#🔌│tech
1 messages · Page 862 of 1
What happens if you shred a device file 
It marks it as overwritable
So dd it is
shred is probably slower, so it may do less damage before you notice and cancel it
i'd assume shred does nothing
'shred - overwrite a file to hide its contents, and optionally delete it'
Overwrite the specified FILE(s) repeatedly, in order to make it harder for even very expensive hardware probing to recover the data.
And modern storage devices have a secure erase function, I've never tried it though
If you want the data off the drive the good old 12 gauge shotgun never failed me
Still recoverable afaik
The best way is to chemically destroy it
Some use thermite.
then load another 50 rounds
Tbh
simplest option is to remove the platters and bring them with you
until eventually you want them gone
The triangle?
they did not confirm the cooler as such but there was a frame of the airflow pattern in the video. and well this pcb pretty much tells you there is a fan there
confirm two pcbs? what second pcb did tehy confirm?
thats from the video nvidia posted today
If they care for cooler design why not up the voltage to 24v to half the amperage?

24v would require a completely new power supply or extra components.
Also I think most vrm powerstages used now do not like 24v but I could be wrong on that one.
Or use the -12v and 12v lines? Oh that might be bad, powersupplies can be uppredictable when power outs are all "suppling"
oh shit my study is failing me,,,,,
ahhhhh
on my psu -12v is rated for 0.3 a
dont know
not gonna go well
powersupplies are not recommended to redefine the ground and have two differend rails being used..
Regardless even if it were fine -12v is a very weak rail on most power supplies because nobody actually uses it
It would mean that current is going the opposide side it would normally go potentially killing it
-12 is for the opamps i believe.
so the audiosystem might use it
seems mostly serial and the real time clock and older pci cards.
According to https://www.acopian.com/acopianPowerSupplies/entry.aspx?nsid=17 it should work but there are some very significant asterist and considerations
We will power the world using 24 volts!
MORE VOLTS MORE ARCS AND THAT IS ARCH FRIENDLY
Over the past 20 years, computer graphics have made huge advancements in performance. It takes a combination of expertise in thermal, mechanical, electrical, and industrial design to make it all happen. NVIDIA presents the art and science of graphics card design today and look...
am i late?
Would be interesting to see what kind of power delivery / motherboard layout we would get if we were allowed to start completely from scratch
ofc
also yes
I would guess that they would use 24v as it would allow for more power, lower noise and compatability with the PLC/industry market.
if they will do it raise it to 48v instead
I doubt they go away from 12v
negative
Dont be so negative 😠
unless you get a AiB card that will just have 2 8 pins as normal
I think my PSU is the only thing ill keep when making a new build
isn't logitcal to buy a new 850w full module
jesus christ you dont need a 850w psu
what if they use a 24 pin to 8 pin? Then they will be able to carry even more of the same amount of current !
do you plan to get a 3080/90 FE edition? if not you wont see the 12 pin on your card anyway
maybe
who knows
probably not though
jesus christ you dont need a 850w psu
I was saying i already have one
So there is no reason for me to change that when i make a new build
and that itll be the only part i keep
If you gift one to me I would accept this, as it would be an upgrade
I dont even know why its necessary for the 3080. Is the sammy 8nm node really that bad
Depends on the difference between the 3070 and 380
I can do 380w on 8 pin
it's not about the powerdraw, its the same as if you would put 2 8 pins on it, idk why people think the 12 pin means higher powerdraw
and again, founders edition only
The 12pin cable isn't necessary. It's nVidia trying to force their new standard.
What else are power connectors for
Which should be obvious from the fact that AIBs have said they use 8pin
More pins mean more ability to carry current not that it must concume more power
Why would you need more current if its not consuming more power
??? where do you get the more current from
Less resistance
how is the 12 pin supposed to draw more power than 2 8 pins if you literally plug it into 2 8 pins
Wait really
2x 8 pin has more ground and less power pins from what I understand
Why on earth is this a thing
If each pin has the same current rating the 2x 8 pins can carry more power.
Because mVidia want to get rid of the two cable standsrd.
Also there are more ground pins than power pins
I personally think the main reason they made this connector is board space, with their weird cooler design on a high powered gpu that becomes pretty valuable.
Too bad their "solution" effs over waterblock users
board realestate, less rma bc low cost psus use 6+2 in low quality, there are some valid reasons potentially for it, idk what their main reason is
https://www.igorslab.de/en/what-is-dran-an-nvidias-new-12-pin-power-supply-no-panic-but-it-is-override-igorslab-investigative/ igor has the most recent tech take on it I think
their 12 pin connector would be a bit more prone to mechanical stress
but yeah, the vertical mounting is, uh, unexpected
Seems to be more of a power consiladation from NVIDIA
I genuinely wonder if the FE will sell noticeably less because of the need for a two slow block. Not that multi GPU users are a huge market.
But they're going to be noticeably bigger than 1000/2000 series blocks.
Considering how widely used 2x 8 pin is I don't think it's a bad move per-se. Sure it makes for an extra connector but in the long run it can be a minor improvement.
It would not surprise me if these end up on amd systems as well over time. Probably not vertical though.
Multi gpu seems to be a niece market nowadays
in the DiY market? maybe, but most sales for FE cards should still be in OEM systems so I dont think that changes much, and OEM might be the reason they do this, similar to intel wanting 12vo for oem mainly aswell
Multi gpu is still used but not for gaming, mostly for rendering / compute
It could just be a BOM thing
bom?
Outside of OEM's including the separate adapter will always be more costly.
Bill of materials (=list of all things on the pcb)
hm maybe, doubt it though
Even for OEMS it's not a big cost anway
One cent is sometimes the reason choose one thing over another
Scale makes cents into thousands
thats why I think, similar to 12VO this might in some way be better for them
It could be that the pick and place machine was cheaper with that connector
I think the connectors on the power cables are more costly.
Also requires 4 less wires
how does it use 4 less wires if you need to plug it into 2 8 pin slots
I doubt thats it
2x8=16 and 16 > 12
for oem's you have a single 12pin cable coming from the psu.
of course if you use an adapter it will practically never be cheaper.
ah you meant if psu's switch to have ja 12 pin directly, I guess
idk, I'm certain we get some reason soon for it and then it's time to look at if it's making sense or sounds like only marketing they tell us
Ah, I see why you got confused
Probably marketing needed some wank to impress the boss with
In the consumer space I think the only real reason right now is board space.
Oh this actually looks nice
They made it vertical which is mechanically not the best move / probably requires a sturdier connector.
if they had room to spare they would not have done that.
But that is effort
and this solution locks the user into us, intensify vendor lock-in
Nah I doubt it will create a lock-in. It's pretty easy to make an adapter. Nothing they can legally claim. Just extra pins on a connector.
Still not sure I will be an adopter of these cards. I'd prefer something more efficient, this gives me comet lake vibes
well its FE edition only aka not relevant for most
AMD: we got stuff, its silicone and does math. Get it
and again, if you heard the 3080 has 2 8 pins, would that give you bulldozer vibes? dont think it would
perfect ad
No just heard the power draw will be high
I actually like the 12 pin connector after seeing it
It is for the cold winters to warm you up 🙂
If the framerate does not match the room temparature then it might be cold or hot
y tho
Are optional mice the ones you can see a light from the bottom?
... optional?
lol
optical just take pictures of the surface super fast or something iirc
so i dont think they would have a light
but idk
Oh ok
My mouse has two version that are identical but one has optical and one has laser sensors
And idk what one I have
i know my mouse is optical and it doesnt have any visible light from the sensor
when sensor does it say the optical one uses
It says "class 1 laser product" on the little sticker on the underside
.... Must be laser then :p
yeah
so i googled it and apparently most laser mice dont use visible light now either
Yeah
Its a laser
Those can be invisible to us
i know my mouse is optical and it doesnt have any visible light from the sensor
@frozen anchor same
First of all, why is this window so obscenely big? Secondly, thanks Windows very cool.
@proud nacelle what window
The check disk window.
It literally takes all the width possible on the monitor for whatever unsane reason.
When TSMC says something like -30% power at iso perf, or +15% perf at iso power, what does it even mean?
Is that for whole chips, single cores, or just best case, like for a chunk of cache only?
It could mean anything without context
Probably thinking too large
Guess it's more something like 30% less power to flip a transistor on, or 15% faster flipping for the same power
@wheat plume laser is red, optical is IR, so you might be able to se a slight bit of red at the right angle
if the mouse came with both than laser might actually be the better option
Technically, all laser mice are optical, so.... :p
Also, is it still technically spelled LASER or L.A.S.E.R.? Or is laser officially correct nowadays?
¯_(ツ)_/¯
didn't expect rebranding when they are doing worse, why not keep the image that most still have still you have something to compete again 
@wheat plume laser is red, optical is IR, so you might be able to se a slight bit of red at the right angle
@old marlin oooohhhh
@zenith sonnet while lasers also are optical they use a diffeerent sensing method
lasers typically detect optical doppler effect
while optical look for surface texture movement
image zooming in discord sucks
He kept arguing iphone boots up in dark theme by default when you first turn it on
So I sent two initial setup videos and a screenshot that you choose what theme too use
The things ppl worry about 
Ikr
He hasn't replied to me since I sent the videos
He went to chat in another channel
Impressive that you spend time on that
4 minutes was too much time
Should be 0
Even MacOS or Android don't boot in dark theme and dark theme made debut in android 10
Typical case of someone is wrong on the internet. (insert relevant xkcd)
Relevant what?
Lel
The only two operating systems that ever booted into dark theme by default was android until version 5
And windows phone
Rip windows phone ;_;
android had no native dark mode until recently
I think before the "Material" look for Android the "Holo" look and whatever was before that was dark by design
:pepeThonk: android had no native dark mode until recently
@old marlin before version 5 did :P
Going to try zen kernel
but don't you have an intel system? 
yeh
I also selected intel-ucode
zen kernel is a fork of the normal kernel to improve desktop usage
... I swear if they named it that after amd announced zen...
no..
its just a fork to improve desktop linux rather than using the same kernel it uses for servers
neglible difference probably?
zen kernel has been around for longer much longer than amd ryzen.
As for performance yea it won't matter that much, only case were it does matter is when running games through wine / proton.
Zen kernel includes the fsync kernel patchset which helps with performance in wine / proton.
Outside of that running a heavy multicore workload in the background will not hurt system responsiveness as much running zen vs the default kernel.
fair then if its around longer than zen uarch
Macintosh Classic. Gonna gut it and turn it into a fun retro PC.
nice
Either a Raspberry Pi or a Nuc. The CRT is like 90% of the insides. So it's going to be ITX ish inside.
With an external HDMI port of course, since the 9" screen is going to be impossible to use for modern UIs 😄
Got a decent deal on a functioning one.
Isn't for me though so not paying ultimately 🙂
when running a pi with emulators you probably don't need an external screen
Older consoles? Definitely not.
Is hooking up the crt hard?
Ah makes sence so you are only using the shell
Yep.
Besides the CRTs in these are known to slowly die at this point.
They just can't keep doing bright and clear contrasts.
Going to be a lot of room with all of the insides taken out. And sold on for parts 😄
still nice
So how are you going to fit the screen (I don't know the availability of screens with the correct aspect ratio / size.
Others have done it before. Supposedly you can buy 9" 10" (forgot the exact size) LCDs from Ali express.
Will have to do a bit of foam or something to deal with the curve of the front though.
Could be fun to see if someone could do a glass/plexi curved "shield" in front of the LCD.
It may be possible to salvage the glass from the crt itself, I don't know if that would look good / be viable.
I'd rather sell the entire innards for repairs than break those. 🙂
If it's flat it's flat.
besides that, tearing down crt's is not the easiest process
Might just put a black piece of card plastic where the LCD is to begin with and run it externally.
Got my iMac for OS9.
Original hardware is honestly the only way to run PPC.
There IS an emulator but it is mad unstable.
I could definitely run stuff from the era of the machine in emulator though.
I seem to recall that Basilisk is very stable.
I've seen a few of them for like 100$ or less.
Some PPC some early Xeons
uh......battery low.... 100%
smh not overclocking your battery to 200%


Sheepsaver is... more stable than I remember. At least with the handful of games I've tested. Might be able to fake the Classic and slighly newer Macs in the "MansenClassic" after all 😄
https://www.igorslab.de/nvidia-ampere-rtx-3080-und-3090-mit-12-layer-platine-und-backdrill-bios-schon-als-rc2-release-candidate-pilot-produktion-laeuft/ seems to be close to ready
urghhh more and more stuff that makes me want to maybe go with Navi21 but amd drivers ughhh
TLDR?
Okay so why does it make you want to go amd
I usually dont like to pay more for a worse performing product that also is a space heater when I just got an AC
Okay makes sense, so Navi better?
could be
we just dont have enough leaks to be sure, aside from enough info that somewhat proves that it can be better
then again nvidia could obviously do the intel thing and subsidize consumer prices with DC sales etc. but especially with them releasing before navi I dont see that happening
in a way it feels like all the big 3 are tripping over something which makes their stuff worse, which, uh, is nice for competition but at the same time makes it not ideal xD
does anyone else have a habit of storing tools in the root directory of a drive? be it on Linux or Windows?
no
shame on me then
i dont really care for adding a ~ infront of the path
like directly in c: or you mean making a folder with all of it in c:
directly in c:
nah
I guess I should stop doing that then ¯_(ツ)_/¯
I mean do you use a gui or cmd most of the time
if it's not a gui it doesnt matter really
Batch files and GUI. If it helps, all the tools I currently have are self contained, as in, a single exe and no other thing.
nasty
Pricing is going to be super interesting. With how RTX hiked up prices till the 16xx cards came out, I doubt they'll be allowed to do that again anytime soon without massive backlash from mainstream consumers.
ye, they'll need to be careful with marketing and how they compare it etc.
so you're saying that the next gen cards might be worse than the 2080ti or something?
noo
Price wise? Yes. Potentially
nah they aint releasing something that then lowers price by 30% bc of amd a month later lmao
the lower end xbox wont
they already lost because they decided to make two versions so every game will be targeted to series s and not x
what do you think will bring more performance with the next gen gpu? nvidia or amd?
honestly hard to tell
Total most probably nvidia tbh
As much as i'd want amd to get more compeitive in the gpu market
I do think nvidia will win based on the leaks / rumors, however amd will be able to cover a much bigger part of the market, and possibly at a lower power consumption as well.
I don't think the 2080ti was not that relavant when it came out so anything that would cost the same or more could be pretty niche
I doubt AMD will be able to beat nVidia in the 2080TI segment for a while to come.
They definitely seem to focus on the more middle segment with a lot more customers. The 5700XT is a decent card for the price.
nah they are aiming for more
you're not getting supercomputer contracts with middle class hardware
then again, intel got one 
Time will tell.
maybe they can then finally afford a couple more driver developers

ngl if they had better driver history I'd say im getting amd for sure this time
I never had issues with amd's windows drivers, however I have not had an amd card in my system past the 7950
I had at least one card of any gen from r9 2xx to vega
And managed to dodge all driver problems
Except r9 fury though
From what I understand having a clean driver install is still very important and could be a major source of issues if you just use the update function.
I did have the update breaking a bit in the past but nothing ddu couldn't solve
Nah
I just always switched to nvidia before amd had problems
And back when nvidia had problems
idk people still have vega/navi issues that thats recent in terms of uarch and still over a decent time that I think it's worse on amd for now
ofc something like cooler mounting on 2080ti can happen again on nvidia side
My biggest driver issue right now which is shared between amd and nvidia is that both don't have a good low power state for setups with 3 or more monitors .
And that nvidia the linux drivers nvidia makes are a pain to work with.
Overal I have never had too many big issues with graphics drivers.
sure, linux idk about :p
If it's not relevent for your uses it's not relevant for your uses. Overal I'll admit it's a pretty niche use case relatively speaking.
The multi monitor is also not that common but that is not a linux issue
it's just that multi monitors requires the cards to run a bit more on idle, but from what I know last time I looked at it neither amd nor nvidia bothered bother to optimize for that.
so they run clocks at idle that are way overkill
that too is an issue, luckily with a bigger move towards a few large gameengines and less custom lower level code this seems to be less of an issue over time
around 2008-2010 it was common for games to flat out not run until either amd or nvidia made a fix
I tend to play a lot of games way after release or otherwise fairly few different games.
It's just too tempting to play another EU4 campaign every now and again.
i havent really gotten any of the latest big releases / installed them, mostly cause i know my system is / was too bad for it
my gpu is still a bit... overwhelmed
i'm honestly suprised it can run rdr2 on lowest.
you can be thankful its a console port so it had to run on pretty bad chips
ye i pretty much imagine i had to thank the consoles on that one 
though i'm sure like a ps4 has more power than my computer 8 months ago
I just hope amd drops a hint when nvidia annouces their thing so we can at least estimate when their stuff drops
yea. would be pretty shitty if we buy an rtx at november cause amd didnt let themselves be heard for a while, and a day after they say "but wait, we got better"
SUPRISE
i would totally go for amd, but the problem is, they simply arnt better than nvidia. the 5700xt is about equal as a 2070. and thats the best they brought out.
nvidia at the time was like. 5700xt. pff. we got a 2080. forget that. 2080ti. whats that? we can upgrade out chips and rebrand them as supers? well, why not.
wait for navi2x :p
there's reasons for why we see nvidia rush out ampere
that means a 2080 loses to a fury bc it doesnt have TrueAudio™🙃
based on the xbox hot chips talk the BVH math on rdna2 is much better than the tensor approach, at least on turing, so it should not really lack in that regard either, as long as its not done via the vendor locked API calls
trueaudio lol. you plebs sometimes humor me. real audiophiles use scout mode from sound blaster.
control just came out on steam, didnt it?
I wanted to play Thief with that fancy audio
However when I played it that option was long removed from the drivers I guess
doubt they will make control run on it even though they could
pretty sure they are locked in with nvidia
nvidia is not paying them to then make a game that will be free pr to show that amd is better lmao
My understanding is that all the DXR stuff is not vendor locked, just Nvidias RTX marketing is Nvidia exclusive
yes and control uses RTX libraries as well as DLSS
RTX libraries 
nvidia 100% has them under contract to not open it up for dxr
If that is the case I trust hackers to make it work..
maybe? but that wouldnt run nearly as well as direct integration so it's not gonna be that amazing of a pr boost for amd
Unofficial hacks tend to be like that
I doubt It, physx on the gpu also never ran on amd after a few early attempts. Nvidia got that locked tight.
that's different
amd never really added a direct implementation of that, and there wasn't really any incentive to mod that to work with amd gpus
at some point there wa a modified physx driver for amd. Nvidia actively blocked that.
♾ = 1|0.
0 x ♾ = 1.
(0 x ♾) + (0 x ♾) = 2.
0 x ♾ = 2
1 = 2
Infinity is equal to infinity.

this a good build?
If you can find one, yes.
why?
O.o
and cheap
Ehhhh.
gaming
Cheap. Not better. But better than a 3400G 😄
Some games can utilize the cores the 3300X does not have.
its either on par or better
3400g is an apu and thus zen+ not zen2 like the 3300x
3300X or 3600.
Grab some 3000/3200Mhz RAM. This will help a noticeable bit since you're on Ryzen.
And maybe grab an M.2 instead.
wat
What did you WAT at 😄

m.2 SSD is expensive AF
m.2 does not really add cost
so what CPU instead?
3300x
3300X if you can find one.
not there
3100x there?
they have the 3600X
what about 3600
3600 non x?
As for M.2.
You can look for 600P from Intel, A1000/A2000 from Kingston or something like that. They're NVMEs and not expensive compared to their size.
I have the 660p and I can agree with that
ignore the 3400g completely...
o
3300X if in stock, if not either 3100 or 3600
The G APUs are not "bad", but you get so much more longevity by paying a bit more for CPU.
yes if it has a G it is an Apu
what are the prices for 3100/3600?
3600 is the sweet spot if you have budget. Otherwise 3100 is noticeably better than 3400g
Faster and has 4 extra threads.
they both quad core
3100 has "hyperthreading"
imagine
And the CPU is a generation newer.
😄
we should not tell her about intel/nvidia naming being even worse lul
nvidias was fine until the 1660 happened
or monitor names
Oh - The case. Might want to look for a different one.
if you go 3100 you could look at getting a slightly better gpu, 3600 if you want more cpu power
dot IE? Ireland?
yup
How TF does Komplett not have 3300X 😄
{insert potato joke here}
maybe it does but not the irish version
Yeah no kidding. Danish one has it.
LEL
Guess they don't share stock 😦
oh and dont get that psu, get a modular one
O
Hang on. Making an example build on their site.
owo


CURSED
The graphics card you have in your screenshot isn't available.



then why make a list now???
me and my friend were bored
then why ask for feedback since the feedback list might be fully out of stock in 2 months 
Everything except the GPU is what I'd pick.
Better RAM - I use these for several machines. Case with similar price and much better airflow. NVME for same ish price. Same PSU
This will most likely be useless once Ryzen 4000 comes out later this year. And GPU prices change due to RTX.
you do?
I dont have ethernet
just get a board with wifi then
we can spend two hours looking what this site has in stock, doesnt matter if in 2/3 months the stock is not the same anymore xD
this is the only motherboard I see with wifi for that socket type
on that site right now perhaps
doubt its going to change much
this is ireland
bet these have been sitting in storage for months already
then this one will be sold and they have a different one that got RMA idfk
still no point hunting down what a single shop has months before you buy from there
wont be buying anything at all if I need to pay the full price for my accomodation
I’d say it makes no difference, she‘s tethering anyway so that’s where it’ll get bottlenecked
Fastest download I ever got was 18mb/s
@wheat plume y u no on pc 
Fancy
Myarr, go for a 3300x or a 3600
A little late to the conversation but pointing it out non the less
And when they arrive, buy a 4950x
I've been wondering
ASML seems to be the lone supplier of EUV equipment to the semiconductor industry
Why is that monopoly accepted?
https://wccftech.com/nvidia-geforce-rtx-3090-custom-models-listed-online-colorful-vulcan-x-asus-rog-strix-oc/ how do you have this many different pricings in a single article and consider it valid to talk about it, idk why im even linking this clickbait lmao
idk why im even linking this clickbait lmao
And I'm not gonna click it
yea it's wild
I got the MSI B450 GAMING PRO CARBON AC its working fine (so i could have wifi)
nice
intel wifi even.
is there any other wifi chip maker for consumer boards?
broadcom i think
could be I guess, never really looked at it, I know for lan it's only intel and aquanita basically
ye
and broadcom used to be one of the companies for high end mouse sensors 
was a really weird deal
the og broadcom got bought by avago
the combination of the two was called broadcom
even though avago had like twice the capital etc
I think they're somewhat fine nowadays?
We're benchmarking the input latency/lag (end-to-end system latency) of AMD vs. Intel CPUs for Fortnite, CSGO, Rocket League, Overwatch, and more in this test of the best CPUs for gaming.
Sponsor: Get 10% off Squarespace purchases (https://geni.us/BqEpf)
Part 1: AMD Ryzen vs....
my previous motherboard also had a killer nic. if you used its traffic shaping (on by default ofcourse) it crashed quite often -_-
couple years ago a friend had killer software still stuck on the pc, we found his internet issues once I found that on the pc lmao
lol
if you didnt install the killer trash software it was fine
on my Z170A G45 I had killer
if I installed the drivers I had to get some stupid software with them
whenever it needed an update it put all my ram in the unallocated pool
and I had 100MB to update
with the regular windows drivers it worked just fine
idk what the software is even supposed to do
make the system unstable for that overcloked feeling
lol
"feels faster"
ig
my brother had a broken 6700
with a similar problem
if it went in the low power state it never got out
and then managed to have an even lower power state
and so on
until it reached a multiplier of 1
aka 100mhz
and intel consumer support is basically nonexistant
GN talks about plugging the mouse into the USB ports that are directly linked to the CPU, not going through the chipset
Does it really make a difference?
yes
I think I did that intuitively
if you want to go for lowest possible latency it does
But they don't provide any benchmark numbers on that
well they usualy are at the very top
where ps2 used to sit
so ig a lot of people do it instinctively
iirc was about 3-4ms
last time I heard about it
could be different on different architecture
Lol
It's weird
`Integrated in the CPU:
2 x USB 3.2 Gen 2*/Gen 1 Type-A ports (red) on the back panel
- For 3rd Generation AMD Ryzen™ processors only.
2 x USB 3.2 Gen 1 ports on the back panel
Integrated in the Chipset:
1 x USB Type-C™ port with USB 3.2 Gen 2 support, available through the internal USB header
1 x USB Type-C™ port on the back panel, with USB 3.2 Gen 2 support
1 x USB 3.2 Gen 2 Type-A port (red) on the back panel
4 x USB 3.2 Gen 1 ports available through the internal USB headers
Chipset+USB 2.0 Hub:
8 x USB 2.0/1.1 ports (4 ports on the back panel, 4 ports available through the internal USB headers)`
Wait
The IO shield reads 3.0 and 3.1
And that list is all 3.2
lol
How to you do code blocks again?
Yeah don't you just love that silly USB naming
There, better
Do a triple backtick if you want the fancy block ooh fancy!
2 x USB 3.2 Gen 2*/Gen 1 Type-A ports (red) on the back panel
* For 3rd Generation AMD Ryzen:tm: processors only.
2 x USB 3.2 Gen 1 ports on the back panel
Integrated in the Chipset:
1 x USB Type-C:tm: port on the back panel, with USB 3.2 Gen 2 support
1 x USB 3.2 Gen 2 Type-A port (red) on the back panel
Chipset+USB 2.0 Hub:
8 x USB 2.0/1.1 ports (4 ports on the back panel)```
A little cut down
Now to find out which is which
But it seems like 4 of those have direct CPU links
the 3.0 ports are not mentioned lol
Just not the 2.0 ones, which are useless for anything other than mice/keyboards nowadays 🤨
(USB 3.0 was renamed to USB 3.2 Gen 1 because reasons)
But it doesn't mention blue ports at all
And I've got 2 red, 1 blue and 1 blue with white tongue
my guess
What mobo is that?
you mean this one? https://www.youtube.com/watch?v=4WYIlhzE72s
We're benchmarking the input latency/lag (end-to-end system latency) of AMD vs. Intel CPUs for Fortnite, CSGO, Rocket League, Overwatch, and more in this test of the best CPUs for gaming.
Sponsor: Get 10% off Squarespace purchases (https://geni.us/BqEpf)
Part 1: AMD Ryzen vs....
ye
I wonder if we get that in article version a day later so I didn't check it out yet in detail
anyway input lag of amd vs intel masked by framelatency + peripheral latency, more news at 11
looks like some logitech ?
I think you are looking for something they were not looking to answer with this
this is the same camp that still argues one side is "smoother" bc of lower latency whatever
"amd ryzen vs intel system latency" kek
either way I'm certain we get a follow up on this either in text or video form, because reagardless of if this is accurate or not, we know how hardcore some people get ragarding this topic lmao
they basically wasted 10s of man hours for this and they could've just looked at frametime avg charts
god people are stupid
What is interesting though, is that rocket league has half the latency of OW
that's a bit of an embarrassment for blizzard 
RL after all uses a public engine, and blizzard couldnt make that better?
source at least is, uh, am bit older so I can excuse that
tru
Actually GN's method for OW is flawed as fuck?
He didnt say anything about enabling high precision input xD
so if they did not, lol
high precision input is 1000hz sampling for mouse clicks in the game, while when it's off it's running default framerate-dependant input sampling
so basically -4 to -5ms from his test results
if he got 200fps+
Is that a new setting in OW?
about 1 year old
Ye uninstalled that crap before that
Ideally I would’ve like to see them do it in diabotical but I think the beta servers are not up atm
Looks like Facebook might be joining Epic 😄
.. to think I’d say Facebook is on the right side this year
Ooft
@ashen spindle Also imagine not looking when the packets leave network to overwatch server when testing system latency. Client is predicted but server is authoratitive 😄
Does wireshark show when packet leave from your PC?
Im pretty sure the bot training is not a server?
it is in overwatch
I remember playing that at points when internet was down so I doubt it unless that changed too
says I got like 60 ping there so shrug
coud be local
Still what matters is how fast packets leave from your PC.. Which is what you should measure for system latency aswell. Would have been interesting to see results from chipset vs ondie IO
He even talks about it but doesnt test it
intel: mouse -> usb -> chipset -> dmi -> cpu (make a frame) -> dmi -> chipset -> LAN
AMD : mouse -> usb -> cpu -> LAN
no im pretty sure that was IO die as in zen2 dies vs integrated in monolithic chip, not chipset on the baord?
sure you could do all that and that are valid/interresting points, but I dont think for showing amd/intel specific dont have gigantic differences that's what you need to do, you can't make it a 2 hours video lmao
yeah well guess he isn't interested in the topic in scientific manner
but now everyone will scream their heads off intel has lower "SYSTEM LATENCY" than AMD
wut how
just because Overwtch/CSGO/Fortnte ran 10 fps higher on Intel than AMD
then you also get into parts with, if intel runs 10fps higher and gets you lower latency, does that not count? that's like insisting on 480p cpu testing to show cpu power and fully remove gpu need from the system etc., sure that's a nice technical deep dive, not relevant to consumers though
(I think) this video was aimed at people thinking amd or intel are smother and want to buy/recommend one of them bc of that and not people who want to go into the tech specs of how it works, that's not what GN does in any video really
Sure it counts, but you have to deduce the fps advantage from the timestamps of when the packets left your LAN... Because in Online gaming the one whos packets hit the server first win, and if doesnt aling in one cycle of tickrate.. It goes to the next server processing tick..
Maybe
I wouldn't mind a absolute deep dive from like anandtech or some other outlet, but I dont think any channel that youtube-first is out there that could really do that atm
I think all that his testing show is that latency depends on frame rate
Data of which he does not even include
It's great that they out in as much work as they did
I just think he should stay in the benchmarking stuff and leave latency test to people with knowledge
But it's a long shot off the quality Battlenonsense has
He has great benches + reviews
But his stupid latency test do more harm than good for consumers
I feel like for the audience they get this is enough though, currently amd/intel doesnt matter if the rest is good enough
True. But could have at least tired to reduce the variance of measurements. 11.2/28ms (min/max) respectively
jesus the variance is higher than min amount
16.8ms vs 11.2 ms
could do a lot more sure, but then you get into making this about the latency, and not about latency on intel vs amd and that's the focus of the video
and imo we wont ever see that change, cpu choice should not make a difference on latency in that regard, aside from engine that tie it to framerate I guess
choose a mouse with 0 debouncing delay, a monitor with lowest lag, capping fps, using high precision input, making sure the game is cpu bound, etc, etc
i am sure this is insignificant for home usage
@fair gull you're always downplaying significance in everything.
i can smell the "but comptetitive games tho" coming
@fair gull u so cool bro
the point was, if you want to buy a cpu for gaming, intel or amd doesnt matter if you are thinking about latency even as a competetive gamer
@fair gull no need to discuss if not interested
the time were latency actually matters in big tournaments is also not were u as the player worry cause usually the pcs are build for the tournament and all with the same parts
id say for home gaming the skill ceiling of games is the bigger hurdle
@ashen spindle Yeap. That's a nice conclusion tbh. Just doesn't aling with the title of the video IMO.
ehhh, now we get into the pedantic area but sure
Guess I just expected more and am disappoint. But I am not the target audience of the video.
which differences
see now we are already onto something that the video wasnt about xD
the whole io latency stuff
this is smth that enthusiasts care about but i am almost sure the average consumer and even high profile gamers dont
they buy the newest best parts and thats it
sure but it's not like it hurts, even if you can't notice a difference, it's not like you will start to miss more shots with lower latency, even if at best you only hit one more shot every 20 hour bcs of it, not point not wanting to push it better for everyone etc.
exactly
most people care about fps
ik that cause most people who ask me to build a pc just wanna know how good games run
for a given price usually
like "gib pc for 50euro needs to run crysis with 400fps"
the video wasn't about latency itself, the video was about latency in regards to amd vs intel, there it doesnt matter and they showed data for that, not more not less
so then i dont get the point that dag tried to make
gn isnt really the indepth diver
more like average reviews
we already figured out he wasn't the audience for the video, he even said it himself :p

@ashen spindle Do you know if a game can sample an usb input before the CPU has prepared the frame, or does the frame need to be fully prepared before it can sample that?
I think there always has to be 1 frame prepared before a game can show it, but might be wrong
just run input on a sperate thread and its fully independent
ye ik
but it just reduces sampling interval
Where does the game thread get the mouse input from? From a prepared CPU frame or USB controller buffer?
y
I mean, however you implement it
most common its getting input the frame it prepares everything to keep it simple
look up if Diabotical has some tech talks about how they did they input system somewhere
they are running it multithreaded
I know input is sampled just before a frame is displayed usually to guarantee low lag, but I do not know whether it requires a prerendered frame or whether it can be get straght from usb controller buffer just in time to send it off to GPU
Cause apparently you always have to have 1 pre-rendered frame in modern graphics pipeline
fully depends on how the game implements it, it's just an electric signal that comes in, the engine then has to deal with it, there is no set "it has to be x way"
ULTRA low lag is 1.5 prerendered frames and "1" pre rendered frame setting is actually 2 pre rendered frames as per AMD driver designer
if you were to go without pre-rendered frames that would be "racing the beam"?
idk thats too deep in gpu for me
okay fine
The GeForce RTX 3090 comes with 24 GB of GDDR6X memory running on a 384-bit bus at 19.5 Gbps. This gives a memory bandwidth capacity of 936 GB/s. The card features the GA102-300 GPU with 5,248 CUDA cores running at 1695 MHz, and is rated for 350 W TGP (board power).
New leak
Uhhh
I think ive read about that 2 weeks ago
Atleast bits and pieces
Theres leaks everywhere
At least the frequency and TGP are fresh
Also, if some1 can explain. Why does this graphics card have ddr6 while we still got ddr4 ram
Ok
Meanwhile, on the CPU front:
https://www.techpowerup.com/271496/amd-zen-3-based-epyc-milan-cpus-to-usher-in-20-performance-increase-compared-to-rome
Bodes well for consumer Zen 3
gddr is optimized for throughput, ddr for latency, in ram you need a lot of random access, on gpus you need ot move big pictures etc.
its all nand flash but you're not putting your ram stick in your SSD port either :p
i dunno, i did see youtube vids here and there where people used ram memory for disc space
thats ramdisk
Landlord left this tv wall mount, does anyone know how am I supposed to use it? Lol
screw ur tv onto it
Do I have to screw it directly into the tv or do I have to buy railings?
Thank you, I just need to find fitting screws now
hope u can take off the part that screws onto the tv
usually its screwed on and the tv then hooked into the part thats mounted on the wall
I can screw the two holes on the right but I'm not sure how I'm supposed to do with the black ones on the left...
Yes
cause it seems the black ones are the bottom screws
which prob are bigger to have a higher shear strength
Yes indeed
so they can support the tv weight easily
I see
Are the black ones made of rubber/silicone/soft polymer? if so, they could be for vibrations, expansion or possibly even for over-torque mitigation for the screw. In anycase would not worry about them and just screw them in, if instructions dont explicitly tell you to remove them. Also use a washer between the mount and panel. Also handy to measure if the holes of the mount and display line up.
https://videocardz.com/newz/zotac-geforce-rtx-3090-trinity-pictured lul zotac was the one who first had rtx 20 pictures leaked too iirc
is it acidentally leaked or "Acidentally" leaked?
Nvidia is just permanently angry
no 3090ti?
so far it looks like the 3090 might get the Titan die right off the bat
What about a RTX 3090ti TitanX limited edition over usb 2.0
no like, the best die you can get
That would be a quantum coompuutor
the only way to get an upgrade over the 3090 would be them switching to a better node so it's not just a less disabled die like turing super cards
everything is ""accidentally"" leaked now days
nah
Whoops did i drop something? What did I drop? Dont look, cus its secret! It would be soooooooooo bad if you would see this. oh oops just acidentally uploaded it and keep it up untill media picks up. Then hastely remove it again and act like there was nothing.. but dont forget this is secret
ofc there are pr leaks that are on purpose, doesnt meant every leak is something they want
i dunno, sounds like free advertising to me
if you do it on purpose you can select what you leak
so why would you leak something that is not a good look but also not giving you pr from being super bad
Leaking stuff is a great pr as it creates ghossip aswell which people love and its hard to recreate otherwise. Just like creating fake outrage for those rare ghossip points
yes hence why some leaks are definitely on purpose
that doesnt meant there are no actual leaks anymore
i feel like the memory seems a bit small for what i'd expect the high end to be these days
ok tahts bad then
yeah midrange replacing a 2080ti is a bit above what's standard
yeah im not touting it as fact
no like, any leak is useless
so it turns out the new 12 pin connector is only slightly bigger over an old 8 pin connector. i think more companies will adapt that new standard.
any speculation for price makes no sense
i mean the heatsink leak wasnt wrong
yeah but you cant get a picture of a price lmao
price can change very close to release, i know
they can change the price when jensen just wants a different number, you cant just change a heatsink 10 minutes before announcement
https://videocardz.com/newz/nvidia-ampere-ga102-rtx-3090-3080-gpu-pictured much more important, nice pics of a very full board
I just wanna go amd but scared about drivers being bad
same honestly
amd got better node so idk
ehh node means nothing and isnt confirmed yet
turing vs rdna1 was 16 vs 7nm and still won
i have no idea how tsmc is gonna supply a simultaneous next gen console launch, zen 3 and navi2
i semi assume huge shortages
Yeah prob.. Gotta get zen 3 on the 1st day
navi2 might be different node
nah that was nvidia shifting the dies they sell
we aint getting that again till later if at all
imagine the feel when opening csgo/ow first time on zen 3.. when coming from i7 2600k
xD
my guess is still if we dont get hopper next year we get the super version on TSMC over samsung so it's new chips
im sad that navi 2 launch will prob be too late for wow shadowlands launch 🙂
ehh I dont think it will be after
im pretty sure they said big navi will be out before the consoles
are you sure? maybe it's an iGPU setup currently xD
Pricing is hard to predict and NVIDIA would gladly sell them for 1 million dollars if they could. They can also sell them underpriced if they want, (with expensive periodic refuelling[razors and blades model]).
nvidias margins are huuuuuge
the lower RAM sizes makes me think they want to subsidize prices a bit by reserving that to quadros
or rather, makes me hope than think lmao
the mem is supposed to be faster tho, not really sure what mem reqs are these days anyways, im not on 4k so i prob dont need to worry at all, i know skyrim modded used to take up insane memory
higher resolution needs bigger VRAM, speed of it is secondary
God I hate bloated gamesizes
they might move to using the system ram instead of vram
ARK survival was like 350GB on my ssd until i moved it LUL
? a pc is not a console, that doesnt work
they dont share ram/vram so no they aren't xD
2.4ghz is pretty bad for high speed internet
anything under 1000 MB unacceptable
this message brought to you by your local ISP
upgrade today!
US ISPs are a joke
make sure you get the gaming package 
yea, that happens when cartels get formed and they decide who gets what exclusivly
my phones is connected through ethernet
DRMs are a expensive way to annoy paying costumers
that is bad...where do you live?
lmao
I'm pretty sure this was not allowed as part of the first ruling lmao
I thought the ruling was that the other accounts were innocent, but if it's just the one that does Fortnite it's fine to end it
ya
basically
unreal engine is ok to keep running
but all games epic published to the store are dead
ye nvm I thought this was the unreal account
pmuch what i expected
ye this makes sense
so in the end
epic achieved nothing really
they could have
pmuch
just talked about it publicly
like "we cant reduce price on ios due to apple tos" or smth
not lost fornite
and fought the same battle more or less
Apparently, Apple also blocked an update to the fb app
That wanted to include transparency about how much of a cut Apple takes from support of people for Covid struck buninesses
if epic would've only said something, nobody would've cared, microsoft wouldn't have joined in, the wordpress app thing would've not been talked about
and yes facebook also joined their side now
@ashen spindle ya but instead of splitting the rage as it is rn they could have just made people angry for 1 thing
but people would've not given a single shit
seeing they can buy vbucks for x on any platform but ios where its y?
idk
psure that would.make the rounds real quick
just lower the price on other platforms so that the ios price is otherplatforms + apple tax
people get angry
you cant
u point to apple contract
thats against apple tos
Then don't have any in app purchases for idevices, and stealthily redirect them to say the website
thats not allowed
the terms are that you can not have it cost more on apple than on other places
@crimson wigeon all inapp and app related purchases have to be made through apple pay
As in don't have the buy button at all on idevices, unless they don't like that as well
they dont like that aswell
unless you fully remove the mentioning that you can buy something but not via the app
no indirect mentioning etc.
Yes
even then they are pushing for you to include in app stuff
you basically cant circumvent it in any way if your name is not Jeff Amazon
wait
werent the fortnite apps all using the same account data?
as in
1 epic account has the same fortnite data on all devices
yes you can but online and apple doesnt get the cut
but in app you are not allowed to mention thats possible
and not allowed to link to it in app etc.
I mean that just can't work, you could not have a youtube app on ios then either
it said all purchases pertaining to the app or digital items for use in the app have to happen through apple tax machine
Epic appears to be in no fucks given mode anyway, so why not stir the pot by putting Forknife and friends on those jailbreak stores
just got an sms from the "tax authorities". i got a last notice on a bill of 700 bucks. now, i know for a fact that those lazy asses wont work outside of office hours XD
Tell em that you got the 700 dollars ready and to send some coppers to collect it
coppers wont be able to get them. the dutch version of 4chan can probably find out who it is tho
7nm now so not Samsung?
The 3090 might be DC rejects, like the VII
Also, first Zotac, not this
The green/turquoise colour scheme seems to be a common thing among those cards
https://www.caseking.de/intel-core-i9-10900k-comet-lake-limited-avengers-edition-boxed-hpit-676.html
i think every graphics card from the 600 series onwards is gsync ready
IT'S REAL 
no way
yea a fancy box
yup
at least msi and others are bundling it with the avengers game
but not intel itself lmao
same price, same everything aside from the box


yup just a box




























































