Nvidia is at it again: 3060 8GB released, up to 35% slower than its 12GB counterpart

[leadeater]

So timely and relevant heh

 

[mr moose]

9 hours ago, Godlygamer23 said:

Performance and VRAM amount should be two different things.

 

They are two different things, no one has said otherwise. 

[leadeater]

1 hour ago, mr moose said:

They are two different things, no one has said otherwise. 

Yet seemingly you are perfectly happy for this to not actually be the case and happy for the capacity designation to also signify a product and performance difference.

 

If you actually truly believe this to be the case then don't say RTX 3060 8GB is acceptable naming when it has a different memory bus and a different performance as a result of that.

 

Far as I am concerned you are and have been saying otherwise, pages and pages of it.

 

The only possible way for VRAM capacity to mean only capacity and not performance also, aka two things, is to only ever use it for capacity and not a single thing more. Pick a stance, you're flip flopping.

[Godlygamer23]

8 hours ago, mr moose said:

They are two different things, no one has said otherwise. 

VRAM capacity deltas are being conflated with outright performance. To say otherwise is ignoring the pages and pages of posts saying differently

 

You are okay with NVIDIA marketing the same tiered card with differing memory as if the only difference is VRAM capacity. The marketed model is making the customer miss critical information, and quite frankly, NVIDIA's behavior is anti-consumer. 

[mr moose]

22 hours ago, leadeater said:

Yet seemingly you are perfectly happy for this to not actually be the case and happy for the capacity designation to also signify a product and performance difference.

Because if you had bothered to read any of my posts you would have seen me say over and over and over that the ram amount is just a way to differentiate models.  I never made claims about it being an indicator of performance outside of what effect marketing might have.    I personally have never bought a product without doing my research so the name has never fooled me into anything more than looking at that product in greater detail. 

 

 

22 hours ago, leadeater said:

If you actually truly believe this to be the case then don't say RTX 3060 8GB is acceptable naming when it has a different memory bus and a different performance as a result of that.

I already told you I don't care what the details or name of a product is so long as it is easily identifiable.    I.E two different products with the same name is not easily identifiable and therefore bad naming and unethical (e.g 1030), two different names for two different products is o.k. 

 

22 hours ago, leadeater said:

Far as I am concerned you are and have been saying otherwise, pages and pages of it.

Perhaps you can quote me where I claimed performance and Vram amount are the same thing.

22 hours ago, leadeater said:

The only possible way for VRAM capacity to mean only capacity and not performance also, aka two things, is to only ever use it for capacity and not a single thing more. Pick a stance, you're flip flopping.

Sure thing,  I've have repeatedly claimed that Vram amount equals Vram amount.Just because the Vram amount is the way to tell if the card is different doesn't mean the Vram amount indicates performance.

 

This is what I said way back on page 6:

 

Quote

The other thing people are forgetting in this discussion is that the ram size never tells you anything about performance,  it never has and likely never will.  Suddenly expecting or demanding they make it mean something in comparison to other products on the market is not going to change that.

 

 

15 hours ago, Godlygamer23 said:

VRAM capacity deltas are being conflated with outright performance. To say otherwise is ignoring the pages and pages of posts saying differently

It's not my fault if people make that assumption.

15 hours ago, Godlygamer23 said:

You are okay with NVIDIA marketing the same tiered card with differing memory as if the only difference is VRAM capacity. The marketed model is making the customer miss critical information, and quite frankly, NVIDIA's behavior is anti-consumer. 

The model marketing tells you this is a card with X ram,  if you want to know how good that is look up a bench mark.

 

I am getting tired of having to re-explain that I don't think ram is a performance metric and that it is simply a way to identify one variant from another.  If you don't look up reviews and expect a card to perform better just because it has higher numbers then more fool you.

 

[Godlygamer23]

40 minutes ago, mr moose said:

The model marketing tells you this is a card with X ram,  if you want to know how good that is look up a bench mark.

 

I am getting tired of having to re-explain that I don't think ram is a performance metric and that it is simply a way to identify one variant from another.  If you don't look up reviews and expect a card to perform better just because it has higher numbers then more fool you.

You're right. It tells you the VRAM capacity. But it's not telling you that it has a 192-bit bus, versus 256-bit, which does impact performance, which, because of NVIDIA's nonsense, is caused by the reduction in VRAM capacity, and that is the problem. More VRAM should not inherently result in higher performance, and neither should less VRAM, and that's the main problem here. However YOU decide to spin it, they are selling a card(for the same price) at the same tier with lower VRAM capacity without clearly(or at all) communicating the lesser bus width, which can easily lead people into thinking it's basically the same card but with less VRAM. 

[mr moose]

1 minute ago, Godlygamer23 said:

You're right. It tells you the VRAM capacity. But it's not telling you that it has a 192-bit bus, versus 256-bit, which does impact performance, which, because of NVIDIA's nonsense, is caused by the reduction in VRAM capacity, and that is the problem.

It also doesn't tell you power loading, card size, number of ports, feature compatibility or if the card is made with cheaper capacitors, budget resistors or from recycled horse dung.    the problem we have is that people want the name to reflect the performance.  Names don't, can't and never have been a reflection of performance, that is a marketing thing and human response thing.  All I want I for the name to be easily identifiable.  So people don't confuse versions when looking up reviews.  

 

1 minute ago, Godlygamer23 said:

More VRAM should not inherently result in higher performance, and neither should less VRAM, and that's the main problem here.

I agree that it shouldn't, But by the same token people will always buy bigger numbers.  If Nvidia release a card with bigger numbers but it performs worse then we have a problem.  What we have though is Nvidia releasing a card with smaller numbers that performs worse.  So in this case they have not leveraged misinformation to sell a shit product.

 

1 minute ago, Godlygamer23 said:

However YOU decide to spin it, they are selling a card(for the same price) at the same tier with lower VRAM capacity without clearly(or at all) communicating the lesser bus width, which can easily lead people into thinking it's basically the same card but with less VRAM. 

As I said before, those specs can be clearly obtained and for the most part people are arguing about buyers who don't even know what these things are.  Which means they will be looking up reviews,  this card is easily identifiable from the 12G version so when they see reviews showing it to perform way worse yet cost the same they aren't going to buy it.  And if we assume it will drop in price and become cheaper, people will see the reviews and know both that it performs worse and is cheaper.  They will still be able to make an informed purchase.   The problem in this thread is people keep reading me say Vram amount and assume it means I think that intrinsically means performance.  I don't and never have,  I acknowledge that the average consumer will see one lower and one higher and will assume the higher one is better (which they would be right about), but that is 1. just a perception and 2. works out better for the consumer so they aren't being mislead to buy a shittier unit.

 

 

 

 

[leadeater]

1 hour ago, mr moose said:

Names don't, can't and never have been a reflection of performance

Yes they have, they are named that way now.

 

1 hour ago, mr moose said:

It also doesn't tell you power loading, card size, number of ports, feature compatibility or if the card is made with cheaper capacitors, budget resistors or from recycled horse dung.

Most of those are indeed told you via the graphics card model number. The generation part tells you almost all the features the graphics card supports, typically the number of port but that is sometimes also reflected by the performance class of the model number.

 

All those numbers Nvidia uses aren't just for fun, they do mean something, as does Intel's and AMD's.

1 hour ago, mr moose said:

the problem we have is that people want the name to reflect the performance.

It already does right up until Nvidia breaks their own model naming convention because they see fit to do so.

 

1 hour ago, mr moose said:

All I want I for the name to be easily identifiable.  So people don't confuse versions when looking up reviews. 

And that will never work when different versions are released at different times, sometimes restricted markets, thus all common references and colloquial usage is already set. You simply cannot go back in time and change what happened to better suit this new future state nobody ever knew would happen.

 

But your want still disregards the issue of the name being confusing. Being able to identify that these two products are indeed differently name when the name offers next to zero indication of how they are different.

 

1 hour ago, mr moose said:

So in this case they have not leveraged misinformation to sell a shit product.

Yet they allow the misinformation that less vram means lower performance making that more prevalent. Unintended outcomes from bad naming doesn't discount the issue that exists, intention isn't really relevant to the issue unless you want to make a malice argument which I have not done for this story.

 

I think you are fixated on thinking I'm trying to say Nvidia is doing marketing misinformation, I've had to point this out numerous times, no I'm point out ONLY that this spreads the misinformation about vram capacity and performance. Not a single thing more, nothing.

 

And it's not solely on Nvidia for how that happens. But their naming is the cause so carry the responsibility.

 

1 hour ago, mr moose said:

Which means they will be looking up reviews,  this card is easily identifiable from the 12G version so when they see reviews showing it to perform way worse yet cost the same they aren't going to buy it.

If only it were both impossible and unlikely that a person ends up looking at the wrong information due to the way data is displayed as a result of when and how products are released in to the market.

 

If these two cards shared nothing more than the product generation model number and had completely unique product class model numbering then it would indeed be near impossible and unlike that a person would end up looking at the wrong card.

 

1 hour ago, mr moose said:

I acknowledge that the average consumer will see one lower and one higher and will assume the higher one is better (which they would be right about), but that is 1. just a perception and 2. works out better for the consumer so they aren't being mislead to buy a shittier unit.

This does not make your assumption correct. Have the forethought to realize that you are saying this is your assumption and you are just as likely wrong as right as is anyone else's assumption.

 

The difference is with your assumption consumers will be harmed if you are wrong. Though it's highly unlikely Nvidia will change the model name and/or never do such a thing again. That won't stop me from saying they are doing something wrong when they are.

 

4 hours ago, mr moose said:

I am getting tired of having to re-explain that I don't think ram is a performance metric and that it is simply a way to identify one variant from another.

I'm sorry but that is what you are saying. If you say that the RTX 3060 8GB is clearly identifiable as different and thus acceptably named then you are indeed saying just this exact thing.

 

RTX 3060 means something, it is also the sole and official Nvidia model name for all these cards. The VRAM capacity was previously just a product specification, it was not a version or variant identifier. The RTX 3060 8GB is a variant and the 8GB signifies it as such, that 8GB means something and represents something and not just VRAM capacity.

 

Using VRAM capacity for anything other than VRAM capacity is improper and should never be done. "It is being used and a way to identify one variant from another" is simply not a good argument  and simply is not logical at all.

 

You can repeat that as much as you like, it will not make it acceptable to anyone but yourself.

[leadeater]

On 12/15/2022 at 10:42 PM, mr moose said:

Sure thing,  I've have repeatedly claimed that Vram amount equals Vram amount.Just because the Vram amount is the way to tell if the card is different doesn't mean the Vram amount indicates performance.

Claim all you like, as it is being used by Nvidia right now it is being used to indicate performance. You are simply wrong. That is exactly how it is being used.

 

On 12/15/2022 at 10:42 PM, mr moose said:

Because if you had bothered to read any of my posts you would have seen me say over and over and over that the ram amount is just a way to differentiate models.

I have and my reply to you every time is this is terrible logic and unacceptable. No matter how many times you say it using 8GB as part of the model name to mean anything other than VRAM capacity is wrong always. Since that is not the sole and only way it is being used then it is being misused, period. 

 

You have been using flawed arguments and reasoning based on bad logic and you simply do not like being told this. I'm not going to tell you anything otherwise when I believe so strongly that is the case.

 

I simply cannot agree with "it's just a model/version/variant identifier" when the entire problem is that is itself a misusage of VRAM capacity. See the problem here? Notice why I can never agree with you and why I'm saying your reasoning and logic is bad. This is the exact reason we are at an impasse as I cannot ever accept VRAM capacity being used in a way that you find acceptable.

 

On 12/15/2022 at 10:42 PM, mr moose said:

Perhaps you can quote me where I claimed performance and Vram amount are the same thing.

Sure

 

Quote

Everything you have said in this entire topic.

Legit the most easy way to quote it.

 

Or just to appease you

On 12/13/2022 at 11:32 PM, mr moose said:

I've always assumed that when a product has a different spec for anything and difference in price that there will be a difference in performance.  ALWAYS.  no matter if its just the ram size or a GT postfix

Now this is not the only instance I could quote, there are many other times where you are indeed essentially saying VRAM capacity and it's written signification as part of product naming means performance.

 

But before you try and defend yourself about this point I would suggest you accustom yourself to why I object to it's usage as that will make you understand real fast there is likely no sufficient defense you could raise for me.

 

On 12/15/2022 at 10:42 PM, mr moose said:

I personally have never bought a product without doing my research so the name has never fooled me into anything more than looking at that product in greater detail. 

Cool story and that's great for you. Since you are not everyone in the world saying this makes no difference and isn't of much relevance.

[leadeater]

4 hours ago, mr moose said:

I already told you I don't care what the details or name of a product is so long as it is easily identifiable.    I.E two different products with the same name is not easily identifiable and therefore bad naming and unethical (e.g 1030), two different names for two different products is o.k. 

You are aware that the GT1030 that used DDR4 memory was "clearly identified" as a GT1030 DDR4. The DDR4 is just being used as a version identifier.

 

DDR4 was written on the box.

 

Sounds like you are saying the GT 1030 was unethical yet is the same situation as this. Is this not rather confusing?

[mr moose]

@leadeater I am not going 30 rounds again, I have said what I believe.  You are clearly not going to change your mind based on my opinions and you haven't given me any reason to change mine.   I bid you a good evening, even though it is just after midnight there as I type this.

 

 

[leadeater]

20 hours ago, mr moose said:

@leadeater I am not going 30 rounds again, I have said what I believe.  You are clearly not going to change your mind based on my opinions and you haven't given me any reason to change mine.   I bid you a good evening, even though it is just after midnight there as I type this.

That is fine. I'll just say something to help you understand why I think this is an issue and why you haven't addressed the talking points I and others have had.

 

Picture this situation.

 

There is a house on 123 Long Road, it has an address that you can clearly identify. We can all agree that this house has a unique address and we can find it. However the issue is the house is on fire. Pointing out that the house can easily be identified by it's address doesn't address the issue about the house being on fire and how to navigate to it, which is the problem situation being talked about.

[Coaxialgamer]

Let me dust off my shaming hat...