#off-topic-tech
1 messages · Page 101 of 1
there's gonna be such a huge stock market crash from ai
The newest Steve Mould video has a GREAT explanation of how LLMs and Stable Diffusion work
only after all that bullshit fades will more competent models emerge
Probably, but they need money. So we get a lot of PR talk.
Anyways, LLMs don't even see the words.
They just get tokens, basically numbers that correspond to words or concepts
Most of it was ok, but there was like 20 seconds where he was explaining it on camera and some something stupid
Why?
Becuase it feels like PR and expecttions got too far ahead what can be delivered (in the nearest future)
Aka in a year it will be so overhyped that looks like some form of crash is inevitable
The fact people still overhyping it makes no sense to me.
It's mostly the companies itself overselling the product.
Well, nowadays we are way more connected than before.
It kinda makes sense to invest early wven if it isnot golden-eggs-duck
Byt, in context of climate disasters that ... Not just overhype, but looping hype, makes little sense nd gets annoying.
Also, any piece of automated data processing is now labeled AI
As was since 1970s
With dash app i can monitor temperatures from my phone on local network niiice
Crazy idea: put sensors in a bottle somehow and let them freeze to finally measure melting point
Hehe, that reads weird 😄
Well, I didn't get expected result from just putting a freezed bottle into narrow bowl with some water
Just trying to see if i need to calibrtae sensors by reading them with a reference 0C
doesn't really look full to me
Bought a desktop pc on govdeals to flip
From State of Maryland
The place I picked it up from was a social services office. The listing said HARD DRIVE HAS BEEN PULLED AND DATA ERASED. The IT guy I picked it up from didn't seem to know anything. I specifically asked him "hey has this been wiped? and he said "oh yeah I did that" and out of curiosity I asked him "did you do 7 passes?" and he didn't know what I meant so I explained the DoD hard drive wiping standard and he just didn't know? This is common knowledge in the IT/Sysadmin space, and especially a state level government IT guy should know about it.
Well anyways, I take it home, open it up, it has an NVMe installed. And it boots into Windows. And it gives me the funny "this is a government computer blah blah blah" warning. This MF didn't wipe shit
I am disappointed in government workers
not surprised honestly
all about pretending
so long as they tick the checkbox that says they did that
Poor gov worker 😢
The various US government(s) are allowed to pay so little to IT workers that only reason someone competent would be in one is if they absolutely wanted to.
And usually the bureaucracy will drive those out very fast.
Lowest bidder strikes again
Sounds like you are describing croatian government too regarding IT workers lol
Sometimes rigid payscales just don't work if they aren't updated to match the private sector.
And have leeway to adjust the spot some specific work is on without giving everyone with same code a huge raise.
And AFAIK in US, they can pay contractors more, but not the permanent staff...
As with contractors, it is competed contract between contracting companies, which then goes by just contracting regulations.
But then lot of other fuckups on that side.
We just remove the drives from our machines and either keep them for replacement or give them to an external company for wiping
And that was what the listing basically said, that you shouldn't expect the computer to have a hard drive...
But it wasn't done in reality.
IT person removed the drive from the computer and put it back in?
Maybe SSD that they tried to degauss with physical HDD erased?
Maybe they removed just the HDD and forgot the nvme
After all, if they're incompetent they might not know that we have drives directly mounted to the mainboard now
Or just opened case, didn't see old type drive and thought someone else had already removed it.
But basically with the salaries US federal or state agencies etc. can pay, they cannot get competent IT people.
after installing frozen sensors back into breadboard i realised that it's not that easy to work with short wires and tiny solder contacts
Same here in Germany
I'm actually amazed by the amount of competency in my team
Sounds like a recipe for some very silly mistakes
I am actually terrified of this. When you realize government hires unskilled IT workers because the competent ones are already working in the private sector, you pray there isn't going to be a huge data breach where sensitive citizen data is going to be stolen and abused
I'm surprised nobody has stolen citizen ID numbers and hope it will never happen because you can do a lot of bad things if you know someone's ID number
Although I personally witnessed a media company organizing a running marathon, collecting people's ID number (which is legal in this case) and their names/surnames
And then later accidentally publishing on the web confidential excel file with runners and their ID card numbers 💀
Took them 3 days to realize they published the wrong file on the internet
This happened before GDPR though if my memory serves me right
Just look at elektronische Patientenakte (ePA) in Germany.
Basically your doctor's file about you and your entire medical history will be digitalized - with a standard of IT security that has been illegal for credit cards for multiple years now
I guess we'll have our data leaked at some point by companies or government institutions (already happened to me)
Here, an employee of a government agency exported a database of all vehicle owners in Croatia on a usb drive and sold it to the third party (most likely insurance companies)
Although this has been a public secret for years because many car insurance companies would call you a month before your insurance expires in order to sell you insurance
And they could only know when your insurance expires because they bought a database of vehicle owners and their insurance information on the black market
If an insurance provider called me about an insurance running out that I had with them, that'd be normal for me.
If some third party called about that tho, creepy as shit and I'd try to find some court leverage
Fuck them big time with GDPR
Yeah, other insurance companies would call you and try to sell you insurance although they don't directly tell you they know your car insurance expiration date
They just pretend like they are simply selling insurance
Do you have some legislation against unsolicited ad calls?
Cuz I know the US has a list that you can input your number into and if you get a call anyway, you can report them and they'll get fined/prosecuted (iirc it's with the FCC and FBI)
Many people have reported them and asked how they got their phone number but they haven't been fined by the government yet
I bet they have people in the government who protect them from being fined
We have a do-not-call list but companies don't care about it
btw, did you watch Putins Bären?
Dope documentary, covering the Bundestag hack (among other things)
Linus Neumann said that the most surprising thing for him was that the hack has actually been noticed - which on one hand is scary af and on the other I wholeheartedly agree with.
I have reported a lot of companies to the government agency for sales call but got no response....
Does law enforcement care tho?
Nope 
You can go a step further to the EU
You mean European Court of Human Rights? I guess I could try that one day
There might be a special GDPR related institution that you can report to
If your government's GDPR official doesn't do shit
If there are worker unions in Croatia, you could try getting them on board
I know ver.di in Germany are very happy to go to court, even all the way to the highest level
Folks managed to get me 1.2k EUR in unpaid unemployment benefits with a bit of court shenanigans, basically for free
I mean they make them pay the pursuit cost right ?
Not in this case
?
There was an "out of court settlement"
Aka we had a first hearing in court, judge explained the legislation that applies or could apply. Asked if we wanted time to discuss a settlement.
And since the judge basically explained that I have a really strong case, the opposing party very quickly agreed to settling according to the legislation.
idk how paid my court/lawyer fees, but it was either the government or my Union.
Probably ended up as "every party pays their own fees, court cost for the first hearing is on the taxpayer"
So they didn't pay for settlement pressure cost ?
For what now?
I think there's hardly any legal basis for that here in Germany
my guess would be they did pay by their legal insurance
My union provides me with legal insurance for anything regarding social and labor law.
And they offer free legal consultation for rent law
Their wage negotiations are quite lackluster tho - but that's for all unions here, not just ver.di
so after all it turned out to be good idea
Just realized I can control tv and stb through desktop or Android 👀
I can also stream channels from the receiver directly to desktop pc or android
There are so many operating systems, plugins and addons I haven't even discovered yet. Owning a linux STB is awesome ❤️
I personally just watch youtube and liberated anime with gf
Or liberated 80's shows like quantum leap
HOOOWWW ITS LITERALLY BEEN 1 MONTH
i dont even believe it because it sits on 99% for suspiciously long
it literally only has 50 run hours
yea it stayed at 99% for 40 minutes before starting to lower in charge again
could sensor have different rate of self-heating?
battery?
afaik they have less accurate charge level readin at high and low
Depends entirely on the chemistry
For most lithium ion batteries, they can really only get a true measurement above around 90% and below 20%
yea but spending half of its charge at 99% then going down like normal for the second half is insanely odd
Anywhere that the curve is nearly flat you really have no way of knowing what the charge is other than estimating based on power consumed
i bet that the 24% wear is BS and that the 99% hold it does is because its over the 39K full charge it thinks it has usable even tho it actually holds more still
now i think melting point wasn't as usefull after all 😦
(included prior warming up and a bump when i mixed water a bit for reference)
but at least i think i can use this to make them closer to each other
And I did it again!
And different one:
Removed the images, just links this time to all three variants there seems to be:
www.redditstatic.com/youbrokeit1.png
www.redditstatic.com/youbrokeit2.png
www.redditstatic.com/youbrokeit3.png
What did you do?
yeah it's not just decisions being made behind closed doors though
something that code.talks could only dream of
i just bounce browsers until one works
i think doing the "full logout" on one of them is important
yea true, tho firefox has been working for me as far as the latest adblock stuff
oh yea if they manage to lock down the API successfully another big browser is going to take off
or maybe one that already exists gets bigger
it's definitely missing some features that they should steal from chrome
i'd love it if ff added in tab grouping
edge's tab groups is so nice for organizing stuff you wanna sort by subject
"enshittification"
take it away to charge for it later
once google figures out how to completely block adblock, then we'll see ff start to remove that stuff & then magically come out with a "premium" version that they'll charge for
make it crappier to make a paid version, because greed
yea the FBI & CIA say to use an adblocker, so i'm wondering if states or the feds can sue google for stopping its usage
de facto forcing unsafe traffic onto people's computers, might make them liable for the damages from it
i do like how google jumped right on LTT & nuked their vid on that
i mean, both are right, it's just fun to see the fight
eh, kinda
doesn't seem like it brought too much attention to it
The most frustrating thing for me is that they have a similar feature that just fails to be useful
Since I did my 100% switch to ff, I have started to have multiple windows open more frequently
Although to be honest I very rarely need more than 10 tabs
opening more windows is way worse to me compared to tab groups
having to float or click down to hunt down a section just sucks
I mostly alt tab to switch windows
Also @tough owl thanks for recommending Futo, ended up buying it
I'm pretty darn happy with it
And if you want auto updates use F-droid
Or obtanium
Did you know that the photo of the moon some time ago was from eclipse?
Google recommended me article "when to see it", from 2 days ago...
yea probably
i dont see them going away in what is practically 3 years
to be honest i think where back at a point where systems are mostly cpu bottlenecked again in most scenarios instead of gpu bottlenecked
No thanks.
With everyone pumping
code nowadays...

It's slow 
Because everyone loves their 3d mesh bloat with raytracing 
More gpu 
Wheres the 5090
Been running HX1200i for 6 years already. Satisfactory with watercooled 13900K and 4090 is about 600W load on the PSU. All my monitors and server around 900W total. UPS can handle 1,6kW so can run my 3D printer from it too while playing 😄
It is by any now
It's a GPU that was mediocre when it came out and you bought it over a year after its release
30 series GPUs were released in 2020
Tho tbf, chip shortage was crazy, GPU prices exploded
But yeah, I don't think you're gonna get another 4y out of that GPU unless you're using it for basic stuff
Oh, and another thing:
future-proofing in the sense of "spend more now, save money on future upgrades" is a myth.
a 2k PC now with a 1k upgrade later will give you more performance than a 3k PC now
https://cdn.discordapp.com/attachments/197095025157472256/1134967467064688660/futureproofing_remake.png
It's a copypasta, but the gist of it is simple: don't spend extra to upgrade less often. Spend less and upgrade more often
Nope got it for free with my 760
Steelseries rival if it was <70g, fuckton of super reachable buttons
Or a G502 if it had good sensor, switches and was <70g, might consider Basilisk V3 Pro
Right now on a Darmoshark M3, super dope, won't change for a while
Glorious Model O 😍
Bought this set from sale in late 2019. Still in use. Headset needed duct tape repair for head band, and a new wire (its not standard 3,5mm) . Liked the Model 0 so much that when getting new PC here at workshop in 2021, got another one, using it right now.
My DIY head tracking led bracket mounted on the headset. Just simple lasercut and bent steel piece. So much robust than 3D printed ones. If that falls on floor it just bends, and fix by rebend by hand.
Meh, QC issues and wonky firmware
I have no issues whatsoever with the 2 model 0s, just map button for VoiP tangent in software and its been 5 years like that 😄
After the last update my game no longer updates and does not load
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!
I love windows
The clocks on the task bar differ by 45 minutes between two monitors. The primary display one is lagging behind behind
I call it wandows for a reason 
intresting combo ngl
use petg to print it
G502 USER SPOTTED MY MAN
I will forever circlejerk the g502
G305 my beloved
instead of just "more secure connection" I realise now that i need something what will allow washing of sensors. and it seems like detaching them individually is a way to go (instead of having single star-cable with resistor in it). So, need to find solution for that (since re-attaching tiny solder wires doesn't feel good, and can break simply from weight and stiffness of sensor's cables)
But for real, why does the G502 have a non reprogrammable button just for switching scroll wheel from smooth to ratchet? What's the use case where you want that on demand?
I'm just curious why it's so important that they would place it where they did, though. I feel like that should be a toggle on the bottom.
g502 is shitty but I did find it useful for websites mainly and fortnite having it unlocked while having pickup binded to the up on the scroll wheel got me loot the loot every time
anything where you need to pick something up and you can rebind it to scroll wheel is super helpful
imo not worth to switch to a shitty mouse
you're insane lol... 8 years x:
i have my 6600XT sitting around doing nothing
kinda tempted to build another system since i'm just missing mobo/psu/case
I ran a 750ti until 2021
7 years
Googles bs ai stupidfest said that the 750ti came put in 2012 and was the flagship wtf
That is correct, I mean, if the internet says it it must be true!
@inner shale you were the one with the twitch problems right?
After upgrading from my 5700 xt and 5500 to 3070 and 5700x3d twitch hasn't fucked my firefox
twitch might just be taking up a lot of cpu on firefox for some reason
Hw accel not working?
What gpu should i pair with a ryzen 5 5600x
And what gpu should I pair with a 12400f
whatever you want
no but truthfully there are many variables at play here such as the game types, screen resolution, vr or not and so on because that varies the ratio of load between the cpu and gpu
Its for games like minecraft, satisfactory, astroneer, farming sim
Basically recourse hungry games
You could say
most of thoes are mostly cpu heavy games so you wouldent have to go crazy on a GPU because it wouldent matter after a certain amount of progression anyway
Any recommendations
not really because i dont remember the specs of thoes cpu's off the top of my head but im saying that it could vary depending on what you value or do in these games
Definitely would be playing them for a long time
Thing is, someone recommended me a 1080 ti paired with a 12400f
thats a ok choice mostly from a Vram perspective
but like if you care about something like global ilumination or ai optimising such as dlss RTX 4000 would be great last i recall FSR or whatever isent as good but nothing stoping you from going for a amd equivilant
Thing is im also on a budget
In poland so my max would be 3k zł
Like around 720 usd
you could probably get 2 decent AMD cards at the sacrifice of Ray tracing and slightly better ai optimising but they would work fine
wich im sure everyone else has a reccomendation for amd gpu's whenever they decide to pop in next
For now im only planning what parts, so there isnt any rush
What's your budget?
if you're not doing top end, there's 0 reason to do ray tracing. anything that's not in the top 3/4 GPUs just still can't handle it well enough to justify the expense
1080Ti is nearly a decade old, not worth recommending
cheaper & less power hungry options for the same or better performance
i had decent luck with my 5600X/6600XT combo
Yeah, that makes sense as a CPU/GPU combo
was amazing for 1080p
& the 6600XT is on par with the 1080Ti, but like $200 now & 150W
$687 and a pretty decent 1080p gaming rig
Part List - AMD Ryzen 5 5600X, Radeon RX 6650 XT, Fractal Design Pop Air ATX Mid Tower
pop air is very cheap waow
$48 for 32gb of ram also feels cheap
12600KF is <$130, also consider Arc for GPU
P3 Plus is QLC, wouldn't use it as a system drive, especially at 500GB
PSU directly competes with C750, also A650BN
Cras C910 and MP44L exist at a higher quality and lower price
plus imo 500gb is a waste of an m.2 slot, 1TB C910 is $56
its ddr4 and super cheap
im out of touch with drive prices nowadays since i bought all of mine when they were super cheap (1tb tlc pci 4.0 for less than $70)
are they still super expensive?
i mean yea most 2TB M.2's last i checked where like 200-400 lol
went with 6650XT to keep complications to a minimum, arc is still kinda in that "experimental" phase
5600X: $123 with cooler
12600KF: $157, no cooler
my pc crashes and goes offline everytime i try to run satisfactory can somebody help me please
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!
arc is awesome if you want to transcode video
No, they're not. Midrange drives like the ones listed above (TLC IG5220/MAP1602) are sub-$60 now.
2TB sits around $100 for a good drive, you often even find high-end E27T+Bics6 configs at that price (rn: MSI M482, Patriot VP4300 Lite)
Dang, deal ran out then. Quite a shame really.
Anyways:
5600 non-X is $115
12400F is $112, better multicore too
And I wouldn't call Arc experimental any longer.
This is the most recent video I have (from a source I trust) and >90% of games work perfectly fine, the rest usually with minor issues.
And I've talked to other Arc owners who said that games flagged red in the video worked perfectly fine for them too
https://youtu.be/Y09iNxx5nFE
Gigabyte Aorus X16: https://www.aorus.com/en-au/laptops/AORUS-16X--2024/Key-Features
Support us on Patreon: https://www.patreon.com/hardwareunboxed
Join us on Floatplane: https://www.floatplane.com/channel/HardwareUnboxed
Buy relevant products from Amazon, Newegg and others below:
Intel Arc A770 16GB - https://geni.us/FLiCtI
Radeon RX 7600 - h...
And then Arc value is super nice cuz you get XeSS which works better than FSR (especially for Satisfactory) and Quicksync video encoder which rivals Nvenc (tho you also get that with any non-F Intel CPU)
@crystal hound also, correction, 159 base game plugins now in 1.0.0.2.
IIRC 1.0.0.0 had 147.
Couldn't attach on #satisfactory so using wrong channel.
List as attachment:
edited; I don't think people will go through that list, but thanks nonetheless 🙂
Doesn't really matter, but as correction to the number.
Leaving consumers with either Intel or server stuff or both
1080Ti is still a very capable GPU. However, as said, it is starting to get old. You can, roughly, compare it with a 3060.
Yay!
Zen 3 still kicking ass i see
i wrote code to convert numbers in any base to any other base
https://github.com/WerIstLuka/int
The python code does not look malicious, but I wouldnt trust a random .deb file
Also does not follow pep8 💀
idk why make your own base converter when you can just do
int(str(myInt), base)
I had to just today create two exceptions in antivirus.
For ZenTimings, generic detections from it doing stuff direct with HW to get the info.
Set my RAM back to almost minimals to see if I can run the ultimate RAM stability tester properly or not.
The name of which is not allowed on this channel.
why would anyone make .deb for THIS?
I don't know.
im writing an assembler and i needed a random project to figure out how to do debian packages and i needed a program to convert between bases
so i can do things right with my assembler
usually i have to convert a lot of numbers at once so its easier to write this once than to always open a python console and do that
just an example, sometimes its even more so i need it to be easy to convert them
you can build it yourself if you dont trust it, the process is descibed in the readme
did you know about Calculator?
calculator in linux is 💩 tho
what do you mean?
what's that got to do with drives?
there's no program in the world that i hate more than windows calculator
and linux one is just the best out there
Sorry I have to disagree
it can convert bases
Other people ralking about 5600
but not from like 100+ numbers at once
Linux doesnt have a calculator?
ok, fair.
Ubuntu (Gnome?)
Thats just whatever comes with ur de or installed separately
Default ubuntu is gnome i think
with my program i can do
int 0 1 2 3 4 5 6 7 8 9 a b c d e f 10 11 12 13 14 15 16 17 18 19 1a 1b 1c 1d 1e 1f -x
and then i get it all in decimal
I use endeavouros on kde
it does but the ui is crap
gnome calulator feels good on mobile tho
but on pc its awful
The only thing windows calc does better than linux is bitshifting and the binary number togglebox.
it's simpel and it works. also you can type multiple operations which is like the best feature ever
The probability of getting 2 heads back to back if you flip a coin 3 times is 25%
I think that's incorrect
Depends how you calculate.
I think its wrong too
Also, is B intersection (A union B) possible and if so what is it?
My brain is short circuiting trying to calculate that and my textbook doesn't have any examples
That would be B.
time.
So intersection cancels the union and the other value when 2 values are the same?
Or is it another meaning entirely that just happens to have an outcome that can be explained like that
does three heads count as getting two B2B?
if H-H-H counts you get three ways of getting B2B Heads:
H-H-H
T-H-H
H-H-T
Each option has a probability of (1/2)³=(1/8)=12.5%
So in total, you have 37.5% probability
-# assuming each outcome is equally likely
However, if we exclude triple Heads, we get the 25% you mentioned earlier
You can just make a basic binary probability tree:
HHH
T
TH
T
THH
T
TH
T
each option being equally likely, you get eight total outcomes - only two (maybe three depending on HHH) of which are B2B heads.
2/8 = 0.25
That's not even how I calculated it
I did it within the bounds of my stat textbook which was easier than that
P(H) = 0.5. HH = 0.5 * 0.5 = 0.25
That's not accounting for the "out of three throws" portion tho, that's only for two throws
Cuz if we do count HHH as B2B heads, you get 37.5% as elaborated above
Fair
@twin dew how do I set a dns server for the router?
so all devices are on the same dns server?
(er605 btw)
Please help me one more time. I'm trying to calculate PMF for Bernoulli distributions, but I just can't figure out how you're supposed to apply this formula, even with examples in the book. The example included in this screenshot is what felt like the most useful one.
(I'm asking about the P(X = x) formula. I've tried watching videos but they're also confusing.)
I have no idea how I'm supposed to apply this formula/plug in values, and where do the factorials magically appear from? I tried applying the pattern I saw in a problem (P(X=3), X ~ B(10, 0.12)) but according to the book nothing I've tried has yielded the right answer.
In the Network, WAN, WAN.
Optional Primary DNS and Secondary DNS fields.
The factorials come from "n on x" (the stuff in brackets).
This term tells you the number of possibilities you can get the result X in N tries.
Meaning: there are "4 on 7" possible ways for you to get 4 heads from 7 coin flips.
Then you take that number and multiply it with the probabibility of heads to the fourth power (cuz four heads) AND you multiply it with the probibility of tails for the remaining number of tries (in this case, 7-4=3)
Binomial distributions do not care about the order in which things happen, they just count how often they did or didn't happen.
So for any single one possible way of getting 4 heads in 7 flips, for example HHHHTTT, you need to multiply the probabilities of each.
Meaning: p(H)⁴×p(T)³
If we compare this to the formula, we already got p^x and (1-p)^(n-x).
n being the amount of total attempts (7) and x being the amount of desired outcomes (4).
The only thing left is counting how often this can occur in the binary tree - which is where the (nx) bracket thingy and the factorials come into play.
There are n! possible permutations to arrange n variables. From this you need to remove all the undesired outcomes (for example no heads or five heads) which is the rest of the factorial formula
@dire igloo I want to strangle whoever wrote my book. Every example they used happened to end up with like terms multiplying each other (causing exponents to add thus removing one like term), thus making the formula in application look different from what I was actually supposed to use
What you wrote helped wake me up to the fact that it's just a normal formula with weird looking formatting
It's a crux really - either you get some complicated mess or you're oversimplifying.
Hence why I used 4 and 7, easy and low level enough to grasp, but distinctly different
Coin flips are also a bad example imo cuz p(T)=p(H)
Unlike let's say rolling a 6 on a fair six sided die
Happy to have helped
Are you a CS/math major
B.Sc. in CompSci, yeah
But most of what I did here is just me being passionate not only about the topics themselves but also sharing them with others
I find that most people who dislike maths (or science classes in general) mainly do so because of shitty teachers
I typically have okay teachers, though current stat teacher is so ass. I also just really struggle with learning math because I'm not interested in it really at all
I'd rather have the computer do it for me
Ehhh, quite a misconception.
Computers do calculations.
Maths isn't calculations, it's understanding the systems
Kinda like saying "I don't need to learn Japanese, I can use Google translate"
Maths and science are the universal language we use to describe and understand the world we live in
A good way to think about it:
Maths is knowing what to tell the computer to do
Yeah I get it, but chances are in a sysadmin job I don't need to know how to do calculus optimization problems or most of stat
Sure I MIGHT need to use that stuff, but all I need to know is what formula(s) the computer should use, and what numbers to plug in. And/or what to ask the search engine
That's maths right there.
You not only need to know the formulas, you need to understand how the problem at hand works and what the fundamental inner workings are.
Without that, you can't know what formulas are correct and get you the right result.
You also need to be able to verify the computer's output - if it gives you a 130% probability, the only thing you can be certain of is that you made a mistake
Optimization problems are a key topic in CompSci, just look up big O notation.
Knowing how your solution scales with higher demand is absolutely detrimental to providing good service quality.
The actual maths isn't too important, but you really really need to understand how maths itself works.
As a sysadmin I find it unlikely I would need to calculate big O
I'm not a programmer. Maybe the odd script or two but most scripts I'd have to write would be for deployments which don't have much meaningful time complexity anyways

For example, I had a physics class in first semester.
The only purpose it served was a reality check for us - there was zero relevance to the rest of the curriculum.
The reason we had that physics class was to put us into a situation where we needed to apply mathematical thinking processes outside of what we were used to dealing with.
Anyone fit for computer science could've easily passed that class just by applying those core problem solving patterns.
And we had one guy who failed that class - along with many others. Having this class in first semester saved him months of struggle with a major that he wouldn't be able to pass
It's pretty good to know about either way.
To your point - I CAN do it. I just really don't want to do math if it's avoidable (ironic because I play Satisfactory)
Calculate? No.
Know how code and system complexity affect performance? Absolutely.
You don't learn these things to use exactly this. You learn them to get the fundamentals down to a level that is crucial for understanding your environment
Again:
Do you dislike maths or do you dislike calculations?
These are two distinctly different things
I dislike calculations
I do not want to sit down and relearn u-substituion for differentiating/integreation for example
Okay, so your issue isn't with maths
Maths is recognizing the patterns and putting it into standardized notation in order to tell you what you want to know.
Calculating is continuing that notation to the point where you have a number as your answer
Hacking calculations into a machine is tedious legwork, I'm not surprised that anyone dislikes that.
However, being a good mathematician gets you 90% of the way there without too much effort - because you recognize the patterns and are able to precisely formulate them
I'd say being a good programmer is a similar ordeal:
Knowing any specific language isn't the important part. Knowing how to write good code however, is.
Designing the program's functionality is 90% of the work, the remaining 10% are the annoying legwork of making it work in the language - tho sometimes these two overlap when your implementation attempt doesn't work in the language
efficient code is better than simply functional code
Depends.
I prefer function over efficiency, to a degree. But I always try to keep efficiency in mind either way, but it's not my main focus. 95% of the time it's fast enough.
yea same but thats mostly because im dumb and my code barely works to begin with so optimising it isent generally a huge priority and plus a dont code often at all...
Another factor for efficiency is maintainability.
Going into the code a couple months or even years into the future shouldn't take you ages just for understanding it
simple question i have wondered for a bit since laptops generally only have 2 ram slots... is having 2 ram sticks worst than 1 or does it just not mater because XMP and speeds arent a factor
okay thanks, is that ipv6 or 4 or either one?
That one is IPv4, there should be same thing in the IPv6 settings for IPv6 address DNS probably, didn't think to check the manual for that.
Here we go again!
- Unauthenticated RCE vs all GNU/Linux systems (plus others) disclosed 3 weeks ago.
- Canonical, RedHat and others have confirmed the severity, a 9.9, check screenshot.
https://x.com/evilsocket/status/1838169889330135132
okay thanks it worked
Most laptops don't even support xmp, they just run JEDEC without even giving you the option to enable XMP
The second stick is about getting the second RAM channel into use.
With just single stick you only have half the RAM bandwidth compared to two (when on different channels).
You always want to have as many sticks as channels.
Well in general as much bit-width in total in the sticks as the CPU IMC has.
Now that DDR5 has 2 32bit channels on one stick.
And LPDDR5 has 16 bit channels.
Or is that only for LPDDR5X, not sure.
On a completely new topic - When you look at the die for, say, an i5-12500, is the top part that you look at actually the die itself with all of the logical circuitry, or is it the top of some form of package?
All the logic is at the bottom
The top isn't really a package, just silicon with not much going on
Must be done that way at least to an extent so that you can mount things like an IHS to it
But it also has some downsides, since all the heat is produced at the bottom of that die, which has fairly low thermal conductivity
So that creates a limiting factor for substrate temperatures
So then would that mean dies are manufactured upside-down? (Meaning pads facing up after fabrication)
I have a load more questions but they start getting into the minutae of the fabrication process which I doubt is widely available info
Uh
Yeah, basically
It also depends a lot on which specific lithography process you're using
Let's take 20nm for example then, that's a mature process. So the dies are manufactured center->down? Like how does that work? The structures of the actual die circuitry are 3-dimensional right? How does the photomask make that possible?
Also as I'm sure you can tell, I actually know very very little about this
One day I will have the time to sit down and read about these topics in-depth
Flip chip.
Start with clean silicon.
Start depositing stuff on top.
Once everything is done, turn it upside down and attach to substrate.
Before flip chip, that was done with tiny wires on outside.
atomic deposition and eching turn in turn.
Oh wait I think you misinterpreted my question and I misinterpreted your answer. Flip chip is the process for attaching the die to the processor's (interconnect? adapter? it's not a PCB I'm pretty sure. You know, the part that actually connects to the mainboard).
I was asking about whether or not the top part of the shiny die is a glorified heatsink that's attached to the die later, or if the die is created "upside-down" then attached to the aforementioned green thing that connects to the mainboard
VSOC failed, the Asus VRM blasted AM5 CPU sectioned from side:
The black is the silicon.
The tiny silver layer between it and the solder bumps is the active area.
Means any case where the active "silicon" is at the bottom, and not the top like classically.
So you can do direct solder bumps and not wires.
Oh wow why is so much of the space just silicon mass?
Less than 1mm thick.
Or is that empty space? Hard for me to tell
That black.
Bulk silicon.
In cases you can thin that down to 0.3mm if you are crazy.
Closeup
Bulk silicon in white.
Half the scale of previous
The IOD active is those few small layers between the bulk white/grey and the solder bumps.
X3D with the extra second die
Considering you've never seemed to get tired of my random hardware questions, is it safe to say you enjoy these topics?
Sometimes yes, sometimes no, mainly sometimes get annoying to find nice pics you know you saw before.
I know that feeling
@glossy glacier you've been typing for a while you good?
Very Late but I wanted to give my two cents
For normal consumer CPUs its better to have (at least) two sticks of RAM because then the memory controller can run in dual channel mode.
You can imagine (pre DDR5) Channels as huge one-way (as in only one vehicle in a single direction at once) roads to and from the RAM.
Dual Channel upgrades this to two one-way roads! So double the bandwidth of single channel!
DDR5 changes this a bit by dividing each prrviously 64bit wide Channel into two smaller 32bit(+ECC) channels.
Again taking the image of the road, the total width of it says the same but it's divided into two one-way lanes.
Workstation and Server CPUs can have more memory channels, thus more data lanes.
Most non-laptop devices also allow you to use more than one memory module (DIMM) per channel. This does not increase the channel/lane count. It's only to increase the total amount of memory. The DIMMs share the same channel. More than one DIMM per channel comes with its own challanges but that is a topic for another day.
Peachy 🙂
Also GDDR (graphics VRAM) is not the same as DDR (system RAM). VRAM is designed for more parallel access than normal RAM
Yes
Workstations are quad channel, server CPU's are up to 8 channel I believe
more
12 at this time.
Sapphire Rapids is only "8", as in 8x2x32bit.
Per socket.
AMD is at 12 per socket.
Emerald Rapids too is just 8.
so that makes easy 24 dual socket servers
Granite Rapids should get to 12 per socket.
which top at 3TB RAM
For full XCC with 3 tiles.
Everything is "up-to"
There are dual channel Intel Xeon "server" CPUs and dual channel Intel Xeon "Workstation" CPUs.
12channel/CPU is the current max in x86 land iirc.
Don't know what IBM is up to or what ARM-land is useing
Just rebranded consumer CPUs for the 2-channels "Xeon"s.
Exact same silicon as consumer ones.
Same with the "Epyc" 4004, just AM5 consumer CPUs rebranded with more testing, validation and guarantees.
Epyc is better than Xeon rn change my mind
The only practical use case I can think of for Xeon is Intel-optimized applications/databases
But if you're doing any kind of production virtualization, especially if you're virtualizing employee computers/hosting an rds farm server, why wouldn't you go with Epyc outside of bias?
I'm also talking about real epyc and xeon, not what you're talking about above
Nobody is arguing against that
Sad to see the state Intel is in, great to see the state AMD is in
I like having play in my duopolies
It is mostly bias
Procurement guys are notorious for picking favorites
And to some extent, that does make it easier
My company doesn't have procurement lol, we have CTO and 2 sysadmins and that's it that's the IT team
I get to make some quotes for a new VM server and I'm excited
Our servers are from win8 era so with the budget I was given, the server we end up going with will more than likely be powerful enough to replace our 3 VM servers and still be faster
I was told to stick to only Dell and HP which sucks, because Supermicro has good options or even just custom building it as there are great non-proprietary hardware options but still fun. Right now I have about $2500 of breathing room left in the budget given the first quote I came up with, so now I'm trying to decide if I want to put more of that into SAS HDD's to make a new RAID5 array that ISN'T MADE OF 15 YEAR OLD HDD'S, or if I want to put more of it into CPU performance for further futureproofing/scalability
That's not a joke either, our HDD's are 10-15 years old depending on what array you're looking at. CTO never replaces them because the Dell/HP health monitors report that they are "fine" (except the hardware RAID controllers don't have any way to query them for SMART info of individual drives, and I can't just pull a drive out of a prod server during a workday to manually check it SMART attribs for obvious reasons)
So old controllers that they don't support SMART lookups even with the card manufacturers software?
First ensure the safety and integrity of the data, then worry about future Scalability
i mostly meant is the data overload from using too many lanes from 4 channels (desktop) in this case 2 still applied
i was mostly wondering if 2 ram sticks in a laptop was equivilant to running 4 sticks in a desktop in the sense its being limited or if it just acts like normal dual channel
and i mentioned i assumed it probably dident matter because laptops dont have XMP and therefore it shouldent run fast enough to be a problem
I'm pretty sure 2 sticks isn't really a difference*
*on most laptops
Their software definitely does smart lookups, buuuuut it doesn't show any attribs. Only healthy or not
It is
Especially since most laptops run JEDEC or at most 3200, the extra memory channel does provide noticibly better performance
IIRC Ryzen also really likes dual-channel, at least older Ryzen did
Fair, I just haven't even thought about single-channel anything because why the heck would anyone build a PC with just one stick of ram
Zen as a whole is really really hungry for both memory bandwidth and latency
Even single or dual rank sticks made a big difference on Ryzen laptops at least through Zen 3.
And that's its biggest limitation for a lot of workloads
Linus had a video expressing shock that the one AMD Advantage laptop with a 6800M shipped with badly performing RAM.
2 channel on laptop is the same as 4 on desktop because the second DIMM per channel only increases memory capacity, not throughput/latency
also theres no such thing as "data overload" with memory. either your memory is fast enough or not
and to compensate a missing channel the other would have to run more than twice as fast
It's only like that in some laptops
?
Almost all laptop CPUs support 128 bit memory width.
One SODIMM is 64 bit wide.
Some specific laptops are stupid and have only slots on one channel, not both.
And some very low end CPUs are limited by segmentation to only expose one 64bit channel.
With the stupid laptop designs using same motherboard for the segmentation limited CPUs and those that would have the second channel, to save costs, but nerfing performance for the higher CPU models.
As all the laptop CPUs supporting 2 slots per channel, for up to 4 slots in even laptops.
Just that Intel does very heavy segmentation on how much RAM the laptop CPUs support in total.
And AMD did mistakes in past where the lowest end non-Ryzen branded CPU models only had 1 channel.
So even the Ryzen variants then got that bad 1-channel laptop MB.
When the Ryzens had 2 channels.
Two (SO)DIMMs don't make the channel 128bit, it's 2x64bit.
Thought we were past the single channel days, but yeah, apparently lowest end chips do really support only single channel, N100 for example.
That 128 bit wasn't in one channel.
Two 64 bit channels for DDR4 and before.
Four 32bit channels for DDR5.
Eight 16 bit channels for LPDDR5.
And one (SO)DIMM is 64 bits wide.
With 1 channel in past, and 2 with DDR5 on it.
Reason we need to switch to speaking about the widths, not channels.
As we aren't in "one channel is 64bit and equals one stick" anymore with DDR5.
And the CAMM2 will shake it up even more.
As that can be 64 bit or 128 bit wide per module.
Channel width on amd64/x86-64 is 64bits. You can't go larger because the registers for addressing/reading/writing memory are 64bit wide
Not related?
And in smaller channels you need multiple clock cycles to send the data
And like I said, DDR5 uses 32bit channels.
LPDDR4 already used that too, and LPDDR5 went to 16 bit ones.
But that is what the memory controller deals with.
Converting the 64 bit accesses to whatever the memory is.
Remember that you cannot even with DDR4 get just 64 bits from the memory, you get much more in each request.
You can it's just not very effective usage of ram
No, at least with DDR5 you get the requested address and then 15 more after it as minimum.
If I remember the amount correctly.
Either 8 or 16.
Doubled from DDR4.
You can then get more of those chunks with simple request, but you cannot get less.
IIRC it was 8 64 bit chunks in DDR4 and now 16 32bit chunks in DDR5.
For each read request.
And each DDR5 sticks has those two completely independent 32bit channels.
So on "dual-channel" DDR5, you actually can do 4 different memory accesses at same time, each on 32bit actual channel.
That is what defines a channel, being independently usable by the memory controller.
DDR5 IMC can use two 32 bit channels just as emulated single 64 bit wide one if it wants to.
As each 32 bit DDR5 channel has its own address, data and side signal pins.
But it doesn't have to.
But DDR4 IMC cannot split the one 64bit channel into two separate 32bit ones.
- Longer Burst Length
The fifth major change is burst length. DDR4 burst chop length is four and burst length is eight. For DDR5, burst chop and burst length will be extended to eight and sixteen to increase burst payload. Burst length of sixteen (BL16), allows a single burst to access 64 Bytes of data, which is the typical CPU cache line size. It can do this using only one of the two independent channels.
- DDR5 vs DDR4 Channel Architecture
Another major change with DDR5, number four on our list, is a new DIMM channel architecture. DDR4 DIMMs have a 72-bit bus, comprised of 64 data bits plus eight ECC bits. With DDR5, each DIMM will have two channels. Each of these channels will be 40-bits wide: 32 data bits with eight ECC bits. While the data width is the same (64-bits total) having two smaller independent channels improves memory access efficiency. So not only do you get the benefit of the speed bump with DDR5, the benefit of that higher MT/s is amplified by greater efficiency.
So that longer Burst is the normal one, and that Burst Chop is the minimum you can do.
This is for the RAM > IMC part.
Of course the IMC can then discard the extra data it got and not deliver it to the core.
Yeah, seems to depend on the burst length and internal width
DDR4 has a burst of 8, with a internal width of 4, 8 or 16. So you get either 32, 64 or 128 bits per read
Bytes for the latter.
64 bits multiplied by the burst length.
For DDR4.
And 32 bits multiplied on DDR5.
Bus width * burst length.
That bus width between RAM and IMC per (sub)channel.
DDR4 doing two normal back-to-back reads, getting 16 chunks of data and then closing the access.
In real use all those NOPs would be addressing other banks of RAM in same channel.
To keep that DQ having data moving all the time.
(not ranks, ranks just add more bank groups)
And reason why DDR5 doesn't get significantly more performance from second rank, like DDR4 does, is because in DDR5, each rank usually has double the bank groups.
So one rank of DDR5 x8 chips has same amount of bank parallelity possible as two ranks of DDR4 x8 chips.
Without that 64 -> 2x32 extra parallelity.
I wonder if someone already posted this issue that arrises when exiting game in Linux...
It appears everytime, after clicking OK, we need to manually stop the game in Steam, because it sees it as still playing
I would post it on their website, but probably someone already did, and searching for already posted issues is .. not too flawless
I need help.
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!
This my Satisfactori
Bro
Wrong place. Set upscaling method to none.
ok and how to do it
Options, Video, Upscaling Method?
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!
Ok
thank you
Is there a safe way to heat up plastic bottle at home?
Preferably with known energy consumption
Uh
That is incredibly specific
And yet it lacks all of the information I need
Please elaborate
Do you know what i am doing? 🙂
Like, i don't want the plastic to deform like i did once on stove...
Goal is to benchmark PCM
Does the bottle have water in it?
If you're using a water based pcm you could try microwave
Yes, it has water, and some salts (sodium sulfate, table)
Don't have that tool
Ye
Second best bet would be a double boiler
I have a pressure cooker though. But its a no go
? Which is?..
Suspend your sample in boiling water
Boiling water is a constant temperature, you can get a passable estimate of energy
Oh.
What about it is double i am curious
Aka without continuously heating it?
I had a problem once
I was heating contents to 'melt' thickener inside
Bottom of the bottle deformed a lot
But that was on gas stove with small fire setting...
So my best idea was playing around placing into boiling water or inside double removing direct contact and measuring time on stove as it is constant heat, even if its unknown
Just that you have container with substance b immersed in container a full of another substance boiling
That's the idea, you can continuously apply as much energy as you'd like and the temp is a guaranteed 100c as long as you keep it boiling
Like Fight or Flight says:
Normal pot, water in it, then the plastic bottle in the water.
But some plastics already start to soften above 60C.
So you need to see about the specific bottle.
But direct heat will be way higher than what you can get with water bath.
You can pretty easily do less than 50C water baths with thermometer and switching the heat on and off if the minimum setting is too high.
Hello! Can I ask for some help picking some new parts for my pc? Im sure people ask this often but im drowning on an empy glass deciding here
Sure
First a few questions for you
What's your budget, what do you want to play, are you willing to buy used, and where are you located?
And if you're upgrading, what are your current specs?
Thanks, I was going to adress all that luckily.
Im on 400usd budget.
For games is mostly Path of Exile, Satisfactory, Dyson Sphere and multiplayer games that dont really make a difference.
I live in Argentina so prices may vary a lot compared to the rest of the world (for example a ryzen 7 7800x3d its like 900usd here lol)
My current pc is a i5 9400f, rxt 2060 and 16gb 2666mz ram so the only thing usable is the gpu honestly
I dont recall my mobo rn but its 9th gen so is no use anyway
For intel im into a i5-12600kf or a i5-13400f with mobo and ram. The part im lost is ryzen, they are a bit pricier here and dont really know if they are all beter than intel ones
Also I apologize if I misstype too much words, English is not my native lenguage hehe
I'd say wait and save some more.
You current system is not terrible and should stay useable for some time
400USD is unfortunately not enough money to build a PC worth building unless you install used parts (namely a used gpu which is risky)
Your current PC is fine. If you really really want an upgrade now, switch to 32GB of DDR4-3200/3600. Maybe upgrade GPU but you're limited on options. Probably your most economic one would be a 6650XT or a 6750XT
gotta keep saving then. Thanks!
Either way, whitch one would be worth keeping an eye on? In case some sales come up hehe
12600kf is better than 13400f
at this point my opinion is that a 12100F is better than a 14900KS
mainly because it doesn't melt itself
Arc GPU
neither does 14900KS
not after the fixes
i'll wait 2 years to render a verdict on that
🤣
and 2026
oh i'm on a 5600X, i'm good
my hopes are low since it's just regular bottle from soda
hmmm, should i get some rotors to control gas output?..
unless there's some active step i need to take i din't know of, rn it has 240MB free of 537MB /boot partition, and apparently that's not enough to upgrade...
i thought you upgraded
nah, not yet
I also have 5600X and it's serving me well so far 👍
They're fine in 2024
Depends on what CPU you have though
Definitely won't be doing heavy gaming on it, but you'd still be able to at least run most
Tolerable.
This was about 12100F which isn't a worthwhile buy against 12400F and used AM4
Oh yeah fair
I did put a 12100 in someone's custom build like 1.5 months ago
But, guy was on a little budget AND it was just a regular home PC, no gaming at all. On top of that I refuse to give someone a pentium/celeron
It was also still insanely fast compared to what we upgraded from
i was joking about 13/14th gen melting itself when i made that comment
12100F isn't a good cpu, but at least you don't have to worry about it overvolting itself into oblivion
Lying.
The PCB does have vias...
As all the traces would be on the opposite side of the PCB otherwise.
Prime usecase for the used market imo
Also Celeron/Pentium is same shit as i3/i5/i7/i9:
It just doesn't matter across generations
Any LGA1700 Celeron or Pentium will shit on any Nehalem i7
I can't warranty used parts
That's the price you pay for paying a lower price
5 day boot time just for it to probably crash and die in a few minutes 😂
thats still impresive tho lol
Build quality probably on par with other mainstream gamurrrr keebs
Oh, the switches are actually interesting. It's based on magnets.
aka rather mediocre
It's called hall effect and there are dozens of those on the market
How do you know, it's a brand new model 😂
Wooting having the best implementation rn
yeah, mentioned in article
Steelseries
well, that wooting also uses em, not that' wooting is best. Not mentioned.
The exact same models of switches? Or just that technology?
tech
Unsurprising
Anyways, I am 99% certain that that keyboard ain't gonna be worth its money
Guess how much the advice price is 
Plus, there's hardly any reason to go HE in the first place
I'm gonna say 250€
200€ for tkl
260 euro, 240 for tkl, 290 for wireless TKL.
Unless you're a comp athlete
Dayum, I'm good
Yeah, just interesting tech.
Steelseries ain't tho
Can't think of a single piece of tech of theirs that I'd be fine recommending (except stuff you get for free)
I own one product of steelseries and I am very happy with it.
Quite the opposite in fact, I hold a deep grudge against their headsets.
Had to help too many people who got fucked because of that chat vs game audio mixing bs
We've already discussed why that's meaningless in the grand scheme of things when we talked about monitors
It's their Arctic Wireless headset. Comfy. The pads last week died, I can order new pads and just replace the broken ones. You could say: "bad quality" but this headset is 8 years by now. Maybe 7? And instead of buying a new one, I can save a lot of money by just buying the pads itself and click em on. It's a nice detail.
Unless you're knowledgeable enough to be able to tell good from bad (both in terms of knowing enough to have seen good and bad and in terms of being able to spot the difference), I'll treat it like any other comment I'd see on Reddit, YouTube or wherever
Yeah, I know you don't give a shit about personal experiences. But guess what. That's how people work 😉
Two questions:
- What are you comparing it to?
- What's the price?
Comparing it to every other headset I've had and that died instantly and hurt my head. The price I can't remember since it's long ago.
Which is an ironic thing to say really. >90% of people say that they like apple pie, doesn't make it a good diet
Doesn't mean those people are wrong either, it's called an opinion and personal taste 😄
Popularity is not a metric by which anyone should ever judge anything's quality
The price of the headset is 199 atm. But I don't think I paid that much back then.
Opinion and taste don't have a place in a discussion about facts tho
Which is very problematic as source of what is good or bad, when most people have so limited experiences for such products.
That it was better than the other thing(s) the specific person had ever owned isn't often that high mark to pass.
"hurt my head" - fit is subjective, not suited to judge quality
"Every other headset I've had" - that's literally what I'm asking you about
Hey you don't have to like my headset 🤷♂️ That's why I bought it and not you.
headsets 🤮
I am not trying to convince anyone in here, I just said, that I own one product of Steelseries and I am happy with it.
And fireworker said: steelseries always suck.
Not being bad is the bare minimum a product has to achieve.
Which isn't mutually exclusive
Cuz your statement is solely based on personal opinion, mine at least tries to account for as much objective fact as possible
Do you take into account that heads are different? That some headsets hurt people? Do you take into account that you can replace parts of the headset?
Or do you only care for audio quality vs price?
And what about how long the battery lasts?
So about as much as an Epos Game One, HD560S or DT880
tech-talk-circle-jerks are back!
Tech-talk ... never changes. 😄
You cannot factor ergonomics into the equation in a generalized discussion about a product because you can't know the other person's ergonomic requirements.
Which is why adjustability is so utterly important
And part replacements are a tough topic because you also need to factor the replacement cost, the replacement period and the availability of replacements even for the future.
So overall, I tend to ignore it unless I can account for all of these
Anyways, the two main things that are important with a headset are the headphones audio quality and the price.
Every other metric either just filters into one of the two, is too subjective to be a good factor or has too many strings attached to be reliably applied
let me fix that, for you*
I don't care about audio quality, I don't listen to music with my headset. I have audio setup for that.
Example: wired headphones have an infinite battery life.
That means every single pair of wireless headphones immediately loses that debate by an unwinnable margin
Okay, so stuff like positional accuracy and separation of sound queues is irrelevant?
I game for 10 hours+ sometimes, so comfyness is important (to me), battery life is important (to me). And the replaceablility of parts is just something I found out last week and is a nice little extra.
kinda, I don't play competitive either anymore.
Comfort is important but ultimately can't be accounted for due to being too subjective.
Battery life I just explained, get wired.
Replaceability is given on other models too (should be on all three of the aforementioned ones)
wired 🤮 I just don't like it. For a headset.
That's a you problem then
Not really, there is a reason there are wireless headsets. Because people like it.
So if none of that matters. Why pay 150€ when 50€ does the trick?
I like wired because I don't have to think about charging the bloody thing, but hot damn those headphone wires are fucking magnetically attracted to my desk chair's wheels
You keep iterating the same stuff over and over again.
That objective metrics don't matter because subjective opinion and popularity matter more
Same to you.
I've met a couple people with a genuine use case for a wireless headset before.
The rest were just purists that didn't actually need it - either they just thought it was convenient not knowing the cost and quality implications or they were fully aware and just too obsessed
I don't like wires dangling around me 😒
The main use case:
Being on a VC with someone and going to a different place (either for smoking, the toilet or someone else's desk)
from my head, to my PC case. No thank you.
The keyboard and mouse are fine, they don't touch me.
Tell me how subjective factors matter when talking about products.
You can't know my ergonomic preferences/requirements and I can't know yours.
So as long as the discussion revolves around those, it's getting us nowhere.
And I don't see how popularity and uneducated opinions could ever matter
So just sensory oversensitivity
New one but valid
There is no discussion. You said Steelseries is shit. I say, I have one product I like from them.
The end.
Oh, I do have to say. My niece is very happy with her new monitor btw 😛
So thanks for that.
Which one is it?
In that case, switching to your phone is usually fine though
For practicality, maybe
For usability, not really
AOC 27G4X
Unleash your potential and win sweet victories with the AOC Q27G4XN gaming monitor- a sleek battle station with QHD resolution, 180Hz lightning-fast refresh rate and 1ms GtG rapid response time. With Adaptive-Sync, you get low latency, high refresh rates, and incredibly smooth graphics that react...
This bad boy?
What about it?
Anyone knows how to change DLSS preset in satisfactory ?
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!
the 1080p one
Why not just get the asus or otherwise equivalent
AOC is an unknown brand to me
that's a you problem then, they're well established
also, model over brand
and yeah, Q27G4X is dope
27G4X is meh cuz pixel density
gotcha
That technology is so awesome
Fuck it, I'll bite.
Wdym "technology"?
The only technology I see in this pic is the shit causing all the pollution
Well, it's technological process of implanting microbes for soil and water cleaning. In theory, this may clean even pollution caused by microplastic. Microbes with eating plastic abilities already exist
Sounds super hand wavy
Theorycrafting needs to be more thorough than this
However, you gave me a cool idea.
There are bacteria being developed to help with oil spills, they consume crude oil and reproduce until there's no more oil left, then they die.
What would happen to global economy if the biggest oil reserves were accidentally exposed to this bacteria?
Gotta love eBay...
That's just fexed being FedEx
https://uk.pcpartpicker.com/list/GXdMjH can you guys just check over my upgrade list and see if everything is compatible and reasonably future proof?
Part List - AMD Ryzen 5 7600
My budget has also gone up so any new case recommendations would be great
How much has your budget gone up?
Since a month ago, some videos send from a friend of mine through whatsapp show up blank with only audio. I search a bit and find a post on reddit of users with the same issues, they contacted whatsapp and got 0 support. Fucking meta.
Discord voice messages longer than around 2 minutes are simply unplayable and have been for years 🤷♂️
✨ capitalism ✨
Although I'm personally against blaming the system for what still ultimately comes down to individual choice
dont use whatsapp anymore?
I know easier said than done but it can be done
Nah, she only uses whatsapp 😦 And so does my family.
I tend to use the chat app my close ones use so I can stay in touch.
sms 🙂
they also reduced the max file size back down to 10MB after trying out 25MB for a bit claiming no one sent files that big but truth is they probably started loosing money from storage costs 🤣
So that warned about 9.9 security fault was only in CUPS it seems:
www.evilsocket.net/2024/09/26/Attacking-UNIX-systems-via-CUPS-Part-I/
And cannot be actual 9.9 as it needs user to try to print to that printer.
But in general it seems CUPS is a dumpster fire and the devs don't want to fix it.
isn't cups the only available thing for printing under Linux ?
Basically.
But it is 1999 started codebase emulating two much older printing systems originally.
With very bad overall code quality AFAIK even before this.
Only Extend, Not Fix! type of thinking.
Whops, 1997 started codebase, with first public versions in 1999.
And has been in hands of one guy from the start.
Who even left Apple and forked the project few years after Apple bought the project and hired the guy.
This is a move I haven't seen before
I suppose it makes sense, the barrier to entry for even leveling a dispute is made much higher by removing the option for arbitration
At least they're not forcing us into arbitration
Indeed
Unlike discorf
I did the discord opt out
I forgot to 💀
Realistically I won't be using discord all that much in 2 years so I don't care too much
Not in murica so arbitation does not apply to me.
I have ordered all the parts to my new PC
Hopefully this'll let me stream without the weird slowdown in some places
And only an issue if a) you have a public facing cups server... why? b) the bad actor is on your local lan.. which may mean you have bigger problems
I didn't like how he started out by saying that linux is full of security holes.. it does have some, but I wouldn't describe it as full of holes.
Yeah, not as big deal as it originally seemed for that specific thing.
But lot of misconfigured systems out there.
And that attitude of the developers...
And that there seemed to be much more issues, the specific person just threw their hands up on those.
Yeah, but if your it guy made your cups server public facing, cups or not he probably did other shit that'll get you hosed
The attitude of Foss devs is mostly rotten, and i can't really blame them sometimes
Not all, but they do seem to turn bitter after some time
In this case one guy in charge that has been the main developer since 1997 when he started the project.
"You dare to criticize MY CODING?!"
At that point perhaps it's time to fork it.
Including bailing from job at Apple to get full control back.
Fork and rewrite.
Apple still has the original after the guy sold the original project and worked for them as employee for few years, then bailed and forked to get control back.
Apple has their own CUPS for MacOS only, and then everyone uses that re-forked project by same original guy.
I'll admit, cups is neat, I was able to use it to make a network printer or if an old laser printer that didn't have networking abilities
And it even supports ipp for mobile printing lol
Yeah, it does what it should, but the code is rotten.
It's old.. it's a case of if it works why fix it.
But in this case.. if it doesn't get fixed then it falls on the sys admins to probably configure and firewall it
Which should be done anyway. I didn't do anything fancy to mine because it's on a home network
I hate that to an absolute level
Project that old absolutely needs high amount of refactoring at this point.
But that is "useless work".
yes why fix the logic in the core when you can prevent any sane improvement by making it break everything everysingle time, and being a duck tape fest
then there is the naive "but they coded it all, they still have the merit"
yes and ? if they still hold this value to the world, then it means they should still be behaving in a good way
I'm the first to criticize architecture design and code actual quality (not text/looking quality)
but guess what
I DO REWRITE MY OWN CODE A LOT AND IMPROVE ITS QUALITY STILL
either you give up or you dont just play pretend
there is no inbetween unless you should let people call you either a liar or an asshole
Ok, people really need to stop claiming every single machine is powered by AI.... https://www.youtube.com/shorts/G4tkJH6hXlE
There is no AI, just sensors lol
Even FTC is going after companies abusing the term AI https://www.reddit.com/r/ProductManagement/comments/1fpea99/ftc_announces_crackdown_on_deceptive_ai_claims/
now if they could go after all those "gaming" products
what the heck devs? why 150 is not enough to supply 120
This is the wrong channel. Also, it's a bug with power generation. Disconnect machines from the burners and let them power on for a second before connecting
I'm honestly fascinated how people are still ending up here. It has off-topic in the name now, in a section labeled off-topic, and it's at the bottom of all the channels.
I think the last time I pressed someone on it, they just kept vaguely saying this is the Satisfactory server.
reading is a skill
😠
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!
no I'm just dumb
I could see why people come here if their game crashes, they have performance issues or similar.
But it baffles me why someone with an ingame question thinks this is the appropriate channel
I hope they hit their little toe on their bedpost
I hope their favorite restaurant changes recipes for their favorite dish so that it tastes disgusting
I hope the next time they argue the other person pulls out wikipedia and proves them wrong
In front of all their friends
oh yea capacity is 150
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!
one people seem to not have
wrong channel
I know this was a response, it just go even wronger
arch users getting so hard over this
I am
You don't even know
arch will get hyped again
people will stop complaining about secure boot being so tedious to activate
now once uefi key setup you'll be able to run the arch iso
are steamdeck and SteamOS users licensed to say that they use Arch?
I'm just happy that it's not windows
its an arch derivative, so they do kind of use arch
very epic
satis running on my Summit E16 Flip (not designed for gaming) after switching to kryonaut (liquid metal)
This is also on the silent fan preset
I can now play satis in class
I saw someone able to run Satis on an Intel 1185G (or something like that) Thinkpad X1. I was amazed it even made it to the menu, let alone didn't crash after them playing for an hour.
intels igpus are good, just 720p low 🤣
The person running it had no concept of how miraculous it was. They mentioned the laptop has a 4k screen as if that was helping.
This laptop is 2560x1600 (16:10 2k) with 3050ti
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!
How is my 1650TI MaxQ going to perform on satis?
I have an X1 Extreme, 9750H
I'd test but im scared to get dragged back into playing any games again when college has to be the top priority
I did a bit over the summer but was like 4 hours here and there
Id consider playing runescape again maybe
but that thing is a timesuck so probably not 😦
runescape is such a timesuck
I might try playing games tha aren't huge timesinks maybe
ive been having a real problem when I hit the wall and the homework is done.
Mine does fine on low
Although I don't have MaxQ
Just base 1650m
60+ fps at 1080
MAX Settings a big fat 0fps reasonable settings you can defedently make it work
Decent enough, my 3060m handled it well and the 1650 is not that far behind it tbh
Max Q is a designator of a low power GPU
45w baby
Im GF + College + Work + Repair business I feel you
wouldn't change it for anything but yea it is a lot of time spent
ok, the key is also to have the fans, both intake and exhaust to spin at good speeds
It has internal fans but it’s really dusty inside, is it affecting anything?
seems fine to me... no dust on intake filter?
dust will reduce cooling performance for sure
I’ll power it down so I can actually touch it without cutting my finger off
just be careful what you poke at in there when dusting it off
So which is the intake filter?
I got a compressor here to blow out dust at times
not super bad, but it is starting to build some dust
check if you have a dust filter in front
Is this bad?
not really no
but if you get a chance, take it outside and blow out the dust with something
Also in the gpu there is lots of ingrained dust, can I use tweezers to pull it out safely?
the GPU might be full of dust between the fins yes, but poking in there is not a good idea
How should I clean it out other than blowing air in? I don’t really have anything for that
you kind of need a blower that is kind of powerful to blow it out
Could I use a dusting brush?
there are cans of compressed air you can buy, but they don't last long
they are usually intended for smaller dusting
That just moves the dust around inside the PC
you have to get the dust between the cooling fins on both CPU and GPU out
As long as it isn’t on the gpu or fans it’s generally ok though?
Cooling fins?
the GPU has lots of small thin fins behind the two fans
There are two issues with dust:
- Dust clogging up the heatsinks
- Dust about to be sucked up into a heatsink
can't see them on the GPU as its "upside down"
there are two fans on the underside of that pic you posted
Where?
That's just the backplate, you can see the PCB through that and that's all.
The heatsink is on the other side between fans and PCB
Can you point it out on this pic? #off-topic-media message
Wait couldn’t I just point a hoover at it
You can post pics in here.
Also, look from the bottom up
those just suck up dust
On the side of the gpu?
Isn’t the goal to remove the dust tho?
Can work, but generally not recommended for lack of "dust loosening power"
Between the fans and the PCB
yes but you wont get the important dust out from behind the fans on the fins with a hoover
The pcb is the circuit board inside the gpu right
trust me... I've been servicing and building PCs for 30 years 🙂
I can just rub that off with fingers and hoover the rest no?
Yes. PCB means printed circuit board
ideally you'd remove the GPU from its slot and blow the dust off of it
No idea how to do that without destroying the pc
the GPU goes out the way it went in, just in reverse order
It was already in when I bought it
aww man...
then go watch a PC building tutorial
can't like to the GPU website so you can see what it looks like from different angles
unscrew the bracket screw, push down the retaining clip, pull out GPU
This the retaining clip or just cables?
careful giving people instructions @dire igloo
if they have never touched a PC inside before they can damage things if done incorrectly
cables, they need to be removed too
that is just the power cable
that's why I told them to watch a tutorial
please just go watch a video about pc building. its a simple process that you can do right, and doing it wrong is unfortunate
it literally could be a 30s youtube short but it will make your success chances here much much higher 🙂
if you don't unlatch the PCIe clip at the socket you can damage the GPU
I've seen a few repair videos of that tab being broken
its not a gigglebyte gpu so breaking the locking notch off prolly wouldn't kill it.
Best bet is to take the PC to a professional and get it done
I disagree strongly with that. I think a video or two and you can do it fine.
Ive done MUCH worse with gpus and have them survive fine, I think the only real risk here is breaking some of the plastic on the boards cheap PCI-e slot
depends on how handy a person is tbh
wonder... do car shops use oil based compressors for their air?
I’ve just been scraping the dust out really carefully and I’ll hoover the loose dust out, will have to do for now, one of my friends is much more experienced with this than me so I’ll wait until then
if not then ask one to borrow the blower and clean it in no time
disagree. best bet is to just dedicate time to understand how a PC functions hardware-wise so you don't have to pay a professional for somethis as simple as a cleaning job
just don't use an oil based compressor system as it can have residue in the air coming out
I bought a cheap compressor for like 140 USD many years ago
Is there a way to remove the blue fan to get to the stuff behind it?
small tank on it but don't have to last long and I don't need it at full blast anyway
Never mind I see the screws
Don't touch those yet?
DON'T TOUCH THE SCREWS!
They are holding the whole cooler to the CPU.
And after removing them you would need new thermal paste.
I was just about to unscrew it cause I was cleaning it with a hook to loosen the dust
I was going to hoover it once I loosened it
so how old is this thing? I've had PCs for like a decade that didn't get this dusty 😛
I just booted the pc and a bunch of dust blew out real quick, problem solved?
It never is : (
anyways to see the different angles on your GPU search google for Zotac 3060Ti
I can't link the Zotac product website as its on on the safe link list
It should be fine to run for now though right?
its fine, the GPU will throttle if it gets to hot
you'll just loose performance basically
I’m assuming the throttle is 80c, it rarely went above
could be
not sure what the nVidia requirement is for the 3060 cards
pretty sure its they that set this limit, not the board partners
a quick search seem to indicate 83C
that's a very low throttle point - modern CPUs throttle in the 90s
Also the fact it got to 80c while being intentionally as conductive as possible in such a short space means it’s probably using a ton of power
power is limited either way
It’s probably copper as the whole set was £1000, monitor keyboard everything, which is like 300 ish watts I believe
there are many safeguards in modern GPUs to prevent damage
Just checked and it’s using roughly 270-280 watts total
if the GPU gets to hot it will start reducing clock speed and thus power consumption to stay within temp limits
Nvm it’s closer to 350
80c is still loads, two large lights being on is half that power usage
GPUs get hot these days
no?
part temp is a function that involves power draw, heat dissipation, heat transfer and a couple other factors
All I do know is it was 80c and it used roughly 280-310 ish watts
depends on what you define as "hot"
none of the components should be able to boil water
but 80C on a 3060Ti seem a bit much
and how do you know that?
Smart meter
smartmeter?
Turned it off and it dropped about that much
I'm at 70C on my 3070Ti and its using 250 Watts
