/v/ - JACK FROST DIED FOR THIS

you forgot your pink angry wojak

Damn and I just buy Ryzen 3 2200G.

OK I guess.

Console wars?

I've never turned on my 2600x SMT, does the performance increase that much?

>>15821728

Only in a few AAA games like assassin's kangz or vaginafield

this is neo-consolewars

>>15821702

SIEG FUCKING HEIL AMD

(((INTEL))) BTFO AND INTO THE OVENS.

Both IntelAviv and AMD have kike spyware in both of them, but at least AMD doesn't funnel their money straight into some yid fab. Lesser of two evils, the idea of powerful as fuck AMD processors at least puts a jab into the lentil's plan.

>>15821719

Except, he works for Intel now, advising them on the same chiplet designs he gave to AMD.

>>15821728

There is no reason to disable hyperthreading anymore. It helps in basically every scenario including gaming. The only reason anyone ever had to disable it was when developers hadn't yet optimized for it and disableing it let you overclock higher than with it on.

>>15821702

high-end cpus are not for video games

reported for off-topic

:^)

>>15821774

Nothing actually stops you from running Windows 7 on new processors, Microsoft simply disabled updates for them beyond a certain point, and even then there already exists a fix that let's you get the updates anyway.

>>15821772

isn't he working on arm or their version of it?

Stay long $AMD, rumor or not it's going to be a fun ride.

>>15821774

>using windows at all

You need some freedom in your software.

>>15821787

It's trickier than that. You need a bunch of workarounds to get USB working at all, as Win7 doesn't include the drivers for the new mobos' chipset.

>>15821798

I wouldn't really know exactly what he is up to, but I would just assume Intel snapped him up because they would need his knowledge of chiplet design, and there is already some official Intel documentation out about a potential future chiplet-style design they are working on. Also Intel's current make-shift solution to compete with Epyc is going to be two of their current 28 core dies stuck together on a single package, and I'm sure he will have some input on how to make that work well.

Also since Navi GPU seems to be part of the chiplet plot, Intel snapping up Raja provides further insight into that design.

>>15821774

Now that Asrock finally released official Windows 7 drivers for the Taichi X470 I'm buying it and upgrading to a 3600x when it comes out. Though I only use Windows for games occasionally since I run Linux as my main OS. Only the new 16 core chips require the new X570 chipset.

It's gonna take Intel years to catch up from this. With AMD in consoles, taking over servers and leading charge on PC they're in a really exciting position. It's gonna take a lot of Nvidia money to control the ecosystem and Intel's gonna take a long time to move forward.

>>15821832

I had no problems with USB ports on my Windows 7 build with a 1500x.

AMD vs Nvidia threads have always been console war tier

but AMD should be supported over Nyidia if only for the health of the market

no link to article or video?

>>15821864

it's adoredtv's new video

>>15821864

A screencap of the OP watching a jewtube video should be more than enough for you, bigot :^)

the title of the video is in the screencap if you care enough to look for it; I don't

>>15821873

B-B-B-BUT AMD ARE BASED AND /OURGUYS/

Jews Jewing Jews.

>>15821873

>vargposter is a shill

Pottery

>>15821862

Considering nVidia's anti-consumer practices, I rather support AMD no matter what.

>>15821832

>It's trickier than that. You need a bunch of workarounds to get USB working at all, as Win7 doesn't include the drivers for the new mobos' chipset.

I hadn't considered that. Eventually I'll need to build a new pc and no way am I using a version of windows beyond 7. How is this dealt with?

>>15821920

See my post >>15821854

>>15821910

That too

>>15821702

>AMD isn't owned by jews

Sage.

>>15821832

>You need some freedom in your software.

but i want to play my pornographic japanese games!

>>15821922

You think publishers would pay coders for optimization that would make more money than advertising? Not happening, unless you want pajeets optimizing.

>>15821950

That loli looks like she's tired of being told to show you her pantsu again for the millionth time but still goes and does it anyway for you because she cares for you as a friend deep down.

>>15821982

I get mixed messages from the leg positioning, like an effort to keep the viewer away while still showing a little.

>>15822020

Explain to me how it is, retard. A CPU is not a video game.

>>15821702

I don't believe the pricing, nor the boost clocks, nor the High End tier. Likely an Intel leak to over produce expectations while they formulate their counter attack.

>>15822002

Anon, she's told you every time she's shown you her pantsu to wait for her to grow up before you ask her to be more than a friend.

>>15821759

>>15821702

>>15821766

>>15821838

>>15821862

>>15822020

>muh juuuuuice

https://hardware.slashdot.org/story/11/06/02/1730219/amd-opens-israeli-rd-center-hints-at-arm-link

>>15821950

>PCIe passthrough

This. Just run Win10 in a hypervisor and you'll get native performance with airgap-level security.

>>15821910

I'm glad AMD is making CPUs that are competitive with Intel again, but let's not pretend AMD GPUs aren't still an embarrassing joke compared to nVidia. Crossing my fingers that nVidia's disappointing 2xxx series will give Navi room to excel, but I still wouldn't bet even odds.

>>15822071

Same here, I already has a full AMD system built out of spite for Intel after they supported anti-gamer/gaming grifters. 8350/380 to 3700 and a 590 doesn't make sense as of now, the GPU has to improve even have as much as the CPUs did.

>>15822075

A few years ago AMD started stealing the naming conventions of Intel's motherboard chipsets by taking the number that would have been the follow up generation, and if this "leak" is correct then they are going to steal Nvidia's naming convention next. Not to mention the parallels between i3/i5/i7 and R3/R5/R7. It's pretty hilarious really.

>>15822087

The Vega prices are still higher than what they were originally planned to be, they're only as high as they because of mining despite how disappointing they were. A year of process maturation, architecture maturation, refinement, software and a die shrink to 7nm is what is making basically the same foundation so much better. Vega wasn't a bad architecture, it was an overpriced one. This is basically the HD3000 series happening again, although I do not believe they have another HD4000 up their sleeves.

>>15822115

>like last time.

been in a coma for 10 years?

>>15822115

AMD can do it with 7nm, but Intel will still have a 1000fps+ lead in Source engine games.

>>15822113

No, the OP leak is separate from the Reddit leak and has key differences, though they certainly have similarities. Most likely Adored's leak is newer info based on the latest engineering samples.

i'm going to observe this war and maybe rebuild/upgrade in 2~3 years depending on who wins

>>15822075

I'll believe it when they're out.

>>15822051

>Just run Win10 in a hypervisor and you'll get native performance

No you fucking wont I am sick and tired of this idea being peddled by everyone who obviously has never actually set up a gpu passthrough. At best you'll get 10 to 20 percent less than native performance and at worst your performance will decrease pretty dramatically compared to native. I have GPU passthrough set up and the VM installed directly on an ssd so I can directly boot to it to compare the VM performance to native and I lose a big chunk of fps sometimes up to 40 in any game that is CPU intensive.

>>15822126

2500k wins again

>>15822133

the reddit leak is entirely something different, it's used as a point of comparison and Adoredtv does a lot of work to check his sources on these things. He did leak RTX after all.

>>15822137

Sounds more like you fucked up somewhere. You have KVM enabled?

>>15822137

I'm not sure why the VM passthrough meme took off for Windows 10. It's still data mining you and your local network. The whole point of the passthrough is to get around the lack of drivers for new motherboards for Windows 7

>>15822178

Intel's failed transition to 10nm is causing shortages, which is causing price increases. They even had to roll back their chipsets manufacturing to an older process node. The only Intel chip that makes any sort of sense would be the 9900k for maximum game FPS, in every other scenario and price point AMD is the better choice.

If you buy Nvidia, I'd say get it used. I bought my 1070 off an Ethereum miner for a good price. As long as they took proper care of the cards while mining (aka undervolting and proper cooling) there is nothing wrong with them.

>>15822017

It's shit that can be used on a rig to play vidya.

>>15821832

>You need some freedom in your software.

You mean like how sjws have fucked over the freeware community? Nothing in software is unpozzed sadly

>>15822186

The point is that you can restrict the botnet in Windows to gayman tasks only, or just completely shut off NIC access, while having a more secure OS for things where privacy matters (officework, email, etc.), all on the same PC without any duplicate parts aside from the need for at least two separate GPUs.

Even if all you run is Windows, the ability to switch between versions without a full reboot, to run older vidya, is a godsend.

>>15822214

Do you even lift?

>>15822087

salads are not optimal for sustaining the human body

>>15822214

Every video game uses a CPU, not every human uses a salad.

>>15822214

CPUs are a key constraint on vidya, 1331 g8mR f00d isn't a crucial factor in performance.

>>15821719

Why doesn't intel make a dedicated GPU instead of just part of a chipset?

>>15822358

They're working on it. Intellectual property is a major obstacle.

>>15821832

The work around is EZ. if you already have the installer files, you can inject the USB drivers into it. MS even has a guide on how to do it on their site

>>15822366

I heard, for all intents and purposes their GPU technology does a nice job on its own so I was always wondering why not making a dedicated GPU, more options is always good.

>>15822280

>Every video game uses a CPU

I think some very early arcade boards do not actually have a CPU. I'm not too familiar with those boards however.

>>15822423

I genuinely think there should be pinball threads on /v/.

>>15822423

>bailiwick

Such a comfy word.

>>15822432

Make one, I would make a few questions.

>>15822207

I'm saying there's very little benefits of running Windows 10 in a VM then running it on bare metal. People should be using Windows 7 for games regardless. I use two SSDs in my desktop, one for Linux (normal everyday use) and one for Windows 7 (only games). Windows 10 should never be used period.

>completely shut off NIC access

That's not something that's VM specific. If you want to play offline then unplug your ethernet cable.

>need for at least two separate GPUs

Or like me you can just reboot into Windows. I do acknowledge the benefits of VM passthrough but using it with Windows 10 is just about as bad as booting Windows 10 naively. There's no point because you're screwed either way. Using Windows 7 should be people's only choice.

>>15822458

W10 is terrible in terms of botnet, but is W7 really that much better, especially with modded LTSB? And don't forget other devices you can wall VMs off from, like storage that it might read or (even accidentally, due to bugs!) delete, compared to dual booting.

>>15821702

>Bought a used i5 4c for 100 usd few months ago. Not even HT enabled

>Could get brand new ryzen with 6c for that amount of cash

>die internally

>>15822378

Intel's GPUs are way behind the competition, which is why Intel recently used AMD GPUs for some of their NUC and I think laptop products recently, but bringing Raja onboard to help overhaul their architecture could bring some interesting results in the future.

One of the easiest ways to improve performance is to make larger dies, containing more transistors and whatnot, which they could certainly do right now, but it's also an easy way to balloon your power consumption and heat output. I'm sure Intel doesn't want to launch something that runs as hot as the surface of the sun just to barely match AMD's performance level, let alone Nvidia's.

>>15822473

The main criticism of Win10 is about the telemetry and updates, both of which can be completely disabled with LTSB + some programs/scripts.

>>15822484

What generation was it?

ermagerrr so many cores!!!1!

>>15822517

>both of which can be completely disabled with LTSB + some programs/scripts.

Can you prove they are disabled, and that there is no risk of Microsoft releasing an update that gets around the programs/scripts? Windows 7 just seems more hassle-free.

>>15821702

Still sitting here with my OCed i7 2600k. Literally no reason to upgrade yet for games. 8 year old mid-power CPU still holding strong. Wonder if it'll hit 10 years old before I dump it.

>>15822536

I wouldn't guarantee telemetry gets entirely eliminated from any tweaks, but Long Term Service Branch (LTSB) is specifically designed NOT to get updates for years so that machines using it that require 24/7/365 reliability don't get interrupted or broken by updates.

>>15822473

>is W7 really that much better

Yes, it is. No need to play the cat and mouse game with (((updates))) like with 10.

>other devices you can wall VMs off from, like storage

That's true, Windows could read the raw bytes off the drive looking for the superblock to determine what file system it has. But beyond that I doubt it would do anything else in secret to the drive, certainly not perform any writes to it since that would very easily be detectable.

>>15822536

Just to be clear, there's no way to prove everything is truly disabled in Windows 7 either. It's just that Microsoft built Windows 10 with those features front and center. If they were going to (and have already done) do an " accidental" forced update they would do it to Windows 10, not Windows 7.

>>15821719

A-Alex Jones?

>>15822536

Can you prove Windows 7 doesn't either?

>>15822572

I'm thinking not just of snooping, but of Windows bugs just fucking up your system. There are confirmed cases of Windows accidentally corrupting or deleting partitions and drives.

>>15822571

>>15822572

It's not possible, as far as I know, to verify absolutely everything. However, I've looked at my network traffic and not seen any communication with MS servers, and after installing it and using the program Shutup10, haven't had a single update even attempt to install, nor will any update ever attempt to install as the ability to do so is disabled by the program.

Shutup10 is the program I used, but there are many other programs and scripts out there that do the same thing, or you can do it all manually. Just installing LTSB itself isn't enough to disable all updates permanently, nor is it enough to disable telemetry.

>>15821772

yeah i just enabling smt and its like a night and day different when i decide to make some webms while the game is minimized.

granted my last experience with HT was with pentium 4

>>15822595

>There are confirmed cases of Windows accidentally corrupting or deleting partitions and drives

Yeah only with Windows 10. Windows 7 has never had any issues like that in almost 10 years.

>>15822600

>I've looked at my network traffic

Sure you can try to confirm the current blocking status, but the real threat is another "accidental" forced update for Windows 10 through a previously unknown or dormant service. By the time it's received and activated it's too late if you're only taking reactive measures instead of proactive measures.

>Shutup10

>The Software is licensed, not sold.

>Prohibition on Reverse Engineering, Decompilation, and Disassembly. You may not reverse engineer, decompile, or disassemble this software except and only to the extent that such activity is expressly permitted by applicable law notwithstanding this limitation.

>Without prejudice to any other rights, O&O may terminate this License Agreement if you fail to comply with the terms and conditions of this agreement. In such event, you must destroy all copies of the software.

So you're relying on proprietary software that explicitly states you can't try to confirm that it actually works? How backwards is that?

>>15822341

To a degree, yes, but there's a fucking board for this kind of stuff.

>>15822671

The one filled with "muh buttnet" LARPers that refuse to run anything except TempleOS on their original Commodore ThinkPad abacuses because of IME/PSP/TrustZone?

If I want to discuss high-performance 80x86 hardware, especially to run proprietary games with Windows 10 in the system requirements, this is the only board to do it.

>>15822702

>>>/g/

>>>/soyboys/

>>>/coal/

>>>/cuckchan/

>>15822702

>windows 10 in the system requirements

what game does this that isn't dogshit

>>15822528

Gen 4

4ghz 50w 99$

that sounds good to me

>>15822710

Forza. Few now, but I can only expect it to increase, given the rapid shift of Steam users to 2/3rds W8/10. And even if something can run on W7, bottlenecks such as the inability to use D3D12 could become a problem if Vulkan support drops off.

>>15822113

>post crying about op uses image trying to defame a man who accepts donations to his site in the form of dismembered jew fingers

Lot of Intel loyalty for a hired gun

$300 to $500 for a cpu? AMD has almost NEVER broken the $200 price point until now. AMD has lost its touch. It use to really be the welfare hero, now it's just Intel 2.0.

>invest half your CPU into housing backdoors and botnet instead of available computing power

>surprised when competition beats you

>>15822868

>AMD is gunning hard for the compute market as a priority

Yes but unfortunately much of that is vendor locked because of proprietary CUDA software. AMD and Khronos need to either really aggressively push an open source alternative or make a CUDA transpiler or translation layer.

>>15822868

I don't know why but I suddenly got the urge to rewatch the LotR trilogy.

>>15822432

There was one about Black Rose getting pants slapped on some recent pinball simulator. I didn't post in it, but I didn't see it locked/anchored either.

>>15823031

it's not like hell wake up even if I don't tell him not to

>>15822868

Nvdia was in a bad position because of the crypto boom. That probably sounds strange, but booms are hard for companies with a long lead time to handle. They either had to underproduce and let prices go to space where only buttcoiners would go and risk losing their grip on the gay men market or produce to meet the needs of both and have a glut of cards on hand after a crypto crash. They went with the latter, crypto crashed, and they were stuck with a ton of unsold cards and weak demand. But they also had the design pipeline getting backed up. So their solution was to release the GTX 20 series early at extreme prices so that on paper they're still keeping up with the tech curve but they were priced high enough to keep people buying old inventory. The first GTX 20 run was probably very small compared to what they'd usually produce.

Things will return to normal. It's just a recovery from crypto they're going through.

>makes the most GPU known to man

>shit still won't run OG Crysis because it's optimized for Nvidia

The main problem here is that Nvida cards run everything while AMD cards lose some of their performance for more than half the games out there. Since Nvidia has their money in most devs and publishers' pockets to some degree, that's probably going to keep being the case for a while now.

>>15822051

>>15821950

At that point wouldn't it just be easier to just get a properly spec'd gaming PC and a cheaper Linux/*BSD computer for general browsing and media purposes? The non-gaming computer wouldn't need to be that powerful at all.

>>15823112

Ironically enough, now is the perfect time for miners since idiots are selling their used cards for dirt cheap. You're a fool if you thought you would set up a mining rig for a quick profit.

>>15823125

>while AMD cards lose some of their performance for more than half the games

Is that even true any more with all the console ports?

>>15823131

Setting aside other demanding tasks one might perform (compilation, transcoding, art, etc.), modern browsers and sites are getting so incredibly bloated that at least a powerful multithreaded CPU and a lot of RAM are increasingly crucial to mere web browsing.

>>15823157

Also, of course, there is the slow trickle of Linux ports, and improvements to Linux GPU drivers that are making it a superior platform for some emulators.

>>15823196

>the size of Ubi games is like 40-50% DRM

Sort of. It isn't literally DRM software, but the sheer size of the game itself possibly intended as a sort of DRM, though the ability of pirates to repack games smaller can mitigate that. Also, since (especially PC, but even console) games increasingly require massive installs/downloads even when purchased at retail, this has a 1:1 penalty on legit customers, too.

Many games are simply very inefficiently stored, ranging from more arcane goofups like a failure to compress meshes/animations/scripts/etc or use more modern codecs on cutscenes, to large amounts of uncompressed audio or textures, to very embarrassing practices like mandatory install of multiple audio localizations. Some of these are openly justified as an intentional tradeoff between CPU/bus demand vs. storage space, but unwillingness to make these optional is galling.

>>15823291

>>15823196

It might also be referring to cases like this https://archive.is/QMcpz#selection-371.0-371.149, where most of the main .exe's filesize was DRM.

>>15823262

I try not to think about 3DPD whores. I just want a kawaii 2D loli with tsundere/lewd tenancies to hang out with.

I understand where you're coming from though.

>>15823291

You know, for so much uncompressed shit you'd think the game could actually load it without chugging balls and yet I keep seeing minor graphical improvements paired with massive content downgrades for extreme hardware requirements.

For some reason a machine that can run Far Cry 4 at like 40-50fps can't get a solid 30 out of Far Cry 5 even on lower settings for relatively minor updates(it's mainly textures, some lighting and improved facial models) and some pretty big downgrades on effects(their fire effects keep getting shittier with every game).

>>15821832

It's not tricky at all, I myself used a PS/2 keyboard and it worked perfectly. Afterwards you just install the USB drivers and you're good to go.

>>15823328

Insufficient compression can sometimes hurt performance if you have free CPU cores, since the CPU saved by uncompressed assets is sometimes less valuable than the bus throughput wasted pushing it around the mobo. Talented game optimization is in many ways both science and art.

>>15822048

could be an op to make /v/ hate AMD. Think about it, all you gotta do is act retarded and samefag while shilling for something you hate, and the thing you hate will then be hated by all of /v/ due to shilling

>>15823031

Sleep Tight Terry

>>15821702

>Ryzen 5 3600G

>8C / 16T / 20CU

That's more powerful than the PS4 processor. It could be around a rx 560 in performance.

>>15823383

Intel on suicide watch.

>>15822051

>win10

>airgap-level

>security

AAAAAAAAAAAAAAAAAAAAAHAHAHAHAHAHAHAHAAAAAAAAAAAAHAHAHAHAHAHAHAHAAAAAAAAAAAA~

>>15823328

Farcry 5 has some issue specific to it if not something introduced to the dunia engine, the guy from gamers nexus has mentioned it in passing in a few of his last videos, something about the framerate interval in miliseconds which affects every cpu

>>15823383

you're literally describing current Intel housefire CPUs with your post, gg anon.

>>15823383

Be fair, Intel's highest clocked parts since they've become so ridiculously stingy with unlocked multipliers and turbo boost on cheaper parts in the years since awesome parts like the G3258 also have enormous numbers of cores. Simply by shutting down extra cores, such CPUs can clock similarly to smaller devices.

Granted, it's pretty wasteful, but having the option is nice if you want to run stuff like badly multithreaded games and emulators.

>>15823461

>>15823441

>samefagging

>padded 13-digit utc filename

You have to go back.

>>15822031

>intel shills this mad

<K-K-KEEP BUYING INTEL GO–GUYS, OUR PRODUCTS ARE SUPERIOR AND RUN VIDEO GAMES BETTER

<LOL AMD MORE LIKE M-A-D. MAD THAT THEY SUCK

wait why are people so mad about good/decent cpus at a decent price.

>>15823504

Didn't answer my question

>>15823514

CPUs and GPUs are both heavily involved with being able to play video games.

A new high end CPU at a very affordable price is relatable because it opens the door up for more people that can't afford Intel CPUs at their inflated price point. It doesn't help that modern games are programmed by lazy degenerate faggots that can't code worth a shit and push out bloated trash that runs like dog shit on anything one generation old.

So cheap powerful CPUs means that more people can upgrade their computers with better hardware for a fraction of the price which means more people are able to play more video games.

This is very relatable to me because right now money is tight and my CPU is becoming a bottleneck for newer games and if I want to record or stream, and these powerful CPUs for bargain prices will give me the opportunity to upgrade and play more games.

>>15823505

INTEL has a 90% market share and near total dominance of all CPU manufacturing infrastructure and they don't want you to have something good and cheap. AMD threatens this.

How do you not understand this?

>>15823591

sounds terrible well haven't really used a intel in like a decade so can't say

>>15823328

ubisoft sometimes does some retarded shit, remember asscreed unity which sometimes had more drawcalls than most hardware could handle?

>>15823291

> to very embarrassing practices like mandatory install of multiple audio localizations.

fucking titanfall man, and on top of that that shit was all uncompressed because "muh cpu-cycles". their excuse was "we optimized down so even a core2 duo can play it". yeah right.

>>15822423

imo video games is the software, hardware is more /tech/. but then I'm not that triggered by the occasional offtopic thread, so who cares.

>>15823006

>>15822894

they're still around in italy

Best treat to buy when you're visiting parents for a big dinner and pasta and seconds have already been covered by the others

>>15823386

I can see that, that goes along perfectly with the falseflag "raids" that are happening.

>>15823802

intel shills just want you to stop talking about it and just buy it!

>>15823386

Same thing happened with nier automata here, shit OPs like "if you dont think this game is the best game of all time you should kill yourself" etc.

>>15821702

Wake me up when AMD is at the top of this chart.

https://www.cpubenchmark.net/singleThread.html

>>15821774

Use Windows 8.1 with start is back instead.

>>15824603

All I saw about Automata on /v/ was "Wow Taro is a fucking evil bastard who stole some faggot nobody from a AGDG's waifu so the game is shit." As if I'm supposed to care about what happened to some cuckchanner.

>>15822894

>tfw chilean and I can buy Viennetta almost fucking anywhere

Sucks to be you, gringo

>>15822894

Wait, you don't have Viennetta in the US? I thought it was the "greatest country in the world".

Is the RX 580 any good for the price? I want to stay future proof at the very least so I can play the newest unoptimized shit at a decent framerate.

>>15825373

most powerful government. doesnt mean were happy.

>>15821787

>>15821774

there is still remote control shit on new amds

>>15821920

I asked /tech/ last year, the pointed out what was basically Windows 10: Windows 7 Edition to me and I pirated that.

>>15822075

>The 3060 is equivalent to a RX580 and is priced at $130

I've spent $380 for 580 because it was during bitcoin craze. fuuuck.

>>15823031

Sleep Tight Terry

>>15823383

>All that retardation over not understanding that Boost |= All core Boost

>>15822075

The important thing is that if this is true, ray tracing will completely die. There is a massive market of people waiting for cheap 1080s or equivilents. By never releasing an affordable mid-range card, Nvidia is basically trying to coerce people into getting an RTX-capable card. Charging $250 for a 1080 equivilent will shit all over Nvidia and all of their work and give AMD a spike in the video card market. They might actually have to put in some fucking effort again.

>>15825294

I don't know where you saw that, because AGDG shits on that fag harder than anyone else. Not like he'd ever work on his game again after becoming a laughing stock to us.

>>15826784

>Charging $250 for a 1080 equivilent will shit all over Nvidia and all of their work and give AMD a spike in the video card market. They might actually have to put in some fucking effort again.

Nah, what'll happen is everyone will say "wow I can't wait to see what Nvidia puts out in response" and sit on their hands until Nvidia has time to roll its enormous marketing and R&D budget over to crush AMD and then they'll go back to buying overpriced bullshit. It's happened to AMD twice now, don't see why anything would change in the future.

>>15826784

>if this is true, ray tracing will completely die

Reminder that according to nVidia & MS themselves, the RTRT ASIC on RTX only increases DXR API performance by 2-3x over a pure shader/CPU implementation that also works on AMD:

https://www.techpowerup.com/242504/microsoft-releases-directx-raytracing-nvidia-volta-based-rtx-adds-real-time-capability

Reminder nVidia's RTRT ASIC plopped onto the mammoth RTX die, at 6 gigarays/second, performs on par with a literal cellphone GPU's RTRT ASIC from PowerVR several years ago:

https://www.youtube.com/watch?v=EGBbw2DLBEA

Reminder AMD has offered their own implementation of RTRT in Vulkan, that can also run on nVidia, for years:

https://www.tomshardware.com/news/amd-radeon-rays-raytracing-software,36702.html

Reminder that RTRT, frequently on par with or exceeding the capabilities of rasterized hardware 3D, as opposed to the terrible looking interpolated hybrid garbage RTX does, has been implemented by penniless teenage demoscene hackers since the 90s:

http://realtimeradiosity.com/demos/

Reminder the only reason everything still uses triangular Gouraud-shaded polygons with bitmapped textures largely identical to what the 3Dfx VooDoo did is because game engine programmers are worthless ignoramuses.

>>15826792

Because this time the attitude towards Nvidia is similar to the attitude towards Intel - people are finally realizing they have been getting ripped off, and they want competition in the market again. People see how good Ryzen was at shaking up Intel, and if AMD can actually hit hard with Navi I think people will be very happy with it.

>>15827636

>an incredibly demanding task.

Sure, but far from an insurmountable one (compared with, for instance, a pure software rasterizer on even the fastest multicore CPU versus a current GPU).

>radeon rays is not a real-time renderer

It is now, AMD is even using it for that purpose in their own ProRender package.

>120W TDP

At 28nm, which is 2.3x less efficient than RTX's fab node.

>Real-time radiosity is not the same thing as real-time ray-tracing and scene complexity is about ten orders of magnitude greater in the scenes Nvidia is trying to light

The demos on that page use a variety of techniques, with a variety of scene complexities, since they stretch from the present back 20 years. Note that nearly all of them are full-scene raytraced, many near or even surpassing contemporary conventional 3D hardware in terms of scene complexity.

My point is that the RT performance of RTX is easily within the grasp of conventional hardware, that there is enormous headroom for RT ASICs within typical GPU semitransistor budgets, and nVidia's timid avoidance of full-scene RT even with the hardware they're selling does the potential of RTRT a great disservice.

ever played switch? you put one thumb in your ass and another in your mouth, and when somebody says "switch" you switch!