8GB VRAM not enough (and 10 / 12)

Page 149 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

BFG10K

Lifer
Aug 14, 2000
22,709
3,000
126
8GB
Horizon Forbidden West 3060 is faster than the 2080 Super despite the former usually competing with the 2070. Also 3060 has a better 1% low than 4060 and 4060Ti 8GB.
Resident Evil Village 3060TI/3070 tanks at 4K and is slower than the 3060/6700XT when ray tracing:
Company Of Heroes 3060 has a higher minimum than the 3070TI:

10GB / 12GB

Reasons why still shipping 8GB since 2014 isn't NV's fault.
  1. It's the player's fault.
  2. It's the reviewer's fault.
  3. It's the developer's fault.
  4. It's AMD's fault.
  5. It's the game's fault.
  6. It's the driver's fault.
  7. It's a system configuration issue.
  8. Wrong settings were tested.
  9. Wrong area was tested.
  10. Wrong games were tested.
  11. 4K is irrelevant.
  12. Texture quality is irrelevant as long as it matches a console's.
  13. Detail levels are irrelevant as long as they match a console's.
  14. There's no reason a game should use more than 8GB, because a random forum user said so.
  15. It's completely acceptable for the more expensive 3070/3070TI/3080 to turn down settings while the cheaper 3060/6700XT has no issue.
  16. It's an anomaly.
  17. It's a console port.
  18. It's a conspiracy against NV.
  19. 8GB cards aren't meant for 4K / 1440p / 1080p / 720p gaming.
  20. It's completely acceptable to disable ray tracing on NV while AMD has no issue.
  21. Polls, hardware market share, and game title count are evidence 8GB is enough, but are totally ignored when they don't suit the ray tracing agenda.
According to some people here, 8GB is neeeevaaaaah NV's fault and objective evidence "doesn't count" because of reasons(tm). If you have others please let me know and I'll add them to the list. Cheers!
 
Last edited:

Timorous

Golden Member
Oct 27, 2008
1,931
3,741
136
If the rumors about AMD limiting supply of 8 GB 9060 XT (they should have just called it the 9060, but I digress) are true, perhaps they're actually being clever here.

Sample both cards to reviewers and let them point out the massive flaws and shortcomings of an 8 GB card. The few you did release into the market get scooped up eventually anyway, but now you've got consumers paying an extra $50 for what cost you $20 at most and clamoring for that product.

Being able to beat a the 8 GB 5060 Ti in a few games as well is just a cherry on top. NVidia could have done the same and just leaned into it to create a natural way to upsell customers and get them to thank you for it, but they tried to hide the football and look worse for having done so.

Did AMD marketing not completely crap the bed for a change?

Giving N44 the 16 lane bus will also help when it does exceed Vram so even in Vram limited scenarios it may not poo the bed as much as the 5060Ti 8GB and 5060 do. Especially on older gen 3 and gen 4 motherboards.

Still the 8GB model is a terrible product at 8GB but those extra 8 lanes make it marginally less terrible.

Edit. Wait did I really write the 8GB model is a terrible product at 8GB? Ugh. I meant the 8GB model is a terrible product at $300. Mega brain fart there.
 
Last edited:

coercitiv

Diamond Member
Jan 24, 2014
7,156
16,660
136
As some of us keep saying: Benchmarking games is not always the same as playing games. 5060 can play 2077 with RT fine...until it can't. It will look great in bar charts at the same settings though.
Quote from the video:
Console users get bigger VRAM upgades than we do...

hahahahahahaahah... just don't go to Dogtown. Correct settings!
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,380
30,515
146
Giving N44 the 16 lane bus will also help when it does exceed Vram so even in Vram limited scenarios it may not poo the bed as much as the 5060Ti 8GB and 5060 do. Especially on older gen 3 and gen 4 motherboards.
The common use case systems a NIB $300 card finds its way into.
Still the 8GB model is a terrible product at 8GB but those extra 8 lanes make it marginally less terrible.
That's interesting stuff. Depending on the game engine, perhaps more than marginal? I think I've seen games where older cards with low vram did better simply by having 32GB of system memory vs 16GB. Cunningham's law on this one
 
Reactions: marees

jpiniero

Lifer
Oct 1, 2010
16,269
6,750
136
I think I've seen games where older cards with low vram did better simply by having 32GB of system memory vs 16GB. Cunningham's law on this one

I'd kind of expect that to be the case. Slowest part of the process is the decompression of the textures... could leave the uncompressed texture in system memory if there was room.

Granted the game would have to actually do that.
 

Mopetar

Diamond Member
Jan 31, 2011
8,399
7,549
136
I'd rather have 16 GB VRAM on 8x PCIe lanes than 8 GB VRAM on 16x PCIe lanes. Yes it's better than 8 GB on 8x lanes, but I think it would be more fair to characterize it as "less awful" because it's still going to have problems. At best it spares the card from having single digit 1% lows on a bar chart, but the performance is still going to be below expectations and require lowering settings to find playable frame rates.

Bragging about sleeping with the cousin that's a 6 instead of the cousin that's a 4 is making a stupid numerical argument and trying to sidestep the obvious issue.
 

marees

Golden Member
Apr 28, 2024
1,085
1,510
96

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,380
30,515
146
Do you buy an 8GB GPU to play at 1080p, or do you play at 1080p because you have an 8GB GPU?

I'll give it a read. If the source of data for resolution is the Steam survey then it starts with a flawed premise. If we ignore all of the issues with the survey and its accuracy and accept the data, it would have to differentiate between desktops and the mobile users that compromise 80% of PC sales. My personal hypothesis is that 1440p comprises a much higher percentage of Steam desktop gamers than the survey reflects.
 

Mopetar

Diamond Member
Jan 31, 2011
8,399
7,549
136
That and it doesn't distinguish users with multiple systems. I've got some truly ancient hardware that can still play old games. It's not my primary system though, but it would still skew the overall results of the survey because it has equal weighting to the computer I primarily game on even though it only sees a small fraction of the use.
 

marees

Golden Member
Apr 28, 2024
1,085
1,510
96
That and it doesn't distinguish users with multiple systems. I've got some truly ancient hardware that can still play old games. It's not my primary system though, but it would still skew the overall results of the survey because it has equal weighting to the computer I primarily game on even though it only sees a small fraction of the use.
It gives lot of weightage to chinese e-cafes which could distort the survey
 

Rigg

Senior member
May 6, 2020
676
1,672
136
It gives lot of weightage to chinese e-cafes which could distort the survey
Pre-builts and laptops with Nvidia 50/60 class GPUs also distort the survey. I find arguments that use the Steam survey to try and paint a picture of what "most gamers choose" entirely disingenuous. Of course they choose Nvidia 50/60 class GPUs and 1080p displays. The market is flooded with them preinstalled in OEM systems, and they often represent the only widely available option in entry level gaming laptops and desktops.
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,380
30,515
146
It gives lot of weightage to chinese e-cafes which could distort the survey
Massive distortions. Chinese language jumping by an impossible 20-25% in a single month. Yeah, Steam survey is fake news.

The fact that Polaris sold like hot cakes due to crypto but it was never reflected in the survey was the first time I saw it as seriously sus. Since then it has only gotten worse. It fits a narrative and reinforces confirmation biases that all mysteriously favor Nvidia.



Pre-builts and laptops with Nvidia 50/60 class GPUs also distort the survey. I find arguments that use the Steam survey to try and paint a picture of what "most gamers choose" entirely disingenuous. Of course they choose Nvidia 50/60 class GPUs and 1080p displays. The market is flooded with them preinstalled in OEM systems, and they often represent the only widely available option in entry level gaming laptops and desktops.
And since DIY/retail and the article are focused on 8GB dGPUs all of those mobile users would have to be culled from the data before making any conclusions like Frank is doing with his BS talking points.
 

Rigg

Senior member
May 6, 2020
676
1,672
136
I'd rather have 16 GB VRAM on 8x PCIe lanes than 8 GB VRAM on 16x PCIe lanes. Yes it's better than 8 GB on 8x lanes, but I think it would be more fair to characterize it as "less awful" because it's still going to have problems. At best it spares the card from having single digit 1% lows on a bar chart, but the performance is still going to be below expectations and require lowering settings to find playable frame rates.
I totally agree. The PCIE bandwidth effect is interesting and all, but if you're dipping into system memory, more PCIE bandwidth seems to only make things less awful. You'll likely be turning down settings to get a good experience regardless. Maybe there are some edge cases where the VRAM is only slightly exceeded and frametime consistency doesn't completely go out the window though.
 
Reactions: DAPUNISHER

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,380
30,515
146
I was just watching Daniel-San rant about the idiots of Reddit. It brings me to the next fake news he references so heavily in videos; TPU relative performance chart. It's worse than useless if you spend 5 minutes exposing all of the issues with it. My favorite example is the ARC A380 sitting next to the ancient GTX 780.

Then there is the low effort 25 second test passes. The poorly chosen testing places in some games. The variances in vram, driver support, feature sets, games tested. And for that matter, when was the last time he even tested those old cards? Using my A380 vs GTX 780 example, he apparently hasn't tested the 780 in ages. Relying on that chart to help decide what card to buy, makes it worse than useless, because it horribly misrepresents the hardware and its capabilities.

His sycophants call him the gold standard of reviews. How's the weather in fantasy land today? He has become 💩 tier and his data and testing are Nvidia approved editorial direction. One of the absolute last sources I would put any credence in for 8GB testing until he stops shilling and starts doing the hard work.
 

Rigg

Senior member
May 6, 2020
676
1,672
136
I was just watching Daniel-San rant about the idiots of Reddit. It brings me to the next fake news he references so heavily in videos; TPU relative performance chart. It's worse than useless if you spend 5 minutes exposing all of the issues with it. My favorite example is the ARC A380 sitting next to the ancient GTX 780.

Then there is the low effort 25 second test passes. The poorly chosen testing places in some games. The variances in vram, driver support, feature sets, games tested. And for that matter, when was the last time he even tested those old cards? Using my A380 vs GTX 780 example, he apparently hasn't tested the 780 in ages. Relying on that chart to help decide what card to buy, makes it worse than useless, because it horribly misrepresents the hardware and its capabilities.

His sycophants call him the gold standard of reviews. How's the weather in fantasy land today? He has become 💩 tier and his data and testing are Nvidia approved editorial direction. One of the absolute last sources I would put any credence in for 8GB testing until he stops shilling and starts doing the hard work.
I'm not sure I'm ready to consider W1zard a paid Nvidia shill (I'm definitely starting to lean this way though) just yet, but I agree that his methodology sucks balls. The sycophants on the forums are just sad. They all seem oblivious to the myriad challenges present in games testing generally, much less the specific considerations when properly testing VRAM limitations. The Principal Skinner meme from earlier was on point.

Has he ever even mentioned textures getting nerfed in certain games? I find it highly probable that he has bunk data from 8 GB cards out there with nerfed textures if he isn't even looking for this.
 
Last edited:

Thunder 57

Diamond Member
Aug 19, 2007
3,606
6,027
136
I'm not sure I'm ready to consider W1zard a paid Nvidia shill (I'm definitely starting to lean this way though) just yet, but I agree that his methodology sucks balls. The sycophants on the forums are just sad. They all seem oblivious to the myriad challenges present in games testing generally, much less the specific considerations when properly testing VRAM limitations. The Principal Skinner meme from earlier was on point.

Has he ever even mentioned textures getting nerfed in certain games? I find it highly improbable that he doesn't have bunk data from 8 GB cards out there with nerfed textures if he isn't even looking for this.

He's close. To his credit I don't think he did a 5060 "preview". But listing any non Nvidia card lacking DLSS as a con is just, well, crap to say it nicely.
 

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,380
30,515
146
He's close. To his credit I don't think he did a 5060 "preview". But listing any non Nvidia card lacking DLSS as a con is just, well, crap to say it nicely.
It goes hand in hand with the other Nvidia talking point I hear and read constantly now: DLSS is in more games than FSR 4. Most of which are older titles making them largely irrelevant. But since FSR 4 closed the gap, it's all they have left to push DLSS.

@SolidQ

Thanks for posting it, I had forgotten all about this video. I posted it when it came out. Cool team up with Iceberg. I readily admit that video stroked my confirmation bias.
 

Timorous

Golden Member
Oct 27, 2008
1,931
3,741
136
Has he ever even mentioned textures getting nerfed in certain games? I find it highly probable that he has bunk data from 8 GB cards out there with nerfed textures if he isn't even looking for this.

He mentioned it on Reddit. He said he does not include those games. Replied saying it was useful information so he probably should find a way to incorporate it into the bar charts.

I suggested for games where the games does not respect the chosen settings due to Vram limits the card scores a 0 on that test since it technically failed to perform at the chosen settings.
 

marees

Golden Member
Apr 28, 2024
1,085
1,510
96

DAPUNISHER

Super Moderator CPU Forum Mod and Elite Member
Super Moderator
Aug 22, 2001
31,380
30,515
146
Good point raised in this blog

People aren't buying 8 GB graphics cards because they want to, they are doing it because there isn't an alternative.

Only a couple of disagreements with that article.
The RTX 40 series and RX 7000 series did not see price drops as the generation progressed.
7000 series had significant price drops from MSRPs here in the U.S. There were hot deals on 7900XTX under $750, 7900 XT near $600, 7800 XT near $400. I did not keep up well with Nvidia pricing.

"No alternatives" is hyperbole; as there are viable alternatives to 8GB cards in the same price range. And there has been for generations now.

The pessimism about the future of PC gaming is unwarranted. PC gaming contraction has not happened. There has been consistent growth and it is projected to continue. AAA gaming may seriously contract, but I see that as a good thing. Insane budgets and development times, all to deliver badly performing slop. Get lean, mean, and execute.
 

poke01

Diamond Member
Mar 8, 2022
3,470
4,761
106
The pessimism about the future of PC gaming is unwarranted. PC gaming contraction has not happened. There has been consistent growth and it is projected to continue. AAA gaming may seriously contract, but I see that as a good thing. Insane budgets and development times, all to deliver badly performing slop. Get lean, mean, and execute.
AAA games should also not use insane graphics but be smart about game mechanics. Somewhere along the way we lost the focus on that.

OG Minecraft(2011-2015) is the perfect example of PC gaming being at its best and it got me into PC gaming. Nobody needs to have a 4090 to have fun, I hope future developers and publishers recognise this.
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |