Osc8r wrote on Dec 3, 2024, 15:59:

Nullity wrote on Dec 3, 2024, 13:22:
OLED for a PC monitor is not a good idea (currently). Sure, the picture quality is great (save for the low brightness), but the negatives are too steep, and all the babying and hoops you have to jump through to attempt to prevent any of that is too much of a pain in the ass.

Also, this again.

Low brightness? Searing my retinas doesn’t sound enjoyable to me.

For HDR you should have 600 nits at an absolute minimum, with 1000 nits being being much better. It's been a couple years since I've followed OLED tech advancements, but as far as I know, OLED can't currently come near that.

If you're happy with current OLED tech, then more power to you. I may be inclined to get an OLED TV, but a monitor would be more susceptible to the negatives (like burn-in, for example). I currently have the last gen plasma TV made by Panasonic (65"), and while I'd like to move up to 4k and HDR, the picture quality is excellent (again, plasma), which has dissuaded me from spending money that I don't feel is necessary yet.

Hopefully microLED isn't too far off, which is even more promising than OLED.

Nullity wrote on Dec 3, 2024, 13:22:
OLED for a PC monitor is not a good idea (currently). Sure, the picture quality is great (save for the low brightness), but the negatives are too steep, and all the babying and hoops you have to jump through to attempt to prevent any of that is too much of a pain in the ass.

Also, this again.

Low brightness? Searing my retinas doesn’t sound enjoyable to me.

I own three OLED TVs, including the first-gen LG, two Master Series Sonys, and an MSI QD-OLED running in 4K with HDR. None have issues, though prior to moving to OLED i had to return two LCD's due to horrible banding (well, one, but it had to be returned twice to finally get a unit that was somewhat acceptable).

There isn’t much to do to babysit the monitor. Set windows task bar to hide itself, screen to go into standby after a few minutes of inactivity, then most other safety features are set and forget.

For me, the risk of burn-in is worth it compared to the 100% guarantee of subpar picture quality.

Tbh, going from a top of the widescreen alienware LCD to a OLED 4K was the biggest step forward i've seen since gl-quake days

OLED for a PC monitor is not a good idea (currently). Sure, the picture quality is great (save for the low brightness), but the negatives are too steep, and all the babying and hoops you have to jump through to attempt to prevent any of that is too much of a pain in the ass.

Also, this again.

VaranDragon wrote on Dec 3, 2024, 09:01:
Wish I had an OLED TV. Love the colors and contrast. Best I've seen since Plasma TVs.

Yeah, the top of the line plasma TVs were awesome. I'd buy another OLED TV if I was convinced they have resolved the burn-in issue, or I can get at least a 5 year replacement warranty if it does get burn in (preferably 10 year).

E-Core and P-Core scheduling in Win 11 is garbage, on my laptop I sometimes have to reboot to force it off the E-cores (plugged in 99.9% of the time!) and while we're at it, why do I have two separate power plan locations in Windows 11, that seem to work at odds with each other? Hot garbage. Moved to Linux on my desktop PC and I'd do it on my work laptop if I could.

Wish I had an OLED TV. Love the colors and contrast. Best I've seen since Plasma TVs. As for Intel, oh how the mighty have fallen. I do own their i5 13600k CPU, and I have to say it's the best performing CPU that I have ever owned. (I say this as someone who upgraded from an i9-9900k, and the difference was night and day) Yeah I'm aware that people have had issues with 13th gen, I guess I lucked out on the silicone, but then again it's "only" an i5, and was the best bang for buck when I was upgrading.

I seriously doubt I will be upgrading any time soon, even the most modern games run pretty well on my system but if I do I guess AMD is now a no-brainer.

RedEye9 wrote on Dec 2, 2024, 22:29:

Longevity Burn-In Test - Updated Nov 20, 2024
Updates And Results From 100 TVs
https://www.rtings.com/tv/tests/longevity-burn-in-test-updates-and-results

Almost exclusively TVs unfortunately, BUT since the ASUS PG42UQ has a LG panel, one can extrapolate from the INSANE burn-in on the LG panels that the PG42UQ is a prime candidate for massive burn-in coming up. Great. I guess it is pretty obvious why ASUS is shipping these displays with only a two year warranty.

Oh well, glad I waited for the recent price drop. It was "only" €1190 or so including shipping so it won't be too sad when I have to haul it off to the garbage dump in two and a half years.

Longevity Burn-In Test - Updated Nov 20, 2024
Updates And Results From 100 TVs
https://www.rtings.com/tv/tests/longevity-burn-in-test-updates-and-results

Mr. Tact wrote on Dec 2, 2024, 21:31:
I mean, I didn't leave it up for hours at a time. It was just "normal" usage, I opened it when I was looking for a show. Did I look around more than others do? I have no idea. Certainly didn't feel like I was spending excessive amounts of time in the guild. *shrug*

*gulp*

Now I'm officially (yes, OFFICIALLY) scared!

I mean, I didn't leave it up for hours at a time. It was just "normal" usage, I opened it when I was looking for a show. Did I look around more than others do? I have no idea. Certainly didn't feel like I was spending excessive amounts of time in the guild. *shrug*

Mr. Tact wrote on Dec 2, 2024, 18:38:
I got an OLED 6-7 years ago. Got burn in from the channel guide in just under two years, I was not happy. I mean maybe it was the guide's fault more than the TV's fault, but I hope they have improved the OLED in that area...

How extensively did you use the guide then? Did you leave it on the screen for hours sometimes or would you call it regular usage, i.e. did you just bring up the guide every now and then when you needed it to check for upcoming pr0n?

I got an OLED two weeks ago as a monitor (ASUS PG42UQ) and I'm already paranoid as fuck . I'm following all the usual advice like dark mode, no desktop icons, auto-hide taskbar, screen sleep after 10 minutes, yada, yada... this shit better not suffer from burn-in with my bookmark bar tatooed into the screen in a year or so or I will throw the fucker out the window. Before opening the window, I mean.

I got an OLED 6-7 years ago. Got burn in from the channel guide in just under two years, I was not happy. I mean maybe it was the guide's fault more than the TV's fault, but I hope they have improved the OLED in that area...

I've been considering getting an OLED TV (and lately OLED monitor) for years. W-OLED or QD-OLED have pros/cons. Getting close to pulling the trigger then damn... OLED MAX... They mention Samsung Display so maybe 5th gen QD-OLED will have it (2026 displays?) if PHOLED doesn't come first.

So Intel needs an 80s montage done to 80s music?

Very doubtful it was a voluntary retirement. Note how short Intel's parting "thanks" is. With a person of Pat's history and pedigree at Intel, they would have usually written a very flowery love letter instead of a short single mostly factual paragraph (original Intel statement here).

There must be more to it. The only known fuckup of Pat was him missing out on a 40% discount at TSMC because of some not so thoughtful remarks he made about Taiwan/TSMC.

Other than that, he is hardly to blame for most of Intel's misery. He's not been with the company long enough for that. That said, his 5N4Y plan hasn't exactly worked out as planned as the 20A node had to be outright cancelled.
The other nodes were of mediocre quality at best and they produced mostly degenerating garbage on the "Intel 7" (10nm) node for over two years before realizing it (Intel 13th and 14th gen mega fuckup). Two gens for the trash bin. Gratz @ Intel.

18A was supposed to become the savior and Pat "bet the company" on it by his own admission.
Qualcomm was apparently less than impressed with 18A when they sampled chips earlier this year and it's not like customers are exactly queuing up for 18A manufacturing at IFS.

So, my guess is that 18A is a bust like its predecessors. Skipping the line without the invaluable learning process is probably not the way to go or more companies might try leading edge chip manufacturing. Intel tried the impossible. You can't just effectively skip from 10nm ("Intel 7") to something like 18A (that is supposed to rival TSMC's 2nm) within such a short amount of time. You are lacking the experience, the learning process(es), the understanding of the tools and you can't keep up developing PDKs for the potential foundry customers. Mission impossible.

It will be interesting to see Intel's follow-up announcements in the coming weeks and months. They are probably in some really, really deep shit with 18A a bust and still no end of the troubles in sight after almost a decade of fumbling around at IFS...

Can't get the neowin link to work, atm, must be a temporary, but this will be the second retirement from Intel that Gelsinger has taken--so maybe they'll do better with someone whose experience is significantly more up to date than Gelsinger's.

Yeah the crazy number of SKUs has been a real issue at Intel for a long time.

They need to also get realistic about performance and stop playing games. Performance improvements from here on are going to be incremental, and I think most OEMs, enterprises and enthusiasts understand that. Release a slightly better performing chip every year, THAT DOESN'T HAVE MAJOR BUGS, and for servers and enterprise, yeah, keep tweaking efficiency because that IS their bread and butter. Keep exploring innovative ideas like specialized cores, because we're hitting the limits of what silicon can do, so it's all about optimizing operations now, at least until there's some radical breakthrough like the diamond based chips researchers keep talking about.

It should be a serious wakeup call for them that the Snapdragon X Plus CPU in my new Surface Pro can eat a core i5's lunch AND more than double battery life. We may stop seeing any x86 CPUs in any laptops except gaming laptops in the next few years.

Gelsinger retiring isn't going to right Intel's ship. A new CEO isn't going to right Intel's ship.

Intel needs to go back to basics. Being innovative and dumping billions per quarter in to R&D. They sat on their ass for most of a decade doing nothing and it has caught up to them in a massive way. The first and foremost things they need to do is get power consumption and TDW under control and they need to do it yesterday. I know most people here focus on desktop parts but that hasn't been Intel's bread and butter for a long time. Enterprise has been. When I look at my yearly budget, OPEX stands out in bright, red letters. Power costs and cooling costs are significant contributions to OPEX. Right now, AMD is the better deal with their Epyc lineup. The performance metrics don't matter as much as desktop enthusiasts like to think in that space because most of us aren't redlining our servers 24/7. I aim for every one of my servers to sit at around 70% utilization so that I have headroom available in case I need to shift things around. At that level, Epyc spanks Xeon. It's quite literally not even a contest. Intel had Jim Keller on staff. A man absolutely known for high performance/low power designs. That's literally his "thing". And they still can't get their act together.

But then there are the very embarrassing failures on the desktop side. 13th gen stands out starkly. Arrow Lake is currently selling as well as three day old vomit. Why would you trust Intel right now? Especially at their current prices? AMD, here, is the safer bet.

Intel also needs to simplify their SKUs drastically. You don't need four or five submodels of a single generation. You need at most three. "Good", "Better", and "Best". Stop with the gatekeeping bullshit about which technologies are enabled while you're at it. Reduce your choices to core count just like AMD does. I can use ECC on any Zen 3, 4, or 5 CPU if I want (so long as the motherboard supports it). Why is this such a hassle with Intel systems?

Lastly, punch Microsoft in the head and kick them in the ass and get their atrocious scheduler to be somewhat modern. The E cores are phenomenal monsters under Linux (In fact, quite a few homelabs are built around cheap Intel CPUs that are all E cores) yet suffer worse performance under Windows. If Microsoft can't write a scheduler to save their lives (which is incredibly believable), then write one for them. You already have a whole group dedicated to doing just that in the 01.org folks. They're literally spinning out your own Linux distro with optimizations specifically for your hardware including the scheduler. Moreover, this wouldn't even be the first time you've done so. When I worked at Intel we wrote the scheduler for NT that supported Itanium because Microsoft's in-house efforts were complete ass. Despite having had silicon access for more than a year. Yes, Itanium sucked. That's besides the point.