• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

NVIDIA GeForce RTX 5060 Ti 8 GB Variant Benched by Chinese Reviewer, Lags Behind 16 GB Sibling in DLSS 4 Test Scenario

T0@st

News Editor
Joined
Mar 7, 2023
Messages
3,063 (3.88/day)
Location
South East, UK
System Name The TPU Typewriter
Processor AMD Ryzen 5 5600 (non-X)
Motherboard GIGABYTE B550M DS3H Micro ATX
Cooling DeepCool AS500
Memory Kingston Fury Renegade RGB 32 GB (2 x 16 GB) DDR4-3600 CL16
Video Card(s) PowerColor Radeon RX 7800 XT 16 GB Hellhound OC
Storage Samsung 980 Pro 1 TB M.2-2280 PCIe 4.0 X4 NVME SSD
Display(s) Lenovo Legion Y27q-20 27" QHD IPS monitor
Case GameMax Spark M-ATX (re-badged Jonsbo D30)
Audio Device(s) FiiO K7 Desktop DAC/Amp + Philips Fidelio X3 headphones, or ARTTI T10 Planar IEMs
Power Supply ADATA XPG CORE Reactor 650 W 80+ Gold ATX
Mouse Roccat Kone Pro Air
Keyboard Cooler Master MasterKeys Pro L
Software Windows 10 64-bit Home Edition
NVIDIA's GeForce RTX 5060 Ti 8 GB graphics card design received little fanfare when review embargoes lifted mid-way through the working week. Reportedly by official instruction, involved board partners sent out 16 GB samples to evaluators. Multiple Western outlets are currently attempting to source GeForce RTX 5060 Ti 8 GB cards—on their own dime—including TechPowerUp. As mentioned in his conclusive rundown of PALIT's GeForce RTX 5060 Ti Infinity 3 16 GB model, W1zzard commented on this situation: "personally, I'm very interested in my results for the RTX 5060 Ti 8 GB, which I'm trying to buy now." The ever reliable harukaze5719 has already stumbled upon one such review. Yesterday, Carbon-based Technology Research Institute (CBTRI) uploaded their findings onto the Chinese bilibili video platform.

Two ASUS options were compared to each other: an 8 GB Hatsune Miku Special Edition card, and a better known property: PRIME RTX 5060 Ti 16 GB. In most situations the two variants perform similarly. A clear difference was demonstrated when CBTRI's lab test moved into a DLSS 4 with Multi-Frame Generation (MFG) phase. Both harukaze5719 and Tom's Hardware noted a significant gulf—the latter's report observed: "in Cyberpunk 2077, for example, the RTX 5060 Ti 8 GB inexplicably performed worse than the RTX 4060 Ti 8 GB at native 1440p resolution. While enabling MFG helped improve performance, pushing it to 4x delivered underwhelming results, with the 16 GB version providing 22% higher performance than the 8 GB card." Rumors have swirled about the late arrival of GeForce RTX 5060 Ti 8 GB cards at retail; potentially a week after the launch of 16 GB siblings. As evidenced by early results, potential buyers should consider paying a little extra ($50) for a larger pool of VRAM. Team Green's introductory material outlined starter price tags of $429 (16 GB) and $379 (8 GB).



View at TechPowerUp Main Site | Source
 
accourding to Hardware Unboxed they specifically requested an 8gb model and Nvidia told them to contact the partners for one and when they did that those partners said they were specifically instructed by Nvidia not to send out any for review.....

Now if that is true...and I have no reason to doubt Hardware Unboxed....man is Nvidia shady....not that it matters, yall will be supporting them anyway.
We live in a consequence free world.
 
We already knew DLSS (and RT, for that matter) eat more VRAM, no surprise there. 5060Ti performing worse than the 4060Ti, that something worth looking into though.
Delayed availability or not, people will review the 8GB variant.
 
Its been said before and it will get said again, nVidia are not your frined and never will be. They do not give a flying fcuk about you, the customer and havent for a very long time..
 
Its been said before and it will get said again, nVidia are not your frined and never will be. They do not give a flying fcuk about you, the customer and havent for a very long time..
Companies are not your friends in general. They will always exploit a dominant position. AMD or any other company is not any different, they are just underdogs so they're forced to play "nicer".
 
Looks like the 4060 Ti 16GB will be quicker than the 5060 Ti 8GB in some situations then.
 
accourding to Hardware Unboxed they specifically requested an 8gb model and Nvidia told them to contact the partners for one and when they did that those partners said they were specifically instructed by Nvidia not to send out any for review.....

Now if that is true...and I have no reason to doubt Hardware Unboxed....man is Nvidia shady....not that it matters, yall will be supporting them anyway.
We live in a consequence free world.

Which is why people need to wait for Radeon RX 9060 series. I did for the RTX 5070 series and found out RX 9070 XT was a better deal. And it was. Also a bit to spite nGreedia. If AMD plays this right, RX 9060 might be an excellent option. RTX 5060 just feels so underwhelming, it's like they barely put any effort into it and just bank on stupid overinflated charts with DLSS set to Ultra Garbage Quality Mega Performance preset with multiframe generation cranked to the max so you get 400 frames in everything and just ignoring the latency. The charts they posted with latency numbers are meaningless. Even i don't know what frame latency is the point where mouse feels like ass, I just know when I feel it in the game. So, they keep posting irrelevant numbers, but most reviewers always say it feels like hot garbage as is and it just gets way worse on lower end parts because base framerate is already bad and feel is based on that, not the inflated frames number.
 
Which is why people need to wait for Radeon RX 9060 series. I did for the RTX 5070 series and found out RX 9070 XT was a better deal. And it was. Also a bit to spite nGreedia. If AMD plays this right, RX 9060 might be an excellent option. RTX 5060 just feels so underwhelming, it's like they barely put any effort into it and just bank on stupid overinflated charts with DLSS set to Ultra Garbage Quality Mega Performance preset with multiframe generation cranked to the max so you get 400 frames in everything and just ignoring the latency. The charts they posted with latency numbers are meaningless. Even i don't know what frame latency is the point where mouse feels like ass, I just know when I feel it in the game. So, they keep posting irrelevant numbers, but most reviewers always say it feels like hot garbage as is and it just gets way worse on lower end parts because base framerate is already bad and feel is based on that, not the inflated frames number.
Well, I don't mind DLSS and Blackwell does support neural rendering which RDNA4 doesn't (no games take advantage of that today, but it's in DX, that will change). So the sliver of an advantage RDNA4 has may not be a decisive buying factor.
The problem I see with the Blackwell lineup is going from 5060Ti, 5070 is like 40% faster, 5070Ti is 80% faster. Those are huuge gaps. Almost directly telling you: "whatever you buy today, you'll be sorry tomorrow when we fill those gaps with the Supers".
 
As evidenced by the TPU reviews, 8GB is clearly sufficient and will remain so for the next decade or more, regardless of resolution.
Clearly, these tests were conducted by Nvidia heretics seeking to defame the sacred image of the venerable Jensen Huang.
Serious Dog GIF by GifGari


1744929278236.png
 
So Nvidia typically lies on charts and in their presentations. Hope their customers stop buying their product.
 
We already knew DLSS (and RT, for that matter) eat more VRAM, no surprise there. 5060Ti performing worse than the 4060Ti, that something worth looking into though.
Delayed availability or not, people will review the 8GB variant.
Nah, you can’t draw any conclusions between 9 and 13 fps. Anyway raster performance could be close between 8 and 16gb, but clearly worse dlss4 and FG, that defies Nvidia’s advertising. ;)
 
Well, I don't mind DLSS and Blackwell does support neural rendering which RDNA4 doesn't (no games take advantage of that today, but it's in DX, that will change). So the sliver of an advantage RDNA4 has may not be a decisive buying factor.

AMD also demonstrated neural rendering running on RDNA4:

FSR4 is close to DLSS4. RDNA4 is used in the PS5 PRO,so lots of developers will use it features.


FSR 4 isn't just for upscaling, either; it also has frame generation. From our understanding, both upscaling and framegen will use the AI accelerators of the RX 9000-series GPUs. AMD also says RDNA 4 is "neural rendering ready" without really going into further detail. Presumably that's related to Microsoft's new Cooperative Vectors feature, which is something Nvidia also talked about with Blackwell.
 
Just watch as AIBs will raise the prices on 8GB models to take advantage of all the reviewers trying to source the cards we’re not supposed to be reading about.
 
As evidenced by the TPU reviews, 8GB is clearly sufficient and will remain so for the next decade or more, regardless of resolution.
Clearly, these tests were conducted by Nvidia heretics seeking to defame the sacred image of the venerable Jensen Huang.
Serious Dog GIF by GifGari


View attachment 395624
LMAO, using 8GB in 2035? Are you serious?

Why do you guys get so hung up on 8GB? Imagine someone in 2015 on forums bragging about how his 512MB GPU would be plenty for games in 2025. We've already had sites like TechSpot show that there are games that run very poorly with 8GB (Forspoken) or dont run at all (RE:VII) and other games that have rendering issues, like missing or overly simplistic textures, which result is surprisingly good FPS numbers for 8GB cards, because the rendering errors mean higher FPS.

8GB cards came out 11 years ago. The 5060ti is a FAR superior card in terms of compute power to the likes of the 290x or vega 64. We also know using features like frame generation will easily saturate 8GB cards, even 12GB cards struggle at times. 8GB on cards that retail for $500+ are just insulting, at $600 outright offensive! The 3080 came out with 10GB half a decade ago at $700. 8GB should be the territory of sub $200 hardware by now.

I suppose when the consoles move to 32GB of combined RAM the PC mustard race will finally release their hold on 8GB hardware.
 
Last edited:
Just wait. Someone will defend this.
Was that supposed to be subtle?

So Nvidia typically lies on charts and in their presentations. Hope their customers stop buying their product.
Hey, news flash: Everyone lies about their products! This includes but is not limited to AMD, Intel, NVidia, microsoft, Apple, Google etc, etc, etc! That is the very reason tech reviewers exist!

IF manufacturers were honest places like TPU would not need to exist as we could trust what was claimed. They are not honest so here we are.
Tada..
morpheus-real.gif


LMAO, using 8GB in 2035? Are you serious?
2035? You might want to look at a calendar Marty..

We already knew DLSS (and RT, for that matter) eat more VRAM, no surprise there. 5060Ti performing worse than the 4060Ti, that something worth looking into though.
Delayed availability or not, people will review the 8GB variant.
Looks like the 4060 Ti 16GB will be quicker than the 5060 Ti 8GB in some situations then.
What graphs are you two looking at?
 
Last edited:
Crazy that we are getting 5030 class cards starting at $550 now. I don't see why anyone would upgrade to this unless they are forced, given one could have had this level of performance for much cheaper (and with free games) last gen. One can only hope AMD's response is reasonably priced but thus far AMD has not lived up to it's 9000 series pricing promises of reaching MSRP.
 
Nah, you can’t draw any conclusions between 9 and 13 fps. Anyway raster performance could be close between 8 and 16gb, but clearly worse dlss4 and FG, that defies Nvidia’s advertising. ;)

Without most likely physx the card is not even useable for retro gaming.

Crazy that we are getting 5030 class cards starting at $550 now.

I put several cheap nvidia cards with four display outputs years ago in desktop computers at work. They had very bloated graphic card drivers for windows 10 pro which were a problem with those 8GiB DRAM boxes. I think the graphic cards were nvidia 710 or 730. cheap cards, bad drivers, sluggish, and windows update did not install proper drivers.

Well, I don't mind DLSS and Blackwell does support neural rendering which RDNA4 doesn't

A longer argument about why DLSS, neural rengerding and raytracing for myself as a Cheapskate is not worth it. Especially in the viewpoint of a basic graphic accellerator card and not a gaming card: (i hope mods will not delete this post - as it looks a lot off topic now)

I do not get consumers. I bought cheap, entry, low spec graphic cards past years. RAdeon 6600XT, later Radeon 6800 non xt and 7800 xt.
If I were willing to spend 120€ on a single raytracing game I may have demand for a nvidia graphic card.
If I were willing to spend 120€ 6 times a year on games than i would also have a 8000€ computer. I do not buy games anymore because STEAM and EPIC annoy me with advertisement and always internet connection requirement and windows 11 pro restriction. Steam on gnu linux is still not worth it, especially gnu gentoo linux. I do not want to see ubuntu binaries on my box when nearly all libaries are compiled and are newer anyway on my box.


My 7800xt is from november 2023. I hardly see rayracing. I hardly need it. Those AMD giveaway games with AMD hardware like 7600X, X670, Radeon 6800 non xt, Radeon 7800xt do not show any difference with raytracing. I just started to fast play Control (that is a game) again with the worst settings possible for the 7800XT. I just wanted to see ~2 weeks ago how my card does with Control with nearly max settings. Only one or two options are not maxed out. Control was free on the epic game store. I do remember that my RAdeon 6600XT fresh bought at release had struggles with Control in WHQD resolution. 7800XT basically can run max (except a few options)

I only need raster. When I may get a single Raytracing game which needs a card I will just leave it out. Raytracing is like physx something I hardly saw in games. Avatar - Pandora = bad graphics, Star wars Jedi survivor = barely a difference. The difference is the higher power consumption and a heated room in winter while gaming.

--

Anyway: That card is really 50x "the SCAM". I refer to the chart where nvidia showed 50x of something.
 
accourding to Hardware Unboxed they specifically requested an 8gb model and Nvidia told them to contact the partners for one and when they did that those partners said they were specifically instructed by Nvidia not to send out any for review.....

Now if that is true...and I have no reason to doubt Hardware Unboxed....man is Nvidia shady....not that it matters, yall will be supporting them anyway.
We live in a consequence free world.
When has ngreedia NOT been shady ? that's the q.
 
Just wait. Someone will defend this.
8GB is fine! With MFG I can still get >60 FPS! And you don't need ultra! I can also just run 720p upscale.. oh wait I already was. Euhhmmm! Its great! The more you buy the more you save! Herp derp

Its also funny how the 16GB card this time, when actually equipped with more than entry-level bandwidth, now does manage to sprint ahead of its 8GB sibling. Nvidia knows its becoming very hard to hide this immense disparity that is now undeniably just due to VRAM capacity.

But its fine, right! 'Its only an x60, this is normal'
:kookoo:
 
This year is probably some kind of race to see who will release the worst product, the worst card :kookoo: ?
The worst top card - RTX 5080, performance increase to RTX 4080 Super in the range of 9-10% at a price of only $1200:banghead:. And for the worst card from the low/entry level segment will the 5060 8GB with RX 9060 fight? Who will win the competition for the worst overall, the worst for the price, the worst due to the failures, the underdevelopment of RTX 5090? :peace: Or maybe a ghost card?

AMD did show a fairly successful model 9070 XT, but if it means hunting for models that are not significantly above MSRP and we had almost three months of delay?:wtf:
And where are models like RX 9080, 9090 XT, 9090 XTX? After all, with such a "weak" RTX 5080, it's just begging for a repeat of the RX 7900 XTX, only with improved results in RT/PT and FSR 4. This is simply the perfect moment that AMD is sleeping through.

Yes, I know, they were supposed to fight in the lower shelf, hence the three-month delay, hence, like Nvidia, the fun in three models of the RX 9060 regular, XT, GRE. Because like Nvidia, you have to cut performance by 5% so that players don't get too good a card.
I hope the market collapses :pimp: for them, someone other companies shows efficient chips for AI and the profits from this are spread across several companies - which will potentially force Nvidia to sell full chips and faster ones, because such an RTX 5080 is half of the RTX 5090 (Transistors) - once 1060 was half compared to 1080. Today, XX80 is half of a top chip, but at a premium price :confused: Let pray to the corporate gods :respect:
 
It'll be interesting to see what W1zzard comes up with. I wonder how much RAM would actually be enough? 10GB? 12GB?

Most likely 12GB, then when you use G-Assist, that goes out the window, then you need 16GB
 
Back
Top