#off-topic-tech
1 messages ยท Page 33 of 1
i absolutely love this little gadget LOL ... i dont care if its truly accurate against real world wind speeds ... but its hella useful in comparing differences between different setups/fans/boxes n shit https://www.amazon.com/dp/B07NRSYNW2 ..... it even shed some light on something where I had thought one way was 'more flow' vs another way!
spent a couple days swapping fans, and measuring, and swapping, and adjusting bios speeds, and measuring ....... got the right combo which gives me the most air flow at the same audible noise as other combos. and strangely, NONE of my Noctua fans were good =p ended up on a set of 6 arctic p14s
That is one big boy.
Is that an pc case, or an AC unit?? ๐
Also, I spot a user of an actual physical pciex sound card!
SBZ yes, its been great for a decade now lol
still sounds better than onboard ๐
and also ... yes, my "pc case" is basically an old printer drawer, refactored, and sealed, with an intake filter and 6 exhaust fans which keep the entire thing clean and cool .... many years of tweaks LOL
i guess it does resemble an HVAC A/C unit....
that takes a lot more space than needed, but interesting idea.
I mean, basically a desk with built-in pc
just dont shake it too much
its a pretty heavy desk... not much moves
and yes, it may take a lot more space, BUT ... it really cuts down on noise generated by internal fasn like on the cpu and gpu itself
when closed, i can only hear the cpu fans when its on full tilt (which is rare unless im running p95)
do the fans even do anything with the filter? they look like that'd just suck air from everywhere but that
it looks way too thick for fans
those two fans in the front are only there to direct some flow backwards over the ram/front of gpu... they dont 'suck' from the filter
its why they 'float' forward on the bracket
see, cool air is being pulled in from the whole filter (from the fans in the back of the desk cubby)... so then those two small fans up front, just kindly 'direct' that fresh cool air in two areas heading towards the back
Why do I like the look of the phanteks NV5
Tho I'd need a much nicer gpu
Bc my Lil dual slot card will be so underwhelming in a NV5
or these temps with "case" (drawer?) closed or open?
I'm thinking of buying this case. In the sample pics it shows a custom water loop. Should I prolly see if I could invest in it
Rough airflow
Yeah it was noted it's better suited for waterloops by the look of it
Even those need air to go through the rads
It's a nice unique style tho
It should have some airflow
You can brute force the lack of proper ventilation
I could push pull the rads to keep air going
Small slits around the glass
If it allows you to do that
No point in replacing glass with mesh either if you can't properly mount a rad on it
Get strong fans and it's fine
Can you suggest some strong fans.
I'm prolly replacing that big frontal fan with 2 140mm ones
Not in 140mm
Lian Li Unifans P28/AL120v2, Phanteks T30/D30
Or just simpleton TL-C12C-S
Or TL-S12-S
I already have these
I'll use my back stock
I have like 4 extra
are they 140?
Do they need to be?
I'll find some decent 140mm ones for front
160mm fans is where it's really at 
he did ask for 140s no?
yes he did.
But use the back log of 120mm ones on the rad
and arctic p14 are 140
Bc I have a bunch of em
Bet
contrary to the arctic p12 which aaaarrrreeee?
110mm
death

didn't know they had lcd's
I probably should do a full custom loop
maybe your gpu likes that ๐คทโโ๏ธ
you 100% should
No my gpu does not like being asphixiated
Or however it's spelt
do you know that?
closed is cooler, open not so much since there isnt much general 'flow' going on
but those temps you measured... close/open ?
hot air stagnates more with a fully open air case (without fans blowing on it)
my 3080 ran hot to 83C with or without the drawer open/closed... but the 4080 was peaking 56 when it was open, but about 53 when closed ... not big, prolly just variance
thing is, with the drawer open, those two front fans still 'blow over' the board/gpu, so it helps
damn, was 3080 OC'D?
nope! it was just a hot mofo
yup
holy shit
it would run 50 on none aggressive games (like factorio lol), but would hit 82/83 during 3dmark and aggressive games
the 3080s just ran hot.... 4070 and 4080s run much cooler, because 90% of them are heatsink LOL!
I am not going custom loop yet
weakling
Almost a thousand Canadian dollars if I get the parts for alphacool
come to the evil side
all of this .... is nothing but solid heatsink passthrough. there is no board, no pcb, no wires, nothing ... !
stop buying the most expensive parts they have jesus
I'm not
do it!
yup, my 6950XT has the same.
Problem is if I wanna watercooler my next build it's like 760w of heat to dissipate
ouch
heatsinks from the front to the back
with 3 fans on the bottom
What I said
Watercooling a 4080 and 7800X3D along with the ram. Not cheap
BUT AIR HANDLES THAT?
sure
Gotta watercool that ddr5
Not in that case
yup, same... the extra thickness too now is 'more fins' ... its honkin up pci slots
That case is restricted to all hell.
watercooled ddr5 ??? wat ?
So if I do wanna use it I'd have to have strong intake fans and prolly full custom loop
with a good case, sure it can. My case can handle it with air.
Watercool your voltage regulators too.. those actually get hot
I'll have to find something yeah
I might as well and get a car radiator lol
If it was me I'd do the cpu and gpu only. 760w isn't that much heat for a radiator. A 360 shouldn't have any issues with it.
280 coming in
Then a 280 with good air flow.
86mm thiccc
I am not getting a brick for a radiator
Why not
why noooot
One. It looks goofy as all hell. Two how the actual fuck do I fit that in this
with some effort my man
Voltage regulators not watercooled D:
I mean I can try to fin ones that are smaller ๐ฎ
nope
but you should be able to just use those
they're tiny
@edgy hazel
oh fuck that looks nice
apart from the olives I hope you choke on them
god that crust must be so fluffy and crispy
they should be banned by law in general
Thats some pretty good lookin pizza
makes me want to make another Detroit style pizza
I prefer thin and crispy but I love any
Me too but I've been craving that deep dish style, just prefer the crunchy cheese detroit has over chicago style
I will now
black olives, mushrooms, pepperoni, banana peppers ... yum
This dude gets it
what doesnt belong on a pizza?? FUCKING PINEAPPLE
Nuh uh. I disagree
I'm Canadain and this is a hot topic for us
i'll put anything/everything else on a pizza ๐
Wasint it a Canadian that made a Hawaiian pizza
garlic sauce is the king of pizza
fight me about it
but, youre canadian!
clearly not canadian, hasnt said sorry once yet
right... i thought maple syrup was like americans are with ketchup ... put it on everything
Sorry ๐
We are worse
We have Ketchup flavored chips
we truly are
i like those
the doritos mustard ones werent bad either
if i cant get some, i simply dip potatoe chips in ketchup ... (instead of ranch)
i mean, potato chips are just really crispy french fries
whoa whoa whoa....thems fighting words....not ranch?
My go to dip is mayonnaise with hot sauce
lactose free ranch is gross, so i kinda have to steer clear
I'm gonna hork
That
yeah.... fat free, lactose free, sugar free, ranch ... seriously, who would buy that
And what is the best flavor of peanut butter
Nah. It's not. Or is it?!
...peanut butter flavored?
peanut butter flavored peanut butter is the best !
Like what brand
As they all taste a Lil diffrent
clearly its reeses
I actually agree on that
the extra crunchy is good on sandwiches, but suck for spreading on crackers and veggies
Nuh uh. Way too much shugar
I spelt that wrong
Sugar*
reese peanut butter isnt peanut butter, i dunno what that stuff is
99% sugar 1% peanut butter
I still havent found a good use for reeses pb
Jiff is where its at
Planters peanut butter isint too bad either imo
nor suprised, but didnt know they made pb
I actually agree to an extent. It depends if it's available. And if it's not I'll drink coke
And if that's not available I'll drink pepsi
Ooh that reminds me of a candy I havent had in years
Milk duds
Starburst twizlers?
I remeber the last time I had one was when my grandpa was still alive like 11 years ago
Never had one
the most superior twizler type candy.... none of the imitations, nor real twizlers come close
milk duds are good too but only in small quantities
Mhm
i like responding to "is pepsi alright?" in restaurants when I order a coke, with "fuck no, gimmie a water instead"
one restaurant here got so many customer complaints when they switched to pepsi, that they switched back to coke and put a sign in their front window "We have coke back!"
hydrohommies unite! I'm almost always a water person anyways
yeah, at home i drink a LOT of water... treat myself to just a can of coke on occasion
carry around 2 64oz hydroflasks at work
then drink another 40oz or so when I get home.
Good water is the way
Water has flavors depending on where you are and teh quality and material of your pipes
I am a Instant noodles mf
but what brand?
Mr noodles
Who was it that I talked to abt flashing a new OS onto a android phone
I might do it
I'm glad the pizza could spark such a debate
Already decided I'm having pizza tonight
garlic sauce, peperroni, bacon
One day I bet there will be a war over what belongs on pizza
Pretty Sure Franz Ferdinand said pineapple belonged on pizza
one day i believe a war will be solved with pizza
Do many high-profile breaches lately https://arstechnica.com/security/2023/12/hack-of-unpatched-comcast-servers-results-in-stolen-personal-data-including-passwords/
https://arstechnica.com/gaming/2023/12/ransomware-attack-on-insomniac-games-exposes-employee-data-future-games-and-more/
its more or less inevitable now, with insurance taking the risk costs away from businesses they have less incentive to properly (costly) ensure they secure thier end. Especially when it doesnt affect the bottom line
was browsing store for usb extension and found this.
spot what is wrong with this cable
currently feeling like an idiot cause it looks normal
a normal legal cable looks like this
are usb-a male to male cables not a legit thing?
male-A to male-A are not a legit thing, as far as my knowledge goes.
Right, it's the whole reason USB-B exists, they were afraid of A to A
This is to prevent people connecting two host devices, eg two computers together. because that shorts together the power supplies of the two.
there are so many male-a to male-a cables
mobile phones created a bit of mess with micro usb-A (that has same plug as micro B)
I wonder if USB-C to C being allowed is simply a result of the logic needed to prevent shorts becoming cheaper.
they packed a lot of new things into C
but in its basic form, usb requires one host, supplying power, and any number of slave devices.
And there are some insane nonstandard USB-C uses, too, like the Supervooc 240W charging. And that might go to 300W.
Supervooc 240W is crazy because it's 10A, IIRC.
usb has been misused in myriad ways across the ages
Though I have to wonder if 10A USB is coming anyway.
I hope not. Supplying anything above 50w via that cable is an abomination and should be stopped.
The standard does support 230w officially iirc
It's just at 48v
240
exactly 240w
do you legit just multiply voltage times amplitude to get wattage?
Yep
in dc, sure
that... is a lot simpler than i thought it'd be
in ac, it gets trickier
is ac a sine wave or am i being dumb and misremembering
Techincally, any sort of wave
it is
But sine is standard
I will say, I'm starting to like powering my laptop with USB-C.
And it's the only reasonable way to employ it on a grid scale
Now I just need that small Anker 100W wall wart and a cable.
why not go wireless? only 1-2kw supply, can double as room heater and hope you dont wear jewelry
Honestly, that's plausible for a lot of low power devices.
The bar I'm DJing right now is blasting heat, capturing any amount for passive electricity would be big for a smartwatch. It could probably keep the mixer here going if the mixer had a UPS built in.
can't convert heat to electricity, perpetuum mobile 2nd kind has not been discovered yet
Can you dump wireless electricity without aiming it, though? It would accomplish the same thing as a heater, but maybe some device can use it.
Not effeciently
Oh, the efficiency couldn't be good by definition, sure
The efficiency of wireless transfer decreases exponentially with distance
And it's actually already been tried
One sec
about 10% transferred. need 1kw to power 100wat tv
SimpliSafe is awesome home security. It's really effective, easy to use, and the price is great. Check out SimpliSafe here: https://simplisafe.com/actionlab
Get your Action Lab Box Now! https://www.theactionlab.com/
Follow me on Twitter: https://twitter.com/theactionlabman
Facebook: https://www.facebook.com/theactionlabrat
In this video I show...
I remember an Intel engineer making a video at the top of a building powering a lightbulb with stray power.
I can't find it, but it's different than them powering one indoors
I would guess 10 years ago, maybe more
you can easily 'catch' any kind of em waves if you make an antenna with the resonant frequency.
in fact, if you have an atm card, it already work by the principle of resonance. The terminal powers it up by radio wave. atm card has no battery inside.
If you have an 800/900/1800/1900 antenna and stand by cell tower, you most likely can power a small bulb off it.
did the engineer and building in question have a cell tower near?
It's weird, I can picture the face of the engineer because he did a bunch of videos, but I can't find a single one.
guhhhh ....
windows went bonkers on the motherboard swap ... it just wont go through the 'reactivate after hardware change' ... it just keeps bailing. damnit
clearly like if your motherboard dies, you are fucked, gotta buy another copy of windows
i suspect i'll have to buy a fresh copy of win11 if i want to upgrade my mobo
think it's been 1 too many hardware changes now
i think i officially now have nothing left of my first pc in this one
weird thing is, it worked the last time i changed my board from a z87a to a z690
this time though, its just going through the motions, but keeps erroring
so i did a 'remove device' thinking it would free up my digital license ... but now it just wont do jack shit
i've stuck to AM4 since 2017, so i think my OS key has stayed valid because of that even tho it was win8 originally
mine was win7 originally ๐
i mean i got the milegae out of it, and im a tight wad ... i just dont like having to buy windows when ive 100% lost my original windows license ....
kinda wish i had an IDE to usb cuz i still have one of those old clunkers hanging around
i mean, i cant even install it on the old hardware now to run as a second machine like nothing happened
oh yea if this one feels like fucking off i'll finally switch over to win11
i have so much stuff back in boxes i could almost make another pc
but i'm not gonna invest in a mobo just for a 1700X
besides i'm most disappointed that my 6600XT has been sitting doing nothing
oops
although if i do go for a new platform & have to get a new license, all i'd need to actually build a 2nd system would be a case & PSU
so then i'd probably go for it
5600X & 6600XT is still a decent setup, 6800XT feels a bit OP for my 5600X
its for printers
False since at least 2014, nowadays nothing happens (there is a special driver you can get to enable file sharing though)
okay lets try that again
look in dms I have a solution
if you change every single part of your hardware the reactovation tool wont work
however if you change everything but your boot drive, it will work
then if you want to upgrade drive after that you can
yup, that's what my most recent change was
went from old laptop 250gb ssd to nvme right on the board
i just have amazing luck & i know that just transferring the nvme over to a new mobo/cpu will make it brain fart & refuse to stay activated
so i might be wrong
and i probably am
OEM systems have the license baked into the mobo, i think in the UEFI ROM
tho if i do go ahead & build a 2nd system like that it'll probably go to a family member so i'd just get them their own license anyway
Pretty sure the reason it freaks out when you switch mobo is because theres no more uefi license
nothing appears to work for me, i didnt change my boot m2, it had the same install of windows11 on it from the old motherboard. and entering in my old 'key', doesnt accept it, because i had already converted that key to a digital license last time
frigid
yeah, its the serial number on the board
i dmd you a solution
i know, and i dont like punching things into command line that i dont fully understand ๐
yes which is why i encouraged you to look it up tourself
wasnt finding much info
the website goes to the github last i checked
for anyone who is wondering, this is about massgravel's activator script
its the successor to ms toolkit
not literally but it did take its place
the 1 thing i'd like about jumping to win11 is the default powertoys stuff
Those are Micro-USB-AB, have both A and B compatibility in the form.
So you have Micro-USB-A, Micro-USB-B, and then Micro-USB-AB on phones that can be both host and device.
The -AB ones are very easy to break by trying to insert -B wrong way around.
I have like 5 of those for a pass-through
Not allowed by the USB spec.
Still made and used.
Should be A on one side, B on other.
I didn't say the ones you have are, but they should have been, the maker fucked up.
Same with that thing.
Because like said, A-A cable is dangerous.
Plug it between two computers and you burn out both ports, or even both MB.
For example.
Same reason why A->C adapters are not allowed, to turn A port on computer to C port.
Only turning C ports to A ports.
GRAB A BRAND NEW GN15 MOUSE MAT! https://store.gamersnexus.net/products/15-yr-mouse-mat (black/blue/yellow and available in signed or unsigned!) or consider buying one of our GN15 metal emblem glasses to support our work! https://store.gamersnexus.net/products/gn-3d-emblem-glasses
A lot of you have requested that we run power consumption tests f...
Comparing power usage and power efficiency in games between various CPUs.
Greg 2 amazes me. Intel never had a chance in this sort of test because they're on a node that's basically a generation or two behind Zen 4
Well, Intel 14700K did win efficiency in Blender when power limited to same level.
But just against 7800X3D ๐
The higher-core-count AMDs with the same power limit would still have kicked the 14700K.
Well, they did when you look the numbers from the previous charts.
On ECO mode.
In a way it's cool Intel is able to be as competitive as they are by cranking power so much. But that also raises the platform cost for the user.
amd is better
Depends on thing.
But for general power efficiency when not idle, usually.
The multi-die causes high idle power compared to Intel.
IOD will pull 6-10W at all times.
reminds me of how AMD would try to melt their GPUs with power to try to keep up with nvidia
you don't mean that story where users didn't connect the PCIE cables properly and melted their GPU? To then complain it was the companies fault.
lol no i mean the old RX400/500 series that'd suck down power but even the 590 ended up at like 1070 performance
@twin dew this kit seems to be H16A, is it the same with CL38 vs CL30/CL32 that the primary timings are negligible due to high secondary and tertiary timings or does it go so far that it needs tuning to run properly?
https://pcpartpicker.com/product/kp9wrH/
Cuz it sure is dirt cheap
If it actually is Hynix 16Gb A-die, it should clock and tighten on timings like any other, just "unbinned", but with that kind of settings at stock, it can have almost any current production DRAM chips, not just Hynix.
And that someone has gotten Hynix A on that SKU doesn't mean all the production with that SKU uses it.
Ah, sorry, 6000MT/s @ 1.1V?
Then it must be Hynix A
Just no XMP/EXPO profile.
Just at JEDEC default latencies for 6000MT/s. Like the voltage.
So with increasing the voltage to 1.35V, you get the same as with any XMP/EXPO set with that voltage and same chips mostly.
Just need to manually find the limits.
But should hit that CL32 at least @ 6000MT/s.
So probably something like 32-38-38-50 on AMD could be easily doable. And the third might go to 30 or so
On Intel the last will be higher probably, as you cannot often separately configure tRC.
So yes, almost certainly Hynix A-die, and as the current Hynix DDR5 dies don't seem to have big variance, prebinning made by RAM manufacturer isn't that important.
And that it can do 6000MT/s with just DDR5 stock 1.1V means it must be somewhat good A-die AFAIK.
Oh, sorry, bad wording of my question.
Will it perform like any other 6000 MT/s kit when sticking to just xmp/expo?
It doesn't have XMP or EXPO profile, just JEDEC SPD profile.
So no, as the timings will be much more loose, at that 48-48-48 instead of something like 32-40-40-96 XMP/EXPO set would be.
But very easy to set the primaries to what Hynix A can about do on bad day and run pass of memtest.
Or if you don't tune, you get little less performance, but lower voltages.
So it's kinda slower but only a little bit?
Depends if the specific program cares more about bandwidth, or latency.
Big latency difference, small bandwidth difference.
Games usually care more about latency AFAIK.
But tuning and checking for stability for the primaries is easy.
Or you just so happened to have HEMM
First goes around to replace older stuff, and finally to electronics recycling.
Either that or hand it down to family or friends
Or have it sit in the attic because you never know if you might need it
I'd always aim to have a replacement GPU at hand
Could always set them up as a dedicated server for games
Or a private server
My friend. It's Christmas. Wrap it up. Gift it to someone ๐ It's not a matter of what somebody wants. It's a matter of you getting rid of old hardware easily ๐
Can always make him.
Pretty sure there's a game out there that ties neatly into his hobbies and interests
My old PC is going to my niece and newphew. They don't know it.
btw:
Next generation Arc ๐ - 2024. The 2nd generation is called "Battlemage".
Yeah, that code-name for the second gen has been known for long time.
Was just question if the exec at Intel would shut the ARC project down or not.
apparantly not.
It would have been stupid, but that didn't make it less likely.
Alchemist, Battlemage, Celestial, Druid <- should be all in pipeline.
Alchemist - Battlemag - Cryptlord (3rd generation) - Dragon (4th generation) ... ๐
haha Baldur. FFS was thinking exactly the same.
Those are the leaked names you can find on internet.
Ooooh darn it ๐
The ones I listed.
I made em up. I felt a theme.
nah, they going alphabetic with fantasy stuff, because that works for gamers.
I really don't want a GPU called "druid" fuck that. Sounds way to eco friendly.
screenshot I posted is from Intel during event in Japan.
Yup, but that image I posted was from 2021.
Cool cool ๐ I can only root for intel.
So they are finally starting to talk about the actual timeline for the second gen.
But developing new CPU/GPU as 4-6 year timeline.
So those were all on the pipeline already in 2021 to some extent.
Can intel use AMD/Nvidia technology? Like fsr?
Dutch? Thought he was Belgian?
Belgium != The Netherlands.
Both speak funny and use tweakers
cannot denny that ๐
I don't mind if people get it confused. It's all just a bit overcomplicated.
But, I am wondering. Can Intel benefit from Nvidia/AMD technology?
Because both are adding their techs specific for their hardware. Intel is behind on this or ... ?
So arc GPU benefit from ReBAR
The performance is very bad without it.
So that's intel technology? ReBAR?
Oh no. Nvidia. And smart access memory for AMD?
ReBAR is the base standard technology, that was brought forward from being obscure optional part of PCIe spec to forefront by AMD under the name SAM.
Been a thing forever as part of PCIe, but no-one supported it.
And ReBAR lets CPU use full memory of GPU if needed?
Thoughts on "Belgian flag is just vertical German flag"?
There's this one dutch company which makes machines that turn silicon into chips.
Companies like Intel and TSMC use those machines to make chips.
etc
My area is deeper into the chain, mining equiment in copper mines
iโm so happy, iโve ordered almost everything for my PC, just gonna wait until like after christmas to order my GPU cause i feel they will go down in price.
High-end GPU market is all over the place, rest seems quite stable
Does your CPU have an iGPU?
So yes, it has very basic IGP.
Is it enough, to run satisfactory or no?
Runs Satisfactory on Full-HD with everything minimum at about 20 FPS.
HSIZYAJAHSHAH 20 FPS.
๐ i meannnnn i might use that until they go down, there sky rocketed
I wonder if itโll lag a bunch?
Or do you have current computer with separate GPU?
No. this is my first ever PC.
The IGP in Raphael (7000-series desktop parts) is 2 CU, 128 shading units, at 2200MHz max boost.
For comparison Radeon 6600 has 28 CU and 1792 shading units, at 2491MHz max boost.
0.56 TFlops vs. 8.9 TFlops.
Both RDNA2.
No, just the low FPS.
Have run Taros 28MB save on 7800X3D with the IGP fine.
Just with that low FPS.
It's enough to set up and test your PC before you buy the GPU.
oooo amazing! ๐คฉ Iโll use that
Thanks guys! for your help.
am i crazy or is the 253W max turbo power claim of the 14900K and 14700k just a complete lie
i feel like i have seen the 14900k consistently in the 400W range under overclockless benchmarks
Probably whole machine power usage?
Most places don't have the gear to measure the CPU specific one.
Or based on CPU reported numbers?
now im assuming thats probably whitout undervolting anything so you could probably save some watage there but like still odd
well even the intel utility tunning app reads watage so even for the most basic benchmarking its still over
My thoughts: it is. ๐
Yellow/gold stripe ain't in the middle
That 235w is at stock, once you enable the Intel auto-oc thing(similar to pbo) it'll go waaaay up
We 100% just ripped it off though
Printers use Usb-A to USB-B by design. Square one goes into printer. If somebody makes a printer with USB-A, they messed up. (printer may have an usb socket on front panel for usb drive printing, but it connects to pc via usb-b, always, by spec)
By spec? Didnt know there was an inter-brand universal design spec for printers
Jokes aside theres some printers that use A-A
Some newer models use usb-c
I've never seen that
Yup
Weird ones from china made by companies youve never heard of
Any printer with an a port(while it might technically work) isn't spec compatible and might cause issues
Oh, that I could believe
But yeah in general A-A cables are super niche and rarely used
But not entirely unused
Well, they shouldn't exist, but idiot makers have made devices that need them.
Usb equivalent of
Even one Asus MB accessory had USB-A on the device side and came with short A-A cable...
Thats for backfeeding a generator
Thats for connecting generator to house that clearly shouldn't have one connected.
AKA fire hazard
No, that is suicide cable that shouldn't exist again.
But some generator makers are idiots.
And people are idiots and don't want to do it the right way.
Lets say you have a USB2.0 device (4 pins). The cable is attached to it.
Now replace that cable with a female A port wired the same way
Not generator maker fault. cable maker fault.
Now it has a replacable cable
Should have B-port.
B is for device, A is for host.
Of course there are more realistic standards to implement but it is A use case
No it isn't.
Whether or not its correct does not invalidate basic eletrical engineering
Some generators supply those cables though(again, usually off-brand from China targeting consumers that don't know better)
If you have a generator, you either use extension cords, OR you install a transition switch. Not plug a live voltage device into grid.
Basic electrical engineering means that those cables can kill people
WHAT?
USB spec doesn't allow for A-A.
And neither does it for A-male to C converters.
USB-B exists for a reason.
You're really going to tell me a 5v USB cable is going to kill someone?
Both exist, but are very dangerous and idiotic.
Whos gonna put 5 amps through a USB cable
Referencing the suicide cord earlier for AC power.
Ohbyeah fair
No, the extension cord.
If you plug it into a 5k pc. and it fries usb/mb. Then, yes, somebody is going to die.
But A-A cable can kill the ports if you use it between actual host-ports.
And only host ports are supposed to be A, and all device ports are supposed to be B.
On older / low quality mbs yes
lol true
Its like saying 'plug random two haded cords into your house bc fuses exist'
No, because the whole spec isn't taking into account that someone even COULD plug two host ports together, as it is supposed to be impossible because the cables won't fit.
Even on newer boards, you can permanently blow that port
They're typically fused
Modern USB controllers would detect overcurrent
That was my point
Which is why the Male-A to Female-C adapters aren't supposed to exist either.
Ive never seen one of those
There isn't overvolt, there is two line drivers fighting against each other and blowing one up.
Ok, fixed while I was writing.
Might help, might not.
Usually the overcurrent protection is on the power pins, not the data pins.
And often that is just a one time fuse that blows.
Yeah data line is highly dependent on how robust the controller is
Notice one thing.
When you short 5V and ground on one port, thats overcurrent and can just be shut down, or even limited by current supply.
What happens when you FEED 5V into the current supply from another source? Overcurrent? current flows backwards?
If you are lucky, you will burn the 5V rail on mb, maybe even damage the cpu.
Which again low end boards wjll use low end controller
Almost all implementations use USB controllers in the CPU and the Chipset, no separate ones.
having fuses in house installation is not an excuse for causing shorts.
USB-C has redriver, but not older port types.
?, if yoy plug female into female then the physical wiring reverses on either side which would mean 5v always goes to ground on both sides unless it was implemented incorrectly
aka male to male cable
And again theres shady drivers that enable a-a file sharing most likely by telling the controllers to ignore the overcurrent and just feed data
Im not saying any of this is proper
Its not, at all
If wire is connected straight through, 5v should go to 5v. I would consult a wiring diagram to be sure, if this was a proper wiring to begin with.
Assuming they were made to spec, 5vcc and ground should always be on the same pins right?
Yes, and that makes them connect to each other.
If both ends are wired the same.
Same pinout.
So that would mean due to physical reversal by connecting from end to end you are just shorting 5vcc to ground on each side
Which is overcurrent
no, same pins = same pins in sockets
No, Because Pin 1 would be +5V on both ends.
And Pin 1 of end 1 would connect to pin 1 of end 2.
Which as baldur said would result in the drivers fighting & probably blowing something up
FYI, cat5 cables also connect same pin to same pin on both sides.
Lets say you have a female A port (2.0) and ground is on the leftmost pin if the opening is facing yoy
Which is why various cross-over cables have existed for long time.
To connect two host ports together for Ethernet or serial etc.
If you connect that to another one wired the same, then that left pin connects to the right pin of the other one (5vcc)
And before autodetection, switches and HUBs had different pinout to network adapters.
ethernet is actually valid between two pc, just sometimes needed crossover if crad did not feature auto-mdx
And you needed cross-over cable to connect two switches/hubs together, or two computers together.
Oh my god I worked with an okd hub like that recently
It was weird
no, pinouts are physical and always the same. There is auto-mdx negotation that decides which pair is sending receiving.
Thas is NOW.
It wasn't so on 10Mbit/s for most cases, and only some for 100Mbit/s.
1Gbit/s has it mandatory.
And this is where I do not know much
Also how does POE work? I assume it sacrifices two data wires for power
The 10 and 100 only use two pairs, two are not used.
in one of the modes, yes. other is dc bias.
1000 and up use all.
But they use DC and all network ports have always had separation transformers.
So the DC for power doesn't go over those, only the actual signal.
The big Pulse "chip" between the port and the small network chip is separation transformer.
ofc standard ethernet is not supposed to carry any power, so it doesnt need to be wired directly to controller.
and in fact is susceptible to current buildups from nearby lightning strikes.
The ethernet port pins and rest of the card are electrically separated for ground loop, fault current etc. separation since ages ago.
And that blocks any POE DC bias, as long as the bias is below certain level.
And the 48V is comfortably so.
Some versions of passive POE use the unused pairs in 100Mbit/s ethernet, some use the active pairs.
Then for the actual active standards, the first AF as one mode that uses the inactive pairs in 100Mbit/s ethernet, another that uses the active pairs.
Same for the midpower AT standard.
Newest BT has the those still, and then two different 4-pair mode for highest power.
You can also create a 'dumb poe' by splicing your own connectors to inject X-Volts on unused pairs. I have my cameras powered like that.
I also hate that they put DC+ on BLUE and DC- on Brown.
have they had any EE classes about wire coding?
And on the 10 and 100 those RX and TX assignments swapped around depending if the port was on "network adapter" or "network switch"
Those depend on country, there isn't universal standard.
And even in Europe, they have changed over the years.
Finland has had four different color codes, some which had opposite meanings to current one.
pretty sure most countries agree that red/brown/orange means hot, and black/white/gray/blue is ground.
There is now EU standard, but that is Europe only.
For rest, depends.
Changing something like that because some other country has opposite doesn't make it good reason to switch.
for ac. for dc, I'm not sure there is one
So there hasn't even been attempt at world wide standardization AFAIK.
more of an unspoken tradition. definitely not using red for minus and blue for plus
I don't think there's a universal standard, the most typical I've seen are white, black, or blue for ground and anything else for positive
the unspoken standard, yes. and poe uses blue for positive ๐
For Finland, for long time it was Gray for neutral, black for first phase, red for second, blue for third.
again, thats ac, and those are real rainbow around the world
And network cables had their colors long before POE was a thing.
And two of the pairs have alternative locations.
ofc they did, but why would poe standard makers pick the wrong color for +/-??
Unfortunately it's just common enough for something to break that that you have to check in essentially every case but a red/black pairing
Even then its wise to be safe
Because they cared about the pin numbers more than the colors.
Because you don't care about the cable colors for the pinout side.
And having the "lower" pin number as + and higher as - is more common for that.
It should not make much difference of which one is +/- unless there are some advanced intereferences on host side to account for?
oh, that could be the case.
is it standard or just common?
So it is either 1 + 2 for +, and 3 + 6 for -, or 4+5 for + and 7+8 for -
And I really hate AT&T for that mid split...
At least they had to stop after having one pair in center, second around the first, original plan was for the third pair to be 2+7 and fourth the 1+8.
But too much crosstalk.
When they could just as easily at the beginning of time have added second pair to either side of the first, and so on.
Also same for the two color schemes, A is what AT&T used, B is what everyone else used.
US government still specifies that all wiring for them needs to be A, and almost everyone else specifies B.
A is used in usa and B in europe, or other way around. think my house is wired in B
Nah, even in US it was only AT&T (and via them the US government spec) that calls for A.
But the US government one is just for government buildings.
so, america A and america B? like trump said?
Oh, so crossover cables exist solely to connect government buildings to "normal" networks!
[I'll stop the puns now]
No, crossover was for the old stuff that didn't support the current autonegotiation (MDI-X)
[it was a joke]
yea you can enable that but even off i saw 450W somewhere
and tbh the auto-oc is mid everything turns off at 70C so it kinda ruins the point entirely on chips that are barely coolable
Because the nominal TDP is for long term, the short term boost is basically unlimited.
And many MB and with some CPUs the long term boost is unlimited too.
Sorry, 253W is supposed to be both for 14900K as stock.
yea thats why im really confused
At stock, 450w is a misread
But many MBs default to unlocking the turbo limit.
Yeah
And don't follow the Intel spec.
Which isn't stock
i saw 400-450 fluctuation on a constant bench load
So the CPU never runs as stock unless you change MB settings.
whit the stock 6ghz core clock and all cores active
Point was that unless you change MB BIOS settings, it will on MB defaults remove the power limit completely, and only leave the temperature and frequency limits.
Which isn't stock for the CPU, but is stock for the MB.
MB makers have since time long since past just straight up ignored CPU maker guidance and done defaults that OC the CPU.
so this was odd ... when i was moving my m2's from my old board to the new board... the m2 that was 'under the MSI M2 heatsink' ... had this odd 'gel' on top of the chips that were in contact with the squishy padding on the MSI M2 heatsink
clear goey gel.... which i never put there, and the padding wasnt sticky itself (just a soft pad)
did that gel come FROM the pad, or did it come OUT of the chips ??
ah
the new boad has even more of those M2 heatsinks and pads ... so now all my M2s are under them ... ๐ฆ
Heat and pressure.
one M2 has its own heatspreader cap (a thin metal sheet)... which i cannot remove. the sabrent rocket
Why would that kind of sheet need to be removed?
Many of the markings sheets are actually metal backed for that reason in SSDs now.
To act as basic heat spreader.
i read somewhere (prolly fake reddit BS) "you should remove any extra heat spreaders if you put an m2 under a heatsink on the board"
i dont see why it would be a problem.....
Well, the thick ones don't fit...
the chips transfer to the metal sheet, which transfers to the pad ...
Sticker ones don't matter.
yeah, the sabrent is like a sticker, thin metal
There would be very minor difference without the sticker, but doesn't really matter.
As the total transfer distance would be less.
my gammix blade CAME wit ha metal sticker, but i never put it on, since that one went under the MSI heatsink
and it was that M2 that had all that GOOO on it when i removed it lol
And another CPU core ECC error on Core 3.
So Curve Optimizer -10 wasn't stable for it on 4 core load.
Single error during one hours run with Y-Cruncher on three best cores and Prime95 Core cycler on that core. AVX512+HT on both.
And the AGESA 1.1.0.1 has made the stable CO offsets smaller on some of my cores.
I have been very lucky that - 20 has just. Worked.
Might have not, just haven't noticed the seldomly happening errors.
Mostly the larger offsets work, on all-core or single-core load.
But not on partial core loads, but the errors are so minor that they don't usually cause crashes.
Just very hidden data corruption.
I haven't seen a single ECC error on multi core and single core loads
That or it's extremely rare
Like you say
Because by default large part of AM4 and AM5 boards just swallow corrected errors, and reboot on uncorrected ones.
And most of the stuff isn't ECC protected, just caches.
I have mine sending corrected errors to os
Can be for example one error every few weeks in normal usage.
And uncorrected ones are just an instant bsod
Ahh
Yeah I haven't seen one in the system logs
I remeber seeing a lot of errors when I was pushing 4000mhz 1:1 ram
Like 168 corrected errors in 2min
And one critical error on vram
Probably IF bus errors, on AM4, which actually reports those.
AM5 at this time doesn't.
All of them were something with a bus cycle being missed or something of the line
But point was that most Curve Optimizer negative offset machines are just marginally stable, because there isn't easy way to currently test those "some cores used" cases.
So almost no-one tests for those except you then get the "this game crashes when everything else is stable!".
I haven't had one of those
Just pointing the easy to see case from not checking the mid-core loads.
The only game I've had crash was bc my gpu was acting up and the driver had a microstutter causing the game to error out. JC3 so it's a old game
It doesn't like latest Radeon drivers for some reason
That the "CO offset" current meta is completely missing in OC community.
Mines prolly even less stable bc of the 10x scalar and - 20 offset
@twin dew its safe to assume you can't run 4 sticks of DDR5 on a 7800X3D
You can, but the frequency will be low.
Officially 3600MT/s.
When single stick per channel is 5200MT/s officially
6200-6600 in 1:1 if you are willing to pump the voltages.
My old one could do 6000 on 1.05V SOC, current one needs 1.15V.
With pretty tight latencies.
Part List - AMD Ryzen 7 7800X3D, Radeon RX 7900 XTX, Phanteks NV5 ATX Mid Tower
This is a full setup change
(and cmdr:1T)
Any suggestions or tweaks
Well, better to say that Gear Down Mode is disabled, as the default is that on, so 1.5T.
And I'm with GDM off and 1T.
GRAB A BRAND NEW GN15 MOUSE MAT! https://store.gamersnexus.net/products/15-yr-mouse-mat (black/blue/yellow and available in signed or unsigned!) or consider buying one of our GN15 metal emblem glasses to support our work! https://store.gamersnexus.net/products/gn-3d-emblem-glasses
A lot of you have requested that we run power consumption tests f...
2 setup
one with bad motherboard
the other with me messing the ram slots by overloading channels
while the talk is about ram: would it help to grab high clock rated ram with already good timings for that speed and set it to 6000-6400 for even better timings?
"but Photoshop is much more efficient on Intel so shut up they said they wanted production workstation to do 3d blender so that means Photoshop because workstations means Photoshop and Photoshop means Intel is better so they need Intel" folks going to die with that one
GN: "amd is slightly better even in Photoshop"
rip bozos
@night girder you do development on macos right? I need help compiling something meant for x86 macos but im on arm macos
The high-end kits are mostly about making sure you get the best die type with good bin if you go for that type of thing.
But for current Hynix, there doesn't seem to be bad bins at all.
I'm thinking about upgrading from 32 to 64GB (or 96 if 48GB modules are a viable option)
there are way too many options to choose from ๐
so going hynix is a good call either way, thanks
seems so
They got something magical right with the 16Gb M-die and have been able to keep that and improve on it with the later ones.
shouldn't be a problem, some compilers already do multiple architectures. But it depends on what language.
And Micron and Samsung had problems just hitting JEDEC 4800.
Does this mean anything to you, it means nothing to me
The zic.o, part of nx_tzdb, isn't set to be compiled on arm64, only as 64-bit x86.
Also:
Missing libraries: RenderDoc, SimpleIni, stb, VulkanMemoryAllocator, xbyak, dynarmic, cubeb, Doxygen.
Marginal CMake that is depricated for the specific compile because it is too old.
You are trying to compile the program as Apple CPU native, and that specific library isn't set up for it.
my cpu is getting up to 70C just running crystaldiskmark
If only somebody could've told you intel 13th/14th gen is power hungry and inneffecient ๐
lol
Might be possible to set compiler option to compile for MacOS native x86-64.
its efficient in speed... just power hungry ๐
I know for cemu compiling on macos you can set the environment to be x86 and compile as if it was x86 and then use rosetta to do translation afterword
i dont even think a custom water loop would help either ๐
That's not what efficiency means
It's fast but extremely inefficient
i was joshing
From what I've seen, it probably wouldn't ๐
Maybe ln2
why do something effectively, when you can over do it!
Same idea needed, cannot comment on HOW to do it in this case.
honestly unless intel changes their tech in a big way, i dont know how they can progress with faster cpus
And I was wrong about the CMake warnings, it is the CMake warning that the project specifies too old version.
And newer CMake versions might not work with the project anymore.
i guess since they keep saying "runing all the time at 100C is what they were built for, dont worry bout it" .... ๐คท
The native macOS build is currently purely experimental and should not be considered stable or ready for issue-free gameplay. There are also known issues with degraded performance due to the use of MoltenVK and Rosetta for ARM Macs. We appreciate your patience while we improve Cemu for macOS.
"We appreciate your patience" ๐
I wonder if you can even get it stable if you get it to compile ๐
theres precompiled binaries
I am just posting from their github page.
I know cemu started to work on macos because someone was like screw it and took the linux build instructions and made an effort to get them to work on macos
๐
Oh yea I know
So, arch -x86_64 zsh gives you a arm compiled Cemu? ๐
no
sets the environment to be x86 so you can compile the x86 version to use on intel macs. and then for arm macs just use rosetta
Cool cool ๐
also look at that sweet sweet VULKAN
Ugh Vulkan ๐
mmfmph
BG3 was developed against Vulkan. Alpha. At release Larian said: "don't use Vulkan, use DirectX" - So they had 3 years alpha developed against an API they eventually kinda ditched because too unstable.
BG3 uses metal on macos now ๐คฃ
Now, I don't want to point the finger at Vulkan. But still, something to think about.
Might use the same MoltenVK, so Vulkan to Metal translation thingie.
But as MacOS doesn't support OpenGL, Vulkan or DX, so you either need native Metal, or Vulkan with MoltenVK.
oh also snowrunner now supports native macos with arm native and metal
Just bootcamp it all
Doesn't that make life easier?
๐ญ
Doesn't work on ARM Macs.
Ah, it's one of those features I love about mac ๐ฆ
Only the old Intel ones.
Eh, I mean. As a gamer. Bootcamp on macbook is doable to game. (personal opinion)
all the rest is just too much effort ๐
but you got it up and running, enjoy kade ๐
runs smooth
I don't see frames ๐
it even goes around the notch when in fullscreen mode ๐คฃ
I need to figure out how to turn those on, steam overlay is broken
๐คฃ
also, it doesn't mean that if a mac can run a game, it can run a game. It can get so hot the laptop shuts down or something ๐
138% out of 100% used ๐
yea idk how that works
GPU has a little bit of room left, 5% 
doing not toooo bad
poor macbook ๐
I would panick if I see 95 celcius. But only if it's a real windows PC.
for laptop it seems fine
Buddy of mine months ago wanted to buy gaming laptop second hand. I said that it looked like a good laptop for the price. I warned him that the laptop might get hot. Few days ago, "is it normal the laptop get's insanely hot?" ๐
He asked me if I could repaste the CPU 
Oof
I said sure, but that's not going to help much against burning your legs when you put the laptop on it ๐
he doesn't even game on the thing, just netflix and the laptop lifts off haha
close to 4k for this I aint complaning
snowpiercer a heavy game? hardware heavy?
Mm. 30fps.
I wanna say so
I cant crack more than 144 on my 5600+7600 desktop
I wouldn't complain either tbh, giving the context.
well, see what happens when you game for an hour ๐
I mean for any windows laptop ive used temps jump up immedietaly to max temps allowed by the system even at full fanspin
macbook != windows laptop.
Just curious if the mac can keep those temps over longer periods. Probably can. Just curious ๐
let me download warthunder and report back
My computer is finally balanced for 1080p
5800x3D and RTX 3080
previous configurations:
3950x and 3080
3950x and 2060
i5 9400f and 2060
all had some significant bottlenecks
What refresh rate you trying to push? That's overkill
Or building for the future. Maybe Nosey wants a 1440p next year or in two years?
Yeah, I did the same Nosey ๐ Build an "overkill" PC for 1080p with the idea to buy 1440p eventually.
if not mistaken, AMD fsr 2.0 renders to 1440p. So on 1080 FSR gives a bit of a blurry effect sometimes.
im hitting 400+ without FSR...
Fsr works like dlss... can upscale from any input res to any output
I am speaking optimal here.
"So on 1080p monitor FSR gives a bit of a blurry effect sometimes." - Trust me. This is true.
Oh, that I believe
They both struggle at 1080p
You're a lot better off upscaling from 1080 to 1440 in terms of the quality you get
This is what I read today.
In theory with RSR.
But better to just set your GPU driver to do normal upscaling to monitor resolution from any lower resolution input.
I didn't know that FSR was meant for 1440p and above ๐
Yeah, I play my copy of simcity 3k from GOG quite a lot
watching some odd behavior on this 14700 ... nothing is using cpu (except a .8% for msi afterburner itself) in task manager, yet afterburner (and also hwmon) show that cores 5 and 6 are chewing on something around 16% of the core .... its also causing the "balanced power setting" to keep jiggling the boost clock and idle clock up and down every split second
basically it doesnt want to 'rest' and go idle at all cause those two cores are doing SOMETHING all the time
they are also p-cores, not e-cores
never saw this behavior on the old intel
so odd!
That's probably just misc os processes etc.
they hide themselves well in task manager ๐
weird they have also clung to those two cores and never jumping about
That's part of the new advanced scheduling they're doing for the big.little arch
Similar to preferred cores, but opposite
so i guess setting "balanced power mode" is pointless if windows is just going to jiggle the crap out of the clocks all the time huh
should set to full performance mode and just let it sit at max
uboot ? poor kit
lol took me the last hour adjust msi afterburner graphs to deal with all the extra cores and readings
Didn't someone post that GN video here?
what was once e-cores are p-cores, and now a ton more ecores
GRAB A BRAND NEW GN15 MOUSE MAT! https://store.gamersnexus.net/products/15-yr-mouse-mat (black/blue/yellow and available in signed or unsigned!) or consider buying one of our GN15 metal emblem glasses to support our work! https://store.gamersnexus.net/products/gn-3d-emblem-glasses
A lot of you have requested that we run power consumption tests f...
Nah he doesn't need to get a haircut.
๐
heat and power be damned! i live in murica! we aint green!!! shake fist at clouds in sky
Flight, what is that scheduling you are talking bout?
Windows scheduler now has some more advanced logic that allows it to do a better job of putting tasks where they belong in intel's big.little arch
ok but what is this?
wait what's wrong with gamers nexus
nothing
good, good
That's the balanced power profile agressively ramping the core back down during very short idle time
ok, I guess the cores are made to handle this 
Yeah, it won't hurt anything
they are, to some degree, but power droop can cause instability if overclocking
Theoretically, you could see slightly better stability without it
Don't know why, but it doesn't look healty to keep doing this to cores for long times (years) ๐
all that fluctuating power and clock... can get antsy ๐
In anything but a solid-state system, it would be
i mean, typically, when im just sitting here doing nothing, and the pc is running... it should IDLE ......... drop the clocks, power use, and just sit twiddling thumbs
but something in windows11 is just not letting it happen
yeah
Well, if you bought a normal CPU
Try the performance power plan @charred pewter
if i bought a potato cpu, it would be like idling all the time ๐
uhu, we peasants need to think about โก bill
yeah, i just went and set windows to the performance plan
US power is very cheap
where i live, my electric bill is so tiny compared to our INSANE GAS BILL ๐ฆ so i guess its relative
I'm under 10c/kwh
pay 500 for a CPU that consumes more than other CPU.
Gas bill is also very cheap here
with 20 cores and you probably ever use 8 of em ๐
but hey, now you can render videos!
but given that we used to keep a 150 W light bulb (maybe two, or three) on all night outdoors ..... just eliminating those to LEDs, now i can run my PC full blast and still save money ๐
Or what is one of those reasons every gamer keeps buying em? Streaming.
streaming lol
i should start streaming more
and do game giveaways during them
a next year idea
You sure? You might need a 14900K for that. 
๐ค ... IM ON IT! ๐
@charred pewter this is what usage on preferred core looks like on my chip on performance power plan
two machines side by side, one for handling streaming, the other for web browsing ! YES! I see the light now
This is with many applications open though
yeah, but frigid it goes to 5700MHz and then drops down ๐ Your picture looks normal.
Frigid's chip clocks way higher than mine... that's pretty standard behavior to see
Mine's jumping back and forth between base & boost
oh, yeah, ignore my graph shot, i didnt adjust the 'max' on the graph right yet....
Oh wait sorry, that was on balanced power plan
balanced jumps about, performance just slams it to the top
Oh weird
this is what its now after switching to performance:
What is your total cpu usage according to task manager?
1%
Did you forget to install chipset drivers?
Control Panel > System and Security > Power Plan > High Performance
nope, those are all up to date ok
Huh
when i drop to "Balanced" ... then my clocks jump all over from base to boost ...
That jumping around is more or less expceted
yup
It's just weird that it's not using the e cores at idle
right lol
whatever is churning on my two pcores...... when im doing nothing .... is odd
Still, why is your CPU constantly running at 5700? ๐
Frigid, have you updated bios from what came with your board?
yup
This is very common with 13/14th gen
it was a rev behind
This makes no sense?
It doesn't really
its clocks are at 5700 ... but its not 100% using that
A idle CPU, just running at full throtle? And then go down. And then full throtle again. For idle?
but it's runnin at 5700MHz while it shouldn't!
I mean, my CPU is on an average of 240Mhz. Discord and idle more or less.
Set back to balanced power plan, then go to the advanced settings for the power plan
its like setting your gpu to not jump from base to boost clocks ...
What are the min and max values there?
That's as it should be then
yup
Try enabling EIST in the bios if it isn't already and disabling Intel Speed SHIFT
its just those two cores ... keeping it from actually staying at 5% idling ... and i checked, and it is the p-cores, not the e-cores
checked bios?
๐ A lesson I learned the hard way.
Yeah, try enabling EIST & disabling Speed SHIFT
ive also 'quit' steam ... so its not that in the background lol
We all became addicted to USB-C power at this bar where I DJ. I donated the power strip with USB, and someone else gave the cable.
im wondering if its not just simply msi afterburner DOING that much ... lemme see, how to monitor with turning off monitoring ....
Now we just swap the USB-C cable at the shift change. And Ethernet.
another thing you can do, is safe boot and check CPU usage (to exclude 3rd party software, such as MSI afterburner)
seems to still be doing it, not as much, but still doing it on those same cores (watching with different monitor apps at a time) ...
gonna go look at bios
you running some benchmarks? To put some pressure on those cores?
EIST is on
Intel Speed Shift is 'auto' ...
anyways, seems the issue is just windows doing something on cores it probably shouldnt be which just gets in the way of the system properly idling
im not going to concern myself too much with it, its a game machine, not a sit and hum machine ๐
blasting it with prime95 now
Try disabling
gpu is like 'huh, my boy over there is getting warm... you k bro?'
cpu is sweating a 79C max temp right now ....
weird, thats cooler than my 12600k would get with the same p95 test
my 12600k would get to 98 C
What's the relative power draw?
Also could be a difference in die placement under the ihs or better application of thermal compound
The only part that's abnormal is that it's not using the e-cores at idle
Yeah, just saying it so I don't add to the confussion.
heh 500 Watts peak
yeah, i think i need to get a new PSU
although i dont know any game out there that would use all cores at maximum like prime95
Plenty do during level load etc
Yeah, but not that much ๐
games of 2024 will finally be optimized for mega-core cpus
Star Citizen will release too
Side note - currently shopping for a new phone and I'm looking at a oneplus 9/9 pro but I'm a little worried about screen burn-in since there seem to be a lot of them out there with burn-in problems
Any of you have experience with that gen of oneplus devices?
not me, naw... just got a cheapy droid a10e
sorry, i iPhone.

