-one to capture it at the landing area.
Ok, they did have cameras near.
So prepositioned buoy and were able to capture the landing on other side of the world...
efficiency of accessing data isn't much of concern (at least yet) - i bet that even good implementation of tree will crack under 25mil entries (that's just 2 splits deep)
as i see it, there's need for a good way to decide when to try checking deeper
simple case: initial request is processed - ez
still simple case: initial request wasn't processed, checking it's children, all of them are - also ez
harder case: ... some children weren't processed - try their children? how to set limit on splitting: just hard limit or some more complex logic? when to try lower degree split even when higher degree splits where processed?
I just can't grasp a good logic on how to deal with this...
tbf, 25mil is far from real amount that would be request/stored/checked.
but 1mil is good estimate of how many requests will be cached when it's done
Project Borealis: Prologue
Return to the snow-covered streets of Ravenholm in this reimagined continuation of the beloved Half-Life saga. Project Borealis: Prologue is a fan-made gaming experience set before the events of Project Borealis.
Fall 2024
https://www.projectborealis.com/
https://twitter.com/proj_borealis
https://www.reddit.com/...
Hey man if it works it works
Nah its a fan thing
If you still run on w7 it's a skill issue.
To answer your question, if I use the same settings for lumen as I do for not lumen, the difference is 70-90 fps vs monitor capped at 144 fps. When I'm compensating for the extra power, my typical settings have me around 120 fps. 120 is my target smoothness.
I've got a 165hz monitor and over 60fps in a game like this I settle just fine
That is a preference. I do not like the disjointed feel of turning my head at high speed and having to see the world updates in slide shows.
Yes I'm calling 60 fps a slide show. Because I can definitely feel the difference between it and 120, but not between 120 and 144.
I can barely feel the difference between 120 and 165
it's there but rather subtle
Yeah its something like that for me as well. 120 is very smooth to me, and if I'm having to balance looks against framerate, that's my target for any game.
90 is still noticeably less smooth to me.
I've been playing a lot of Star Citizen and Satisfactory and since both run between 30-75fps It's been a while since I played something at 120
so direct x12 is "locked" behind windows 10 because microsoft is using that to force people to upgrade. But vulkan is catching up fast, and mass adoption is getting even better. Vulkan is a lot better about how it detects and utilizes the capabilities of cards, and doesn't rely on directx12 telling it what it can and can't do. It is true that you won't be able to take advantage of some of the newest rendering methods and microcode if you're using an older gpu and your frame rate will definitely be potato but that's not what's actually holding you up. It's microsoft and their finger wagging.
Yeah that's the only thing about dx12, fancy shit like RT and so on
That Epic decided to use Windows APIs (outside DirectX12 etc. graphics stuff) that don't exist in Windows 7 yet?
I am pretty damn sure you can run vulkan on windows 7.
At point when Windows 10 hadn't been supported for multiple years anymore?
I specifically said OUTSIDE of graphics APIs.
What are you saying
It just makes sense to do little thing to make it playeable on win 7 since a lot of ppl don't use it...
I do not follow
There are new APIs to other stuff the OS does with each OS major release.
And Epic chose to use the newer better Windows APIs to some file system stuff and network stuff, that don't exist in Windows 7 yet.
So?
As Windows 7 was way outside of support at that point, so Epic decided to not support it either.
And Windows 10 was free upgrade, and could run on all computers that could run Windows 7.
Epic isn't supporting or not supporting windows 7. Satisfactory doesn't much care about your OS. It just needs the proper directions to the hardware, and the proper methods for communicating with it. That is what the OS does.
Epic with Unreal Engine 5 is specifically not supporting Windows 7, as they chose to use Windows APIs that Win7 doesn't support, but Win10 does.
API set that Windows provides is not static target.
Even outside graphics like DirectX12 support being added etc.
Point being?
You got the point already, he didn't.
I mean steam being the largest PC platform, only 0.3% of users still on win 7... Epic's numbers would be even lower
So it sounds like you're saying this game doesn't run on windows 7 because the devs made a short sighted decision. I'm saying you can probably make it work if you wanted. On windows 7.
Epic chose to not support Windows 7 API set, as the OS was out of support when UE5 was released.
So UE5 games don't support Win7.
Why is that "short sighted"?
Not supporting very small userbase that shouldn't be on that OS anyways, and making their work easier and the game tiny bit faster?
Epic's hardware charts would most likely be 99% Windows 10 and 11 0.8-0.9% MacOS and 0.1-0.2% others
People can be on any OS they want. Microshaft forcing people to upgrade like they're doing is bs.
So all games should still support Windows 95?
Let's not strawman.
Same kind of OS support deprication for games has been going on for a long time.
Epic decided to drop the support for dead OS few years after it was dead in new engine major version, first in 10 years.
So that they could use better APIs that the newer OS versions had added.
Selecting what API set to use is always selection between backwards compatibility and features, ease of use etc.
Epic selected not to be stuck with older API set for years after the last OS to need that support had been buried in any sane persons mind for multiple years.
Just run Linux 
A game developer deciding to use a more efficient API that is only available with the newer Kernel version and reducing bloat by not building in backwards compatibility is not forcing anyone to do anything and is not unethical
Its not their fault people are simping for win7
I love win7 but its time to move on
Not to mention it hasn't gotten a security patch in ages
that is what happens when they "drop support"
Yea so simping over it makes no sense
Xp tho 
Which was the point on me saying that no-one sane has been on it for ages anymore.
As the support ended in January 2020.
Microsoft however is trying to use their monopoly and IP (directX) to force people to upgrade. And through these "upgrades" we are relinquishing greater control over our hardware to microsft... every patch.
Maybe im too young but I never liked xp
the gui of explorer was ass
I didn't like XP either. I stay on win2k pro right until windows vista... which was also garbage. But At that point it was time.
What? Competing API's exist look at Vulkan. Besides nobody implements DirectX according to spec. The reason we dont have other renderers is because theres not really any readon to
Yes. Yes. If you say so. Disagree on the last bit. Linux is growing... far faster than some people realize.
I think that's extreme.. but they did say windows 10 world be the last.. I'd like to have windows 7 back lol
W10 is my last ๐คฃ
Yeah, the Win10 Win11 split is just stupid on most part.
They didn't even increment the minor version, only patch level.
As mostly it was just UI replacement with very strict official hardware requirements.
Now it has diverged more.
I spent 6 months recently running wholly on linux. Gaming. On linux. All my productivity, and everything. I did eventually hit a hard wall on what I could and couldn't do vs being on windows. Would you care to guess what it was?
w11 is mainly because tpm2.0 which my pc does not support ๐ so yay for replacing all hardware just to use windows.
Linux has its own problems
Win11 should have been 10.1.xxxxx
But it is just 10.0.xxxxx like Win10, just starting from 20000 for the patch level.
Except for my steam deck, all.. of my Linux systems are cli only
What I couldn't do was run games that required anticheat software for which there wasn't a maintained linux db.
Everything else worked. Flawlessly. (once I figured out how)
The forced tpm requiement didn't/doesn't really effect me, but it's still annoying.. it makes for an easy way to stop the automatic upgrade though
you have no idea how directx works
or how graphics drivers work
Alot of games use directx...
There's alternatives now
DX12 and DX11 are extremely different and operate on different philosophies
I know its a set of instructions that talk to the hardware through the graphics drivers about what the software is wanting to do. That's the short version.
there always have been alternatives. OpenGL predates D3D
I was referring to Vulcan, but that too
its a set of specifications to be implemented in a graphics driver
DirectX is written by Microsoft and implemented by AMD, Nvidia, and Intel
theoretically, it can be implemented in Linux, and it kind of technically has since DXVK just translates the calls into vulkan calls
But in DirectX:s case, there is middle-layer in the OS too.
For versions up to 9, it was separate installable.
DirectX and D3D are technically different, yes
DirectX is the whole encompassing system, while D3D is the part in the graphics driver
DirectX also has things like audio stuff
i don't know DX but it think it also contains the D3D equivalent of GLEW or other competing libraries
New features in new versions of Windows makes perfect sense, why on earth would you keep developing an obsolete product that nobody uses?
i dont like microsoft's choices (why i dont use windows), but this is not one of the problems
Yes I understand all of that. You are missing two important points though. The reason AMD and Nvidia and Intel have to communicate so closely with microsoft about what they are making their hardware capable of is because the keykeeper / gateholder of mass adoption is microsoft. And they own directx entirely. AMD, Nvidia, and Intel do not need directx to exist or be functional. They need it to be purchased widely. It's one big bucket of palm grease.
That is why vulkan is better.
Not really?
Many of the DirectX features were in HW first, then in SW.
Whole DirectX12 was first AMD project.
Which was rewritten by MS later.
Vulkan was also an AMD project
Both came from that same seed, DirectX12 more closely.
this is also false. AMD and Nvidia implement things in their drivers outside of DX all the time
they are called extensions
it got really bad in OpenGL
hence why Vulkan exists
Further evidence and demonstration that they do not need microsoft's support to do their thing.
yeah they don't, hence why they don't
Microsoft has DX because Vulkan didn't use to exist
It didn't need to until it needed to.
AMD got fed up with OpenGL and DirectX11 holding them back and made a project that took the world.
Just not under the original name.
before Vulkan it was OpenGL and DX, DX surpassed OpenGL so it got used more. Then AMD got mad and make Vulkan/DX12
OpenGL as organization was just so slow and almost dead for long time.
DirectX just makes life easier for MS
since the Xbox also uses DirectX, which is what is was named after
another reason DirectX became used more
DirectX was just easier to use than OpenGL.
As the OpenGL standard was and is so broken.
From the very slow consensus approach without rewrites to core stuff.
yeah RIP OGL
annoying vendor specific extensions
From that too slow main approach.
For the future of gaming, Vulkan and linux need better mass adoption. Otherwise, microsoft is going to continue to be a control problem.
So everyone created their own extensions to do stuff they couldn't get into the actual standard in sane time.
not really
DX just makes development for windows/xbox easier. You can use both
hell, satisfactory has both
i never ever going to try my code without tests, again...
fuzzer ftw.
Welcome to tech talk
If there's a * to what you're saying, it'll be brought up lol
Linux support partly.. or majorly thanks to the steam deck has come a very long ways.. the elephant in the room is anticheat, which I think should be addressed server side with a trust no one model rather than client side trust everyone model..
Trust nobody model leads to false bans
Anticheat is hard
Problem is you build too much logic into the client.
Probablity based model seems neat, perhaps this could be an actually good use for "AI"
It's not ai
Yeah thats what the quotes are for
But a neural network could be good at checking the probablity of events
Just make it a requset model. Server should do all the logic either way.
Want to move, request it.
Problem is latency.
Yeah ping could ruin things
and it will be suspicious of any emergent gameplay, potentially killing all e-sports
That is true
Did I read something about how neural netwroks are now better at predicting weather than proper simulations?
It wouldn't surprise me if anticheat starts to work via know it when I see it ML intuition of some sort.
Make Esc work in the blueprint designers!!!
??
This is a major part of my argument with a dude irl
Serverside anticheat introduces enough latency to be noticible, as the server has to dictate every microsecond of player input from all players. If we take it a step further and somehow remove enemy positions from memory (that is, until they are immediately needed (you and/or them has a line of sight on the other), that introduces even more problems
Actually it's something of an upside of game hiding enemies based on some discovery mechanic.
Like in WoT (visibility range and camouflage) or Dota (fog of war)
And some tricks like keeping only probable enemies on client side gets very tricky outside of predictable games like CS (close quarters, limited space)
Honestly, the idea of cheating is gross outside of reverse engineering game systems
But, what can i say, irl cheating is even worse
Hiding enemies is fine, I'm talking about enemy locations/whatever not even being present in client memory until the server decides the client needs that info
@halcyon moss So my question to you is are these compatible with each other (in the amazon cart) and are they compatible with the processor
i dont see why they wouldnt be, thats a very common combination
mainly wondering if the GPU is with the cord
i think it is but dont quote me on that, id look up what specific port it is
and stop fucking pinging me
will you stop fucking pinging me
well what freaking end are you using if its DVI then fuck no a 3070 isent gona support dvi
probably has 1HDMI and 3 display port up its rear
if your using the dvi for the monitor end and the dp for the gpu then yea it will work
client doesn;t have to "hide" what it doesn't know
NASA's Europa Clipper mission is now targeted to launch to Jupiter's icy moon Europa no earlier than Monday (Oct. 14) atop a SpaceX Falcon Heavy rocket from Pad 39A of NASA's Kennedy Space Center.
Broadcast courtesy NASA
falcon heavy in fully expendable mode
apparently it takes a bit to send 6 tons to jupiter
The official NASA stream later today:
https://www.youtube.com/watch?v=lQToTWKwtuw
NASA is set to launch the Europa Clipper spacecraft to explore Europa, an ocean moon orbiting Jupiter.
Europa Clipperโs launch is targeted for no earlier than Monday, Oct. 14, 2024 at 12:06 p.m. EDT (1606 UTC). The spacecraft, the largest NASA has ever built for a planetary mission, will launch on a SpaceX Falcon Heavy rocket from Launch Comple...
15:00 UTC
Too bad, working at that time.
The entire reason wall hacks work is because enemy positions are stored in client memory
There are some wall hacks that can trick the server into sending the data through various requests, but that is a lot more advanced
sure
So are you agreeing with me, or are you disagreeing without elaborating
you wrote a fact, there's nothing to (dis)agree on.
You responded to it with sure, which is easily perceived as sarcasm or otherwise disagreement
My bad
Ddr4 + super old intel
So no, the ram is not compatable
You need ddr3, but i would honestly upgrade the whole pc since that CPU is going to hold you back hard
Also, whatever DVI-only monitor this is probably isn't gonna do the 3070 justice
If you have $500 to spend on upgrading a PC with a 4770, you should not be spending $400 on a 3070
Agree
the most improtant thing for computer confort is having longterm good keyboard mouse and monitor
then a good tower and psu
even if its laggy and slow
it will be much more comfortable
Safety > comfort = performance > looks
Good luck playing Cyberpunk on your shiny 4K OLED with an i7-4770 and a GTX 960
Performance IS a comfort factor
that's not the same comfort we are talking about at all
and go play cyberpunk collector edition on your dvi monitor
I'll laugh
with flickering ofcourse
go get your daily dose of expensive eye strain
Our management board member told us a story today about how a competitor ISP sent a salesperson recently to knock on people's doors and offer fiber internet. The poor salesperson didnโt have a clue he just knocked on the door of a management board member of the competitor and our MB member started asking questions about what CPE's are offered to the customers and what wifi standards are supported.
The poor guy froze and just said: let me give you contact details of my supervisor because I don't have any clues about the technical aspects 
Out of 800.000 people in the city, you just knocked on the doors of competitor ISP CTIO, talk about bad luck lol
seems about right
you can take your cpe and shove it where the sun doesn't shine.. I hope the fcc passes a ruling like they did with cable providers.
I shouldn't have to buy a $200 sfp that can mimic the isp's stupid white box just to get the rated speeds I'm paying for.
Wait, you have to buy for an additional equipment so you can get the speeds you are paying for? ๐
To not have to use the ISP box.
I think you can directly connect your own router to an ISP ONT here (you could also buy your own ONT if the specs are the same even though I don't see the point)
That is replacing the ONT.
having good monitor is the best upgrade for visual fidelity and eye comfort
User owned fiber module that can be inserted into SFP port, that fakes it is the ISP ONT.
Because with many ISPs, they don't allow bridged mode on the ONT for example.
Oh, in that case you have to find an sfp module that is compatible
And pay for it ๐
And just have integrated ONT, NAT-router, WLAN box in permanent NAT mode.
It's possible with my ISP afaik but things get complicated if you need a phone line and IPTV too
Although IPTV could be solved by turning on "pppoe passthrough" I think
Yup, their nokia modem is a shitbox that offers piss poor passthrough performance. so I can either embrace the comfort that my isp will keep me so very safe with their "active armor" protection.. or suffer by letting my UDM gateway handle that.
so I'm being punished by having my own gateway.
somebody have problem with gray screen with blue vertical lines while launching the game?
Hello there fellow Pioneer!
You're currently chatting in off-topic channels.
Please note, the entire off-topic category is reserved for not Satisfactory related chattery. To comply with your FICSIT Inc. contract, please make use of the appropriate channels in the Satisfactory category. Thank you!
it's bullshit.
I like how the technicians glued a transparent fiber strand on the wall corners and ceiling in my sister's apartment. You can barely notice it
I noticed something on the walls and thought it was a spider web or something, later realized it's fiber lol
Fun fact: a technician will ask you if you own cats before running fiber strand through the apartment
Because cats can see the laser and they will play with the fiber so the technicians have to protect it
I blame @inner shale
Oh, when my house was built in 1988, it came with a phone line covered in cement...
They also get vision damage lol
I am more worried if my mother comes to sister's apartment with a vacuum cleaner. I bet she is going to damage the fiber considering how clumsy she is
Not to mention the fiber is transparent and hardly noticeable
Get cat6 and you should be safe
*Not to be confused with owning 6 cats
you can be happy your isp is allowing you to do that, most fiber lines are ont serial locked here
Good thing I could use my own ONT if I wanted, you just have to change the serial and input the one from the ISP ONT
ours is serial locked.
so the sfp you have to get is a programmable one that looks like their provided modem.
that's why it's $200 for a SC fiber sfp :/
it might be sfp+ I didn't look at the speeds it could handle.
real question if one know here
I know sfp+ socket can handle sfp modules right ?
what about sfp28 etc
can't sfp28 socket handle sfp+ modules ?
yes
Wish Ubiquiti had some basic dream machine with sfp port and not something expensive like UDM pro or similar
but depending on the device there may be limitations if it has multiple sfp28 ports
I know about variing software compatibility
like all ports running in 25gbps mode if at least one is connected
but physically speaking sockets are retrocompatible ?
yes
so sfp48 can support physical sfp28 sfp+
(I know there are some problems with sfp, but sfp+ is the baseline iirc)
(normal sfp-coding shenanigans may apply)
what have bobbing to do with factorio?
do we need factorio-anatomy channel?
this is funnier now that I know how to code.. might cop this one for my next project to mess with the prof
Long thread:
There's a highly inaccurate article about Pixels from Cybernews making the rounds everywhere in privacy communities.
grapheneos.social/@GrapheneOS/113275934184024856
That's efficient but takes some time to understand.
return x%2==1 would be more legible
And I'd even argue that if(x%2==1) {
return true;
} else {
return false;
}
Would be even easier to intuitively interpret
This morning SpaceX performed the 5th flight of their Starship/SuperHeavy launch system and once again made a clear step forward as they demonstrated precision touchdown with both the booster and the Starship. The booster touchdown demonstrated a radical new recovery method with the rocket carefully maneuvering next to a tower to be caught by a ...
Good coding isn't all about cramping as much processing into as little space as possible.
It's about taking care of future coders (including yourself) that have to go back and look at the code you wrote
Tesla is in the news again. With their Optimus-robot. That is controlled by a human behind the scenes. ๐
๐
bool(x%2)
and it's explicit
just like amazon go?
compiler must optimize %2 out (&1)
(and btw compilers can't do magic, but I've been specific, this optimization it must do)
was some robot event flooded with Tesla bots. Someone asked if they were automonous and the robot answered it was controlled by humans.
idk how bool() works
it takes the value and evaluate it to a boolean (true/false)
boolean evaluation is well defined
bool() is explicit
and %2 is the semantic of what is done too
apparently there is no this writing way in C# ?
C# instead use things you must absolutly not use in almost any other language sanely ?
Microsoft pulled a microsoft again
Im confused... why is sha256 for the same input different
Between windows and linux??
Oh yeah, completely forgot that modulo is a mathematical operator, not a boolean one, it returns a number value
Tho I did do that properly with x%2==1
the ==1 is a bit overloaded but yea
what are you sha256 ing and with what function ?
Input text had a newline... love that...
my customer is stuck on 4gb of ram
idk how but he managed to get a pc running 32 bit win10
remember times when you could run a game for hours and that hit a story point that caused game to either crash or just not trigger progression? lovely
probably felt genius and picked an old install media thinking it'll do it
(if the pc have more than that in hardware)
win7 32b or something ?
Because Windows uses both carriage return and line feed for newline.
Unix uses only one of those, don't remember which.
And then something else at some point used only the other.
Officially there are separate characters for returning to start of line (carriage return), and going to next line.
From how typewriters worked.
Not to sound rude, but I know that... it's just I didnt expect the command to add newline automatically for a text file containing no newline...
Ah, sorry.
Yes, most echo etc. type commands add newline to the end if not suppressed.
Because otherwise the result would be this in shell etc.:
Testing<redacted>@server-irssi:~$```
Same with files:
<redacted>@server-irssi:~$ cat Testing
Testing<redacted>@server-irssi:~$```
I did sha256sum x
Ah, the enter at the end to run the command counts.
hello, does anyone know how to check which version of nvidia driver i had before updating? i updated to the latest version today and its causing satisfactory to crash
When using appended string.
So would need to do that echo -n "String" > sha256sum probably
Are you sure?
Most text editors don't allow you to create text files without newline at the end.
And add that automatically.
But that "cat file" is easy way to check for example.
If the output doesn't do that "stupid" thing that is on same line as the next command prompt, then the file contains newline.
Yeah, VIM adds newline as default, can be configured to not do it.
I hate windows server and managing groups/users folder permission....
Couldn't access a specific folder and realized I added my testing account to two different groups which had access to the same folder resulting in permission conflicts
od -t x1 filename
if it ends in 0a you have a newline:
x@hotblack:~$ od -t x1 x
0000000 68 69 0a
0000003
x@hotblack:~$
I did solve the issue.
But I'm speculating my SD card driver (embedded) is reading some parts wrong so I get checksum mismatch upon reading the file into ram.
But that's for tomorrow...
card driver should not care about files
the filesystem is
how are you loading what function are you using ?
it shouldn't give conflict that's a windows moment
Both groups had a mix of allow/deny permissions on specific folders and I solved the issue by removing a testing account from another group since I really don't need it there for testing
At least there was no issue for vendors accessing the server since each vendor has it's own group and there is no vendor username which belongs to multiple groups
That's what happens when you add a testing username to multiple groups and cause a mess lol
Last thing I need is my boss calling me a day before going to teambuilding and asking me to fix vendor access to File Share 
Not to mention so many authorization steps in order to allow vendors to access our server inside the company
I just request data from card, start block, num blocks, which is the "file" data. So if the data I get back is wrong, the file data isnt read propertly. If it's a file or not does not really matter...
then I dont get why the behavior would be diferrent as long as you read raw binary same file
unless you are editing it
Because card or card reader can be corrupting the data being saved or read?
now I also want a hydrofoil chair boat thingy
My bad
i'm currently doing a full clearcut of the map to see how bad the lag gets
The wrong channel thing by switching between my phone and laptop is killing me today.
I really hope Valheim figures out how to get save times down like Satis did, that's been miraculous. They're on Unity, though.
valheim's weird to me
how can a game have such crap graphics but be absolutely awful on your gpu
They've got a new patch on the public test branch, some of the patch notes allude to improved performance. They really got out of hand with the ashlands. I think now all the devs have upgraded their PCs since initial EA release and lost sight of the initial performance target.
i thought it was more CPU
There are large downsides to procedural generation and deformable terrain in performance front.
procedural is ok deformed terrain is more of a headache
Depends on how much of the procedural generation you do early, and save, and how much you run at runtime.
But Valheim probably does almost all at the world generation step and saves.
tile based seeded can be highly optimized
always rendering fully defined zones
it's the best for both performance and immersion
Valheim gets plenty of deformed terrain though
DRG manages both
With pretty small levels in the end.
Relatively perhaps. Have you seen last update deep scan? It's quite large
From first person it feels on edge of getting lost, considering that caves have details, more narrow, than in subnautica for example (like, we probably all got lost a hit in shallow small caves, but much less so in lost river)
Destructible terrain is in a weird spot since it seems like something every sandbox wants to have, but in practice 95% of usage is digging tunnels and flattening terrain.
It gets ironic when in game about dwarfs, even players for class with drills don't dig that much.
When potential is huge, both defence and offence
Being able to setup proper network of caves around caretaker is just one of greatest strategic moments if shooters
does anyone know which app causes this overlay?
looks a bit like steam's fps counter but can't say for sure
thanks, will have a look
was going to suggest it was nvidia shadowplay but probably not if you have an amd card...
Could be Xbox game bar in compact mode
Game Bar doesn't seem to have anything network related in it.
For the performance thingie
Ctrl+Shift+Q as default toggle but that thread didn't tell what causes it.
Yeah, that is from Razer software.
Under Game Booster part
Razer ๐คข
So Razers software.
Of course put behind key-combo that you absolutely never will be accidentally pressing when playing a game, that Ctrl+Shift+Q...
Because it does not follow manifest v3
you must abandon chrome now
There is UBlock Origin Lite for Manifest v3 IIRC.
But that is pretty neutered.
Amazon is adding more and more ads into their prime service too.
So when you want to watch something on it, in some countries, you get ads now ๐ฆ
feels like we are entering an era of ad wars. Tech doing everything so consumers can't block ads. And more and more consumers getting tired of ads.
streaming became cable
Yes, more and more.
Seems v2 extensions are getting auto-disabled, but they can still be enabled.
The removal of the support is only coming in early 2025.
Come join the Firefox side
thank you so much!
I can see again!
Go and disable everything under that Razer Game Booster Ruiner?
do you still need to login to an account to use a razer mouse
unsure but you do for basically any adjustment to any mouse nowadays its just razer's software is particularly shitty
Cant believe windows got sudo finally
stupid to have to open a WHOLE new elevated terminal every time. i hate it
y'all stoked for all the fun things we can do with windows sudo?!
Eh, when something nice comes out of MS, it feels like they've added a cool new thing to the prison yard. In the end, we all still want out.
I've always enjoyed the platform. It's nice and open. As a platform. People often think i mean open source when i say that. I mean anyone can publish on it without restriction or license at all.
wsl , the linux subsystems, aren't so bad. they make bridging tools so good. If this is a prison it's like one those dutch ones.
Paper Boi finds himself in a Dutch prison, but things could be worse. Watch new episodes of #AtlantaFX Thursdays on FX. Stream on Hulu.
Atlanta Season 3, Episode 2.
Subscribe now for more Atlanta clips: http://bit.ly/SubscribeFX
Taking place almost entirely in Europe, Season 3 of Atlanta finds โEarnโ (Donald Glover), โAlfred โPaper Boiโ Miles...
@sudden ivy - I'm on Arch
valid thank you for the channel lmao
ah okay im just on kubuntu
re: dx12, presumably just a gap in either driver or proton support. It's possible that since 1.0 released there's been new versions of each which would let me play the game w/ dx12
Though everything's getting translated to Vulkan on Linux anyway
I was using Vulkan pre-1.0 anyway, actually, 'cause it was more stable for me than dx12 was. (dx12, pre-1.0, would occasionally crash for me. Vulkan's been 100% stable)
im just scarred from vulkan because on my windows install, vulkan was incredibly choppy and didn't really function at all
but im trying satisfactory for the first time on linux so who knows
Even if you do use dx11 or dx12 on Linux, Proton/Wine will be translating it all into Vulkan anyway. So I figure the Vulkan implementation just cuts right to the chase. :)
Would be worth sticking to the default and seeing if it works for you, though. For me, dx12 on 1.0 literally wouldn't even launch the game, though
Anyway, headin' AFK; good luck!
yeah i would prefer to be on vulkan anyhow, i just feel like it being newer technology and innovations it should be inherently better
but dx12 is an oldie but guaranteed stable (at least on windows)
Oh, though before I go: technically the Vulkan support in Satisfactory is unofficial and unsupported; it's also not been touched in awhile. So if dx12 (the default) does work for you, then you'd probably be best of sticking with it
But as I say, for me I had to use Vulkan anyway, so it was a moot point. :) Okay, AFK for real
DX12 and Vulkan are the same age
And are both based on the same design principles
UE5 (what satis uses) is more optimised for DX12
Proton may do a better job at translating dx12 than the oroginal vulkan implementation
May have been a GPU/driver issue
Just make sure you have AMDGPU and RADV installed and enabled if on amd, or nvidia proprietary if on nvidia
never tried vulkan on linux but DX12 work pretty well
i dont get why you want vulkan over DX12
The Vulkan selection inside Satisfactory is marked as Experimental for a reason.
What it does in any system is pretty random from instant crashes to working perfectly.
And that is for specific system, not just based on GPU and driver version.
And there are plenty of games where using DX12 and getting converted to Vulkan is faster than native Vulkan.
Those DX to Vulkan coverters are very well coded, unlike most games.
iโm running ubuntu, iโm sure my drivers and kernel are in play for why itโs not running smooth
i donโt want to touch the kernels bc iโm a baby with linux still and ill definitely mess my install up lol
im pretty new too
mint work fine for gaming
have you checked if you use the latest graphic drivers ?
yea ubuntu ran like garbage for me too especially for games
this is so sad, i was just trying to look for an easy to use distro, so i went with kubuntu
Wonder at what time that 16:00 is...
Completely useless time marking.
Me letting a 5700xt slip out of my hands 
Hawaii
Kappa

My timezone in 15 years.
or which year or day
That was clear that it is in the next 24h.
But without timezone, it could be over 24 different times... (Some +xh 30 min zones out there)
He is coming over Saturday ๐
One of the first things I learned was that JavaScript != Java.
Branding excercise worked again.
Android operating system Gmail and JavaShit in same paragraph
Google lobbying spotted
the classical "detailed learning resource" of academics
what you know before reading :
what you know after reading :
also academics: javascript was not created in Java, Indonesia, and is unrelated to the coffee production style
in 1975, Edsgar Dijkstra famously proclaimed that โThe use of COBOL cripples the mind; its teaching should, therefore, be regarded as a criminal offence[sic]." This undoubtedly led to the decline of teaching COBOL in universities, but it remained the dominant business language.
little bit of cobol hate from dijkstra I found funny.
"In Plato's Phaedrus, Socrates calls into question the propriety and impropriety of writing. Throughout his discussion with a colleague, Socrates insists that writing destroys memory and weakens the mind."
oh wouldn't ya know, he had his own programming language. not biased there at all
Because everyone should have just used Fortran, the programming language of scientists!
Just like Fortran was designed for engineers and scientists to crunch numerical mathematics, COBOL was designed for accountants and admins to automate business processes. And it does a surprisingly good job at at that.
Just like Fortran today is a different language from the 1950's original, so is COBOL. Many of the oddities and shortcomings actually make sense when applied to the specific domain it was designed for. Don't think of it as a general purpose programming language โ it isn't โ but as a pre-Visicalc alternative to spreadsheets.
automating processes with a language not designed for engineers
sad stupidity
"but it's for business" look like a Microsoft argument
something ultimately made to programming and calculating must be good at doing that
no matter what title are the user
Completely different use cases back in then.
That COBOL is still used is weird today.
That it was used still in 70s wasn't.
Because it was very fast for what it was intended.
So you could actually get the job done in time available in the day.
Wikipedia:
COBOL was designed in 1959 by CODASYL and was partly based on the programming language FLOW-MATIC, designed by Grace Hopper. It was created as part of a U.S. Department of Defense effort to create a portable programming language for data processing. It was originally seen as a stopgap, but the Defense Department promptly pressured computer manufacturers to provide it, resulting in its widespread adoption. It was standardized in 1968 and has been revised five times. Expansions include support for structured and object-oriented programming. The current standard is ISO/IEC 1989:2023.
Academic computer scientists were generally uninterested in business applications when COBOL was created and were not involved in its design; it was (effectively) designed from the ground up as a computer language for business, with an emphasis on inputs and outputs, whose only data types were numbers and strings of text.
The Department of Defense was concerned about whether it could run the same data processing programs on different computers. FORTRAN, the only mainstream language at the time, lacked the features needed to write such programs.
Funny quote from Jeremy Keith: "Java is to JavaScript as ham is to hamster." Never heard that one ๐
Good comparison 
COBOL is the hack they donโt want you to know to get a job in CS and keep it until you die ๐
If your head can take the work.
i'm looking to gamerip the Satisfactory music files
-# โข๏ธ
Why do I get this crap? I never installed this on my PC...
Auto-install by Windows because of UEFI hint.
That Asus has on by default.
Baah
Can be disabled in BIOS/UEFI
Which means you need to remember to disable that specific setting after each BIOS/UEFI update.
Advanced mode, Tools tab.
I love the before Halloween snow Storm
I didn't get a snow storm just a bunch of rocks EVERYWHERE
rocks falling out of the sky?
I don't think so? Just heavy wind blowing rocks everywhere
Geologists love it
its over now. It stopped early in the morning for me
snow storm? try heatwave
Sounds lovely
We got 5/6 inches in 5 hours
Trees still have leaves branches are just laying on the ground because they broke off
@tough owl checked out the recall-file explorer dependency a bit. Fuck windows
we don't love it, we just find it gneiss
I want to slap everyone who says "200 MW per minute"
I think it comes with windows if you have ROG hardware (or Asus in general). I don't like it much though...
Btw, am I the only one who doesn't see much difference between Android and iOS? My sister gave me a spare iPhone 12 Pro Max and I've tested it a little.
Didn't find that "wow" feature which would force me to swith to iOS
i honestly like android because of apks
apples redeeming quality is I can reply to my girlfriend yelling at me with pool
Speaking of apps, I see android users are complaining most iOS apps are paid (even the free ones on android)
You have to have a payment method that works to download ANYTHING or update
even if its free
It's like one of those websites that offer free trial but still ask you for credit/debit card info hoping you will forget about the trial expiration and get charged automatically lol
I would also prefer iPhone 12 over Pro Max since Pro Max is too big for my hand
Btw, I have visited Nikola Tesla's birthplace yesterday ๐
And tested the tesla coil in his experimental station with fluorescent tube
I jumped out of my skin when my coworker pressed the test button on this coil lol (not my video) https://www.youtube.com/shorts/r1UnC_N6gJs
Tesla's birthplace is a perfect place to develop fear from electricity as a kid lol
If I visited his birthplace as a kid, I would be terrified of all the electric stuff
My mind is like "If it's noisy, it's dangerous" lol
Or maybe that just hidden fear of thunder
https://www.youtube.com/watch?v=mntddpL8eKE "Because our fluid, in this case, is air moving past a nuclear-powered manhole cover at Mach middle-finger!" is the science sentence of all time
#ad Thanks to New Scientist for sponsoring todayโs episode! Right now, New Scientist are offering 10 weeks for $10: this is just $1 a week for full digital access on the New Scientist app and website: https://bit.ly/KyleNS3
In 1957, a high-speed camera caught a cast iron cap moving at 66 kilometers per second. It was being accelerated by a nuc...
It's because of apple's fees
Publishing a free app on ios doesn't just cost time and effort, apple charges you for the privelage of contributing to their ecosystem
Always
im i the only one who thinks hp hinges is just tape
and laptops from hp are help togheter using tooth picks, hot glue, string and duct tape
no
Any of y'all have experience developing with firebase?
one of my hp laptops literally used the plastic housing of the speaker as the main clamp for the lid hinge
Spirit was selling NASA jackets, so I grabbed one for Halloween so I can say I'm a rocket scientist. Now I just need to find someone to make me a circular SLS patch with a diagonal strike through.
my d keycap fell off after i put my nail under the massive gap and began pulling
i AM the keycap puller
i finally gave up on logitech when their $120 keyboard had 2 scissor switches just snapped off
within warranty
your in warranty?
my acutal reaction to opening discord and reading that:
so far i havent had any issues with razer
best keyboard i've ever owned so far
its just poor paint quaility on the razer deathadderv2 pro
had it dangling off my desk as i waited for it to charge
and it hit the wall and scrapped it
i had one really weird requirement for my keyboard that no other brand had
what was it
a big fat wrist rest that's hard attached to it & doesn't just rest up against it
wdym hard attached
i can say im hard attached to my wall if i put hot glue on me and hug the wall
like integrated
like
not one you just put next to the keyboard but it's still seperate
send ss
i used to love my K800
look that one up
they stopped making it like 10 years ago, so when it died, i couldn't find a direct replacement
then i found this one
backlit, wireless, & the wrist rest, which has been my 3 big desires on a keyboard
annoyingly specific, but hey, someone makes it ๐คฃ
yea, wish they made something with macros
the wrist rest is detachable on the X5, but i wish it was fully integrated so it could have a big fat battery in it
loved that thing
wish they would make an updated version
too bad MX keys is what we got from logitech instead
i've never cared about rgb on a keyboard, just backlit for night/low light use
yup
and the k800 lasted like a week on battery
lol the flex in the case was hilarious
ah
i mean the k800 was all plastic, so it flexxed like a cheap laptop
that and the scissor switches
for $40 it was great
yea, logitech K800
weird
yea K800 was definitely not the K100 ๐คฃ
but look at that chunky wrist rest!
it was great
oof
the 12 button?
i loved their 6 button one
until the switches died
ah