[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/g/ - Technology

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • You may highlight syntax and preserve whitespace by using [code] tags.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


[Advertise on 4chan]


File: 1757740892894573.png (184 KB, 750x500)
184 KB
184 KB PNG
Everyone hated this product at launch, saying it only had 12GB of VRAM and wasn't fast enough. Now everyone has been buying this GPU in droves. What gives? Is this really the GPU people are going to be settling on for the rest of their life?
>>
I boughted the upgraded version which is called dgx spark
>>
it's the only option between the $250 8gb price bracket at the $1250 16gb price bracket

8gb is insufficient as of last year
12gb is insufficient for the future
>>
>>108091107
Pedantically bitching about the market probably isn't useful. Figure out how to make the IRL AI robot work faster so we can have all of the silicon, silicone and plastic trash we want.
>>
>>108091120
>12gb is insufficient for the future
People were already calling 12GB obsolete today given 12GB GPU's have been around for a decade in mass market, while 16GB were also available for a decade albeit in a limited scope but are seen as the new minimum by many. Yet many still chose the 5070 over the 5060 Ti. Why?
>>
>>108091142
5070 has a 30% performance advantage in games with extra cuda cores and memory bandwidth. When the 12GB ram starts to choke due to resolution, 16GB 5060ti wouldn't fare much better at that point.
>>
File: 1770546368895.png (202 KB, 512x512)
202 KB
202 KB PNG
>>108091107
Nobody is buying this. Everyone is trying to get the Ti variant but they're already cucked paying 300$ or more compared to september kekypow
>>
>>108091107
because Njudean goyslaves are retards. There is no function reason to buy a 5070 unless you have brain damage, are a completely uninformed consoomer, or have been taught to hate AMD by other goyslaves you bootlick.

If you're gonna paypig for Njudea, get a 5070ti jesus....
>>
>>108091107
It's a cope because that's all most people can afford now. 16GB is the bare minimum these days. I don't even game that much and still bought a 4070 Ti Super back in August 2024 because I saw the 50 series being a pile of shit coming. With a 20% retailer discount and two overlapping ANUS cashback programmes I only paid base 4070 money for it too. Best GPU purchase I've ever made.
>>
>>108091268
Thought it was funny how trying to explain this before always just got drowned out. The 16gb is more useful for gening porn though, which is probably the main reason most people were buying them anyways
>>
>>108091282
>phenom
>dabbed all over intel
What kind of revisionism is this, Phenom couldn't into 3ghz due to 65nm and Phenom 2 barely reached IPC parity to core 2, then immediately got murdered by Nehalem.
>>
>>108091107
People buy it because it's the new thing and they want to play their favorite AAA slopware that is so unoptimized it can't run without memescaling and fake frames.
It's why I play factorio on a, by today's standards, toaster.
>>
>>108091275
>Nobody is buying this.
Steam charts disagree with you, Microcenter prices disagree with you. These were the last GPUs you could buy at MSRP sans taxes and for some things like Blender the price to performance just made sense.
>>
We're going back to 8GB VRAM as standard, 12 GB is the new GPU nobility
>>
File: 1770550423343.png (266 KB, 496x783)
266 KB
266 KB PNG
>>108091486
Said nobody ever
>>
File: not very clever.jpg (41 KB, 370x472)
41 KB
41 KB JPG
>>108091306
im not sure how you didn't get this far without knowing this, but a ~8900k at 5+ghz is still wildly inferior on every front, in every way, to a 9800x3d that is being performance limited to base clocks at 4.7ghz. The Ghz stat only matters in regards to the individual cpu in a vacuum. Cinebench, in-game stress-testing, chrome-stress testing, and winzip benchmarks are what really matters for cpu scores, not numbers on the page.
>>
>>108091518
but it's happening, the future is here and is 8GB VRAM and 16GB RAM
>>
>>108091486
12GB is the new 6C/12T to the 8C/16T of 16GB.
>>108091518
>anime
kys srsly
>>
>>108091486
There's no way in hell that the 6060 will have less than 12GB.
>>
File: 144727-2005253962.jpg (30 KB, 384x384)
30 KB
30 KB JPG
>>108091537
>>
File: 1759464977401284.png (48 KB, 841x755)
48 KB
48 KB PNG
>>108091532
Phenom had the same or worse IPC as Phenom II (latter having more L3 cache) and clocked lower, Phenom II had the same IPC as Core2, how in the fuck is Phenom at 2.6ghz going to beat an i7 which had better IPC than Core2?
>>
>>108091107
it's all that's left. woulda absolutely necked myself had i gone for that instead of the Ti jesus christ what an absolute scam of a product.
>>
>>108091616
5070 Ti was like 900 dollars after tax in many places despite being worse than the preceding generations 4080/4080 Super in several areas. 5070 could be had for 550 at its lowest, 610 after tax in many places.
>>
>>108091622
Shit I saw the 5070 drop under $500 during bf and cyber monday sales. Paid less for mine than I did for the 2070 super which it replaced that was also a bf/cyber monday deal
>>
>>108091107
It is because it is only Blackwell option in Nvidia town that isn't massively inflated at the moment.
>MAH 12GiB
The buyers in this demographic are running at 1080/1440 at most. 12GiB is sufficient for this. Nvidia knows how to price its product and make it move. 5070Ti, 5080 and 5090 are still moving even at their current price points.
>>
>>108091518
You will gobble and swallow gabe's cubes and you will be happy.
>>
>>108091107
Most online discussion about hardware is driven by a loud minority of enthusiasts and influencers who spend way too much time reading about high end products and their specs and have completely lost touch with the kind of compromises actual budget buyers are willing to make.
>>
>>108091124
this, with robots everything except real estate and some artificially costly "luxury" i.e. handmade things will deflate massively
>>
>be me
>for the past 15 years I've upgraded my GPU approximately every 3 years
>have always been able to double or triple my raster performance every 3 years for the same price as my last GPU
>haven't upgraded from my 3060 Ti since 2020 though
>"Hmm, what can I buy today for the same price as I got my 3060 Ti?"
>it's all 3060 Ti equivalents
Even inflation-adjusted you're getting scammed. When will I be able to get a 2x boost over the 3060 Ti for a 3060 Ti's price? In 50 years, right?
>>
>>108091770
>Shit I saw the 5070 drop under $500 during bf and cyber monday sales.
Why would you buy a 5070 when you can buy a second hand RTX3080 for $250~ and it's barely 10% slower? I got a RTX3080 10GB for $200 a few weeks after nvidia starting shilling 5070 = 4090 lmao.
>>
>>108092176
There's a reason why Nvidia is aggressively pushing AI upscaling
>>
>>108091107
I bought a 5060ti 16gb

Not my problem
>>
>>108092155
The niggers will just breed more niggers.
>>
Youtubers and other online retards seem to think people spend hours looking up price/performance fps per dollar charts and other such nonsense and faggotry

Most cards are simply sold by "I have X dollars, what's the best Nvidia card I can get for that"

In 2025 X was $500 and the card was the 5070.
>>
>>108092250
uninformed buyers fuck themselves over, a tale as old as time. Don't tell us not to care at all, this entire generals point is to mutually educate each other on not making bad choices.

its obvious we're a tiny minority compared to the gay orgy going on at reddits PCMR subreddit but they're all fags who deserve to suffer so meh.
>>
>>108091834
>12GiB is sufficient for this.
Doesn't ray tracing use WAY more memory if you need to use it in a game?
>>
>>108092176
>"Hmm, what can I buy today for the same price as I got my 3060 Ti?"
I made the mistake of buying a 3060 Ti in 2022 for like $750 so in my case this was the 5070 as the upgrade for the same price.
>>
>>108091107
because the others options cost at least $400 more if they are available.
>>
>>108091107
people are too focused on being brandcattle to realize that a 9070/xt is the superior option if you aren't spending the big nigger dollars on 5080/5090. there's literally no reason to buy a mid-low tier nvidia card.
>>
>>108092617
Yeah those who bought 6900xt and 7900xt are real happy about fsr4 support right now.
>>
>>108092241
that is already a problem
>>
>>108091532
amd cpus were trash for like a decade until ryzen came out. full stop
>>
File: moving goalposts.jpg (262 KB, 1056x594)
262 KB
262 KB JPG
>>108092955
....and?
>>
>>108092655
Even if those cards had FSR4 it wouldn't matter because no games have FSR4.
>>
https://videocardz.com/newz/rtx-titan-blackwell-or-rtx-5090-ti-nvidia-reportedly-is-developing-a-new-rtx-50-series-halo-gpu-for-q3-2026

NVIDIA WON
>>
>>108093210
>msrp: $7250
Calling it now.
>>
File: 8319.jpg (857 KB, 1330x1440)
857 KB
857 KB JPG
>>108091622
5070 ti dropped to $650 before new year at microcenter. Just live someplace relevant enough to have one.
>>
>raughs in rtx 3090 I bought for 500 bucks
>>
>>108092655
You can inject it / mod it in anyway
>>
>>108093210
Cool. They just have to actually release it this time instead of only keeping it in the lab like the 4090ti and titan ada
>>
>>108093133
because everyone was talking about phenom and you started talking about 9800x3d to make a point that was already addressed by, verbatim quote, "Phenom 2 barely reached IPC parity to core 2" you dumbass. you fucking goober. learn how to read
>>
File: 1756701451425399.png (1.01 MB, 976x1149)
1.01 MB
1.01 MB PNG
>>108093210
on god we do not need a 5090 TI Super Titan
They are taking all the scrapped 50 series VRAM and going to sell 8K roided up 5090s instead

The K-shaped economy marches on. I saw saw a tweet that said 60% Taco Bell customers earn like $100,000 so they are going to raise the prices, they know their customers can pay it. It might not have been Taco Bell but it was some fast food slop joint.
>>
>>108091120
>it's the only option between the $250 8gb price bracket at the $1250 16gb price bracket
Or you could stop sucking Nvidia cock and buy AMD. Paid €389 for my 16 GB 9060XT.
And you can get a 9070 XT for 800.
>>
>>108095038
>Spending flat with inflation
Gee I wonder what the reason for this is.
>>
Don't know what to buy next
Could be a monitor
Could be some more storage
Dunno
>>
>>108091120
5060s are $349 now
>>
>>108091120
>$250 8gb price bracket
What the fuck is wrong with you retards? And the OP as well, do you fucking mental patients live in 2023?
>>
>>108095113
definitely monitor. this is the year all the good shit comes out.
>>
File: 1589540230792.jpg (329 KB, 1080x1350)
329 KB
329 KB JPG
>>108091107
>70
It's always the 2nd fastest selling card after *60. No matter what "influencers" might have you believe.
>>
>>108091301
The only people complaining are malcontent influencers and khhv tranny chatbot aficionados.
>>
File: 1598378485324.jpg (127 KB, 800x1131)
127 KB
127 KB JPG
>>108091536
This is an anime website.
>>
>>108093261
>for 500 bucks
When was this? I never saw them go below 670-700 dollars.
>>
>>108091107
>Now everyone has been buying this GPU in droves
The fuck are you even talking about? Shits not even available.
>>
>>108093210
blackwell pro 6000 already exists, this is pointless
>>
>>108095434
Technically a 5090ti is possible.
>>
>>108095038
Debt has alot to do with this.
A huge portion of society have all of their income rapped up in debt from acquiring their possessions. Most people will look at you like you're retarded if you spend money to buy things.
>>
>>108091142
>Why?
The fell for:
> Nvidia Neural Texture Compression delivers 90% VRAM savings with DXR 1.2
https://overclock3d.net/news/gpu-displays/nvidia-neural-texture-compression-delivers-90-vram-savings-with-dxr-1-2/
>>
>>108095454
yeah but it is gonna cost like $5500, maybe more. the msi lightning 5090 that was just released a few days ago was like $5200. with these prices, you might as well just get a blackwell 6000 for like $8500.
>>
>>108091107
The same people who refuse to use anything other than Steam refuse to get a 9070xt. You get what you deserve
>>
>Everyone hated on my guy the rtx 5070 at launch!!1
People shat on the 5070 because the leather jacket retard went on stage and claimed that it was a 4090 for $499
Shut the fuck up you consumer brain moron
>>
>>108095481
People hated it because it was still yet another 12GB GPU with barely any performance improvements over the 4070 it replaces, let alone the 4070 Super.

It about matches 3090 performance, which every xx70 GPU and up from the previous generation also did.
>>
>>108093210
Nice! I will reserve $5000 in my hobbies budget. No new wheel set for my race bike this year.
>>
>>108095434
> blackwell pro 6000
Not for under 10k if even available. A 5090Ti would have significantly less VRAM and be "only" 5k or less.
>>
>>108095547
$8150 on newegg.
https://www.newegg.com/p/N82E16888884001
>>
>>108095479
> people who refuse to use anything other than Steam
I refuse to use anything other than torrents. Can you guess my GPU?
>>
File: lol.png (1.25 MB, 3840x2160)
1.25 MB
1.25 MB PNG
>>108095481
seriously jensen doesn't get grilled enough for this. its almost cartoon villain tier.
>>
>>108095556
Mmmmhh... that's actually tempting.
I just don't like the cooler. Are there good waterblocks for it?
>>
>>108095547
>>108095556
It will be priced somewhere around an rtx 5000.
>>
>>108095576
B.. b.. but 6x frame gen.
>>
>>108095606
not that i know of. i have the max-q edition, and its temps are pretty good with just the normal air cooler.
>>
>>108095576
If you judge "performance" by sheer FPS no matter how shitty the image qt and input latency, THEN he was right. And you know how stupid gaymors are.
>>
>>108095612
> max-q edition
The blower? What about the noise?
>>
>>108095641
yeah the blower. it is pretty quiet. i never hear it.
>>
>>108095612
>>108095641
> Features...
> 300 W Max Power Consumption
This explains it I guess.
>>
>>108095653
>>108095659
Imagine it had to dissipate 600W instead of just 300.
>>
>>108091107
I didn't because I'm not an NPC techlet comsoomer and thus I didn't fall for the nVidia meme.
Gaming at 1440p >100 FPS, running LLMs, doing all kinds of great things and for $300
Enjoy being a sheepbrained consoomer.
>>
>>108095623
he's implying 192bit bus = 384bit bus, which is grade SSS bullshit.
>>
>>108092617
I bought a 5070, friend got a 9070XT, he has a worse experience in 90% of the games simply because DLSS works better.
And he has to inject FSR4 into most games as it seems abandoned by AMD & the developers simply do not care about AMD GPUs.
Also good luck using 3rd party injectors for FSR4 in games that have aggressive Anti-Cheats.
>>
>>108095698
> he's implying 192bit bus = 384bit bus
Is he? I thought he was only referring to the MF generation of the 5070.
>>
>>108095738
> DLSS works better
Who would have thought?
>>
File: 2025-11-01 23.07.47.png (292 KB, 774x435)
292 KB
292 KB PNG
>>108091107

You know dummies fell for it thinking they would get 4090 perf.
>>
>>108095797
They did! But only 50% of it. Where did Jensen say it was 100%?
>>
no housefire connector for me, thanks
>>
>>108095797
It was a typo, he meant 3090 performance... though he did immediately correct himself saying it was impossible without AI fakery.
>>
>>108095797
For some reason I'm thinking about the GTX 970 3.5GB looking at this image, when thinking about the RTX 5070.
>>
>>108095946
> "only possible with AI"
He clearly intended to push the AI hype.
>>
>>108095797
It was 4090 mobile performance.
>>
>>108095454
There is zero reason to make it. 4090 are the subpar GB202 yields that can't make the professional cut and they are selling at $3K+
>>
>>108093210
100% non-sense. Nvidia has ZERO reason to make any gayming SKU. They don't have the GDDR7 supply to throw around and their current Blackwell inventory is still moving at current price points. Gayming market is just not sexy. The next actual gayming SKUs from Nvidia will be Feynman-era silicon. Rubin is the successor to Hopper which is enterprise only.
>>
>>108096116
They also don't require ecc memory and studio drivers.
>>
>>108095176
MILCH
>>
>>108092176
Moore's law has been dead for years now. Nvidia's AI shenanigans is an attempt to get around that.
>>
>>108091107
im playing cdda rn on my 5080
straight gaming
>>
>>108091120
or you can get a 9060xt because the 16gb is more important than the cootie cores
>>
>>108091142
>16GB were also available for a decade albeit in a limited scope but are seen as the new minimum by many
you need at least 16gb to generate futa. 12gb can only generate dickgirls



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.