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

[leadeater]

17 hours ago, LAwLz said:

Thinks that this may potentially be illegal or should be illegal.

No I think it's garbage and we should not be willing to accept it. I have not made a single comment about legality other that  if they did call it a RTX 3050 Ti but re-used all the RTX 3060 product page and performance information meaning it would not represent the product that would be illegal. There's no debate around the legality of that situation, but I'm not saying Nvidia would do that I'm saying if they had called it a RTX 3050 Ti then they would have been legally required to make sure such a product page have the correct information which is a better outcome for consumers.

 

17 hours ago, LAwLz said:

Says that people should not have to rely on suffixes to understand performance. Only the first part of a name matters.

No I said established and known naming conventions are highly important and sticking to those should be done. Those naming conventions also exist for a reason to represent a product, different parts configured to be the same as used to make up the product. That could be different memory suppliers, different silicon die and other board components. So long as the SM count is the same, the memory bus is the same, TDP/TGP reference is the same then you have the same fundamental product and should be name so. Suffixes can and should exist but only for the right reasons to represent the correct thing and not foster the spreading of misinformation. Directly equating VRAM capacity to performance is misinformation, it is factually and objectively not true and you know it. Using it in that way is improper.

 

Ti and Super are suffixes that represent performance, I do not like those however like I said they are existing established naming convention that is understood in the market and they also do not result in spreading misinformation as they are a unique thing to Nvidia for the expressed purpose of representing performance and a product itself.

 

XT, XTX etc are just as bad and said the same thing in the RDNA3/RX 7000 topic. I however do not feel the need to drag another brand in and complain about them in a topic that is not about them. I will reference others as and where necessary and the only reason talking about another brand were even necessary was to counter bad arguments about VRAM capacity, of which both companies have examples of doing it correctly and transparently as well as not.

 

8 hours ago, LAwLz said:

Maybe I am completely wrong, but couldn't it be the case that the 8GB card simply has fewer memory chips connected, and as a result not the entire memory bus is used?

Yes that is literally the situation, resulting in it having the same configured memory bus as the RTX 3050 and the same memory capacity I might add. Both the RTX 3060 and RTX 3050 use GA106 based dies and if you want to make the argument that only the configuration of the die matters aka the SM count then you'd be wrong.

 

An RTX 3080 with a 128bit memory bus would not be a RTX 3080, it would not even be remotely close to the same performance regardless if the memory capacity were the same or not which would actually be possible.. You can very clearly see where I am drawing a line, memory bus matters and if you change that then it's not the same product anymore.

 

You've just been ignoring the memory bus and wanting only the GPU die configuration to be allowed to denote a product name where as I include memory bus and consider it a major part of a product. You can have that opinion but do not tell me it doesn't matter and no not reduce my argument to just mere "it's performance" when I have made the effort more than once to explain and outline this and specifically step in a state "no I do not just mean performance".

 

If you want to summarize my arguments then get it right. I've given you the feed back on some of those points and yet you ignore it and get it wrong. Don't summarize what you cannot get right or understand if that is the issue.

 

8 hours ago, LAwLz said:

The GPU remains unchanged.

The GPU die may be the same or based on the same but the graphics card is not the same. It's not the same memory bus nor the same performance and those are traits that are used to create product names following the naming conventions as mentioned. If it were just a mere capacity only change then I have no issue as I've said. Since the product name no longer matches the product anymore I deem it improper to use it and I do not consider "8GB" as sufficient and proper to designate what should be an actually different model name.

 

8 hours ago, LAwLz said:

Considering the lack of press release and seemingly lack of availability, this might just be an OEM card that some vendors requested

It's plenty available in the regions it's released in. Does a retail Galax box purchased within Australia look like an OEM product to you? I can also buy this card from 3 different AIBs at least here in NZ retail. It's not an OEM product, it just may not be being sold in EU etc and it wouldn't be the first time Nvidia has done a China only or Asia only product. Such a product being region specific however does not make it acceptable. I hope you aren't arguing something is more acceptable because it only potentially effects 50% of the worlds population (I'm assuming you can get this in all of Asia).

 

And none of this addresses the very real problem I mentioned of trying to go back in time and re-write history. Everyone talks about just the "RTX 3060" as there was only one at release. The reviews you will see will simply be just that, all the information will be written like that. If one were to go looking for RTX 3060 information then they will get the 12GB variant even if they were to put 8GB in the google search, that will happen. People will get directed to incorrect information.

 

 

This will happen. This is a problem, and it's been a problem in the past.

[leadeater]

9 hours ago, LAwLz said:

The thing is though, both of these cards (the 3060 8GB and 3060 12GB) seem to have the same die.

Both of them uses the GA104 die, which is why I think it makes sense to call them both the 3060.

I think you mean GA106? Major die code like that isn't a good argument btw. GA104 is used in RTX 3060, RTX 3060 Ti, RTX 3070, RTX 3070 Ti.

 

The difference between the RTX 3070 and RTX 3070 Ti is just 2 SMs, ~%5 difference, while both being 256bit bus however GDDR6 vs GDDR6X so a very large difference in memory bandwidth thus graphic card (product) performance.

 

So why not just call both of these, significantly different performing products, the same RTX 3070 with different VRAM capacity stuck on the end? Because they have the same capacity. But what if one were 8GB and the other 16GB?

 

Does RTX 3070 8GB & RTX 3070 16GB make more sense if that were the situation? I really think not. This is simply not something I think should be normalized.

 

A graphics card is more than just the GPU die. I suggest you have a stronger think about your reasoning and logic around that.

 

VRAM capacity has no place being on the end of a graphics card name other than to signify it's capacity and only the capacity, nothing else.

[leadeater]

On 12/7/2022 at 10:56 AM, LAwLz said:

You sound pretty emotional and upset about this. Maybe you should take a while off, cool down and come back to this topic when you have calmed down a little?

Have a think at why I would be upset at you. How many times do I need to point out to you over and over that you have been ignoring what has said, correcting your generalizations and summations of my points that are false and incorrect. And keep argue points irrelevant to what was said because you lack the courtesy for what I know perfectly well you understand.

 

It's your deliberate and knowing behavior I find abhorrent. I'll treat you with exactly as much as you show in return. If you cannot abide by the points I just raised then I have no other option to knock you back for it.

 

On 12/7/2022 at 10:56 AM, LAwLz said:

I don't think your argument of "this is okay to do because of history" is a valid argument

Yes it is a very valid argument. It's as valid as all knowledge that exists because guess what, that's history. We know we we know and we carry that forward through time, this knowing of things known helps us understand. As soon as you start putting in arbitrary things in to the equation that are not defined or established, or in this case worse, improperly used then understanding get lost.

 

If you want to change historic practices then it's on the one wanting to drive the change to justify it and show that its a better way and support that change. Just like these manufacturers do when they go through a major re-brand of product modeling i.e. GTX to RTX.

 

On 12/7/2022 at 10:56 AM, LAwLz said:

because at least the average Joe will most likely reach the correct conclusion (for the wrong reason) in that regard.

Which is unacceptable and that is my position on that, obviously. That is literally one of the ways misinformation spreads. And as a reminder everyone plays a part in that, I'm not saying it's just Nvidia. In fact for the most part that is spread by others and not Nvidia, they just created the situation. 

 

While those people might be correct on this occasion they may not be correct always, just like they wouldn't have been GTX 1060 3gb vs GTX 1060 5GB. You are actually saying it's not a problem to set a flawed precedent and I don't know why you are ok with that.

 

On 12/7/2022 at 10:56 AM, LAwLz said:

For example AMD only started using XT on their 5000 series of GPUs, and yet nobody had any issues with that. I tried looking on this forum and after 15 minutes of searching I could not find anyone who raised even a single objection to that poor naming scheme,

Well two points here. 1) I did. 2) This topic is about Nvidia and the RTX 3060 8GB so trying to drag in others is both hardly relevant at all and whataboutism. And my response to point 2 is what about if I don't give a damn about other past examples from other brands or even Nvidia themselves. This is this, those are those. Unless you have an actually good reason to bring it up then you'll find I won't care about it, neither can I change the past. What I can do is point to the past like I have and say it was bad then and it's still bad now.

 

The only hypocrisy here is your determination to say it exists. Strong case of confirmation bias is applying.

 

Now you might get some semblance of an apology if you rescind the below because zero of it is accurate or true. Had you done your duty of care as a debater you'd have not made these errors.

17 hours ago, LAwLz said:

Leadeater thinks that Nvidia are:

• Trying to mislead and rip people off.
• Says that people should not have to rely on suffixes to understand performance. Only the first part of a name matters.
• Thinks that this may potentially be illegal or should be illegal.

 

 

And this bares pointing out again. I have said this is not a huge issue, you can quote me on that more than once in this topic. Most things we agree on. What I do not agree on is your acceptance and reasoning given for "8GB" is sufficient. Now you can hold that opinion and I'll obviously disagree but what I will not stand by is all the other nonsense you've tried to use to support it which are just wrong. All you've got is 8GB and 12GB are different numbers/letters, it's not a good argument but it's true and accurate one, and you'll find I have never once denied that they are indeed different numbers/letters.

[Mihle]

I would argue that 3060 8 and 12gb both being called 3060 with so large performance difference, is worse than the "super" or "XT" naming scheme, tho those aren't ideal either. (XTX is very bad tho)

 

It would be better if they just started using more of the numbers. For example a 2070 super could be called 2065. Or the 3060 8gb that this topic is about, could have been called 3055 or something.

[mr moose]

I just don't see how anyone is going to create a naming scheme that accurately reports performance. It has not been done for any other tech product.  Should we ignore all the numbers after i when buying an intel processor and ask them to name them according to raw performance?  what about when said products perform better at specific tasks, should we have two names for each card so that one can represent games using engine X while the other represents workloads using openCL?

 

The card has a 3060 chip, so it should be called a 3060, it is not a 3050 or a 3070 or even a 3055.  It has a unique amount of ram which identifies it as not being the 12G version.  It has benchmarks that clearly place it's performance against other cards including the 12G. This is how we know how it compares to the 12G.   It would be very hard to blame Nvidia because someone didn't read the benchmark properly and confused 8 for 12. 

[LAwLz]

11 hours ago, leadeater said:

No I think it's garbage and we should not be willing to accept it. I have not made a single comment about legality other that  if they did call it a RTX 3050 Ti but re-used all the RTX 3060 product page and performance information meaning it would not represent the product that would be illegal. There's no debate around the legality of that situation, but I'm not saying Nvidia would do that I'm saying if they had called it a RTX 3050 Ti then they would have been legally required to make sure such a product page have the correct information which is a better outcome for consumers.

Maybe I misunderstood this post from you then:

On 12/4/2022 at 1:00 PM, leadeater said:

On 12/4/2022 at 12:43 PM, mr moose said:

At the end of the day it doesn't matter what they call these things

Yes it does. As a consumer protection law advocate yourself you know this.

 

 

Mr Moose says that it doesn't matter if they call it the 3060 8GB.

You replied by saying it does and brings up consumer protection laws. The logical way of reading this post is, in my mind, that you are saying that the way they have named these cards could be at odds with consumer protection laws.

If your statement is read in a vacuum then I could understand you making a grand statement about how sometimes, the name of products matters, but when looking at this context and the person you were replying to my interpretation was "leadeater thinks this might be against the law".

 

 

11 hours ago, leadeater said:

No I said established and known naming conventions are highly important and sticking to those should be done.

But hold on. I feel like we are once again flip flopping between different customers here.

Now we are once again back to the idea that customers have been following hardware news for years and are familiar with certain suffixes, but we are not allowed to change or introduce new suffixes because that might take some time before people get used to.

I thought the people who were "at risk" of being fooled by this naming scheme were people who didn't follow hardware news and may not even be familiar with suffixes like Ti or XT. 

I feel like you are moving the goalpost here because before you said:  

On 12/4/2022 at 12:27 PM, leadeater said:

People know the first part, rely on the first part, talk about the first part. How many people ever really talk about more than that

But now all of a sudden you are saying people do read suffixes, but apparently only if it's "Ti" or "Super"? 

 

You can't have it both ways. Do people ignore suffixes? Then all suffixes are bad. If they don't ignore suffixes, then we should expect people to read all suffixes. We can't say that people only read Ti suffixes but not 8GB suffixes. That makes no sense.

 

 

11 hours ago, leadeater said:

Yes that is literally the situation, resulting in it having the same configured memory bus as the RTX 3050 and the same memory capacity I might add. Both the RTX 3060 and RTX 3050 use GA106 based dies and if you want to make the argument that only the configuration of the die matters aka the SM count then you'd be wrong.

Except you are deliberately leaving out a lot of information here.

Yes, the 3050 and 3060 use the same base die, but they are not using the exact same die.

The 3050 uses the die named GA106-150-KA-A1. The 3060, both the 8GB and 12GB models, uses the GA106-302-A1.

 

There is a very big difference between "using the exact same die" and "one card has a fully enabled die and one has the same die but partially disabled".

 

In the case of the 3060 8GB vs 3060 12GB, it seems like they are the exact same card, except two memory dies have been removed. The only thing that has changed is the memory, and as a result I think it makes sense to differentiate them with a suffix that specifies the memory difference.

 

 

Another point I would like to raise is that we have this issue everywhere when it comes to computer hardware.

For example a lot of SSDs will perform differently depending on capacity too. For example the Samsung 980 Pro 500GB is roughly half as fast as the 1TB model in sequential writes for example (possibly because of more NAND chips, so more chips can be written to at once). The caches are also different sized so performance varies quite a lot.

Does this mean we should advocate for Samsung to rename their drives so the 980 Pro 500GB is now called the 975 Pro? Of course not. That would just make things needlessly complicated and would probably not help any customer.

I think the same deal applies here.

 

I would love if a system existed where it was obvious to anyone, regardless of how well read they were, to instantly understand exactly how a graphics card performed in any given situation, but the fact of the matter is that no such system exists. Of course, you can do a better or worse job at it and I would argue the whole "Ti" and "XT" suffixes are worse than just using slightly higher numbers (3065 for example instead of 3060 Ti), but then again, I am fairly sure people were mocking Intel when they reached 10th gen because "lol so many numbers". The more I think about it, the more my joke about how gamers can't focus long enough to read more than 4 character seems true.

 

 

 

11 hours ago, leadeater said:

I think you mean GA106? Major die code like that isn't a good argument btw. GA104 is used in RTX 3060, RTX 3060 Ti, RTX 3070, RTX 3070 Ti.

Yeah sorry, I meant the GA106.

I don't have to use the major die code though. It seems like the 3060 8GB and 3060 12GB uses the exact same die, the GA106-302-A1.

 

The GA104 is not used in the 3060 by the way. At least not according to TechPowerUp. They list it as the GA106. The 8GB and 12GB models seems to be using the exact same die. It seems like only the memory chips has changed between them. The 8GB has two fewer chips.

[MageTank]

7 minutes ago, LAwLz said:

The GA104 is not used in the 3060 by the way. At least not according to TechPowerUp. They list it as the GA106. The 8GB and 12GB models seems to be using the exact same die. It seems like only the memory chips has changed between them. The 8GB has two fewer chips.

It is, but only certain SKUs which is what confused me earlier when I referenced the GA104: https://www.techpowerup.com/gpu-specs/geforce-rtx-3060-12-gb-ga104.c3832.

 

Two of my cards in the lab are GA104, but looking at my older samples, they are all GA106. Very interesting, nothing on the box art changed to clarify this and I don't recall them performing any differently at all.

 

EDIT: Don't know if this matters, but my two GA104 cards do have V2 or LHR in their naming, but I am also aware of GA106 LHR cards so I doubt this matters.

[Blademaster91]

10 hours ago, mr moose said:

I just don't see how anyone is going to create a naming scheme that accurately reports performance. It has not been done for any other tech product.  Should we ignore all the numbers after i when buying an intel processor and ask them to name them according to raw performance?  what about when said products perform better at specific tasks, should we have two names for each card so that one can represent games using engine X while the other represents workloads using openCL?

If Nvidia wants the VRAM capacity to be an indicator of performance then they should add the memory bus bandwidth to the product naming, like RTX 3060 (192 bit) 12GB. What Nvidia is doing with their 2 different 3060 models is like if Intel made two different i7 cpu's, one of them has higher clocks but doesn't clearly indicate that higher performance in the naming. And IMO, Intel's naming is confusing enough with their 5 digit numbering, and it gets even more confusing with their mobile cpu's.

10 hours ago, mr moose said:

The card has a 3060 chip, so it should be called a 3060, it is not a 3050 or a 3070 or even a 3055.  It has a unique amount of ram which identifies it as not being the 12G version.  It has benchmarks that clearly place it's performance against other cards including the 12G. This is how we know how it compares to the 12G.   It would be very hard to blame Nvidia because someone didn't read the benchmark properly and confused 8 for 12.

The 3060 8GB shouldn't be called a 3060 because it doesn't have the same 3060 die as the 3060 12GB, the 12GB version is the GA106-300-A1, 8GB version is GA106-302-A1, and the 8GB version has a lower 128 bit memory bus bandwidth,compared to the the 12GB having a 192 bit memory bus.

Although you wouldn't know any of that unless you watched or read an in depth review, most people don't do that, they just want to upgrade from the same tier of card and want to get the best product for their money. If they accidentally buy the 8GB instead of the 12GB because the VRAM capacity on the box is intentionally in fine print, Nvidia controls everything the OEM's can do,even box designs, so this whole confusion is on nvidia and I think that blaming this on the consumer is part of the problem of Nvidia getting away with anti-consumer nonsense. Nvidia tried to do the same thing with the 4080 12GB and they got massive backlash for it, yet people here still defended Nvidia for it. It seems like Nvidia didn't learn anything from that so they did it again with the 3060 8GB, and quietly as to not tell any reviewers or offer any review products, I find that a lot more scummy as people buying the lower range cards aren't always going to be as informed as the enthusiasts spending over $1000 on a graphics card.

[xAcid9]

18 hours ago, mr moose said:

I just don't see how anyone is going to create a naming scheme that accurately reports performance. It has not been done for any other tech product.  Should we ignore all the numbers after i when buying an intel processor and ask them to name them according to raw performance?  what about when said products perform better at specific tasks, should we have two names for each card so that one can represent games using engine X while the other represents workloads using openCL?

 

The card has a 3060 chip, so it should be called a 3060, it is not a 3050 or a 3070 or even a 3055.  It has a unique amount of ram which identifies it as not being the 12G version.  It has benchmarks that clearly place it's performance against other cards including the 12G. This is how we know how it compares to the 12G.   It would be very hard to blame Nvidia because someone didn't read the benchmark properly and confused 8 for 12. 

How about putting a relative performance graph or some sort of graphic on side of the box comparing both 8GB & 12B performance.

Do it for all GPUs that have 2 or more variants with the same name. Too much too ask? Or too much information for the customer to make sense of it? 

[leadeater]

On 12/9/2022 at 12:37 AM, mr moose said:

I just don't see how anyone is going to create a naming scheme that accurately reports performance.

It doesn't have to  accurately report performance, it has to accurately differentiate product configuration. You know, literally how its done now bar these edge cases where companies get lazy and can't be bothered going the full effort so just make it a "variant" of an existing product while being as different as any other named SKU.

 

On 12/9/2022 at 12:37 AM, mr moose said:

The card has a 3060 chip, so it should be called a 3060, it is not a 3050 or a 3070 or even a 3055.

The card has the same die from the same wafer as the RTX 3050 and RTX 3060. The only difference is final binning, fusing and microcode which makes it whatever die SKU designation required. Nvidia isn't getting separate wafers and dies manufactured for each of these products, it's just how they end up final configuration wise.

 

And no it can be called whatever the hell Nvidia wants to call it and whatever name they choose we can like, or dislike. I find these arguments reductive and ill-informed because what you just said is literally counter to how Nvidia is doing the entire product creation process going back decades.

 

On 12/9/2022 at 12:37 AM, mr moose said:

It has benchmarks that clearly place it's performance against other cards including the 12G.

Correction, it has benchmarks from so far one, maybe a couple of reviewers who picked up on this product configuration discrepancy and the existence of the product and choose to test it. As yet Nvidia has not officially recognized or identified it as having a difference in performance, you should find this problematic.

 

Literally the first and second hit on Google for RTX 3060 8GB review is actually the RTX 3060 12GB review and it's not in the URL title either so easily missed and it's not identified in any of the graphs as being 12GB either. So unless someone is paying proper attention and actually reading most of at least the opening to the review and not just graph hunting only then will they pickup on it. Those that are just skim checking and graph searching will be looking at the wrong information easily being unaware of that.

 

Have you actually tried looking for reviews of the graphics card or are you just going off the linked source in this topic? Which btw is only the third result in Google.

 

And I can blame Nvidia because they were simply being lazy and created this in the quickest and simplest way for themselves and I'd bet without much consideration at all to any of these potential problems. I also see no benefit to consumers for this product existing so I can't weigh that against this either.

[leadeater]

21 hours ago, MageTank said:

It is, but only certain SKUs which is what confused me earlier when I referenced the GA104: https://www.techpowerup.com/gpu-specs/geforce-rtx-3060-12-gb-ga104.c3832.

Wikipedia lays out the products a lot better with the table on there. It's actually quite well maintained. If you want to get a good overview of what is used across the product line then that is a better place than techpowerup.

[leadeater]

22 hours ago, LAwLz said:

Except you are deliberately leaving out a lot of information here.

Yes, the 3050 and 3060 use the same base die, but they are not using the exact same die.

The 3050 uses the die named GA106-150-KA-A1. The 3060, both the 8GB and 12GB models, uses the GA106-302-A1.

RTX 3050 has 2 different dies of the same configuration, RTX 3060 has 3 different dies of the same configuration supplied by 2 different wafer/die sources (GA106 & GA104).

 

And I note you haven't addressed that RTX 3070 vs RTX 3070 Ti situation, nor RTX 3080 128bit bus possibilities. Die configuration is not the be all and end all of a graphics card and it's not the only factor Nvidia uses right now or in the past to create named product SKUs re: RTX 3070 Ti & RTX 2080/2080 Super & GTX 1080 Ti/Titan X & GTX 780 Ti/Titan Black.

 

22 hours ago, LAwLz said:

The 3050 uses the die named GA106-150-KA-A1. The 3060, both the 8GB and 12GB models, uses the GA106-302-A1.

You know you just repeat exactly what I said right?

 

" Both the RTX 3060 and RTX 3050 use GA106 based dies and if you want to make the argument that only the configuration of the die matters aka the SM count then you'd be wrong."

 

What do you think GA106-150-KA-A1 & GA106-302-A1 represent? SM count and sometimes other sub SM component changes.

 

22 hours ago, LAwLz said:

In the case of the 3060 8GB vs 3060 12GB, it seems like they are the exact same card, except two memory dies have been removed. The only thing that has changed is the memory, and as a result I think it makes sense to differentiate them with a suffix that specifies the memory difference.

And now you are leaving off information. If you remove memory dies then you reduce memory bus and if you reduce memory bus you have a different product as per what Nvidia already does right now to differentiae named product SKUs. I'm not going to let you not acknowledge this actually important fact. If it were just a reduction of memory die density then as I've said fine to name it this way but it's not that is it.

 

So no they are not the exact same card, the mere fact you had to say "except" in that sentence degrades your argument from the outset. It's the same "except" it's not the same.

 

22 hours ago, LAwLz said:

But now all of a sudden you are saying people do read suffixes, but apparently only if it's "Ti" or "Super"? 

 

You can't have it both ways. Do people ignore suffixes? Then all suffixes are bad. If they don't ignore suffixes, then we should expect people to read all suffixes. We can't say that people only read Ti suffixes but not 8GB suffixes. That makes no sense.

Did you not bother to look at the example used, RTX 4080. I'm sure you can figure out where this is going...

 

I'm not having it both ways, you're just intentionally being obtuse, ignoring or not reading what is said and then when clarified later you want to make these misinformed or worse comments.

 

Right now I can say literally anything because it seems you aren't reading anything. Seems you are too busy trying to prove this point that you haven't actually bothered to evaluate what has been said.

 

I literally said Ti and Super have an established and known meaning, like actually come on now. Really do we need to cover this again? REALY? Do we, actually. Do we.....

 

RTX 3070 Ti [Insert stuff after that gets ignored]

RTX 3070 [Insert stuff after that gets ignored]

 

When people recognize known things and naming commonly used that is what they use to identify the product, other things can and do get ignored. That's that last time I'll spell it out for you again, this is REALLY basic and obvious. And I don't for a single second think you were ever confused by this nor didn't know what I meant so making me cover it yet again I do find offensive. Just fyi. Choose wisely how you want to continue from here, wasting my time with things I know you understood just fine is pretty much the least acceptable thing to me.

[leadeater]

22 hours ago, LAwLz said:

Mr Moose says that it doesn't matter if they call it the 3060 8GB.

You replied by saying it does and brings up consumer protection laws. The logical way of reading this post is, in my mind, that you are saying that the way they have named these cards could be at odds with consumer protection laws.

If your statement is read in a vacuum then I could understand you making a grand statement about how sometimes, the name of products matters, but when looking at this context and the person you were replying to my interpretation was "leadeater thinks this might be against the law".

My comment applied to exactly what was quoted. Product names do matter and are covered by this. Saying it "doesn't matter what they name it" is wrong. If you want to make a leap and make out that I was commenting about specifically the RTX 3060 8GB then that is your mistake to make. I simply object to that being said because I know he knows that is not true, it wasn't a well thought out comment by him, that is all.

 

Making an NZ/AUS etc CGA or FTA law argument is a waste of time and nothing would come of it so I just won't do it. Btw just as a bit of extra information for you intent to mislead or deceive is not a requirement to be in breach of these, that might seem a little unfair but the laws are there to protect consumers not business and put consumers first.

[mr moose]

16 hours ago, Blademaster91 said:

If Nvidia wants the VRAM capacity to be an indicator of performance then they should add the memory bus bandwidth to the product naming, like RTX 3060 (192 bit) 12GB. What Nvidia is doing with their 2 different 3060 models is like if Intel made two different i7 cpu's, one of them has higher clocks but doesn't clearly indicate that higher performance in the naming. And IMO, Intel's naming is confusing enough with their 5 digit numbering, and it gets even more confusing with their mobile cpu's.

So how do you propose that they name all their products so all the information is contained within the name? I personally don't consider a name to be an indicator of performance.  I know lot of people fall for the basic marketing of naming and such, but at the end of the day, so long as each product is unique in it's identifiability and the name isn't intentionally obfuscating something (which I don't consider it to be in this case), then I don't see the problem.  Is an falcon XR6 faster than a Falcon futura?  no body goes by the name to work that out, they look up spec sheets and reviews.  Same here is the 8G slower than the 12G model?  yes, but like every other product on the market you need to look up the reviews to know that.

 

 

16 hours ago, Blademaster91 said:

The 3060 8GB shouldn't be called a 3060 because it doesn't have the same 3060 die as the 3060 12GB, the 12GB version is the GA106-300-A1, 8GB version is GA106-302-A1, and the 8GB version has a lower 128 bit memory bus bandwidth,compared to the the 12GB having a 192 bit memory bus.

I know that difference, that doesn't make it not a 3060.   Again this is why we have product reviews and unique identifiers. At best the name should only give you a ball park idea of performance.  Don't forget that there are a lot of specs the name does not tell you about in the tech sphere,  not all i5's are the same and you have to look up reviews or specs to know that the difference between the i513600k and the i513600kf is that one supports ecc. 

16 hours ago, Blademaster91 said:

Although you wouldn't know any of that unless you watched or read an in depth review, most people don't do that, they just want to upgrade from the same tier of card and want to get the best product for their money.

Again, if people are not going to bother reading a review then it doesn't matter what the card is called,  they are buying blind anyway.  Honestly how is anyone going to know what performs better, an RX6700XT or an RTX3060TI if they don't read a review?  The name means diddly squat by itself, and it only is honest if it is unique to it's product and not a lie.  

16 hours ago, Blademaster91 said:

If they accidentally buy the 8GB instead of the 12GB because the VRAM capacity on the box is intentionally in fine print, Nvidia controls everything the OEM's can do,even box designs, so this whole confusion is on nvidia and I think that blaming this on the consumer is part of the problem of Nvidia getting away with anti-consumer nonsense. Nvidia tried to do the same thing with the 4080 12GB and they got massive backlash for it, yet people here still defended Nvidia for it. It seems like Nvidia didn't learn anything from that so they did it again with the 3060 8GB, and quietly as to not tell any reviewers or offer any review products, I find that a lot more scummy as people buying the lower range cards aren't always going to be as informed as the enthusiasts spending over $1000 on a graphics card.

Who is going to "accidentally" buy a GPU?   I don't think there is anything to blame on anyone.  Consumers haven't been lied to and aren't accidentally buying a subpar card because of a name and Nvidia aren't lying about the cards specs.    When there is a problem on any sort of scale that can't e attributed to a handful of people doing stupid things (because there always will be a few) then we'll start discussing if the box information was to blame.   But as of right now it is not a lie, it does not insinuate it is faster and we only know this because it is a different product with a different name and was reviewed as such for us to discuss.

 

 

1 hour ago, leadeater said:

It doesn't have to  accurately report performance, it has to accurately differentiate product configuration.

It does that. it is a 3060 with 8G ram. 

1 hour ago, leadeater said:

You know, literally how its done now bar these edge cases where companies get lazy and can't be bothered going the full effort so just make it a "variant" of an existing product while being as different as any other named SKU.

It is a variant of an existing product, it's a 3060 with less ram.  You know this when you look at any GPU listing because all brands are listed by GPU chip, Ram size, Brand model then maybe some brand postfix.   In this case you cannot confuse it for the 12G model unless you are ignoring the ram size. 

1 hour ago, leadeater said:

The card has the same die from the same wafer as the RTX 3050 and RTX 3060. The only difference is final binning, fusing and microcode which makes it whatever die SKU designation required. Nvidia isn't getting separate wafers and dies manufactured for each of these products, it's just how they end up final configuration wise.

In that case you should be arguing for all 3060's and all the 3050 to be called the same thing. 

1 hour ago, leadeater said:

And no it can be called whatever the hell Nvidia wants to call it and whatever name they choose we can like, or dislike. I find these arguments reductive and ill-informed because what you just said is literally counter to how Nvidia is doing the entire product creation process going back decades.

they can call it the bootlicker 8G for all I care, so long as it isn't a name they already use to label a product that is obviously inferior.

1 hour ago, leadeater said:

Correction, it has benchmarks from so far one, maybe a couple of reviewers who picked up on this product configuration discrepancy and the existence of the product and choose to test it. As yet Nvidia has not officially recognized or identified it as having a difference in performance, you should find this problematic.

Have you considered that their motivation for testing it was due to it being a new variant? I mean that is why nearly every (if not all) GPU is tested and reviewed, so we can see how they perform against not only their own variants but the competition as well.

 

1 hour ago, leadeater said:

Literally the first and second hit on Google for RTX 3060 8GB review is actually the RTX 3060 12GB review and it's not in the URL title either so easily missed and it's not identified in any of the graphs as being 12GB either. So unless someone is paying proper attention and actually reading most of at least the opening to the review and not just graph hunting only then will they pickup on it. Those that are just skim checking and graph searching will be looking at the wrong information easily being unaware of that.

So the card hasn't been out long enough for google search to update with 8G reviews, you yourself said in just the past paragraph that there were only one or two reviews so far.  this doesn't mean anything.

 

1 hour ago, leadeater said:

Have you actually tried looking for reviews of the graphics card or are you just going off the linked source in this topic? Which btw is only the third result in Google.

 Not having enough product reviews does not equal deceptive advertising. 

1 hour ago, leadeater said:

And I can blame Nvidia because they were simply being lazy and created this in the quickest and simplest way for themselves and I'd bet without much consideration at all to any of these potential problems. I also see no benefit to consumers for this product existing so I can't weigh that against this either.

 

You are more than welcome to any feeling on the matter,  I just don;t agree that the name is deceptive.  It is clearly identifiable as a product other than the 3060 12G and reviews can be found if buyers are interested.  Just like every other product (except the 1030).

 

[mr moose]

6 hours ago, xAcid9 said:

How about putting a relative performance graph or some sort of graphic on side of the box comparing both 8GB & 12B performance.

Do it for all GPUs that have 2 or more variants with the same name. Too much too ask? Or too much information for the customer to make sense of it? 

If they are honest and only compare to other variants in the same product line up then I think that would be good for the consumer. However if it becomes the usual marketing wank performance graphs we see from Nvidia and Intel then I personally think just a simple and unique name is enough and consumers can go to independent 3rd parties for performance comparisons.

[leadeater]

1 hour ago, mr moose said:

Not having enough product reviews does not equal deceptive advertising. 

I'm not arguing that it's deceptive advertising, I'm saying the name is crap and people can too easily get information for the wrong product and using memory capacity in this way is misleading people in to thinking it has actually anything to do with performance directly.

 

Not only that Nvidia's own RTX 3060 product page has information solely, I repeat solely for the RTX 3060 12GB. Only the full specifications table has any mention or reference to the 8GB existence.

 

So you have no official information about this "variant" and a single review of it I know of so far that isn't even the first result when you search for it and it's not the first result because of how it's been named.

 

I wonder how that issue could be avoided.... I wonder how hard it would be to have made this not a problem...

 

1 hour ago, mr moose said:

So the card hasn't been out long enough for google search to update with 8G reviews, you yourself said in just the past paragraph that there were only one or two reviews so far.  this doesn't mean anything.

I said the first two reviews were for the RTX 3060 12GB when you search for the RTX 3060 8GB specifically. Do you not think that's a big problem? I said there is currently only one I know of, the linked source of this article (and it's written Techspot version of it which is the 3rd Google result). There could be more reviews, I have not found them and but that does get past the issue of them not being readily found.

 

You know what wouldn't happen if it weren't named with RTX 3060? Yea exactly, no mistakes in finding the product.

 

You know what has information readily available at release, every other graphics card. The only ones that do not are these poorly named, mid cycle actually different products carrying conflicting names. I don't mistakenly get RTX 3080 reviews for the RTX 3080 Ti when it's released, have a think about why that is.

 

1 hour ago, mr moose said:

So how do you propose that they name all their products so all the information is contained within the name?

I'm not saying all the information need to be contained in the name, why would you think that. I'm saying for the core components/factors that make up the product that are ALREADY used to name the products should be how they are named. So if any one of those three are different then a new name is to be used. Change memory bus, new name, Change GPU die configuration, new name. Change memory type, new name.

 

This is not a complicated thing to do, and to remind you again, is happening now. It's how it's always been done. It's just sometimes these badly named products get thrown in to the mix because I dunno, screw it who cares...

 

1 hour ago, mr moose said:

I know that difference, that doesn't make it not a 3060.

Ok then by your logic every single RTX 30 series GPU can and should be named RTX 3060. Great got it, superb reasoning you have.

 

If you want to get back in to reality about how things are actually named and have been named then I'll be waiting.

 

1 hour ago, mr moose said:

It does that. it is a 3060 with 8G ram. 

False. False, millions times false. No, not it is not.

 

1 hour ago, mr moose said:

Same here is the 8G slower than the 12G model?

And I'll repeat this again just once. Coming to the right outcome for the wrong reasoning supporting a wrong concept about VRAM capacity that allows misinformation to spread is wrong always. It's always wrong. I do not care if this time it comes to the right outcome it's still improper to use it like this.

 

Previously it's been wrong, in future it could be wrong. Why not actually do the right thing and use proper naming and avoid something that is actually improper.

 

1 hour ago, mr moose said:

Have you considered that their motivation for testing it was due to it being a new variant? I mean that is why nearly every (if not all) GPU is tested and reviewed, so we can see how they perform against not only their own variants but the competition as well.

And what exactly is your point? Does this in any way excuse it having a conflicting name where the two products are not configured the same resulting in different performance. Does this excuse Nvidia not having official information available for it?

 

No it does not..

[leadeater]

1 hour ago, mr moose said:

It does that. it is a 3060 with 8G ram. 

An RTX 3060 has 28 active SMs, 192bit memory bus, GDDR6 at release and all original official information and reviews reflect this.

 

If it does not have 28 SMs or 192bit memory or GDDR6 then it is not an RTX 3060, End of story, period, period, period. It's something not that, I proposed RTX 3050 Ti but is could be anything not named RTX 3060 since it is not that product.

 

Also this is not the only variant or product refresh Nvidia did in Oct 2022:

• RTX 3060 8GB, reduced memory bus to 128bit
• RTX 3060 Ti, upgraded to GDDR6X
• RTX 3070 Ti, moved to GA102 (GA102-150)

 

RTX 3070 Ti GA102-150 configuration is identical to the GA104-400 die and no other product changes were made, still an RTX 3070 as it exactly matches the product specifications.

 

RTX 3060 Ti upgraded from GDDR6 to GDDR6X, large increase in memory bandwidth. No other changes to product made. Should no longer be called an RTX 3060 Ti however anyone that gets this card gets a better product than advertised. Consumers not adversely affected.

 

RTX 3060 8GB, already talked enough about this one.

 

1 hour ago, mr moose said:

In that case you should be arguing for all 3060's and all the 3050 to be called the same thing. 

Um no, you haven't been reading much of what I have said then. I've specifically called out the die does not matter so long as it's configured the same. Could be a GA102 with 28 SMs, or a GA106 with 28 SMs or a GA102847251564374573262 with 28 SMs. So long as they combine a GPU die configured as per the product specifications with memory of the type as per the product specifications and at the bus width as per the product specifications then it should be named.... as per the product that is reflected by these specifications.

 

1 hour ago, mr moose said:

they can call it the bootlicker 8G for all I care, so long as it isn't a name they already use to label a product that is obviously inferior.

Since RTX 3060 is not contained in the name then this is acceptable.

[MageTank]

2 hours ago, mr moose said:

So how do you propose that they name all their products so all the information is contained within the name? I personally don't consider a name to be an indicator of performance. 

The problem isn't necessarily naming a product to contain all of the information in the name. That part is actually very easy. The hard part is doing so without making it sound like you are summoning Cthulhu. Let's take a look at the cheap SSD's I buy from Micro Center (shoutout to their BYO team for actually explaining this, it's pretty neat):

 

https://www.microcenter.com/product/651302/inland-qn322-1tb-ssd-nvme-pcie-gen-30x4-m2-2280-3d-nand-qlc-internal-solid-state-drive

"Inland QN322". Obviously Inland is their in-house brand. Q stands for QLC, N stands for NVMe, 3 stands for PCIe Gen 3, 22 is their speed rating of 2200MB/s on read. So assuming you stick to this trend, we could assume that a drive named TN450 would be a TLC NVMe Gen 4 5000MB/s drive. You'd likely have to have some kind of online decoder ring, but it's definitely a nice way to name a product.

 

The problem is trying to do this with a GPU. You have so much information to convey when discussing performance. SM core count, VRAM capacity, bus width, memory type, clock speeds, this is all ignoring TMU/ROP counts as well. Now assuming the SM structure is the same across the product stack, you could use the SM count for a naming convention as the TMU/ROP counts should scale accordingly. For the RTX 4090, you could do the NA128G6X24 which would denote an Nvidia Ada 128SM GDDR6X 24GB card, but man that is a mouthful to say out loud compared to "RTX 4090". 

 

Now back to the debate everyone is having with the VRAM capacity changing, I want to remind everyone that this precedent isn't exactly common even in older Nvidia cards. The GTX 770 for example (God I love coming back to Kepler), had both a 2GB and 4GB card. The bus width did not change despite having two different memory capacities. Maxwell's GTX 960 2GB and 4GB had the same bus width as well. The first time I saw a card with different VRAM capacities change the bus width was an RX 580. Though, to be fair to the other side of this argument, I don't remember everyone being this vocal about that back then.

 

Still, I am firmly in the camp that this should have been called a different product name. Mostly because there is nothing stopping Nvidia from calling this an RTX 3055. I'd even go as far to say that I would gladly live without "Ti" and "Super" series if we wanted to change everything to xxx5 intervals. RTX 3050, 3055, 3060, 3065. RTX 4000 series has resurrected the Fermi thermals and power consumption, give me the Fermi naming scheme! 

 

[MageTank]

Whoopsie, double post. Please delete.

[Kisai]

On 12/6/2022 at 2:56 AM, WereCat said:

RX 6800 vs RX 6800 XT implies they are different tiers of cards

RTX 3060 8GB vs RTX 3060 12GB implies they are the same tier of card with different amount of VRAM.

 

 

This is true, but also not. People who are not tech enthusiasts, will see "recommended RTX3060 or RX6800" and buy whatever is in the store if they want to play the game.

 

The difference between -no designation-, XT, XTX, etc isn't immediate apparent any more than 8GB and 12GB. Same with "Ti" and "Super for nvidia.

 

The problem ultimately comes down to there not being a standard way to benchmark a card that can't be cheated (because we know GPU AIB's do that.) If you want a specific tier of performance, you pretty much have to wait for someone to rank it along side all the current generation and previous generation cards and decide if that's what you should upgrade.

 

[ewitte]

The only thing to say is nobody should buy this product.  There are much better options.

[Blademaster91]

23 hours ago, mr moose said:

So how do you propose that they name all their products so all the information is contained within the name? I personally don't consider a name to be an indicator of performance.  I know lot of people fall for the basic marketing of naming and such, but at the end of the day, so long as each product is unique in it's identifiability and the name isn't intentionally obfuscating something (which I don't consider it to be in this case), then I don't see the problem.  Is an falcon XR6 faster than a Falcon futura?  no body goes by the name to work that out, they look up spec sheets and reviews.  Same here is the 8G slower than the 12G model?  yes, but like every other product on the market you need to look up the reviews to know that.

I didn't say all of the info needs to be in the name, I am saying if the product changes in a significant way, the naming of the product should change. The 3060 8GB has a 128bit memory bus and shouldn't be called a 3060. To use the car analogy it would be like if Ford just called both of their cars a Falcon, with no additional naming to know what trim level you're buying unless you know what the differences in specs mean.

23 hours ago, mr moose said:

Again this is why we have product reviews and unique identifiers. At best the name should only give you a ball park idea of performance.

And the 3060 8GB doesn't have any unique identifiers,  the naming of it sucks because it doesn't tell me anything about the performance, and without looking at in depth reviews you wouldn't know its any different than the 3060 12GB because of the naming Nvidia chose to use.

23 hours ago, mr moose said:

Again, if people are not going to bother reading a review then it doesn't matter what the card is called,  they are buying blind anyway.  Honestly how is anyone going to know what performs better, an RX6700XT or an RTX3060TI if they don't read a review?  The name means diddly squat by itself, and it only is honest if it is unique to it's product and not a lie.  

So  you're still saying its fine for people to be misled because of a terrible naming convention, and @leadeater already explained this one, if you google RX6700XT vs. RTX 3060 you're going to get 3060 12GB reviews first,  the naming is important as other Nvidia cards tell you enough with the naming of performance, cards with a different amount of VRAM or the memory bus get a different card naming or "Ti" or "Super" added.

23 hours ago, mr moose said:

Who is going to "accidentally" buy a GPU?   I don't think there is anything to blame on anyone.  Consumers haven't been lied to and aren't accidentally buying a subpar card because of a name and Nvidia aren't lying about the cards specs.    When there is a problem on any sort of scale that can't e attributed to a handful of people doing stupid things (because there always will be a few) then we'll start discussing if the box information was to blame.   But as of right now it is not a lie, it does not insinuate it is faster and we only know this because it is a different product with a different name and was reviewed as such for us to discuss.

Again, if someone googles RX6700XT vs.RTX 3060, they're going to get reviews for the 3060 12GB which isn't even close to being the same card, consumers aren't necessarily being lied to, but they're being misled enough to end up buying a worse card because Nvidia chose to call the slower one a 3060. Also not even Nvidia wants to show the difference between the 3060 8GB and 3060 12GB on their site so they are purposely trying to mislead consumers, and I still find it weird that people are still wanting to defend Nvidia on their terrible naming with this card.

Edited December 10, 2022 by Blademaster91

[mr moose]

10 hours ago, leadeater said:

I'm not arguing that it's deceptive advertising

 

10 hours ago, leadeater said:

 using memory capacity in this way is misleading people in to thinking it has actually anything to do with performance directly.

 

So are you arguing it is misleading or are you arguing it is misleading?

 

 

Until you clarify what your actual problem with the name is (because the ram amount is a part of that whether we accept it technically or informally or not.  They are always listed with the ram amount) this discussion will go no where.

 

10 hours ago, leadeater said:

 

Not only that Nvidia's own RTX 3060 product page has information solely, I repeat solely for the RTX 3060 12GB. Only the full specifications table has any mention or reference to the 8GB existence.

 

So you have no official information about this "variant" and a single review of it I know of so far that isn't even the first result when you search for it and it's not the first result because of how it's been named.

 

I wonder how that issue could be avoided.... I wonder how hard it would be to have made this not a problem...

 

I said the first two reviews were for the RTX 3060 12GB when you search for the RTX 3060 8GB specifically. Do you not think that's a big problem? I said there is currently only one I know of, the linked source of this article (and it's written Techspot version of it which is the 3rd Google result). There could be more reviews, I have not found them and but that does get past the issue of them not being readily found.

 

You know what wouldn't happen if it weren't named with RTX 3060? Yea exactly, no mistakes in finding the product.

 

You know what has information readily available at release, every other graphics card. The only ones that do not are these poorly named, mid cycle actually different products carrying conflicting names. I don't mistakenly get RTX 3080 reviews for the RTX 3080 Ti when it's released, have a think about why that is.

 

I'm not saying all the information need to be contained in the name, why would you think that. I'm saying for the core components/factors that make up the product that are ALREADY used to name the products should be how they are named. So if any one of those three are different then a new name is to be used. Change memory bus, new name, Change GPU die configuration, new name. Change memory type, new name.

 

This is not a complicated thing to do, and to remind you again, is happening now. It's how it's always been done. It's just sometimes these badly named products get thrown in to the mix because I dunno, screw it who cares...

 

Ok then by your logic every single RTX 30 series GPU can and should be named RTX 3060. Great got it, superb reasoning you have.

 

If you want to get back in to reality about how things are actually named and have been named then I'll be waiting.

 

False. False, millions times false. No, not it is not.

 

And I'll repeat this again just once. Coming to the right outcome for the wrong reasoning supporting a wrong concept about VRAM capacity that allows misinformation to spread is wrong always. It's always wrong. I do not care if this time it comes to the right outcome it's still improper to use it like this.

 

Previously it's been wrong, in future it could be wrong. Why not actually do the right thing and use proper naming and avoid something that is actually improper.

 

And what exactly is your point? Does this in any way excuse it having a conflicting name where the two products are not configured the same resulting in different performance. Does this excuse Nvidia not having official information available for it?

 

No it does not..

 

9 hours ago, HenrySalayne said:

No, it's not. It's so unidentifiable that even Nvidia is incapable of calling it 3060 12 GB on their official homepage. They just call it 3060.

 

https://www.nvidia.com/en-us/geforce/graphics-cards/30-series/rtx-3060-3060ti/

 

Their specs page for the 3060 range clearly shows the memory bandwidth and ram size for the ti, the 12G and the 8G

At this point there seems to be a very concerted effort to make this into something it really isn't.

 

 

9 hours ago, MageTank said:

The problem isn't necessarily naming a product to contain all of the information in the name. That part is actually very easy.

I am simply countering the arguments that the name they have is not unique and somehow should reflect the products performance.

 

9 hours ago, MageTank said:

The hard part is doing so without making it sound like you are summoning Cthulhu. Let's take a look at the cheap SSD's I buy from Micro Center (shoutout to their BYO team for actually explaining this, it's pretty neat):

 

 

9 hours ago, MageTank said:

https://www.microcenter.com/product/651302/inland-qn322-1tb-ssd-nvme-pcie-gen-30x4-m2-2280-3d-nand-qlc-internal-solid-state-drive

"Inland QN322". Obviously Inland is their in-house brand. Q stands for QLC, N stands for NVMe, 3 stands for PCIe Gen 3, 22 is their speed rating of 2200MB/s on read. So assuming you stick to this trend, we could assume that a drive named TN450 would be a TLC NVMe Gen 4 5000MB/s drive. You'd likely have to have some kind of online decoder ring, but it's definitely a nice way to name a product.

 

The problem is trying to do this with a GPU. You have so much information to convey when discussing performance. SM core count, VRAM capacity, bus width, memory type, clock speeds, this is all ignoring TMU/ROP counts as well. Now assuming the SM structure is the same across the product stack, you could use the SM count for a naming convention as the TMU/ROP counts should scale accordingly. For the RTX 4090, you could do the NA128G6X24 which would denote an Nvidia Ada 128SM GDDR6X 24GB card, but man that is a mouthful to say out loud compared to "RTX 4090". 

 

Now back to the debate everyone is having with the VRAM capacity changing, I want to remind everyone that this precedent isn't exactly common even in older Nvidia cards. The GTX 770 for example (God I love coming back to Kepler), had both a 2GB and 4GB card. The bus width did not change despite having two different memory capacities. Maxwell's GTX 960 2GB and 4GB had the same bus width as well. The first time I saw a card with different VRAM capacities change the bus width was an RX 580. Though, to be fair to the other side of this argument, I don't remember everyone being this vocal about that back then.

 

Still, I am firmly in the camp that this should have been called a different product name. Mostly because there is nothing stopping Nvidia from calling this an RTX 3055. I'd even go as far to say that I would gladly live without "Ti" and "Super" series if we wanted to change everything to xxx5 intervals. RTX 3050, 3055, 3060, 3065. RTX 4000 series has resurrected the Fermi thermals and power consumption, give me the Fermi naming scheme! 

 

 

The 3060 page on Nvidias website clearly shows the difference between all 3 cards.  All three cards have a different name one is ti one is 8G and the other is 12G.  I really don;t see why calling it a 3050 would make this any different to the end consumer. They will still have to look up reviews in order to know how it performs.

 

 

 

 

6 hours ago, Blademaster91 said:

I didn't say all of the info needs to be in the name, I am saying if the product changes in a significant way, the naming of the product should change. The 3060 8GB has a 128bit memory bus and shouldn't be called a 3060. To use the car analogy it would be like if Ford just called both of their cards a Falcon, with no additional naming to know what trim level you're buying unless you know what the differences in specs mean.

Except this card has a postfix that separates the models.  Just like the falcon.  It's even on their website.

 

 

6 hours ago, Blademaster91 said:

And the 3060 8GB doesn't have any unique identifiers, 

Yes it does,

https://www.nvidia.com/en-us/geforce/graphics-cards/30-series/rtx-3060-3060ti/

 

 

[leadeater]

4 hours ago, mr moose said:

So are you arguing it is misleading or are you arguing it is misleading?

 

 

Until you clarify what your actual problem with the name is (because the ram amount is a part of that whether we accept it technically or informally or not.  They are always listed with the ram amount) this discussion will go no where.

It's misleading for many reason and we the community and the market as a whole are part of that. I'm not making an argument to target Nvidia specifically and saying they are doing deceptive advertising. That case could be made but I'm not making that. I'm saying with how it's named people can and likely will be mislead in to what they get but MORE importantly they are being mislead in to believing and "learning" that VRAM capacity has a direct relation to performance which it does not. There are two different problems involved.

 

I have more than clarified what my problem is, if you are not capable of understanding, likely due to your own bias, that is a you problem not mine. You seem to be the one unable to figure it out and I cannot correct a problem with you and your own mindset, that is impossible. No amount of re-explaining the points given will change this. Go back many pages, read them all again, figure it out for yourself.

 

And when I say read it again, I mean read it in full every word, start to end all of it! 99.9% sure you and others are not and have not.

 

And no they are not always listed with ram amount. Either you are lying now or not even bothering to look at the very pages you are on and being told to look at. Any claim otherwise invalidates everything you say. I will not accept lies nor ignorance from you as I do except better as you are more than capable of that.

 

4 hours ago, mr moose said:

Their specs page for the 3060 range clearly shows the memory bandwidth and ram size for the ti, the 12G and the 8G

At this point there seems to be a very concerted effort to make this into something it really isn't.

I said the spec pages had it. Did you not read anything else? Go back to Nvidia pages and tell me what they are advertising performance wise on there. Tell me what information is that for.

 

Is it labeled RTX 3060 12GB?

No

Does it have a performance graph for the RTX 3060 8GB?

No

 

Your point is not relevant to what was said and are replying to and does not say anything different to what I already said to you.

 

You're making a very concerted effort to ignore the issue because why? Because you for whatever reason think 8GB is enough distinction without any valid and creditable reasoning to back it other than "it's different" while ignoring actual real legitimate problems with it's usage in that way, past problems I have evidenced of why it's a problem, and also any future times where using it again may or may not "accidently" lead consumers to the same outcome.

 

If you cannot see any information about a product on the product page that is a problem. Since I know for a fact that zero information about the RTX 3060 8GB exists on that product page other than the full specifications link for it you saying anything otherwise is false and lying.

 

And the discussion is going nowhere because you don't want it to be. I don't too at this point. I'm made myself ultra clear, change the GPU die configuration, memory bus or VRAM type and it must have a name change as per every other product. It must contain only existing naming conventions used in that generation or past recent generation that signify it is a different model and it must not use important suffixes that have an actual real meaning that is not related to model naming ergo VRAM capacity is an improper model designator and is only to be used to show VRAM capacity and not a single thing more.

 

I will agree to nothing else. Will I get my way? No of course I won't. This crap will happen again and you're part of why it will, but don't worry Nvidia's desire not to have to create an actual new product model and everything associated with it is by far the biggest factor. It's just sad to see people not care about making sure technical terminology and meaning is used correctly.

 

Nvidia agree with the above and does so unless they choose to not do it when it suits themselves and their own motives. As you noticed I don't give a damn about what they want if I can see it creates a problem.

 

And no VRAM capacity is not part of the official product names and even when it is Nvidia doesn't even use it. They also can't keep their own official pages up to date either.

https://www.nvidia.com/en-us/geforce/graphics-cards/30-series/

 

Note the lack in any VRAM as part of the product name, because it's not part of it. Note here RTX 3060 being listed as only having 12GB.

 

Not here too.

[leadeater]

4 hours ago, mr moose said:

I really don;t see why calling it a 3050 would make this any different to the end consumer. They will still have to look up reviews in order to know how it performs.

Google RTX 3060 8GB review and click on the first result, go on I dare you