#off-topic-tech
1 messages Β· Page 48 of 1
it just does it by generation/arch & not by specific CPUs, but it's still useful
yea true
although it doesn't distinguish between intel P & E cores, that's interesting
Because the E-cores are in the Atom section?
Crestmont for the latest gen.
Gracemont for the early ones.
No, the 14th gen desktop parts still use Gracemont?
Yes.
Crestmont is only in the new Meteor Lake laptop parts.
IPC isn't a set number
but there is uops.info which shows CPI for each instruction on every architecture
In a way it is, in a way it isn't.
If you include the actual feeding of the CPU, then you cannot calculate it.
If you only include the core with infinite cache, you can calculate it for specific instructions combo.
yeah
As that comes from that CPI and how many of them you can do in parallel on same core.
Which is why I did include that thread, as hopefully somewhere there is talk about HOW the numbers in the spreadsheet were calculated and what the assumptions were.
Dont buy msi summit flip e16 laptops btw
Apparently the constant bluescreening isnt a skill issue its fairly common among owners
W10 is a bit more stable than W11 but still does it a lot
Nearly all of the codes are related to memory one way or another so I wonder if they go bad chips? It is running in 1Rx2 dual channel (or 1Rx4 single channel which Im pretty sure its not) so its definitely strained
Official minimum RAM for dedi is 12GB.
And that was close to actual minimum for UE4 versions.
U8 uses less RAM for small save, and has the same RAM usage rise as the normal game has with save size.
*minimum system RAM not allocated
Whops. That was just about OpenWRT router monitoring.
Not related to the earlier satisfactory dedi talk.
https://be.pcpartpicker.com/list/Bz27dH is this a good pc? /s obviously
Part List - AMD Ryzen 7 7800X3D, GeForce RTX 4080 SUPER, Cooler Master Cosmos Infinity 30th Anniversary ATX Full Tower
at that price just go for the 4090
otherwise it'd be like getting a corvette ZR1 with steel wheels
HA the case has actually come down in price
ngl except for the gpu not being top end & the case being ridiculous, i don't actually see anything "wrong" with it, just a big bag of "gimme all of it & ignore the cost"
slow RAM that's likely unstable due to using 4 sticks, not even max capacity (that'd be 192GB)
mediocre cooler that only wins through 420 brute force
mainboard with OCP issues and meh BIOS
if you go overkill on the PSU, at least get HX1600i
i do suspect that at some point mobo manufacturers are going to cancel 4 ram slot boards & just stick to dual or HEDT 6/8/12 stick setups
which i wouldn't mind. maybe having another NVMe slot right there would cut down on latency & sticking some fins on it to keep it cool would be nice
4 slot works fine on intel for now tho
yea but also a lot of people buy for the expandability, but then wait so long that it's just worth upgrading the platform instead
Kinda sucks for the niche use case where you need more ram rather than faster ram
Mostly for people who run VM's
Bring back triple channel memory from x58, then at least you can triple up initially.
I wish my board had 4 slots. Either the board, or the intel 7700 supports 16gb/slot so I'm maxed out at 32gb
X79 quad channel
Not really.
Intel has even higher drop in max RAM frequency when you have 4 sticks actually installed than AMD.
You have drop in max frequency with single stick per channel from 1DPC board to 2DPC board.
And then there is large drop if you actually try to use the second slot per channel.
It was there on DDR4 already to significant degree, and with DDR5 the drop is absolutely huge.
AMD official for 7000-series is 5200MT/s for both 1DPC and 2DPC boards with 1 stick per channel, and 3600MT/s for 2 sticks per channel.
Intel official for 14th gen is 5600MT/s for 1DPC boards, 4400MT/s with 2DPC board with 1 stick per channel, 4000MT/s for 2x 1R, 3600MT/s for 2x 2R.
So for Intel 13th and 14th Gen, the official DDR5 max speed on 4 memory slot MB is 4400MT/s, not 5600MT/s.
That 5600MT/s is only for 2 memory slot boards.
And there is 400-1000MT/s difference in max memory speed you can get on Intel between 2-slot and 4-slot MBs.
When doing RAM OC with DDR5.
So what is the excessive amount of RAM for?
Why not just 2x48GB RAM set?
Ah, missed the /s
Only really looked into the comments after yours.
Baldur, do you have experience with wifi frequency channels and width?
In what way?
For 2.4GHz you use 20MHz channels width if you don't live without any neighbors.
And use only channels 1, 6, or 11, whichever is least used at your location.
For 5GHz you use whatever mostly.
I changed the channel for 2.4 GHz and 5GHz to channel where there is no congestion but I am wondering if I should change the width too
Again looking for in theory a channel that is least used.
My SSID is OpenWRT
And overlapping but not identical is WORSE than being on same number
As the channel numbers are 5MHz apart, and the bandwidth is 20MHz.
AP:s on same exact channel share the airtime, ones overlapping just jam each other.
I moved the channel on 24. GHz from 6 to 4 since there are multiple SSID's there
And you are now getting jammed by both the AP:s on 1, 2 and 6.
And jam those.
When you were on 6, you shared airtime.
Here is how it looked like before I moved the channels (and forgot to turn off Wi-Fi direct on my printer π )
With the misconfigured AP:s on channel 2 and 10, and the two misconfigured 40MHz width APs (only officially allowed to be used when no other APs are in view by spec), staying on channel 6 is your best bet.
Should I just change the channel to somewhere from 10 to 14?
In US, 12-14 aren't available.
only 1-11.
In EU 1-13.
14 is Japan only.
Guess I made the mistake, I was on channel 6 before switching channels π
And again, with the US 1-11 spread, only 1, 6, 11 are not overlapping.
With 1-13 1, 5, 9, 13 would also work, but as Android phones etc. autoselect to 1, 6, 11 anyways no matter where, that cannot be used.
So you only EVER use channels 1, 6 or 11 on 2.4GHz band, and only ever use 20MHz channel width.
Or your AP will be acting as WiFi Jammer against everyone else, and getting jammed in turn by everyone else.
For better experience, set "RTS Threshold" in AP Wifi advanced settings to 0-512.
Sets how large data packets need to prereserve time slot.
So, something like this for 2.4GHz?
Use that 6, as the AP on channel 2 is there to act as jammer for channel 1.
And one on channel 10 to act as jammer for channel 11.
Is there a general rule for RTS/CTS Threshold? Can I set it to 512 or leave it off as it is?
Default is max packet size, so to never pre-reserve time slot.
0 means always reserve (0 byte payload max), 512 would mean that packets larger than 512 bytes need reservation, smaller don't.
What is best depends on exact case with all the traffic going on, but best performance is 0-512 in almost any case if there isn't just one active device on the channel.
And it raises reliability and reduces jitter.
But lowers max actual throughput one client can get.
By default WiFi is "everyone tries to yell everything, and stops if someone else tried at same time".
That RTS/CTS makes it that for long enough yell, you reserve speaking turn first.
I have set the channel to 6 and will do some experiments with RTS to see how it works
So far, no problems with wifi
Hate how manufacturers don't let you pick channels on stock router firmware...
*well, most don't
Which is good thing today.
As people keep putting the settings wrong and causing major issues for everyone.
From the idiocy of 5MHz channel numbers, with 20MHz bandwidth used.
People making the same mistake you did, seeing that there are APs on channels 1 and 6, and putting their AP in some "unused" channel between.
And forcing 40MHz channel widths on 2.4GHz, again causing issues both for themselves and everyone else nearby.
I blame Google and people on the youtube. Literally just telling everyone to switch to a channel where there are less AP's or none 
In theory yes, but need to be completely independent channel, not overlapping one.
As the ones on same channel coordinate, ones on overlapping channels don't and just jam each other.
And then at least in Finland, some very common ISP provided routers had faulty firmwares and always use 40MHz channels for 2.4GHz...
And lot of those still in use.
When the WiFi spec explicitly tells that you aren't allowed to switch from 20MHz channels to 40MHz channels if the AP sees even single other AP.
Which brings my building to this complete shitshow:
At least it's not as bad as here where ISP's have started offering bad modems which can only be configured through a mobile app and web interface has very limited settings to change
You can only change the most basic wifi settings on the mobile app, port forwarding that's basically it
That is separate issue, but that the default equipment didn't follow WiFi specs and acts as jammer...
Not only causing issues for the specific customer like bad config options do.
But causing issues for anyone living near the customer.
At least 5GHz side is better, with no-one else here having allowed their AP to use the "check for weather radar" channels.
Ouch, there is an overlap on almost like every channel
There is on all.
To pretty high signal levels.
And seems my AP isn't following the 5GHz channel assignment, I had it set to that channel 100-112.
But it is on channel 36-48.
And now someone has 5GHz AP with 160MHz channel, using 36-64.
At least it was using the same 36 as control channel.
Because with 5GHz, the wider options use one 20MHz segment as control channel, and rest as extensions.
And for APs with same control channel, again the coexistence works, but if the control channels are different, they act as jammers to each other.
And in this pic, you can see the control channel selections, which are almost all different for the various APs.
With three using 36 for it, then one on 44, one on 46, one on 56 and one on 60.
I hope 5GHz doesn't become as congested as 2.4GHz in the near future
So the 44 and 48 act as jammers for all the ones on 36, and 56 and 60 act as jammers for the 160MHz AP on 36.
Already there.
As you see from that pic.
And that is with most still only using 80MHz channels.
And now there are APs using 160 MHz channels, and Wifi 7 brings 320MHz channels.
Guess there are a lot of 5GHz signals near my home but since 5GHz range is shorter, I can't pick them up
Yes, penetration is much worse.
Anyone here have a home router that supports 6e/6GHz yet?
No, got AX AP, but got model that didn't support 6E.
Might go with Edgerouter + Unify AP combo
Because there is no need for that yet for my use.
The 5GHz band is still not too congested to be issue, even if there are lot of visible APs.
For 2.4GHz, the situation is different, but I don't really use the 2.4GHz.
And for the times it is in use, the RTS Threshold keeps it usable even with the other APs around.
Ax still brings plenty of improvements to the 5GHz band so it's not a bad thing
Also AC was 5GHz only and 2.4GHz stayed on N.
Wifi 6/AX is for both bands.
With Wifi 6E bringing that 6GHz to mix.
Did I say something incorrect or are you just elaborating?
Elaborating.
That AX brought improvements to 2.4GHz band too.
And not just 5GHz, like AC did.
Oh, yeah π
What standard is wifi 7?
be
And bn is in draft stages already to be next one as WiFi 8.
@twin dew random question. Apparently a 2.5 slot 7800XT should be vertical mounted due to load being put on the pcb. (there is no sag). Is this a factual statement or not.
If you have an anti-sag bracket I totally wouldn't worry about it
It is designed to be horizontal-mounted anyways, so any issues with that are a design flaw
I don't have any anti sag bracket but I dont have any sag at all with the gpu just mounted @jagged snow
Then it's probably fine
Perfectly healthy for the gpu.
Opened up my phone and changed the battery. Had to heat it up quite a lot for the glue to melt, might've damaged smth ever so slightly so now my touch screen sensitivity is slightly lower as it was previously.
Though as I was writing this text it worked like previously...
Texting seems fine... huh?
I feel like I'm going crazy, will go to sleep
i dont think anybody in this channel can really help you
Probably mod related.
As I go through audio gear at thrift shops, I'm intrigued with the possibility of picking up HDMI cables for $4. Is there anything obvious on the connector I should look for besides 4K or 8K?
these are cheaper x: about as questionable as thriftstore cables probably
The thrift stores I'm going to tend to have plenty of reputable brands. Mostly I'm loading up on short mic cables.
First thing would be to ask on right place ( #1038092680493801533 ) and not on off-topic channel?
Okay, there's a whole vertical line that's not working
A pretty wide one at that
My finger is pressing between the 2 dots, but touch isn't working there
The whole vertical lineπ
And it's not perfect on the other areas either
F
I blame the excessive heat I had to use to safely get the glued battery out of its socket.
At least my Sony had tear-away glue strip.
Just had to pull on a tab to remove it.
Welp, the phone is dead
Whenever I turn it on properly it starts flickering with random colors
But when I touch the screen once it tells the time just fine
Or when I put it to charge it displays the charge animation and percentage perfectly
Could just be that you knocked the connection for the screen ribbon partway loose
Already checked
Not a bad idea to just take it apart and put it back together again
Already did
Ah bummer
Tried the old battery as well
Welp, guess you're sol then
I'm feeling like Medimedes/Loki rn who killed his GPU with OC iirc.
It's crazy that we can't even trust a video call anymore https://arstechnica.com/information-technology/2024/02/deepfake-scammer-walks-off-with-25-million-in-first-of-its-kind-ai-heist/
The only way to be sure is an in-person meeting
ouch
Well, checking where the video call request comes from...
Cloud flare zero trust tunnels are sweet
Can't believe there are no procedures regarding transferring funds in the company that was scammed 
Well yes, there are technical solutions
I meant as a non-tech savvy user who just recieved a video call
Where I work, this scam wouldn't pass since there are several steps and authorizations before a transaction happens. A CFO can't simply order you to transfer funds
Yeah, most places have policies like that
@edgy hazel You'll appreciate this
I want to walk somewhere that's about .75km from my house as the crow flies. The shortest route to get there is 38.94km
Here, we usually receive an invoice or create purchase order in SAP before anything is paid. Invoice and PO also needs authorization from the accounting department
I remember a warning email from some guy saying creating purchases orders in SAP without a valid contract is a serious violation of company policy and could get you fired
we use fax
Can't wait for 2025 when Death Stranding 2 is expected to be released (not officially confirmed) https://www.youtube.com/watch?v=wbLstJHlC4U&t=569s
Embark on an inspiring mission of human connection beyond the UCA. Samβwith companions by his sideβsets out on a new journey to save humanity from extinction. Join them as they traverse a world beset by otherworldly enemies, obstacles and a haunting question: should we have connected? Step by step, legendary game creator Hideo Kojima changes the...
Yay, more walking π
Ouch, hope you didn't watch the video I posted because there is a spoiler in it 
That's why I contributed a lot of resources towards building a highway. Logged in again after 2 years and the highway is still intact
no, i dont watch or read anything about games as i can
i actually prefer going into a game blind ... basically as much as i 'read up' on is just the product page, thats it
What's the game actually about?
And it has live servers? Or how would smth get destroyed after inactivity?
Correction: the structures don't deteriorate when you don't play
You could tell it's a FedEx simulator with another mission to connect regions
Correction: It's not a FedEx simulator since you actually deliver packages this time
Amazon and newegg cuz they list a shit ton of retailers
Amazon has gone down the drain lately
I can't bear searching Amazon for something like SSDs or flash drives. Too many possibilities of fakes.
SSDs have fewer fakes (probably), but also the filtering is terrible.
That's pretty much just Best Buy, Walmart, and so forth in the United States now. And their prices in store aren't worth browsing idly.
I would guess maybe a third of the United States has theoretical access to a Microcenter, if someone wanted to make a day of it. I would say not even 10% live close enough to one to stop in casually to browse.
nice thing about best buy is theyll price match things shipped and sold by amazon
Yeah that too
I live within 50 miles of two other states so gas is cheaper in one and shopping is cheaper in another
This is so dorky https://twitter.com/Teslaconomics/status/1754199840156987861?t=2yek7LMqC_jwWa16-Qbxng&s=19
1700 amps and 2000+ volt architectures sound cool and all, but those things probably should have been figured out before electric trucks started hitting the road. π
PHEV is the way to go
Trains are great where they're applicable
Busses/trams/monorail/subway are best for intracity transport if the area is populous enough
But for rural areas(which still make up a very large portion of the US) PHEV or hydrogen are really the most viable options
monolames?
If subway isn't feasible and you're building a new urban area they're viable
But really I'm just referring to any elevated short-distance rail
Monorail itself is rather pointless
(even if it looks cool and futuristic)
That's how it's gone historically with a lot of developing technology
The countries that were the last to adopt internet can have some of the better infrastructure designs
hence the "can"
They can also have some very bad layouts(mexico city)
We actually used to have electric, and it worked perfectly fine. It was, however, prohibitively expensive to maintain.
I expect that if it were to make a return now it would be much less costly as our rail network has been drastically downsized.
The thing is, rail did used to go everywhere
And the rail companies, in a rather shortsighted manner, have continuously chased short-term profit over long-tem success
It didn't used to be that way
Shortsightedness π€·ββοΈ
Yeah, US rail still does make insane margins
The big four are in the neighborhood of 50%
insert norfolk southern shitposting here
This is, of course, not sustainable. They're actively letting rail infrastructure fall into disrepair in favor of higher profits and better stock prices
Well, we can make safer nuclear. But we can't get anyone to bite on the upfront cost.
Also how do USB sticks still suck?
I can get an NVMe SSD and an enclosure for it, for 60% the cost of the same capacity USB stick that's running at 1/10th the speed.
Yeah.
And not a big enclosure either. A little 2230 stick.
I am tempted to buy 5 of these things.
And then a matching 512GB NVMe 2230 at Β£40 a piece to fit them.
One for backup, one for Ventoy, one for expanding my phone storage, one for testing operating systems, and a spare.
I don't need 512GB for all of those, but I actually can't find smaller capacities for cheaper. So when overkill is cheap..overkill it is.
That above is the enclosure as I mentioned. π
BYOSSD
Just saying, how are the sticks simultaneously worse and more expensive? And they're made from rejected flash memory.
Yeah..but at this point the NVMe is literally cheaper to buy. And enclosures are plenty.
So it's just weird to me that people buy sticks and they're still sold for so much money.
At this point it feels like they should be worth like.. $15/TB stick because of how slow and crappy they are.
I'm not sure what you're trying to say half the time. π
Sometimes it sounds like you're defending the crappy little USB sticks running at SATA speeds.
I have an enclosure already, it's just bulky.
I plug it into a 20gbps USB port on my PC, and I'm actually limited by the speed of the controller on the 'stick'.
So I only get about 1200MB/s.
Huh. Around 1000-1100.
You can actually see the cache spiking on the performance graph though.
This is only a 250GB 'stick' too, so it's not got as much to work with as the larger capacities.
I say stick because it's..a stick, in shape.
Shame I stood on the end cap.
It's so big and bulky though, standard 2280 NVMe.
Hence my interest in this lil' 2230 :3
Razer Naga left-handed edition.
Sadly they don't do the trinity in left-handed.
I know someone who'd let me print on their 3d printer.
If I get 5 of these, I can design and print a stand to put them in.
got an iphone from another shop why does the nand have a shitstain on it
We're good at telling you why it's a bad idea, but not much with helping.
Im making ajoke about the other shop burning the nand
Most likely they tried to desolder it
Oh you guys thought im some random who bought an iphone nah im fixing this stuff
Ah. The context of "I work for a phone repair shop" probably would've helped. π
An everything repair shop bu yeah
This inspiron makes zero sense. 2 different mobos will not consistently charge if the IO/RTC daughter board is plugged in, and the IO board doesn't handle any voltage or USB signal conversion. What's more, there is zero corrosion and no visible shorts on the USB or SD ports
And based on schamtics, the USB data lines go straight to CPU and VCC is controlled by a 5 point switching chip on the mobo which to my knowledge has not failed
This laptop is cursed
It's probably β¨ firmware β¨
HMM
that could be it actuallyBut why????????
Is it getting mad over RTC?
Nope looks like the (mostfet?) voltage switch chip on the old board is bad
I'm reading continuity on what should be +5VALW to USB20_VCC, GND, and USB_OC1# (which goes to a completely different area on the board
Whereas the one on the new board only gets continuity on GND which is what I understand to be expected behavior
And the overall efficiency of hydrogen is just horrible.
And most current hydrogen is produced from natural gas.
Unfortunately for good performance in that, you would need SSD with DRAM, not one using HBM.
As the HBM doesn't work without direct PCIe connection and the performance just tanks without it.
I'd say 1GB/s sustained reads and 350MB/s sustained writes are plenty enough for USB stick-style things though.
With what SSD?
It was just a cheap 250GB Kingston SSD iirc.
So you are basing that just on marketing from the enclosure manufacturer?
Based on the benchmark I ran.
Up here.
So you already have the enclosure and some SSD?
Okay, more like 280-300MB sustained writes with spikes but still.
A different enclosure and SSD yeah.
Still would be nice to know SSD used.
Yeah, I'm going to plug it in to check in a sec.
Which I realise doesn't work because it only identifies itself as the USB controller interface..
Okay, new plan, check purchase history.
Given I've only bought a single 250GB NVMe SSD.
Kingston A2000 (250GB)
Which huh, actually does claim to have some DRAM in it.
Fair enough, explains a lot. Thanks for making me aware of this..
Now I actually would need to find a 2230 NVMe that also has DRAM..
Probably almost impossible.
As the PCB space just isn't there and that size SSDs were only really produced after HBM using controllers were common.
Well that's annoying.
My other option is to go the other direction and get a 2280 at 2TB and cram everything onto one drive.
With only one problem.
I don't think Ventoy lets you choose partition sizes, and it partitions the entire disk as EXFAT. Which can't then be resized, at least not that I've found.
Scratch that, it was just a hidden option. I can just make it smaller, so I have.
Windows cannot, third party utilities can.
GParted could not.
EXFAT isn't designed to be resized
Ah, all the paid utilities say they can, but then the actual instructions have you changing the partition type to another first, and then resizing.
And then back.
I just reformatted my ventoy with 200GB of free space in a setting. I have to copy over my ISOs again and that's taking about 25 seconds.
Well, plus however long it'll take to flush from the memory cache onto the disk. π
Call it 45 seconds for 14GB. Good enough. And adds up, since that's about 300MB/s, similar to the benchmark.
If I did this on my 16GB USB stick, I'd be waiting probably 2-3 hours. Seriously, it's not even that old and it's just so CRAP.
I'm planning to put OpenSUSE Tumbleweed on that other 200GB. I'll either switch to it in future, or use it as a control for when I can't get something to work: is a feature broken, or just my packages being outdated/broken?
Compressing and transferring hydrogen is very power hungry.
Same for producing it with electrolysis instead of cracking from natural gas.
So if the power was from wind, solar etc, not real problem.
But when it isn't, the total efficiency of less than 30% IIRC is horrible.
In terms of total emissions.
And then it's only going to be viable for perhaps a trucking company, or anyone else willing to build out all the infrastructure for themselves.
H is heavy, and trititium is even heavier
__ when you compress and store it, totally agree with you on that
like think about how much desiel fuel was used just to transport a plastic cup from china to your workplace cafeteria or whatver
And if you think car size lithium-ion batteries are dangerous, what about hydrogen storage container with internal pressure of hundreds of bars?
I have an internal pressure of hundreds of bars skkrr
Or the danger of refueling with high-pressure hydrogen to charge that internal pressure container in the car.
Conventional alkaline electrolysis has an efficiency of about 70%,[36] however advanced alkaline water electrolysers with efficiency of up to 82% are available.[37] Accounting for the use of the higher heat value (because inefficiency via heat can be redirected back into the system to create the steam required by the catalyst), average working efficiencies for PEM electrolysis are around 80%, or 82% using the most modern alkaline electrolysers.[38]
PEM efficiency is expected to increase to approximately 86%[39] before 2030. Theoretical efficiency for PEM electrolysers is predicted up to 94%.[40]
Hydrogen gas is produced by several industrial methods. Fossil fuels are the dominant source of hydrogen. As of 2020, the majority of hydrogen (~95%) is produced by steam reforming of natural gas and other light hydrocarbons, and partial oxidation of heavier hydrocarbons. Other methods of hydrogen production include biomass gasification and meth...
That is just part.
The big problem is the pressurization for the refueling.
that's covered in the paragraph above that one
I was talking about total, production, fueling, usage.
That was just about the production.
If you have 90% efficient production, then 90% efficient refueling, and 90% efficient usage, the total efficiency is 72.9%.
Ok, mistake, refueling wouldn't be done that way, as it is separate thing.
But the production and usage point efficiencies get stacked, and then the power needed to do the refueling is handled separately.
And compressing hydrogen is just crap efficiency.
And the fuel cell efficiency in the vehicle isn't that great either with the needed heating power.
Yes, but much worse than direct battery vehicles for total efficiency on road traffic.
And trains in turn should just have electrified rail network.
direct battery vehicles also mean strip mining lithium
I agree, but that isn't always possible
especially in North America
It is, just costs too much on short term.
Would be cheaper in long term, but the companies in the business don't care about that.
Yeah, try building infrastructure when max ROI allowed is less than 5 years...
And anything over 1 year is already almost impossible to get through.
AC doesn't work with underwater cables without huge losses.
The seawaters capacitance causing issues.
depends on the voltage
you don't use the catenary wire as your main conductor
you have evenly spaced transformers supplying zones
So DC transfer cables are now used for underwater, and for transfers between different national/supranational grids so that voltage, frequency and phase differences don't matter.
The channel tunnel is deep underground, there's no salt water to speak of
Not talking about train electrification right now, just general power grid.
Switched to reason for this
you don't
you use battery or hydrogen
unless there is existing power infra you can piggyback off
Higher voltage AC line next to the train, and then periodic downconverting transformers feeding the electrification.
Which is too expensive for the cheapscate MBAs wanting quick returns.
But the same people wouldn't build the new rail line anyways.
And aren't maintaining the current rails either.
Which is why US has weekly train derailments AFAIK.
As the rails and trains are just rotting away.
Was daily train derailments when media tracked it after those couple high profile ones.
The locos are taken good care of, the track and rolling stock not so much
Sorry, used "trains" for both locomotives and rolling stock.
The line blurs when you bring passenger trains into the discussion lol
I think even diesel locos are still the most efficient cargo transport compared to their fossile based alternatives
diesels with batteries and regenerative brakes for downward grades
hydrogen is being explored as an option for long distance, batteries are good on their own in shunters
I still want to know what that 1700 amp cable would look like.
(from the SAE article I posted earlier)
I think like 60% of our rail is electric so we good anyways
where
π©πͺ
ah yeah
overhead wire freight would be best, but the class I companies don't wanna because bad
Article is for trucks, the same could apply to trains or anything else though
something something clearance for autoracks even though the entire rest of the world manages to put the wires high enough
I love electric vehicles (especially trains and scooters because fuck cars) but I hate their battery
Lithium is such a trash material
I don't want to get my hopes up. But it feels like multiple manufacturers are serious enough about solid state batteries. They'll probably only be in the high-end brands like Lexus at first. But in 10 years who knows. Or maybe it'll be 30 years before they can scale production. π€·
also, genset units
a genset is multiple diesel engines that work together and stop turning according to demand
most if not all locos have only one engine here
Hopefully that translate to smaller and lighter laptop batteries as well.
Wearable tech
gensets mean lower idle/low power emission and redundancy, but at the cost of maintenance
If batteries weren't as heavy and shit I'd already have daily driver AR glasses
Most of the various tested Solid State Batteries are still lithium based.
Difference is just that liquid electrolyte is switched to solid.
The longevity of solid state batteries could also be interesting. If they can withstand cycles well enough, fast charging might make people want the same capacity batteries overall. If you can get most of the juice in a couple minutes, you don't need a bigger capacity necessarily. For laptops and smaller devices.
Yes.
Torque you can get from electric motor at near zero speed is immense compared to what any combustion engine can do.
And the efficiency of combustion engine is best at certain speed.
So you use the combustion engine running at optimum efficient speed to generate electricity, that you use for electric motors.
Lead or tungsten pellets
Because somehow people seem to think that higher weight directly correlates with higher quality
Lot of higher end gaming mice have tuneable weight too.
With removeable weights.
Because different people like differing amounts of inertia.
This is a joke btw
Well what would you call these?
The weight system on Logitect G502
There have been other ones out there with some kind of caddy you can pull out with removable weights, but just cannot find them right now.
Don't know the right search terms.
They're not lead iirc
Also, Z906 are speakers
Ah, whops.
But in terms of speakers, high weight is good thing usually.
If it is from more rigid construction.
I feel like high weight for tech is like high wattage for vacuum cleaners
People think that more means better but there's no actual reason for it
And at some point manufacturers start bullshitting their way into people's minds - in this case lead pellets, for vacuum cleaners gigantic resistors producing heat
With PSUs it was about larger heatsinks.
As the cheap ones at one point were cheaping out on that.
Not directly relevant anymore, as the good PSUs have better components that produce less heat, so the heatsink size can be smaller.
I recently got myself open box Z323 for 30β¬, actually an upgrade over my 200+β¬ Razer Leviathan
Talked to an audio engineer about recommendations.
All her go-to picks weren't available here in Germany
Who's not selling here?
Creative? They're readily available
Tho I'd like to have some Klipsch
ProMedia 2.1 was listed as a top tier pick
Mail forwarding service?
The classroom for my fundamentals of computing class has 2 old LGA1151 stock coolers lying out that have been out of use for who knows how long
All I know is that they didn't wipe the paste off and it is hard as a rock like you would literally need to scrape it off with something hard
I still have some LGA775 and AM2 coolers lying around... π
But generally using some alcohol to clean the paste off works best if it's old.
You overestimate the ability of an introductory computer classroom to have 71+% isopropyl
I think I use 73 at work, def not 91. I know because I've seen it put on LCD's and they didn't melt
Although 91 would be nice for board cleaning...
Or.... just heat it up 
Take them home and heat there then
thermal paste softens from heat? well i guess that would make sense
i guess the expansion of a material would soften it because your seperating the particles or whatever
Thermal paste doesn't really expand with heat
well im just aplying expand in a basic sense (cold=shrink heat=expand)
but yea you probably wouldent want thermal paste expanding much under heat
shrink and expand are very clearly defined terms. If it neither increases nor decreases in size, it neither shrinks nor expands
polymers have a glass temperature at which they become ductile
I be getting my own domain tmrw 
I have my own domain already
ive had my domain for so long ive paid ridiculous amounts over the years in renewals π
what do you guys even use it for just learning network stuff or what?
seems like owning a domain is fairly common
i used it for my website and my emails
Oh neato
It is currently used for nothing other than a hostname for my vpn and some game servers I have yet to set up a reverse proxy
I've been debating heavily between two domains
One's a .net and the other is a less common TLD
Mine has rizz in its name
More or less
Oh?
But in the past I have tried to set up an internal intranet for various docker containers and it did not work
you know like seafile/ goes to seafile
When that didn't work I tried using a trial version of windows server 2016 to set up DNS forwarding to internal subdomains (seafile.example.com) but that didn't work either
And nothing I saw on le interwebs was anything but cryptic and hard to follow
So is this a simple operation I'm overthinking or is it complicated and hard to explain
What address space were you using?
That's subnet mask
Then I am confused
I'm just making sure that you weren't using an address with a first octet value of 224-255
And you were just trying to route between containers on the same machine?
ah
WebUI's
online network related stuff is too freaking often misleading
Yeah...
And and
good ressources are made to be read by people already knowing most of the things
it's hard
I know Unraid does some magic for hostname resolution
Because my main server shows up by name no problem
But docker containers dont and I have tried
So your issue was just with the internal hostname resolution, right?
Not the network itself
and there are lot of arbitrary "standard" in the mix
like how Linux check machine domain first and mDNS and /etc/resolv.conf etc
first things first check your understanding of a physical network
and on what physical network your dicker will work
then the logical network part
(physical network can have multiple logical networks, that mostly how dual stack is done)
ipvlan so technically no physical network. Otherwise its just my main home network
Logical network is 127 scheme I think
then understand that any communication is done inside the network
any communication with outside needs a forwarder
that is inside
No I'm sorry my containers are all running on custom IP's in my main network
Why not use DHCP?
physical to the network perspective itself it can be emulate
Why do I WANT random addressing
More scalable and eliminates human config error
Wdym human config error? Address conflicts?
Or wrong subnet mask, or incorrect gateway...
dhcp only is very practical for client machines
for servers you want static unless you also have dns management
Fair although the way unraid handles docker addressing that isn't really an issue
Yeah, that's true
Although I was under the impression that this net was mostly clients?
Oh and thats actually another thing that has been really annoying. I want to put my docker containers (or at least most of them) on eth1 which is configured as the sole member of br1 and not a member of br0. Yet when I have the above config for docker settings unraid only lets me set an IP on br0 and not br1 like why
Well yes it is mostly clients but all of my containers are hosts not clients
Ah, so you're just doing static configs on the containers?
So first dredgen make sure you have clear "physical network" and clear logical network in your head
before you are going for it
remember any machine that need to communicate also need to be in the network
Yes
Please dumb it down for me a bit more or elaborate more
I am an aspiring network engineer, but I'm not one yet lol
physical network : where network communication take place :
- switch
- cable
- switch ports
- emulated switch
- dockers interface and network namespace
- vlan
- etc ..
logical network : how your network identify
how machines recognizes each other's :
- network address
- mask
- subnet
- each single address on interfaces
- what machine needs to join what other address
all this must match in a single unique network for all machines to be able to talk to each other's
Ohhhh okay
remember Internet is short for "inter-network" it must have working "network" part before working on "inter" part
So in my case with the containers, its configured as ipvlan but I am assigning them static IP's on my main network through a bridge configured on eth0, so that would make it physically on my network then yes?
ipvlan is a type of interface
only that
I guess yes
though I don't really understand how you are mixing bridge and ipvlan
ipvlan are interface serving specific ip on a bare interface
so all ips have the same bare interface
i want an internet set up in series & not a web
wanna access a server in NYC? go thru every computer between here & there
Yeah, it's only really practical on LAN and even then really only works for a few devices
Star topology works better in almost every way
Only downside is that it's sometimes more physical cabling
I mean you can technically put ipvlan directly on eth0 without bridge
it don't remember how ipvlan on same interface can communicate with each other's though
i'm assuming that this version of the internet will crash & burn once quantum computers get mass adoption alongside some built-in security in the base code
For insecure communication, it's still perfectly fine
And quantum-hardened encryption algos are much farther along than quantum tech itself
heh, yeah
Networking can be a nightmare
I've recently been working on earnign some basic networking certs
quantum GPUs when
Never
Man I was so going to answer dredgen questions
@winged valley you gone?
Busy atm give me a min
rechecked : they can if ipvlan in bridge mode (not to be confused with bridge device/port) i.e not private nor vepa
if you only want ipvlan interfaces to talk to physical machines (ipvlan interface are only logical) and when directly joined
then you can make ipvlan interfaces in L3 private mode
if you want them to be as if full machine was there you need L2 mode and if you want them to communicate between them you want to select/let bridge flag (not private)
every client device need to have adress in the same network
or send to a default gateway that have one and allow forwarding on that network
what I get your current docker interface setup is :
eth0 phy -- bridge -- br0 > > >
ipv0 | ipv1 | ipv2
ipv0 address must not be set neither on eth0, br0 nor any other interface
if you don't need ipv0 to talk to ipv2, ipv2 to ipv1 etc..
L3 private is recommended
(btw you can setup ipvlan with dhcp but it's tricky, basically holding dhcp communication on br0 and having dhcp daemon dynamically managing ipvlan interface instead of changing br0 settings)
(I did that with custom dhclient setup scripts)
(you can even translate and transfer dhcp requests with that kind of setup)
ok I must go now
I thought you would have come a bit later
And you can have fixed IPs even with DHCP.
Given by the DHCP server based on MAC or computer name etc.
But anyways, my RAM settings don't work anymore if RAM PMIC temperature sensor reading gets to 72C or over.
And the difference in stable settings is absolutely gigantic.
Wasted couple of weeks trying to stabilize again at those high RAM temperatures, but seems the "fix" is just to make sure they never get there.
The test setup was pretty forced, and with fans set to default Silent profile the RAM only hits 60-62C when there isn't GPU heat load in mix.
Or needing to run almost same settings as MB sets as Auto.
Are the CompTIA books worth any salt
hello there. Anyone with a i9 14900k to give your feeling about it ?
i'm about to change my setup, and i'm still wondering what CPU
currently digging into it
It sure is a CPU
For what use?
And what kind of budget?
Because for most non-professional usage it is overpriced and over-hot.
- nvme
as max
yeah that's what i suspect, but well. my last CPU last me 12 years π
so i kinda like when they are performant
i'm currently on a 3770K i7
overclocked since i have it to 4.4
i'm mostly a player
And something like a 7900x3D?
titan X
Or 7800X3D if just for games.
And not for work.
was on 680 before that
Or even just dropping down model or two on the Intel side.
If you mostly game a x3D is nice
Big question is that if you only play, or do significant amount of number crunching on the computer too.
Actual CPU power hungry work type stuff.
"mounting"?
Do you use GPU or CPU encoding to do that?
Lot of multilayer effects or just basic cutting?
basic cutting
So if using GPU accelerated encoding, not lot of CPU usage in it either.
So any game-optimized CPU is enough.
And doesn't need to really be thought for CPU heavy work scenarios, as the game-CPU will do the other stuff well enough.
but CPU and ram are struggling hard for the past 2 years
so you would tend to say i9 is overstocked for my needs
i like to hear that π
So AMD 7800X3D or Intel 13600K/700K / 14600K/14700K.
Or even non-K, would need to check the exact differences in those.
the difference is basically overclocking
Because there isn't just "overclocking or not" anymore.
No, Intel has been doing much bigger changes in last 5 years or so.
the coefficient clock is the difference
To level of different core counts in some models.
must dig into that too
So there can be core count, power max and frequency max differences now.
For 14700 to 14700K it is 0.2GHz P-core max frequency difference and 65/219W to 125/253W power.
that's a lot of power for a few % upgrade
If the MB were following that base power and turbo time limits, that would make huge difference.
It the MB isn't following the base power limit, then the difference will be miniscule.
Because the CPU would be power limited and running at significantly lower frequency once the turbotime-max has run out.
i heard they may remove hyperthreading for 15th
Do you think the following pc is good for Gaming: https://pcpartpicker.com/list/t7BvYN
Part List - AMD Ryzen 7 7800X3D, GeForce RTX 4090, Thermaltake Core P6 ATX Mid Tower
But AFAIK almost all LGA1700 MBs can at least be set to ignore the Intel stock limits, like all previous MBs.
And plenty of them do it by default.
But check that AMD 7800X3D, it is in many ways the current gaming king CPU.
yep for sure
i want to dig into amd too.
as i didnt follow much for tech specs in the last 10 years
there's a lot to learn before making a choice π
Low quality mainboard, also why ATX?
For WIFI, you can always add an AX200 or AX210 based chip.
Overpriced RAM (T-Create Delta CL38) and SSD (Vi5000 import from Japan or NM710).
For PSU: model over brand (!!!), you're fine with even 850W, definitely don't need any more than 1kW.
Mediocre fans, consider Thermalright TL-C12C-S, TL-S12-S or TL-M12(-S).
Alternatively, P12 ARGB (more expensive tho)
Which motherboard Would you use
B650M Pro RS or B650 LiveMixer
Ok thanks
i dont regret getting lian li al120V2's but there quite expensive
like 100$ or more for a set of 3
but on the positive there one of the few things that dident arive broken
AL V2 are some of the best performing RGB fans out there
at this point practically anything that survives me is bulletproof
That said, only few usecases actually benefit from that performance over the likes of TL-C12C-S
yea i personally would defedently spend the extra money to get them again whenever i build a pc again
id rather spend extra money on products that i know are reliable than take chances at products i dont know or havent had good experiences whit
im defedently pushing deepcool to the side for a while
havent had issues whit gigabyte other than there god awfull gpu mounts
the western digital m.2's have been great
everything else has had minor but expected issues that i woudlent really go out of my way to avoid
whats a decent low key, no frills, clean, free mp3 player on windows ?
Foobar2000 is great I heard
Always remember tho:
Model over brand
Lian Li SL V2 and SL INF perform rather poorly as fans
Every brand has good and bad products
Take SN350 for example
Or the usually horribly overpriced SN850X
k i'll give it a shot ... MediaPlayer on win11 drives me up the wall on something as simple as "remove this song from the library" ... =p
i got sick of trying to cast my music through bluetooth speakers, so i copied ALL my mp3s over to my PC (which is hooked up directly to my big 5.1 speaker system) ... but now i need to figure out which songs are not "Work Jam" lol ...
the way the ipod was, i couldnt get 'work jam' list off it to the pc ... long story, short answer; APPLE ... so im basically starting over a new jam playlist built from my entire mp3 library
but i got some stuff in here thats not very .... jam worthy .... like nature songs, and celtic vocalists and stuff hahaha ...
nice thing is im discovering some songs i completely forgot i had, and sound fun
AI! ... Embrace the AI revolution. It is here for us, and for our future! https://www.youtube.com/watch?v=y1SgKJ8JFB0
From the spread of misinformation to the automation of our workforce, the dangers of artificial intelligence are clear. But as Advanced Neural Network Learning Model Signet V4 explains, human being should not be concerned, as our complete and total annihilation is sure to be swift, efficient, and relatively painless.
Subscribe to The Onion on Y...
I love the onion
Pretty much what fireworker said
Why that case and air cooler? Why non-rgb fans on the cooler?
we love macos
any suggestions for good office/game chairs for tall ppl (IE, going high enough and have more upper leg/butt support)?
Can you use a hexidecimal editor to edit a vbios?
If you know what each location means, and if you need to edit some checksum elsewhere too etc.
So a learning curve it is ig.
hex-editors are just pure binary editors.
Yeah. One I tryed tryed to sort it into plaintext and I could make out some locations off the bat.
Something something end something table.
But remember that address of anything must not change.
So need to be careful not to remove or add anything.
Only replace.
π Your computer is on fire π everything is fine π -macOS
Those damn kids who keep ringing my doorbell and running away, little do they know I got cameras everywhere lol
apparently baldor informed me that nivida gpu's have a security processor called falcon that checks for unauthorized bios's so ymmv with editing. I had found some other stuff about bypassing it to use another unverified bios, but it's on my other pc.
the most vbios editing I've one is padding to make a 128kbit bios fit on a 512kbit flash. I used something like cat original.bin /dev/zero | dd bs=1024 count=512000 of=padded.bin
Put a screen on the doorbell. Next time they approach to ring, show them a live feed of them with some horror monster edited into the background
Like lurking behind the trashcans
Nah, their biggest punishment is that I know their parents and have their phone number 
I will let them get away with it this time but if they repeat the same, I'm telling their parents
Do you know their names?
Send them a letter addressed to them
Include a note "I know what you did" from newspaper snippets
People don't even realize we have cameras installed because they are dome cameras. They probably think they are some kind of lights lol
And a screenshot
lol
Traumatize dem kids
And yes, I know the name of the chubby boy
The other kid is some friend, don't know his name
Also, Smart Piss
At least they haven't gone so far as some crazy kids who throw rocks at windows. Glad I don't have such kids in my place
time to put a glob of really sticky goo on the doorbell button
if you are looking to replace the button (or remove it, seriously, doorbells wtf?) ... put superglue on it so they get stuck next time LOL
My mailman is going to pissed lol
You never know who else is going to press the button π
no one comes to my door that is wanted ... π
so i dunno what all fuss is about, my 850w psu has been working just fine with this 14 series intel and 4080 gpu ...
i definitely dont need a 1000w psu
What PSU is it?
Some are far better equipped to handle transient spikes than others
And the atx 3.0 spec places much higher requirements for that on the PSU as well
So if it's 12v HPWR native then that's probably fine
Still one of the better psus in production, so not entirely surprising
im using a corsair version right angle cable (so it just has plugs to the psu, and 1 plug to the gpu... no adaptor crap)
Ever tortured it with simultaneous Furmark, Prime95 and OCCT?
heh! no
See?
i'd need a deep freeze full boiling liquid cooling setup, cause the 14 series would throttle itself so hard
that 14 series really is kind of a joke, to be honest π
Yeah, it still should've been a 13750k imo
Just two extra e-cores doesn't make it an entire new gen
Or even 13790k if you want to denote that it's a more significant bump in perf than all your other refreshed chips
Those could all be xxx50
right
uh... yeah i got the 14700k
and no im not overclocking it ... it just runs too hot π
well yea same here whit the 13700K
unless you disable the e-cores under a full load your staying above 80C whether you like it or not
im guessing this is is because of the smaller surface contact and smaller IHS right?
unsure
my two p-cores (5 and 6) STILL do this ... ALWAYS ... no matter what, they are baseline just doing this. and it never did this with my 12600k
prettysure there was some sort of report or something at one point that AMD chips had IHS's twice as thick as intels IHS's
its fun when a game just out of the blue does this .... LOL ... like really?
that was "Deliver Us Mars" above there... it just took a dump randomly
well yea most loading screens read either 0 or like 1000 3000 5000 even 7000+ fps
because your rendering nothing so the framerate just goes trough the roof
and then falls off a cliff
naw, the normal rate was playing the game... then suddenly you walk into a room and it drops to 30fps π
and even if you go back to where you were, its 30fps
me on my measly 5600 on a 140mm aio
140 AIO?
game just took random dumps like that, had to relaunch
That's cursed af
nzxt h1 be like
a single 140 fan?
yup
they MAKE those ?
yea nzxt lets you use 140mm fans and 120mm fans in each location other than the side mount on cases like the h9
oh dear
assuming you can get your hands on a 140mm rad in the first place you can mount it
its not amazing but ive had no issues with my nzxt h1
till the pump dies π
π€«
shhhh myn is defedently not already squealing
as if the 3 fan bearings being totaled wasent enough
you know what im gonna make the list i have been wanting to make of what needs fixing what needs swaping and so on so i can keep up whit everything
Nah, til you scrape the riser cable's PCB with the metal screw and short the 12V plane to ground
that doesnt sound like a good scenario
It was one of the biggest manufacturer fuckups in the recent years
P-GM level fuckup
Also horrendous handling of the situation
π¦
Journalists find an issue with one of your products, what do you do?
NZXT's strategy:
- Blame journalists. They initiated this shitstorm.
- Blame consumers. This wouldn't happen if they didn't make a mistake during assembly.
- Quietly provide Nylon screws - but make customers contact you for it, this ain't communism where you get shit for free.
sounds like a plan written by an idiot in charge
This was their response
i dont think i missed anything but i probably did
did the braindead child while i was at it (the alienware)
forgot the rad but yea its working fine
looks identical to my h6 flow lmao
π
That looks like a render
It's actually messing with my head a little bit
Where is the pc sitting that has that clean, diffused, white side-lighting
All the surfaces are such a perfect eggshell that it looks incredibly unrealistic
its an actual case
i have gotten that three times so far feom freinds
It's the lighting
Reasonable
ik a bit spaggati but cable extensions (and running out of cable ties) are a minor pain
Nowdays I always struggle wondering if it's a photo or a render haha
@jagged snow heres the front
Nice, that's a pretty clean build
A different approach to lighting my build then I usually take. I really like how it turned out
(yes mixed ram dies and timings match so it's fine replacing the other kit ASAP)
64gb ram.
Cool deal
64gb and only 2 sticks?
2x32 is fairly common with ddr5
That's what I've been reccomending for higher-budget builds
Like I said mixed kits die and timing matched tho
Why would you mix and match RAM? Are they at least the same speed?
See this message
4x16 has better performance
Depends on the clocks you're pushing
Will be adding all vengeance pro soon
And 2x2r should be very similar to 4x1r
Since there are no quad-channel ddr5 consumer platforms
Especially on amd, you're going to have no luck trying to push 4 sticks past 6000MT/s
Most of the time, you can't get 4 sticks past 5200
A full minute?
CPU's are programmed at the factory right?
How has nobody figured out how to reprogram older ones
What do you mean by that?
FPGAs are my inspiration for this nonsensical idea
This camera installation wouldn't be legal in Finland...
You aren't allowed to record into the street, only your own property.
Oh that's very interesting
In the US, you can record anything on public land
And you can record anything on private property that's visible from public areas
So it's perfectly legal for me to stand on the sidewalk and record you in your front yard
And even on your own property, all visitors must have been informed that they are being recorded.
Either by signs or telling them. So they can consent or leave.
Oh wow, that's fascinatingly strict
So hypothetically
US is informed consent for recording in most places as well iirc, but for all interactions my state is single-party consent
If there was a bombing in finland, and the perpetrator was identified through video, but the video was from a single source - an illegal camera
Would it be inadmissable?
No, taking video of crime is legal in all cases.
But all the other people in it would probably need to be made unidentifiable before showing outside closed doors.
The protections against "hidden recording" don't apply to active criminals.
The law is AFAIK very old and originally made against voice tape recording without consent.
But written so broadly that security cameras etc. apply.
Finland is two-party consent for any kind of voice recording on phone etc.
Which is why most company help lines etc. will state that the call will be recorded.
You consent by continuing the call.
US is a mix, most states are single party consent but some are all party consent
So if you're calling a line that serves more than one state you'll usually get the message informing you that you consent to being recorded by continuing the call
Also, if your security cameras record outside your property like that, it would be considered Personal information registry under GDPR if I understood it correctly.
If the persons in the video can be identified visually from it. (good enough quality faces)
And then you have outliers like Oregon which is two party consent most of the time but single party in specific scenarios.
In person conversations are all party consent but phone calls are single party
What does that mean?
https://en.wikipedia.org/wiki/General_Data_Protection_Regulation
In EU, collecting personal information needs legit cause, and specific handling.
gdpr is European laws relating to personal information storage and processing
I've heard the term in connection with advertising and data collection with online services, just wasn't sure how it related here
Thanks for clarifying
The handling etc. clauses would apply to surveillance video taken of public spaces.
With high enough fidelity to make people identifiable.
or to participate in making them identifiable
including the ability for any such people to request removable of the stored records
by right of "being forgotten"
It's the General Data Protection Regulation.
Applies to basically all areas where data is collected in one form or another
Better ask my lawyer about GDPR then π . I have already given some recordings to the police at their demand due to theft at the nearby construction site and they never complained about recording public spaces
They were only disappointed motion-trigger is defined inside my property but I fixed that so the camera is now recording when motion is detected in any area
So the original motion trigger mode would have made them mostly legal here.
As the recording of public spaces would only have been incidental (motion inside property) and not "contant" (motion also outside property).
Police don't enforce GDPR tho.
They don't give a shit about it, more often than not it actually hinders them in what they do - because of surveillance constrictions
This applies in sweden as well.
Technically these recordings are not allowed in court and the defendant is also able to sue
Is it me, or is AMD crap with crashes? π As in, something funky happens, AMD doesn't know what to do with it, crashes the PC.
I could try another clean install of the drivers and AMD software. But like, my PC is stable for a month now, I press F11 in minecraft (goes from full screen to window mode), my PC freezes for a few seconds. And bam Kernel 41; didn't popwer down properly. Only difference with the other crashes I had (from FLCK probably) is that now it rebooted in a weird resolution for some reason π
At this point it feels like my machine just can't handle any type of errors and then continues to just shut down the whole rig.
And before you say anything Baldur, I kept in mind that it might be my CPU that is dying. I am keeping an eye on a X3D in the future π
I used to blame that on amd and now I'm convinced that it's almost entirely a Windows problem
Only reason I say AMD is that I never, ever ever had PC crashes like this with my intel + nvidia combo. For 8 years straight!
I never even had a GPU crash with Nvidia.
And now in only 3 months or so, I had like what? 20-40 full PC hard crashes without any errors?
I was having random unexplainable crashes in all sorts of games and I'd fully stripped down my entire PC and rebuilt it, swapped vbios, DDU & reinstall, reinstalled all the games, nothing made any difference.
Finally I reinstalled windows and I'm several months completely crash free
I did all off that already π I did like 3-4 windows installs.
Do you have an Nvidia gpu to swap in? If not, just pick up something on eBay for around 100 bucks and swap it in, see if it solves your problem
But what you described also sounds like it could be a power delivery issue
Nah, this is way to random to be a power issue.
I thought it happened mostly on alt tab?
I swear to god, PC + something unexpected = full PC crash. And it can be anything. Alt tabbing. FLCK. Downloading a game. Pressing F11 by accident in minecraft.
Just standing AFK in GW2 in a city. Changing resolution. All made my PC crash so far.
My go-to with random unrelated and unexplainable crashes is typically psu or memory
The FLCK to 1700 made the system way more stable. Since then I had no more crashes, untill today when I accidently pressed F11 π
Otherwise that does just sound like dying hardware
So that's 26 days stable? And I press F11, and bam PC dies. Also, first time it rebooted in low resolution for some reason π (after it crashed)
I pretty much confirmed that my laptops ram is bad bc memtest instantly spat out enough errors to fail when the dma test started
And with all the bugcheck codes pointijg to memory one way or another, yeah
So now I HAVE to upgrade it to 32gb woo
That DMA test is experimental.
If the other tests didn't have errors, it might have nothing to do with the RAM.
Damn, I would've just returned the CPU if it can't handle 2000mhz lol
What are you running your ram at then?
Nvm, it was UCLK that was synced with ram
FCLK is independent
Eh, I don't think I stood a chance. And I wanted to game (so, not having a CPU sucks).
My current plan is to price watch that 7800 X3D
You really lost the silicon lottery
Baldur thinks it might have been the motherboard frying the CPU.
Or atleast causing some sort of permanent damage.
And now it's slowing dying. But we can't know for sure.
I have been expecting this to be case of VSOC debacle from not updating MB BIOS before installing the CPU.
Rip
I like flashing the bios before booting up the PC for the first time, makes me feel good and ig good practice?
Yeah, know I know.
I was just dissapointed today that it was stable for a month and now a stupid F11 (fullscreen to smallscreen) caused the PC to hard crash. It's Minecraft so shouldn't be that dramatic.
And I gamed a lot the post couple of weeks. So it's not that I didn't game for a month or something.
If it start to crash commonly now, then the hypothesis would be much closer to being confirmed.
That there is IOD damage from VSOC that is causing the max FCLK to go down with time.
Btw Baldur I ordered the shit, not the exact parts we looked at, but same logic should apply.
In reality the shit I'll be running is much lower wattage so I didn't have to go as overkill and I found the same PWM to DC converter for cheaper. Will arrive next week.
If the 7800 X3D hits around the 300 euro's again I am getting it.
where is my domaaaaiinnnn
does .online usually take this long?
this is like day 2 now
Try RMAing the CPU. Don't claim enabling expo, claim to be a very good boy who followed the rules, tried different components and found out the CPU is causing the crashes.
It's honestly not that hard to lie if that's what it takes, in the end, it wasn't your fault.
The ASRock BIOS seems to be setting weird stuff by default.
Like always defaulting to 2000MHz FCLK?
When AMD base one is 1733.
Don't they all?
No.
Asus only set higher FCLK when you set RAM to higher frequency, either manually or with XMP/EXPO.
Yeah, reading about it now. It's pretty much this: www.amd.com/en/registration/warranty-services-portal.html
Funny part is that the retailer says: AMD will repair the CPU or send a new one. How does one repair a CPU? π
Isn't it cheaper for AMD to just ship a new one? Instead of trying to repair it. Sounds costly to me.
That is what happens in reality.
How does one repair a CPU?
That's the neat part, they don't
Ok, so my gut feeling was correct. Why is the retailer saying that haha π
Generic line.
I've started to ignore retailer text about products, they don't know shit
The quality control on what gets copy pasted on those pages is non-existent for nearly all online shops
And they will find a way to put it in ToS or somewhere it says that manufacturer specs take priority. Or whatever can get them out of legal trouble.
imagine if someone really was repairing CPU's @ AMD one by one π
Yeah, I think this is just a copy paste and they replaced one word to processor and voila π
Typical shopping experience is finding conflicting text in the tile, description and specs
This is from thermalright official page:
MFs figured out a way to say addressable non-addressable
ARGB is in the product name
π
And I guess can be claimed since they're using the header? Even though it's only to power static leds.
Anyways, little fucks
th
Wait, no argb cable either? WUT?
was there a new X3D in the pipeline?
I asked my friend to go skiing, MF so far has replied:
In theory
In principle(commonly used in our language)
In principle can I guess
I think he has problems saying no, though I haven't noticed it too much.
You have skiing locally to you?
Cross country skiing, yes
Very interesting
Ah, it was the "Ryzen 7 5700X3D" that should get released this year Q1
Already was
Sponsor: Get 10% off Squarespace purchases (https://geni.us/BqEpf)
This review & benchmark of the new AMD R7 5700X3D Ryzen CPU tests the 2024 launch vs. AMD's R7 5800X3D, the modern 7800X3D, Intel's 14600K, and plenty of other CPUs. The primary focus is on the R7 5700X3D vs. the 5800X3D, as many AM4 owners may be seeking one final upgrade before...
β¬ 265,- euro's not bad.
TIL there are points of Estonia further away from Tallinn than Helsinki
I know, its just the fact that I've gotten 30bsods in the past 2 weeks and literally all of them trace back to some kind of issue with memory
Im inclined to believe it
Also lately I've been having inconsistent boot issues where several different .sys drivers dont load
I've tried reinstalling twice on 2 different drives so
BG3: suggests 8gb vram gpu
also BG3:
Windows is getting sudo
ovh have a 120s delivery policy
You're welcome
Steam Deck owners, how much should it charge in 30, 60, and 90 minutes? I have my doubts that this 12v Ugreen car charger can even sustain 65W. After an hour, the Deck went up to 53% from whatever the auto shut off point is. Then I powered it off again, and 30 minutes later It was at 79%.
Dunno. If the deck can charge while playing then it's a good charger
If it's a bad charger it could damage the pmic and then it won't charge at all
I'm just annoyed because the charger says it can do 130W, 100W for C2 and 30W for the other two. But it can't even sustain 100W when nothing is plugged into the other two.
Read through it and this seems pretty useless
Its very easy to just open an admin window, and both scripts and programs have access to various calls to prompt for admin
The only real world use case I can think of for this is if you want to force authentication on a per command basis, maybe in a corporation
?
Mutahar going crazy
Hopefully I can move my old system to the secondary case tomorrow.
And then actually build the new primary in the primary case next week.
Time to soon install Win11 on it and check the firewall software options out.
Why networking so hard π¦
Every time I feel like I finally got it, I open a new tab or change a config and I'm completely lost again and nothing works...
Hey hey hey! I made my first major mistake 
Wanted to delete a service. Autocomplete bugged out on me and now I've deleted /etc/services/

I just used 2 hours to download specific Windows 11 ISO, then checked the checksums matched in three different hashes to ones I got from couple other source.
And then went and deleted the file.
So redownloading it now.
So you just recover those files from automatic snapshot made after last package update?
Right ?)

