#off-topic-tech
1 messages ยท Page 95 of 1
I always gamed 1080 on it. And she already tells me fornite is lagging. But I haven't looked at it yet.
She thinks a better monitor will fix the lag ๐
but at the moment she is gaming on big as TV screen, so it's horrible.
im sure il get an answer but uhh what are the squished/mini/half sized ethernet ports called
so i can see if i can get an adapter
Ah yes, and a super outdated piece of pixelated tech will fix that.
What's her current monitor then?
Bro, she is 15 ๐คฃ
"big as[s] tv screen" is not a monitor
She has the PC for a year.
๐คฃ
Tell that to my sister, she is cheap skate.
even a 10 year old would know better
And "lag" is not clear
and I have to solve the problems. And pay for it.
oh hell nawh
I know, but I didn't have time to look at it, she told me when I was going home.
she should have to pay at that point
That's 20% of her life that this monitor has been out for - MINIMUM
my sis? Nah, she had an old TV in her basement and that's now my niece her monitor, it's really bad. I've seen it.
What's the CPU? What settings?
i7 7700K
specs?
4 cores - 4.20 GHz
im saying total build
Intel i7 700K - Asus Strix Z270E - Asis Geforce GTX 1080 Strix- CoolerMaster PSU 750Watt - Samsung 850 EVO 250GB
I also added a extra m2 of 1TB to it.
i mean listen
it a lil old
but they dont make them how they used to! - some old guy on the streets
whats with monitor names tho
Don't ask me, it hurts my head ๐
If the screen has a curve, some add a C to it etc.
Asus 1139URC91GP
27 mostly refers to 27"
acording to what you said
you are smarter than 90% of pioneers
Aah, fixed it. I shouldn't have filtered min 144hz and max 144hz too ๐
Can just as easily be CPU or RAM bottleneck on the lags
Because you're artificially limiting to 144hz which is stupid
No, I think it's the settings.
Go 144Hz+ and sort by price ascending
I did ๐ I noticed.
โ๏ธ
Stuff looks much sharper
Console kiddies.
ikr
They grew up with console, so they never had to mess with "graphics".
There won't be a single good monitor with those settings
1080p is good for game like fortnite
Based on what? Please link data.
1440p is if your just playing regular games
Pixel density
It's about the combination of size and resolution
That's what she plays sometimes.
You know, I have a 27" / 1080 monitor. And it's really good imo.
It's a AOC 27G2U
"imo" is the key part here
me pretending i understand
The only indication this gives you about other 27" 1080p monitors is pixel density and only pixel density
I don't see any pixels from where I am sitting.
Ever heard of AUO, BOE, Innolux or KTC?
So, the whole reason you would say 24" 1080 > 27" 1080 is because of pixel density?
Yes
me on 1920x1080:
- Your viewing distance isn't hers
- Your eyes are older then hers
- What works for you may not work for others
Resolution isn't the issue. Pixel density is
I too use a 1080p display
On my phone. Where it's less than 7" in size
imagine if flash bangs were hdr ready
i think we would all be in a walking stick by now
I have a 24" 1080p to my left (iiyama). My main is 27" 1080p (AOC).
She should come over here and tell me what she prefers ๐
?
Bro, the pixel density should be? You literally posted it here.
If she can't see any pixels on my 27" and she likes that size. Then it's done deal.
It's about viewing distance
me on a laptop:
If she likes the 24" more, then it's also a deal.
And then she's at her desk and sits much closer
You go and sit closer
That's why you go bigger.
You never go closer to bigger, always further if it's an issue ๐
Tell her to sit closer if she complains about size
She's going smaller either way
So get her smth nice
That's what I am trying to do...
but I don't want to dissapoint her either.
AOC 27G2SPAE/BK Zwart for 139 euro's ... ๐ค
@dire igloo I called my sister.
And told here everything we said ๐
That 24" is better because of pixel density, but that I personally, don't mind that and use 27" because of the size.
Then stop insisting on outdated 27" 1080p shit
That's just your opinion.
No?
It's not because data shows, that pixel density isn't optinal, it's unusable.
It's anyone's opinion who stayed up to date in the field of monitors
If we talk about a perfect monitor, than 24" - 1080 seems, based on data, better.
You believed that a three years old "recommendation" from a purchasing guide was anything to hold value to
When I selected this monitor, it had good reviews. (27" - 1080p)
Define "good reviews"
The monitors I bought also had good reviews, majority of them were still shit
Pixel Density: The 24" monitor has a higher pixel density (pixels per inch, or PPI) than the 27" monitor because both have the same resolution (1920x1080). This means that the pixels are more tightly packed in the smaller screen, resulting in a sharper image. The 27" monitor will have larger pixels, which can lead to a less sharp image, especially if you sit close to the screen. Viewing Distance: If you sit further away from the 27" monitor, the difference in sharpness may not be as noticeable. Larger screens can provide a more immersive experience, especially for gaming or watching movies. Use Case: Depending on what you plan to use the monitor for, the larger screen may be beneficial for multitasking or gaming, while the smaller screen may be better for tasks that require precision, like graphic design or photo editing.
Cuz people are incapable of knowing what black smearing it, how and where it pops up and why it's there
and I tell you, this one is really solid. Believe me or not.
- Pretty sure that's ChatGPT
- There's zero actual information contained in this, just 1+1=2 typa shit
there are probably better ones out there ๐
I don't believe you because everything you've been saying about monitors so far leads me to believe that you don't know how to judge monitors
Well, it gives the same information as you. Only with two extra points. A more nuanced answer ๐
The only thing I trust you with is spotting obvious flaws like pixel errors or a bad case of ghosting, neither of which I suspect this monitor to have
Which two extra points?
The "immersive experience" is just "bigger FOV = better" which isn't always true. Image sharpness and a whole lot of other things contribute to immersiveness too.
And the usecase blabla about productivity? Use a second monitor
You have the budget for two good 24" 1080p monitors
And no, smaller doesn't mean better for precision tasks either
It appears more nuanced because it's just throwing shit into the mix which doesn't matter - and also requires A LOT more nuance than ChatGPT is giving it
So how does it handle blacks? How bright is it? What % of the color gamut, how accurate is it out of the box
Watch this and never ask ChatGPT for anything that requires nuance or factual accuracy ever again
https://youtu.be/QrSCwxrLrRc
Thanks to Fiverr for sponsoring this video! If you are curious about AI and are looking for a place to bring your project to life, head to https://fvrr.co/answerinprogress to check out the services available to you and use the code ANSWERINPROGRESS for 10% off!
How intelligent is artificial intelligence? In this video, Sabrina answers that ques...
That's the review I read back then. Sadly, the website stopped writing reviews ๐ฆ
Pixal density is a function of resolution and screen size... which makes it utterly useless information
Good questions, if only there was a place to check, like idk, a professional review??
I normally use rtings for tv, but they do have some mointors
Those would be my questions if I'm looking at displays.
but they did a "full" review of the monitor. Oh, professional.
At least they had data, it looks at least solid for all I can tell.
They're transparent about methodology, give you all the data and they don't judge based on "feel"
If you cared about color accuracy, you'd calibrate manually no matter how it is in the box.
Then the question becomes: how well does it calibrate?
but I get you wont believe me it's a good monitor. But atleast they say the same. The professionals.
Anyway, that's besides the topic anyway. Niece will call me tomorrow so hopefully I will know more.
Stuff that gamers care about:
How's the motion handling? (ghosting? Black smearing?) What are the response times? What's the input lag? How's image quality (color, contrast)?
Prices change. But yes, they did a thorough test (not an unboxing with an unfounded opinion).
That's what you want from a good reviewer.
This is in no way comparable to the tweakers "review" that you linked
I would use hardware.info if they didn't quit.
I really liked the website.
The monitors* you posted? based on what were those? MSI G244F / AOC 24G2SPAE
- That review is almost 5 years old, tech changes.
- "Although the combination of the full HD resolution and the larger diagonal is not ideal in terms of sharpness, the 27-inch version does offer the best experience as far as we are concerned: the cons are less distracting and the larger size is fine for gaming." (Translated from the conclusion)
" offer the best experience as far as we are concerned" - is what I take from that.
"the larger diagonal is not ideal in terms of sharpness" - is what you take from that sentence ๐
You are not wrong about pixel density, but that doesn't mean it can't be a pleasant monitor like I said.
Support us on Patreon: https://www.patreon.com/hardwareunboxed
Join us on Floatplane: https://www.floatplane.com/channel/HardwareUnboxed
Buy relevant products from Amazon, Newegg and others below:
AOC 24G2SP - https://geni.us/wRQh
AOC 24G2 - https://geni.us/xSBIpLw
Pixio PX248 Prime S - https://geni.us/UYAPON
BenQ EX2510 - https://geni.us/NYaB
...
It's about viewing distance ultimately
If you sit far enough away that the pixel sizes diminish, it's perfectly fine to use 27" 1080p.
After all, 55" 4K TVs are similar in terms of density - a bit worse actually
MSI one is just because I've seen people I trust recommend it
And again, a good experience five years ago may not be a good experience today
This review is as old as the RTX 2000 series
The PC is older ๐
I gamed with that PC + this AOC monitor.
But you can't find the model anyway. All sold out.
And yet here we are.
You're shopping for a monitor in 2024, not 2019.
Anyways, 27G2 and 24G2 were awesome monitors at the time (in their respective class) and are still good today.
This doesn't translate to other monitors at all tho
Not just sold out. Phased out too, no longer being made
It's up to my niece if she wants 27" or 24".
All I can do is tell her the difference.
Then we're still in the territory of finding a good monitor at each size
That AOC one you posted seems leggit for 24".
There's far more to it than just size, resolution, refresh rate and panel type
It is, it's 24G2 but upgraded - some say for the better, some for the worse. Ultimately a great monitor still and one of the least expensive good options at 1080p
Just compare Dell G2724D against Gigabyte G27Q
Super similar on paper, miles apart in testing
Well, as far as it goes with IPS, but still. A lot
Anyways, a good monitor is definitely in order, but I'd rather save the difference for a PC upgrade
https://www.youtube.com/watch?v=seyvYETWf34 i love chevy
https://www.youtube.com/watch?v=uONAbvX_KRg
https://web.archive.org/web/20240311090514/https://www.nytimes.com/2024/03/11/technology/carmakers-driver-tracking-insurance.html
https://www.wyden.senate.gov/imo/media/doc/signed_wyden_markey_letter_to_ftc_with_attachmentpdf.pdf
https://www.wyden.senate.gov/imo/media/doc/wyden-markey_auto_privacy_lett...
Bigger number bigger better, with better bigger number number better big bigger. Bigger number means number bigger, while number bigger does not equal number bigger
i cant wait to put my old pc against my new laptop
because they are basically tied on performance other than the laptop chip having 2 extra cores
that is once i get off the dang "updates are underway" screen thats been updating for like 6 hours
you know its a good start when the ONLY thing in the bios is a boot manager
like i tought that means XMP is off
and fan curves are locked too
the day laptop manufacturers start getting sued for locked bios features is going to be a magnificent day
because its basically scaming
finally got my dad a halfway decent laptop
acer nitro 5, ryzen 7535/4050
huge jump up considering his previous one was a 7 year old dell with a dual core
the thing would pin 100% running task manager
Why? You select the monitor size based on distance, and resolution then based on need.
There isn't universal solution.
So the universal line that 27" with "only" full-HD is "shit" isn't valid.
Depends on setup and person.
So larger monitor at farther away with lower resolution is valid solution compared to smaller monitor with higher resolution up close.
But the tweakers "review" linked was absolute crap as source for information if the specific monitor is good or not.
honestly that only applies to tv's
to an extent it can but you reach a point where gettign further you still see pixelation and now everything is so small you cant see what your doing anyway
Which is why in theory the windows DPI scaling is a thing.
But that causes other issues as the support isn't universal.
Which again makes it so that smaller resolution can be a bonus, not a minus.
Again same if the GPU just cannot do games in higher resolution, as the upscaling, monitor, gpu or via DLSS etc. to the native of the higher resolution screen can be worse than just running native on lower resolution screen.
am i weird for thinking having no fan sensors on a laptop is insanity
yea and i hate it because i cant set a fan curve yet alone anything fan related at all
fair
honestly i wouldent buy anyone a laptop unless it is literally mandatory
Younger eyes do benefit more from higher DPI, but the tolerance for "pixelation" is personal, not universal.
id rather literally build a m8's pc for free than have to deal with repairing there laptop when it has no bios and half the features they paid extra for outright disabled permanently
I mean if specific person is annoyed or not if they see "the pixels".
Some people just cannot tolerate that, some don't care at all.
For first, high DPI is a must
For the second, doesn't matter at all as long as stuff fits.
a good pc configuration shouldent need to be cleaned to not overheat is the main rule i try to follow
for that reasson
Any air moving thing needs to be cleaned from time to time from dust builtup.
Question is just if that needs to be constant thing or yearly.
And with proper case with intake filters, the cleaning just means cleaning those filters.
my filters dont catch anything
i checked them after noticing all the dust inside and they where dead clean
Then either they aren't really filters, or your setup is negative pressure one instead of proper positive pressure one.
negative is what when there is more exhaust than intake?
If you have more output than intake, air is drawn in from every source into the case, negative pressure inside.
Need to have little more intake than output so that extra air is pushed out through the other holes, positive pressure inside.
yea i have a 3/7 ratio of intake to exhaust
But the output isn't just the fans, it counts all those extra output paths too.
So you have very bad setup for the fans, causing that dust builtup inside as the nominal intake fans aren't doing much.
good thing the side panels make the dust invisible
oh man i need to clean out my tower, it's been way too long
maxywell
its one problem after another i swear
4070ti is acting up now and its not fully generating some frames
im not even gonna bother i have other things to do
truth is its probably a lack of power because im defedently drawing 1000W+ out of a single outlet right now
You overclocked things didnt you
Also does this mean you broke and replaced 4 computers in as many years or 5?
I lost count a bit
its not overclockable so i actually havent
1
and its no longer broken because i fixed it
individual parts on the other hand i have lost count
... Is what I explained here
#off-topic-tech message
Thing is: people usually sit close enough for 27" 1080p to be noticable
Pinged the wrong message, but still
But if that matters is up to the person.
And the upscaling artifacts from Windows DPI scaling and ingame scalings can be more distracting when you cannot run native for one reason or another.
And if the Windows DPI setting isn't 1:1, some games get really annoyingly scaled by Windows.
As UE for example doesn't correctly set the "I'm DPI aware" flag on the executable.
So Windows does scaling on the rendered frame after rendering, but before display.
check if it's a CRT
I did say that most likely the whole drive is toast and would need to be replaced.
You never answered what even caused the computer to stop booting, but I would expect it to be some kind of hit while the computer was on.
Drive.
Hard drive replacement.
So the laptop needs to be opened, old HDD removed, new one put in its place, and OS reinstalled from external media.
Need a computer to create install USB.
So would need SOME computer to do it.
i know i dident need a install drive to install my latest copy of windows
Doesn't need to be in same place or same time as the install.
But you need external install media on USB stick or DVD.
In format that can be used with the specific laptop.
Yes?
But for that second step you need the external installer USB stick or like.
Which was what I was talking about.
if you select a empty drive enough times as the boot device eventually it will just let you install windows no idea why but it let me do that
Yes, if you do have suitable Windows install/recovery DVD.
Because you had Windows Recovery partition on the computer on another drive.
You probably had, and had only reformatted the normal partitions, not the hidden recovery one at end of OS drive.
Disassembly as in getting access to the SATA HDD.
On some laptops just remove small trapdoor in bottom.
On some remove the whole bottom plate.
On some remove the keyboard, not bottom, etc.
Just the amount to get access to swap the SATA drive.
why would two brand new 2T SN850X's come with windows recovery partitions that randomly bring you to the windows install page after 10 or so boot attempts
What laptop model?
The actual model string not branding.
The installer is currently about 4GB, there isn't one included on ANY computer outside the harddrives.
And new retail drives don't come with recovery/install partition.
That is the branding, need the submodel.
PN: <something>
Or type: xxxx
Ok, that was probably enough:
Hardware maintenance manual for that chassis:
download.lenovo.com/consumer/mobiles_pub/lenovo_g_z_50_series_hmm.pdf
So battery out, bottom cover out.
Battery should be easy, put one side lock tab to unlocked position with red showing, then the other side to outward and it should pop out.
And then three screws at front to remove the bottom cover.
Pages 37, 38, 41.
Of the PDF numbering, not the on-page numbering.
And then lift from the seam up and back.
But can probably be done without battery removal.
That seems to be in this case only for safety.
So that the MB can have no power in any case.
Not enough voltage to do anything to you.
Just don't stick metal bits between various components on the MB.
And create random short circuits
It's more about accidentally dropping a removed screw on the MB and that shorting something etc.
But if one of the battery release latches is broken and bit is stuck in locked position, not a lot that can be done remotely.
Any SATA drive that is same thickness or less than the one in the computer.
So probably 9.5mm, but possibly 7.0mm.
There is also 11mm, but those are very rare.
So anything that physically fits.
SSD or HDD in 2.5"
From the spec part off the HMM:
Hard disk drive: 2.5-inch, 7mm / 9.5mm SATA
SSD if any way or shape you afford to get big enough one (in storage space) .
That Hardware Maintenance Manual
In here 1TB 2.5" HDD is 60e as the cheapest you can get.
But is probably shitty SMR one.
And 1TB SSDs start from about the same.
With lower capacity SSDs for cheaper.
And then 4TB SATA SSDs starting from 260e, with 2TBs in between.
GDDR: Graphics Dual Data Rate.
Just memory type optimized for GPUs.
Compared to normal DDR without the G.
And DIMM is Dual Inline Memory Module.
Just technical name for the memory stick form factor.
DIMM being the fullsized desktop ones, and SO-DIMM being smaller laptop variant. (Space Optimized?)
It was long time in past when we had others, SIMM in two variants etc. (Single inline memory module).
Yes.
Normal maintenance thing for laptops.
The heatsinks will be very full of the stuff.
Field Replaceable Unit.
Just fancy name for component that can be swapped out in the field, and doesn't need visit to the repair center/factory.
There is section in the manual about removing the fan.
PDF page 46, manual page 42
Why 7900X3D?
It is not really good for either gaming (7800X3D is better), or creation (7950X is better), or both (7950X3D is better).
I'm sad about the ryzen 9000 ram support state
was hopping the iod would change
does not make much sense to change that much the core cache bandwitdth management and not the iod pipeline
Yes, laptop fans and heatsinks tend to completely clog up over time with dust.
Fan cannot really run without all the blades.
As it will tear itself apart from the unbalance.
Heatpipes from the CPU to the heatsink next to the fan.
Flat copper heatpipe from the CPU to that side heatsink the fan pushes air through.
And no, you don't want to remove that.
Just straight out, straight in one in that.
With couple of locking protrusions on one side that mesh into the MB side housing.
The fan connector is white and the MB side housing is beigeish?
Still from that manual, from the Fan removal page:
And you don't even need to disconnect that in all cases, there seem sto be enough cable to remove the fan to dangle on the cable (don't let it dangle, but enough to clean).
Not really. In theory from the static electricity from the airflow.
But much less likely than normal vacuum cleaners etc.
And again even compressor compressed air is more likely to cause damage.
The canned air stuff is all various kinds of completely evaporating hydrocarbons.
Normal air compressors leak oil into the compressed air in tiny amounts.
So in theory the stuff in the air can can be something that can damage some materials.
Just that computers don't usually contain any of the materials that can get damaged in the working parts.
Screen is different and the coatings could get damaged.
Yes, the fan should be prevented from turning.
How doesn't really matter, I just usually keep it from turning with a finger if explosed.
Just causes bearing wear.
Much more than normal use.
So no immediate effects, but can shorten the fans life significantly.
Like I said, it probably don't need to be removed.
And that type of connectors are pain to remove.
Just removing the screws and raising it up, possibly from the round back end first?
If there is enough cable slack, which there seemed to be in the pic.
extra cable in this case.
Point is to get the fan out of the position, and then you can move it much more freely in the cable range to clean it.
Without disconnecting.
Just need to be careful not to pull on the table and connector while doing that.
Only reason to disconnect the connector if is that cannot be done because of too short cable.
Just get the fan frame out of the slot it is in inside the other components first, and think then.
And the blades don't matter that much, the actual cleaning you need to do is the heatsink, after the fan is out of the way.
No, that is something else.
Moment.
Pull tab for the display connector.
Just aid to removing that connector.
Yes
That is the display out to the internal display, the white one is backlight power and control.
But you don't need to touch either of those?
They have nothing to do with the fan or fan removal?
So those two circled in the manual, the fan connector at bottom.
You really don't want to disconnect those two.
The fan screws and fan connector location, as you seem to be incapable of actually reading the manual on the page I said to look at.
So remove those three screws and try to lift the fan.
Mechanical hard drive and the laptop got hit while it was on.
And now the drive doesn't read properly, but in some cases after very long time can read something.
So almost certainly physically damaged drive.
Head hit the platter(s) while drive was running.
Yes, but not from "minor" bad handling.
Laptop with SSD can tolerate more extreme physical stuff than laptop with HDD, while running.
The French
Jackpot, the USB-SATA 3.5" drive enclosure I bought uses ASMedia ASM1153E chip \o/
Working UASP support in Linux, TRIM might not be supported but it will be for CMR HDD anyways, so that doesn't matter.
And not JMicron chips that just don't work.
Firmware specific if TRIM is supported or not for that chip.
And the physical drive installation can work as there is enough extra "clap" in the drive mounting to put the drive into position with thermal pad on top, and then turn the enclosure upside down to stick that, and then shim the bottom to force the drive against the thermal pad and enclosure top surface for cooling.
Does anyone got a Ryzen 7800x3D? How many FPS are normal at tier 8+?
Depends on save.
But still around 80-100 FPS in Taros 30MB save that needs 40GB of RAM to load.
Depending on exact spot and view direction.
For the CPU limit on my 7800X3D.
That's massive, what about a 10 mb save?
They sent here, but that was wrong, as this is Satisfactory related.
7800x3d might be overkill for satisfactory, I only need like 100 FPS at most
Not really, as anything else will probably only give half the FPS.
Satisfactory REALLY likes the cache.
Yeah, sorry.
smh baldur you should know better ๐ฆ
Replace satisfactory with games coming out in 2024 ๐
And for that I don't have answers.
Oh god you're so smart and quirky hihi
https://www.bleepingcomputer.com/news/security/hackers-are-exploiting-critical-bug-in-litespeed-cache-plugin/
Aaaaand another WordPress plugin
is there an extra metallic frame around the hdd?
which screw doesnt come out ?
i'm gonna try, no guarentees tho
have you mperhaps mixed some of the screws (from the cover maybe ?)
if you don't have a specific space/desk/deskmat, always get a couple of clean cups or glasses to get your screws parked
also, you're trying to get the hdd out, right?
try to remove the black screws on the left of the hdd
then see if you can slide the metallic frame out of the laptop
then pins will come out as you slide out the carriage/sled
can post pic slightly zoomed out ?
remove the 4 screws, then slide by pressing on the hdd and move it like the arrow
ok
can you wiggle the mettalic frame or the hdd a little and try to lift the screw at the same time ?
you might need to remove the other 3 screws
ok, remove the other screws
are you sure the last screw is loosend all the way?
that's weird, can you try to re-screw it into place then unscrew it ?
hmmm, try re-seating the hdd by pushing it toward the connector
then aside from driving a drill bit into it / stripping the threads, I do not know how, to help you, sorry ๐ฆ
depends on the context, but for computer electronics, yes
do you have another small slit screwdriver or something other small and flat?
apply pressure from under the screwhead upwards with it and continue to unscrew with your normal screwdriver
usually AMD uses the term "APU" to denote CPUs that have better GPUs than their other CPUs. not excellent GPUs, just better than a vega 11 or smth
Intel doesn't usually use the term
i always found that funny. they should have went with EPU for extended processor unit. so C < E < G instead of A > C < G type thing
Put a small flat head screwdriver under/to side of the screw that doesn't come out, and try to lift it and push it to one side, while unscrewing.
To help the screw come up, and to try to engage it to one side of the stripped counterthreads in the chassis.
The official Accelerated Processing Unit is not good name in your mind?
maybe EGU for Extended Graphics Unit?
When that name was made, AMD was betting on automatic Heterogeneous Computing becoming a thing soon...
That code that had been CPU only so far, would start using the IGP for acceleration of some tasks automatically.
OpenCL hype was strong at that time.
Pixel 9 Pro XL is performing nicely for me. It's doing things that made my Pixel 7 Pro throttle
Things which my LG V60 could do without throttling, so not necessarily impressive?
why say "at that time" when nvidia just stole the offloading hype with cuda marketing
I'm all for hybrid 4 tier computing
- CPU with low latency masked data processing smd
- iGPU for low latency big data parralelism efficiency (number crunching)
- NPU/FPGA for more specialized small footprint fast processing hardware
- dGPU with complex fetch pattern, cross unit mixing (in sm), and low latency mixed scalar processing capabilities (small but there, rnda3 like, for masking, address resolution, mixing control etc..)
the big problem this days are locked layer of abstraction, with "proprietary" software control layers
USB-C still only uses its normal 4 power pins for 240W power delivery, right? I wonder if they'll develop an alt mode that changes more pins to power? Isn't that what Oppo is doing?
https://www.oppo.com/content/dam/oppo/en/mkt/newsroom/story/flash-charging-technical-pape/OPPO Flash Charging Technical Paper.pdf
Different kind.
Not programmer made switch to GPU acceleration, like with CUDA.
But theoretical programs written in one language that would automatically adjust to whatever they run on.
With OpenCL as the base at that point, with OpenCL code running opportinistically on CPU, IGP, GPU, separate accelerator cards etc.
I see
but the opportunistic part still needs explicit code support
because offloading does not have same behavior than direct execution
The hype at that time was that the framework would just abstract that away, but no such thing actually was possible of course.
You coded once for OpenCL, and then that just ran wherever.
the cross ccd latencies just show these things straight away
the data paths are very tight
you can't afford somewhat random offloading of some part of the code without explicitly preparing for it
Yup, but the executive and marketing don't care about the reality :)
tbh
the "general scheduling" behavior of current OS is already a problem for that
(and yes I knew it was a problem for that even before chiplet architectures went mainstream)
Itanium like think, where the problems of making the compiler to take advantage of the Itanium features was just whisked out of view.
AMD has now made same mistake twice with GPU architectures too IIRC.
With RDNA3 in 7000 as the latest.
we have a expression here
saying that is a bit like putting the wagon before donkeys
Well, the first one was started when ATI was still separate?
Meaning the whole VLIW TeraScale.
No GP-GPU yet back then, but still problems filling the cores efficiently.
I mean it already is pretty much a problem with smt/ht / e/c-cores
Or even virtualization irq management
and I despise the trend of patenting even the stupidest part of software
when patenting software is already borderline
there is freaking not enough author protection
and patent is not one
US patent office is so undermanned and has no way to just straight up reject idiotic patents.
And the actual patent examinators don't have anywhere near enough time per patent to do review.
And if they try, the management will just cause issues.
And on rejection, the applier just makes some minor change and refiles, again and again and again.
I'm trying to find parts to make a custom build for satisfactory (not themed, just good enough to play the game) with a budget of ยฃ2000. I would like an amd CPU and an nvidea GPU that could support playing the game on ultra or high settings with global illumination set to medium or high. Any part suggestion or tips???
Not for that budget without most parts already existing?
ยฃ200 will buy you an artist's impression of a gaming pc
I mean ยฃ2000 sry
oh
Ok, with that, 7800X3D as CPU, almost any GPU you can buy.
I thought you were a fuckin' troll
My Radeon 6600 can do 100 FPS at all high for most saves on full-HD.
Dropping to 60 FPS in VERY large saves (30MB, way over uobject limit).
my 7900XTX plays it real good
For most cases.
That 30MB save needs 40GB to load, 32GB is enough to run it.
ok
But I would look at 48GB kit options too.
i believe DDR5 can be stretched to 48 or 64GB for not much extra cash
Would a 4070 be able to run high with global illumination on medium?
a DDR5-6000 EXPO kit is what you want
32GB is the minimum you can buy DDR5, as the 8GB sticks use shitty slow x16 chips.
And 16GB stick is the minimum you can buy x8 chips at.
That x16 vs. x8 can mean 20% difference in actual performance.
yeah, global illumination doesn't use RTX stuff
So 2x 16GB stick is the minimum you should ever buy as DDR5 currently.
also get a graphics card that does not use that 16 pin power connector
Ok
And 24Gb chips aren't that much more expensive than 16Gb chips, so I wouldn't expect the 24GB sticks (and then 48GB kits) to be much more expensive than 32GB kits.
64GB kits need double the chips.
48GB seems like the good way to go
Not possible if staying on Nvidia.
And it isn't that cursed.
Don't use third party adapters.
Check that it is seated fully.
definitely fits in your budget
I use PcPartPicker for compatibility issue checks
PSU makers own cables are fine.
don't non-founders use regular connectors
And the Nvidia adapters are fine.
No, Nvidia doesn't allow any maker to use 8-pins anymore.
There are PCBs with the pads, but still using 12-pin connector, as the maker didn't get permission from Nvidia to use 8-pins.
This good so far? Also, should I downgrade the GPU to a 4070?
Part List - AMD Ryzen 7 7800X3D, GeForce RTX 4070 SUPER, Corsair iCUE LINK 6500X RGB ATX Mid Tower
4080 super would be better for the extra VRAM
Just automatic RAM "OC" profile with AMD optimization.
XMP is Intel tested, EXPO is AMD tested.
EXPO DDR5 runs better on AMD systems, XMP DDR5 for intel
But neither guarantees that the set will work at that in your system.
Just that the RAM can do it on golden sample test system.
But 6000MT/s set of either should work fine on AM5 system.
What are your setups???
yeah no matter what pc you buy these days there's a fair chance you'll need to fuck with the RAM settings a bit
Depends if you go overboard.
That 6000 is almost guaranteed on AM5, 6200 probably works, 6400 gets iffy.
At 1:1.
i got a 5800X3D, 7900XTX, 32GB DDR4-3600; short on cash so I'm keeping the 5800X3D for a while
And you want 1:1.
EXPO DDR5-6000 is the meta for the 7800X3D
7800X3D, 64GB of 4800MT/s ECC (can to 6000, don't have enabled right now), Radeon 6600 (needs replacing badly).
For AM5 in general.
People keep buying higher, but that gets more iffy, even when the later AGESA versions have upped the chances of them working.
So the 6000 is the "this should work without tweaking in 99+% cases".
i absolutely do not have time for RAM timing fuckery
tried it once, i still got grey hairs
Any good DDR5 ram recommendations
Just using the XMP or EXPO profile.
Anything with Hynix chips.
16Gb M-die or A-die.
24Gb M-die.
Not Samsung or Micron.
my instinct is to find a gskill kit but I haven't looked at DDR5 for a while, I suspect @twin dew knows
Why not micron, I heard its good?
Not for DDR5.
Should I go with Gskill then?
Samsung 4Gb B-die for DDR4 was the ultimate chip to get.
For DDR5 Samsung is shitty.
The RAM manufacturer doesn't matter much if you can just get the right chips.
@olive hamlet the main problem you'll have is it's very hard to find out which chips a RAM kit uses
Goes then into warranty, visuals etc.
With Corsair you cannot know without seeing the exact kit unit packaging.
they'll advertise speeds and capacity but obfuscate what actual chips they're using
As they use various chips interchangeably on same unit code, and just stick sticker on the actual packaging to specify.
F5-6000J4048F24GX2-TZ5NR <- That seems to be Micron.
And the very loose latencies would point in that direction too.
With GSkill at least the parts after - specify the chips used, but there isn't any form of system.
Just have to find the exact kit in some source that specifies.
The part before specifies the performance at XMP/EXPO, the part after speficies then variants of that spec like the chips used.
no chatgpt variant is ever going to replace the likes of baldur
So I would check how much more expensive any 48GB 6000MT/s DDR5 kit with CL32 or lower is compared to that CL40 kit.
As that would make it certain the kit is using Hynix.
no pre-trained generative model is ever going to replace actual knowledge and especially intelligence
ye
Just marketing.
But like we said, EXPO is tested on AMD system.
XMP is tested on Intel system.
And EXPO has couple of additional features over XMP, but those are optional and most kits don't use them.
also isn't Expo open and XMP isn't? smth like that
And seems that no-one is bothering to make low CAS 48GB kits of 6000MT/s, only the faster speeds...
Other than Corsair.
But it doesn't really matter as much with X3D CPU that the latencies are down.
And if not doing manual tuning, the exact chip doesn't matter either.
It just sets the actual settings you can use, below the XMP/EXPO.
This looks good
Yes, that is good kit if the price is right.
Loses some overclocking headroom over the 48GB kits, but again doesn't matter if just going with the XMP/EXPO and 6000MT/s.
From being dual-rank vs. single-rank on 32GB and 48GB kits.
Good boots SSDs?
But that also gives tiny bit more performance at same speed.
Are samsungs value for money?
I use them.
They had firmware problems at mean time, and depending on time the value isn't there.
Not sure about current pricing.
And the firmware issues have been fixed.
I personally rank them very high for most of them.
With the actual relevant low queue depth performance being among the highest there is, even when the headline sequential speeds can be lower.
But perform very well in actual usage, vs. synthetic tests on empty drives.
Have all the common problems of DRAMless drives, even when HBM fills in some.
below what I consider bare minimum to recommend, assuming you mean 980 non-Pro
I didn't say "not the best"
I said below bare minimum
Is this a good boot drive (I'm gonna add a gaming drive as well) https://uk.pcpartpicker.com/product/FsqPxr/benchmarks/samsung-990-pro-1-tb-m2-2280-pcie-40-x4-nvme-solid-state-drive-mz-v9p1t0bw
The real world numbers on the 980 non-Pro seems to be fine.
as high as 980 Pro, but only about 25% lower.
QD1 Random read and write.
good drive? yes
for OS/software? overkill
priced like a 2TB drive tho, hard pass imo
my 980 pro ended up with some unrepairable bad blocks, Samsung replaced it with a 990 pro that's been solid
also goddang but it was a pain to clone my OS off of the 980
As long as the SLC cache isn't gone over probably.
Need to look closer.
But that isn't problem for normal use.
even dd would error out
https://uk.pcpartpicker.com/list/JgVtPF Good so far???
Part List - AMD Ryzen 7 7800X3D, GeForce RTX 4070 SUPER, Corsair iCUE LINK 6500X RGB ATX Mid Tower
1TB:
https://uk.pcpartpicker.com/product/gDTZxr
https://uk.pcpartpicker.com/product/YVytt6
2TB:
https://uk.pcpartpicker.com/product/qKdG3C
https://uk.pcpartpicker.com/product/kDzXsY
my main drive is now a WD black
Samsungโs 980 absorbed roughly 100GB of data before degrading from an average write speed of 2.5 GBps to an average rate of 315 MBps for the remainder of the test.
It also recovered roughly 30GB of cache space within one minute of idle time.
For gaming?
I really need to look into those...
Because what mostly matters is those QD1 random speeds.
So where those go in that.
awfully overpriced needless AIO
overpriced overkill mainboard that gives you zero benefit
overpriced RAM
awfully overpriced drives, see above
horribly overpriced case
that system can easily fit a 4080 if you just don't overspend
and basically everything else too
thx, I'll try and make the price better (altho I want to keep the case and AIO bc I love corsair)
- brand loyalty is stupid
- Corsair stuff isn't even that good
I like the look
not really
what about it?
48GB kits with same specs were 190, when that kit is 210.
depends what you do or what you play
EVERYTHING
that's never true and you know it
I'll switch to a different company if you give me one that looks as good
I can't show you something that "looks as good" if you don't tell me what makes good looks in your book
example: Thermalright Aqua Elite V3
never go 2dpc on DDR5
No-one seems to have reviewed those 250S:s, and the 250H is shitty.
So what were these recommendations based on?
And the makers of that sheet base the selection into it in what?
Because they only list the almost meaningless max sequential speeds in there.
they list controller and NAND
which allows you to find different drives with same config
Having 2 sticks per channel on DDR5 causes major loss of maximum frequency.
for quad channel, yes. for dual channel, no
Mostly to 3600MT/s in specs.
or a gigantic increase in system instability
Or was it even lower.
4800 cl40 jedec last time i checked
That is the AMD default for 1 stick per channel, for 1R and 2R both.
Intel one is just 4400MT/s for same on 2DPC MBs with 1DPC sticks.
The 5200MT/s is for just 1DPC MBs.
AM4 only the G ones, AM5 all except the F ones
Ok, AMD has upped that to 5600 for 9000-series in 1DPC.
And 5200 for 7000-series.
But for all AMD AM5 CPUs the 2DPC at 4x1R or 4x2R is 3600MT/s.
And the performance of that Fanxiang S660, even in the 4TB config, is just bad at low QD random.
Marketing optimized firmware.
60MB/s read, 230MB write at QD1 4kB
looked to be MP44L config
same performance as SN770/NM710/MP44L - last two aren't surprising as they literally use the same config
Crystaldiskmark performance of E16 drives looks similar too
So the same actual real-world performance as the Samsung 980 non-Pro in that sites testing.
With the higher sequential speeds raising the score.
for random 4k q1t1 yes, everywhere else 980 non-pro falls short
including price
besides, 980 non-pro is only available up to 1TB
But point was, are there cheapist options that actually have good 1QD performance?
Because the sequentials don't really matter to anyone while using the drive.
define "good 1QD performance"
Because they only matter if you are copying large amounts of data on the local computer.
Either inside the same drive or drive to drive. (moving on same drive doesn't care).
80+MB/s reads?
When 100+MB/s can be done.
And no, even the 980 Pro doesn't make the cut in that sites results for that it seems.
But ok, that drive is usable for the actual specs based on those tests, just misremembering stuff on my part on what is the current state.
But seems that that specific test program or the way that site tests has the QD1 results lower on several drives from what I remember other sites testing.
70 MB/s I can find easily, 80 drives up cost significantly
But that is basically the most important speed spec for OS / game drive.
And that there aren't times when the drive goes unresponsive for moment, that some way old drives could do.
QD1 4kB read IOPS / MB/s.
IOPS would be the better way to measure, or average & max latency of a that kind of request.
the only drives that get 80 MB/s on the site I checked are E26 drives
not even 990 Pro does 80
Sequential reads and writes to SLC buffer are completely meaningless.
The sequential writes after SLC buffer exhaustion can mean something.
Yeah, that was what I was commenting that their QD1 results seemed to be lower than anyone elses.
you can simply reverse engineer: find drives that use standard hardware (like Phison E18 or Maxiotec MAP1602) and check which are good and which aren't
But seems I was misremembering, as the 980 Pro results I did are around 74-75MB/s for the QT1 4k Read.
So their numbers are little higher.
then you can use the spreadsheet to check for drives with equal tech
My 1QD 4k writes are much higher than theirs.
But then the sequential read is much slower for me and writes are the same.
as long as they are consistent, it's still usable
anyways, Transcend 250S uses same config as Adata Legend 960 which they did test
not really
Those low QD reads are what affect loading times etc. somewhat.
The others don't really matter at all.
Low QD small writes are also used, but only really matter for synchronous writes.
But we seem to have finally hit the point where the differences on that aren't meaningful either for most drives.
let's also not forget the longevity aspect, especially with people using Shadowplay left and right
I meant the actual performance differences, not the extremes or endurance etc.
So endurance, and write performance once SLC cache is full, can still matter.
Because that "only" 300MB/s sequential writes, once SLC is full for 980 non-Pro isn't anywhere near the worst out there IIRC.
And don't think anyone tests random writes at that point.
MAP1602 my beloved - full drive pSLC
I was just wondering: can you do procedurally generated Minesweeper?
I would expect that it has been that since start.
*infinitely procedurally generated
Remember, the game just has certain amount of mines, and rest comes from that.
Just randomized placement, then count the numbers for the other squares.
Mainly the problem with infinite would be if there is some very long ass way around cluster of mines that don't connect on screen or near the current position.
you can do a mine density over a certain area
How large area around the current location would you pre-generate to check for that, and would you force additional mines to block routes that later generation would have allowed.
these sound like solvable issues
That "place additional mines to block routes between already opened blocks after new area generation" would solve the only real problem.
Well, from already opened, to non-opened but already visible blocks.
And generating large enough area around the viewscreen to not have that as realistic issue wouldn't be heavy today.
So that the few extras could easily be added for the few cases that might happen with purely random placement if not using very low mine density.
And at very low mine density, the open areas would probably link anyways closer.
The AOC 24G4X/24G4 & 27G4X/27G4 sport a flat Full HD 180Hz IPS panel at an extremely affordable price. Find out how they compare to rivals by watching my review!
Buy the AOC 24G4X from Amazon: https://locally.link/ECKb
Specs: https://aoc.com/uk/gaming/products/monitors/24g4x
AOC 24G4 specs: https://sg.aoc.com/product/24G4
AOC 27G4X specs: https:...
@dire igloo So my niece called, apparently they (sister and niece) are leaning more towards to 27". So I did a little bit of research for 27
There's still technically 1440p, but I wouldn't do that if your niece doesn't even know how to use the settings properly
Seems to be the LG 27" Full-HD Nano IPS panel.
Which is good one.
Not confirmed, but based on the specs.
But the larger than sRGB color space can cause issues.
fireworker suggesteded a few, one that fancied my eye was AOC 24G2SPAE. But now I am trying to see if they have 27" model.
I would suggest trying to one that doesn't have color space larger than sRGB if possible.
If overblown colors are something that would be minus and not a plus.
you mean this?
No.
That is overdrive overshoot.
Not the color space, as in what color is shown when signal of 255 is sent.
If it maxes at sRGB max red, or goes to redder.
Same in other directions.
sRGB as the grey triangle, human vision as the outer grey edge.
One of my old HP ZR24w:s (sRGB) in color in first, one of my newer monitors (120% sRGB, 90% P3) in second:
but according to reviewer, only issues when you go outside sRGB mode.
The sRGB modes on most monitors suck ass and prevent using most controls to fix any issues in it.
not here, according to reviewer ๐
So very bad whitepoints too, and not tuneable.
The Measured whitepoint whouls be 6500K, and was 6000K.
I can live witht he whitepoint
it doesn't need to be a perfect monitor for 140 bucks ๐
That is the thing that is most visible.
That sets what color white is.
It's also adjustable in height, which I find important
if she get's taller she can adjust the height of the monitor.
Basically with that 6000K it is much redder than it should.
I can look a bit more, to see if I can find any better
Point is that the whitepoint is tuneable in the monitor menus always, but most sRGB limit modes disable that tuning.
And can have as badly tuned whitepoint as the default unlimited mode.
Thanks
So if you want sRGB monitor, much better to buy one that doesn't need limiter for that.
Or just live with the expanded color space, some prefer that.
Most phones today use overblown color.
And you don't really notice unless you compare to something else.
Same with monitors.
If you aren't doing image editing etc. it doesn't matter that much if you don't notice it as distracting thing.
No, gaming monitor, 1080p, 27", IPS, adjustable in height. Max budget 250. First monitor for my niece. PC is old Asus strix GTX 1080 and i7 7700K. Currently she used a TV screen as her monitor since my sister had it in her basement. She isn't a hardcore gamer either.
So if the extra color space matters or not depends on person.
But don't ever trust a limiting mode in monitor menus to be usable without checking the exact monitor unit.
Because that wrong whitepoint is much more visible than the extra gamut volume in most cases.
Its what is controlled by the monitor menu "Color" settings that often has something like "Reddish, Neutral, Blueish" or various numbers around that 6500K, and then Custom.
You can look how visible that is by opening something near pure white, and changing between those.
Eyes do adjust to whatever is used somewhat.
But then if there are multiple displays or like near, it will look very weird when switching between them if they aren't somewhat closely matched.
thanks for input Baldur
So if the extra color space doesn't matter, that should be fine monitor after by-eye whitepoint correction, that is needed on basically all not-precalibrated monitors.
I'd assume that overblown colors are a benefit here, kids love flashy vibrant imagery
Even if it comes at the cost of color accuracy
I meant that you could still go for a 27" 1440p panel. but with what you said about your niece, I don't think that's wise - primarily because of the graphics performance impact
If Iโve applied liquid metal to my cpu, is it safe to take the cpu block off then put it back on without having to reapply so i can check the underside application?
everytime you take it off then on it will change the underside application
treat any type of thermal interface material as one-time use unless it specifically states otherwise
Liquid metal can be reused, but can need respreading.
If it wasn't used on wrong surface materials and is now contaminated by things it dissolved.
Which also destroyed the surface.
Idea(probably stupid) i want to apply liquid metal to my gpu waterblock. What if i used liquid metal on the silicon and put thermal paste on the other components on the die to protect them?
LM needs to be applied to both surfaces.
I know but i mean put the TP on the resistors to protect them incase LM gets on them
IIIRC.
thermal paste is no-go for that, you need suitable nail polish as minimum, preferrably some purpose made stuff.
Not all nail polishes are suitable.
Dont have any. Gpu comes later then as i dont have anything non conductive to protect them.
Unless i can get something from the local walmart
Then get some before you try to do it?
Even if you have to wait for weeks?
Or just don't use liquid metal.
I wanted to try direct die with liquid metal for my cpu but thats a completely different and new territory for me
Mainly cuz of how amd made the am5 ihs so damn thick
And if you don't have the things needed to do it, don't do it.
To not get warped CPUs like Intel, and then less than 1mm extra to make AM4 height.
Most of the thickness is to provide physical strenght.
And โam4 cooler compatibilityโ only some coolers tho
The height is the same, the mounting holes are the same, the keepouts are the same.
If some cooler isn't compatible that is on the cooler maker being idiot.
Stock backplate is the same.
Some cooler makers just replaced the stock backplate for some weird reason instead of just using right thread screws to the stock one.
How? The am4 backplate for the cpu frame wasnt connected to the cooler mount so you could use your own backplate
How are some makers stupid?
That they didn't use the AM4 stock backplate and shipped their own without any good reason?
But that didnโt mean amd had to make it all one backplate
Not having it all one backplate would have retained all cooler compatibility
So AMD should have made two separate stacked backplates?
To support the socket itself with one and another for CPU coolers?
And then that CPU socket plate would have had to be very thin to not lose the cooler backplate compatibility anyways.
Intel has had standard socket backplate for a long time that the cooler backplates need to fit over.
But their cooler holes don't have stock backplate.
AMD has the longtime standard cooler backplate which they just modified to act as socket backplate too.
And any cooler that used the very longtime standard mountings just worked.
And as the non-backplate mounting hardware couldn't be shared between Intel and AMD because of that Intel socket backplate anyways, there was no reason why the makers kept using their own backplate for AMD.
Beside inertia and not wanting to do actual redesign instead of tweaking dimensions of same parts.
AM5 is the first AMD socket where the forces for CPU mounting aren't just sideways on each pin.
For consumer socket I mean.
Threadrippers and Epycs have been LGA for long time already, but need torque screwdrivers to install.
So AM5 needs MB backside support for the socket, to not break the MB PCB.
And either they would have needed much more complex socket mechanism, or that thich IHS they selected, to not make the CPUs bend in socket like Intel has.
Because while the third party Intel contact frames solve the issue same way as the server CPUs do, that just isn't suitable solution for general consumer issue.
some suspect the thick IHS is to allow more advanced packaging
It was specifically said to be for physical strength, then thickened slightly for the AM4 compatibility, 0.6-0.8mm IIRC, in interview with AMD engineers.
Would in a way been better solution, but like I said, not suitable for general consumer use.
I've seen people sour on them and get inferior contact pressure/have to downgrade RAM speed
And they aren't placebo for Intel CPUs.
Just that if they have big or small impact depends on the exact cooler.
though this is all anecdotal
Like Noctua did three different ones for the new one.
One optimized for use without one, one optimized for AMD (very flat), and then third "normal" in between.
Because the installation is finicky.
If you don't exactly follow the instructions and do the tightening in wrong order, you get very bad results.
And that is the reason why the original thermal grizzly and the server ones need torque screwdrivers to install.
As they aren't "tighten until contact with MB" type.
As that distance isn't static between MBs and CPUs.
So this would point to too thick contact frame on MB with socket lower than average.
With the Thermal Grizzly one, that isn't designed to ever hit the PCB that can be accounted for, but all the others are for "average", as desided by the maker.
Elon Musk was just forced to reveal who really owns X. Hereโs the list:
fortune.com/2024/08/22/elon-musk-x-twitter-owner-list/
From another source based on that list:
Elon Musk was forced to reveal who financed his purchase of Twitter.
Amongst owners of Twitter there are two Russian oligarchs who are close to Putin and we're sanctioned over the Russian invasion of Ukraine: Petr Aven and Vadim Moshkovich.
Also Saudi "Kingdom Holding Company"
And couple of other entries from Saudi Arabia.
That sounds weird.
How so?
Also this entry ๐
Series N Dis, a series of Atreides Special Circumstances Fund, LLC
"Who really owns X" ... just sounds weird.
Ah, blame Fortune.
You can say that about Microsoft, Google, Apple, ... any company with investors.
I am more curious, about the lawsuit from X vs advertisers tbh.
Against advertisers.
๐ He went from that. To lawsuit.
Not investors, advertisers that stopped advertising on Twitter.
Yes typo. Don't panic ๐
Because Musk at that point was annoyed and now he is scared.
my point is, he literally said "Go fuck yourself" to advertisers.
Very strong language.
Yup.
But that was then, and now he is going under from that.
I don't give a rat ass. Think before you say something.
And instead of rolling back the changes that caused the advertisers to go away, he is trying to litigate.
Even now he must realize how weak this is, making strong bold statements and then crawling back.
doubling down.
No, that is doubling down in a way.
A good leader, a true CEO, wouldn't have said it in the first place.
And if they did, admit their mistake, take PR hit and move on.
Nah, Musk is idiot, always has been, with no filter.
Some great ideas, some very bad ideas, and he cannot differentiate himself.
Like the CEO of Crowdstrike, say what you want, he takes it like a champ.
And honestly, a lot of people are saying it's a smart thing to do, just accept that you fucked up. Learn from it. Move on.
But the great ideas did something great with Tesla and SpaceX.
And with SpaceX he still has enough outside forces inside the company to keep him in line.
But Tesla is currently getting fucked by him.
Because there isn't any balancing forces there anymore.
And the company is too big to benefit from Musk.
where did I read about Tesla not doing so well.
Oh, right sorry can't go into that. Politics.
But due to political reasons, Tesla is taking a hit. Or so I heard in the news.
And like previously talked, Musk went completely crazy during COVID time politics.
Tesla is suffering from Musks demands, at time that it isn't in "way ahead of everyone else" position anymore.
And then Musk is also holding the company hostage at this time.
I once defended the man, but he too far gone now.
With that "You must reinstate my giant payday or I take my toys and leave"
I heard other things are going on with Tesla.
But basically he was unstable before, and needed careful control and counterbalancing.
And went completely nuts because of COVID lockdowns.
COVID really made him butt heads with the politicians.
With X he has no counterbalance, and it is purely post-nuts thing.
With Tesla he today has almost no counterbalance it seems.
With SpaceX there is enough counterbalance still outwardly.
he saw it as a personal attack, the working from home regulations etc.
Closing the California Tesla factories.
Don't forget Neuralink...
But I have to head to bed.
I had already closed Discord, but hit that information.
But explains some of the content moderation etc. decision in Twitter/X.
WIth Tesla he has the opposite of counterbalance, plenty of superfans just feed the echochamber of praise for his stupid ideas.
I've seen it plenty in online Tesla owner communities
wifi is good [:
That cost too much.
well uh
I'll take donations though. I'll need about 1k usd to upgrade to 6e.
I'm on 300/300 xgs-pon fiber.
so if I wanted to go to 5 I could. most of my stuff is local network. which is gbe.
man
i only get 200
and people here have 1 gig
and their dorms could have done 10 gig.
i feel like a peasant
200/200?
that's probably cable then.
i cant do ethernet
"ooh look at how eco freindly and how much plastic were saving by not including a disc drive and an ethernet port!"
your wifi should get you the full 300 down if you're on 6e tho.. unless other people are using it.
you're on 5ghz for that speed.
huh?
well, butch & suni aren't coming home on Starliner
OH
wifi is 2.4, 5, and 6 ghz
2.4 is 144mbps with 20mhz 2x2 ( I think ) 5 is 600mbps 1x1 or 1200 mbps 2x2 and 6ghz is just fast lol
I wish I could've got to where our pc's are so I could've wired them.. or just spent the extra cash for 6e and a 2.5gbe switch.. until I realize how much extra that switch and hardware would've cost me
ethernet solves everything
putting holes in the wall makes more work xd
shush
hide this from teardown players!
How does the EU law work? Apple won't be allowed to sell lightning phones starting 2025?
It'll be interesting to see the places they keep selling the 14.
?
Iphones sold, within the EU, need to have, by then, universal charging. Which will be USB C.
I feel double about this law.
Yes I don't want 500 different USB cables to charge stuff with. On the other hand, I don't want to stop innovation either and be stuck with USB-C until the end of times.
Right, but they'll probably clear out iPhone 14 models in the rest of the world. I wonder how many EU folks will be trying to import the 14 in 2025 or some shit.
but I think this is more a case of Apple trying to get a monopoly and more grip over it's eco system than wanting to innovate ๐
iphone 14 will have USB C?
Oh, fuck. USBโC-naar-Lightning-kabel. Huh.
I think Apple's monopoly results in them stagnating rather than innovating in a lot of cases, lightning being one of them. Same with Google, etc.
Apple will say it's for innovation. I think it's one of the arguments I read the most regarding this topic.
What do you think is the latest **big **tech related innovation Red Maw?
'This will stifle innovation' is one of the most boilerplate arguments made by companies like Apple, it's almost impossible to take seriously companies have cried wolf over that so many times.
Otherwise generative AI is a huge innovation that's going to make a lot of waves. There's no predicting how that will play out at this point.
I was thinking the same.
And to be fair, AI is a innovation, not coming from the big Tech companies we know (Google, Microsoft, Apple, ...). They are even "behind" and buying AI tech from the ones who made it big (like OpenAI).
What generative AI is going to do to music is a big unknown all in itself, and that's just one arena.
Google was focusing on glasses and stuff. Apple was focusing on ... I don't know what. Thinner hardware I guess? Microsoft is focusing on ads and W11. And bam then openAI and other came with real innovation and caught em off guard.
It's already doing a lot. Read a few articles this summer about summer hits probably being AI generated.
And I think Sony made a deal a while back? Let me research.
Max Martin is just going to become even more productive now. He can use all his own work, plus everything pre-1930 for training.
Oh, they didn't make a deal. They are sueing ๐
The legal landscape will eventually settle, but in the end you won't be able to stop generative AI from being trained either in places overseas that don't respect IP, or by companies big enough to obtain a training library that's unambiguously legal.
Now Sony and co are scared because they have big contracts with artists, and they paid those artists and want their profit.
But once Sony, Universal, can use AI themselves to make some ๐ฐ, it wouldn't surprise me they do in the future. When contracts are expired.
Let's just hope there will always be room for real human made music on the mainstream platforms ๐
I wouldn't mind to also have AI music, aslong as we can pick ๐คทโโ๏ธ
The arts industry stands a chance of getting some decent money for the talent, but other places not so much. It's generally always been the case that when you're on company time, the company owns the results of your work. So all the scientific/industrial work now using this stuff isn't going to make anyone but the shareholders profit.
If 2pac concerts continue to do well, that's troubling though.
I still wonder what will happen to creativity if hypothetically AI starts to learn from AI and there is no more human input.
Ourobouros
If the learning pool is [Song 1 (made by AI), Song 2 (made by AI), Song 3 (made by AI), ...] what result will that have. Will creativity stagnate. Or maybe not.
Steve Burke talked about that with the gaming assistant Nvidia is playing around with.
Interesting
The assistant would need the online community and resources for the game to develop before it could work effectively. So it would be bad on day 1 because there's no wiki to crawl yet.
community driven AI
My own speculation is that EA games like Satis, or any game that evolves greatly over time like Anno 1800 will be dysfunctional with such an assistant. because the web will contain a lot of out of date info that never gets removed.
it will, at least with current ai tech. if inputs are a+b=c then c+a=ac and so on.
ecxactly
it won't make anything "new" that's not already in the model.
Ofc, the training pool is way bigger, but that's my theory.
That plagued me a lot when I used to play Galactic Civ 3 more. The game has had so many big up dates that overhauled major mechanics or tweaked others. And it has a relatively small community, so that are many, many, pockets of out of date information. Compounded by the fact that it never received as much polish as it should have, so there were unknowns about whether some things had bugs.
Unless they program in that the AI "experiments" and may deviate from what it has learned from other songs.
the pool is currently fixed, it can't make something that's not already there.. I say currently, because at the rate it's moving ai combining models into completely new models based off of, but different than the data in the first two models is probably soon.
but as is if I made a model for stable diffusion based off of the output of a model that was used to generate something, It's either a scrambled mess, or it looks the exact same.
we invented turn it off
100% not what we are talking about ๐
just the general impact of AI on society, I am not afraid of it.
so am i
like the industrial revolution
btw
they choose to call it AI and it's not.. it's llm's and lama's.
it changed society drastically, but we still here and better for it.
Yeah, but this discussing has been going on for 2 years in this channel. AI != ML.
My opinion; I use the term AI to keep it simple. Main stream media uses it.
very different from 2000's AI lol
I know, it's not true AI. But just to talk with my friends/family, I use AI. ML will confuse them.
All they hear/read is "AI" on the news
I agree.
and I cannot blame the news either, the companies sell it as AI.
it's companies called like: openAI
I do think, human input AI has still a chance to create something creative.
like the time it was all "bitcoin this and bitcoin that", then it went into "NFTS NFTS NFTS" and now it's AI ๐
IKR
if it's used as a tool, it can be amazing... although it'll certainty be used as a tool to replace jobs.
like even apple and google is shoving it down our mouth
its word is overused
I hope for a combination, I think AI is powerful in human hands.
you know, customer service when you call them.. when you get past their automated help assistant.. and you get to a person.. it's another computer.
not sure if that makes sense?
Google answers. โMix about 1/8 cup of Elmerโs glue in with the sauce. Non-toxic glue will work.
i love my pizza with that
yum!
Example: If AI gives a dumb answer, a human can reform the question, to get a better answer.
nah bro trust me
Aslong someone is in control, it's all fine. Even for art.
glue in pizza hits
That was funny ๐
its so tasty!
You know what AI will be perfect for?
Making subtitles. It will get near perfect.
I know YT has something that makes subtitles, but it's far from perfect.
true
WE SPONSORED OURSELVES. Buy our coaster pack! https://store.gamersnexus.net/products/3d-coaster-pack-4-component-coasters - these have been in high demand and sold out rapidly last time we stocked them. It's been about 2 years and they're finally back. Directly supports our purchasing of review hardware. Also consider a copper-plated mule mug: h...
i can't remember the last time i saw a positively-named prebuilt review from steve ๐คฃ
nope, hidden buyer
you don't believe GN is legit?
i'd love to see the evidence
like
cuz he roasted the absolute FUCK outta them last time
8 videos about 1 product come out at the same time
and jre didnt even do anything to that new samsung
he was just yapping
he didnt test its durability
Cuyar, random question. Did you draw your profile picture*?
your evidence for being sponsored by Starforged is.........not reviewing a phone?
lol nah, i ain't talented like that
im not talking about starforged
then what was this in reference to
talking with the privlage of talking
wut
now your feeling it
ah k, got it
yup
whats the rod for?
thermocouple
they "mil" a channel into an integrated heat spreader to validate a curious microcode thermal behavior
๐ That's literally what he said, and it confused the hell out of me ๐
2 wires of different metals or alloys that lets you measure temperature really accurately
something about bugs in Intels microcode
details on that part at about 15m mark
i refuse agree that intel cpus run that hot they need that
it's independent verification of claims
especially when there's evidence of unreliability
I see now. Anyone knows what "mil a channel" means?
Oh, never heard of it ๐
That looks satisfying
definitely
So just finished watching it. The drawbacks are, Intel and temperature overhead
yea, i'm actually impressed
especially for such a tiny case
that also means that if you go AMD with it it'll be way cooler
Oh, and there was paint missing.