Post in thread '10 years of Nvidia Video cards. Ultra high end ,high end, mid range, lower end, and

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

linkgoron

Platinum Member
Mar 9, 2005
2,331
854
136
single card dual chip gaming cards stay.

http://www.nvidia.com/gtx-700-graphics-cards/gtx-titan-z/
GeForce® GTX TITAN Z is a gaming monster, the fastest card we’ve ever built to power the most extreme PC gaming rigs on the planet. Stacked with 5760 cores and 12 GB of memory, this dual GPU gives you the power to drive even the most insane multi-monitor displays and 4K hyper PC machines

By nvidia themselves: "to power the most extreme PC gaming rigs".

http://www.geforce.com/whats-new/articles/announcing-the-geforce-gtx-titan-z

GeForce GTX TITAN Z is a gaming monster, built to power the most extreme gaming rigs on the planet. With a massive 5760 cores and 12 GB of 7 Gbps GDDR5 memory, TITAN Z gives you truly amazing performance—easily making it the fastest graphics card we’ve ever made.

This is a serious card built for serious gamers. TITAN Z is designed with the highest-grade components to deliver the best experience – incredible speed and cool, quiet performance—all in a stunningly crafted aluminum case.
 

linkgoron

Platinum Member
Mar 9, 2005
2,331
854
136
next you'll want me to include $5000 Quadro cards that can play games.
The answer is no, I will not include a $3000 prosumer compute/gaming card, you must draw the line somewhere..
You should obviously draw the line where evidence does not align with your conclusion. You should obviously ignore what people are telling you about single-gpu vs dual-gpu, talk about die sizes or gpu code names. You should ignore the fact that Nvidia's site claims that the Titan Z is a gaming card.

Here is your definition of ultra high end:
ultra high end: the fastest card in the stack, max performance, usually the price is out of reach for most.

Now for real, this is my last post. I'm sure the baiting will be strong though. Good luck with the thread.
 
Last edited:

happy medium

Lifer
Jun 8, 2003
14,387
480
126
You should obviously draw the line where evidence does not align with your conclusion. You should obviously ignore what people are telling you about single-gpu vs dual-gpu, talk about die sizes or gpu code names. You should ignore the fact that Nvidia's site claims that the Titan Z is a gaming card.

Here is your definition of ultra high end.


Now for real, this is my last post. I'm sure the baiting will be strong though. Good luck with the thread.
goodluck. I'm happy your taking your thread crapping derailments elsewhere.
 

nurturedhate

Golden Member
Aug 27, 2011
1,760
747
136
If I left out all the Titan cards it would make the later generations look much cheaper and make the difference smaller compared to the older generation.
That would cause to much backlash from the left wing. I mean some are already asking me to include a 3000$ Titan Z compute card.

I think both leaving out all the Titan cards and/or including a 3000$ card would throw the average results to far off , in both directions.
This is called omitting and manipulating data to validate your preconceived conclusion. You can call it what you like.
 

Pariah

Elite Member
Apr 16, 2000
7,357
20
81
Adjusting for inflation, the 8800 GTX $830 MSRP in 2007 would be equivalent to $966 today. The more things change the more they stay the same. From Anandtech's review of the 8800 GTX:

We do know NVIDIA has wanted to push up towards the $1000 graphics card segment for a while. Offering the top of the line for what almost amounts to a performance tax would give NVIDIA the ability to sell a card and treat it like a Ferrari. It would turn high end graphics into a status symbol rather than a commodity. That and having a huge margin part in the mix can easily generate additional profits.

This from 2007. History has shown us the graphics card Ferrari became the Titan line and that Nvidia was able to almost perfectly execute this strategy. Q3 results are a culmination of a strategy 10 years in the works.
 

lixlax

Member
Nov 6, 2014
184
158
116
Not that I'm in the target market, but it seems like we are paying more for a card with single enthusiast class chip now than we did for a card with 2 such chips ~5 years ago (with Titan Z being the exeption). As AtenRa pointed out the prices have risen about 2x for mm2 this decade. I know that the wafer prices increase with almost every new node but certainly not 300+$ for a single 300-400mm2 chip. Unfortunately there is more air in todays prices than ever before.
 

Elfear

Diamond Member
May 30, 2004
7,105
676
126
Before we get started ,let me give a general definition for each category.

1. ultra high end: the fastest card in the stack, max performance, usually the price is out of reach for most.
2. high end : The fastest card under the ultra high end card, very high performance, high price tag
3. lower high end: the third fastest card, a cheaper alternative to the high end card offering high performance.
4. Mid range:
the fourth fastest card, at the sweet spot price, usually offering good performance.
5. lower mid range:
slower than the mid range card and cheaper, usually offering adequate performance.
6. lower end: the last gaming card in the stack, low price, usually plays games at lower settings.
7. lowest end: not a pure gaming card, cheap price, and adequate performance for older games.

In order of performance. Links for release day price quotes.





Now we can compare cards and their price/performance tier, for the past 10 years.
Are we paying significantly more now than in the past, for the same tier of performance?

One thing I noticed is # 4, the mid range ,stayed right around $250.

As others have mentioned, you've created these categories from your own opinion and not from a set of universally acknowledged set of facts. I assume this thread is in response to the Nvidia Q3 revenue thread where you laid out your definition of high-end, mid-range, etc. If you go by price, then you'd be correct. For example the 680, 980, and 1080 would all be "high-end" cards. From a manufacturing (i.e. die size, memory bandwidth, etc.) and Nvidia family code names standpoint, all three of those cards are mid-range. The argument that others have made and you seem to ignore is that we are now paying significantly higher prices for Nvidia's mid-range chips than we used to, mainly due to how they release their product stack. In the past we had the following releases:

High-end (big die, 529mm, GF100) GTX 480 - Mar '10 --> mid-range (medium die, 332mm, GF104) GTX 460 - July '10
High-end (big die, 520mm, GF110) GTX 580 - Nov '10 --> mid-range (medium die, 332mm, GF114) GTX 560Ti - Jan '11

Nvidia's big die, max effort chip was released first and then the mid-range chip came later. The high end chips were $500-650 while the mid-range chips hovered around the $250 mark.

In 2012, Nvidia changed their game (partly because they could now use their mid-range chips to compete with AMD's biggest die chips and because they weren't ready with their big die max effort GPUs). The mid-range GTX 680 (294mm, GK104) was released in March of that year for $500 which moved it a full tier up the product stack. The GTX 780 (561mm, GK110) was released in May '13 for $650. This was in essence the 2nd tier high-end card since it was cut down from the full die 780Ti. Nvidia's 2nd tier cards (GTX 470, 570, etc.) typically sold for ~$350 but that got bumped up big time for the GTX 780.

We now see Nvidia continuing the same trend with the 900 series and 1000 series. The mid-range GTX 980 (398mm, GM204) was released in Sept '14 for $550 and the 2nd tier high-end GTX 980Ti (601mm, GM200) came out in June '15 for $650. The mid-range GTX 1080 (314mm, GP104) released at $600 and we're still waiting on the mainstream big-die chip.

Nvidia's chips have moved up a full tier in the price stack. The ****4 series chips we historically paid $250-350 for we now pay $500-600 for. The ****0/1/2 chips are even worse. We now get the Titan series cards that are essentially the 1st or 2nd tier high-end chips for $1000-1200. A massive jump in price. Even the mainstream cut down high end 980Ti was expensive at $650.

Nvidia is a corporation and will work for their shareholders to maximize profit, just like most other corporations. If I were a shareholder I would be ecstatic at Nvidia's revenue and profit margins. As a consumer it's the opposite and I now pay a significantly higher price for the same chip in their historical product stack.
 
  • Like
Reactions: Head1985
Aug 11, 2008
10,451
642
126
I knew this thread would turn into the usual bashing of nVidia for their prices and claiming 1080 is a midrange card. So absurd. With the exception of Titan which is a niche ultra high end card, it is the most expensive, highest performing card on the market: i.e. high end,(BTW, there can be more than one "high end" card). Doesnt matter what the die size is, how it matches up with previous generations (although it is faster than the high end cards of previous generations obviously) how much profit nVidia makes (that is a separate issue altogether), or if a more powerful card comes out in the future. Right now by definition it is high end. But for the sake of argument, lets call it a mid range card. What does that make the 1070? And in that case, AMD only makes low end 14nm cards, since they, along with the 1060 are a notch below the 1070 and clearly in a lower category than the 1080. Then what does that make the 470, 460, 1050Ti and 1050? Sub-low end? Seriously??
 
Aug 11, 2008
10,451
642
126
As others have mentioned, you've created these categories from your own opinion and not from a set of universally acknowledged set of facts. I assume this thread is in response to the Nvidia Q3 revenue thread where you laid out your definition of high-end, mid-range, etc. If you go by price, then you'd be correct. For example the 680, 980, and 1080 would all be "high-end" cards. From a manufacturing (i.e. die size, memory bandwidth, etc.) and Nvidia family code names standpoint, all three of those cards are mid-range. The argument that others have made and you seem to ignore is that we are now paying significantly higher prices for Nvidia's mid-range chips than we used to, mainly due to how they release their product stack. In the past we had the following releases:

High-end (big die, 529mm, GF100) GTX 480 - Mar '10 --> mid-range (medium die, 332mm, GF104) GTX 460 - July '10
High-end (big die, 520mm, GF110) GTX 580 - Nov '10 --> mid-range (medium die, 332mm, GF114) GTX 560Ti - Jan '11

Nvidia's big die, max effort chip was released first and then the mid-range chip came later. The high end chips were $500-650 while the mid-range chips hovered around the $250 mark.

In 2012, Nvidia changed their game (partly because they could now use their mid-range chips to compete with AMD's biggest die chips and because they weren't ready with their big die max effort GPUs). The mid-range GTX 680 (294mm, GK104) was released in March of that year for $500 which moved it a full tier up the product stack. The GTX 780 (561mm, GK110) was released in May '13 for $650. This was in essence the 2nd tier high-end card since it was cut down from the full die 780Ti. Nvidia's 2nd tier cards (GTX 470, 570, etc.) typically sold for ~$350 but that got bumped up big time for the GTX 780.

We now see Nvidia continuing the same trend with the 900 series and 1000 series. The mid-range GTX 980 (398mm, GM204) was released in Sept '14 for $550 and the 2nd tier high-end GTX 980Ti (601mm, GM200) came out in June '15 for $650. The mid-range GTX 1080 (314mm, GP104) released at $600 and we're still waiting on the mainstream big-die chip.

Nvidia's chips have moved up a full tier in the price stack. The ****4 series chips we historically paid $250-350 for we now pay $500-600 for. The ****0/1/2 chips are even worse. We now get the Titan series cards that are essentially the 1st or 2nd tier high-end chips for $1000-1200. A massive jump in price. Even the mainstream cut down high end 980Ti was expensive at $650.

Nvidia is a corporation and will work for their shareholders to maximize profit, just like most other corporations. If I were a shareholder I would be ecstatic at Nvidia's revenue and profit margins. As a consumer it's the opposite and I now pay a significantly higher price for the same chip in their historical product stack.
Well, again, I have to agree with HM. The "cut down" "mid range" 1080 offers higher performance, than the full size, "high end" chips from previous generations. If I buy a card, I look at the price and performance. I dont care how big the chip is, how many chips AMD/nVidia can make from a wafer, how it stacks up to previous generations (as long as it is faster), etc. I dont have data at hand, but I am pretty sure the "cut down" 1080 offers much more performance per dollar than any of the dual chip cards from previous generations. And this is in absolute dollar amounts, not even accounting for inflation.

Edit: classifying cards into classes by price or performance comes much closer to the conventional definition of "high end" than trying to classify by die size.
 

lixlax

Member
Nov 6, 2014
184
158
116
If not components then what makes currunt cards so expensive then? I don't believe card manufacturers are to blame since the "founders editions" have even higher price tags.

Or if the price should increase almost as much as performance for each generation then yeah- the price increase is justified. But why haven't it been the case in past 30 years? Because the cost per perfomance goes down with new technology.

Nvidia is asking what they want since there is their brand value and next to no competition above GTX1060/Rx480 class.
 

Pariah

Elite Member
Apr 16, 2000
7,357
20
81
If not components then what makes currunt cards so expensive then?

Market demand. The consumer determines what companies can charge for a given product. There is an idiotic line of thinking on this forum by some that a statistically non-zero percentage of video card buyers determine how much they will spend on a given video card based on the size of the GPU die. The actual percentage is probably comparable to the percentage of car buyers that base their purchase primarily on engine displacement. Any sensible buyer focuses on performance, features and price. How the performance is achieved is irrelevant to the user experience.
 

Elfear

Diamond Member
May 30, 2004
7,105
676
126
Market demand. The consumer determines what companies can charge for a given product. There is an idiotic line of thinking on this forum by some that a statistically non-zero percentage of video card buyers determine how much they will spend on a given video card based on the size of the GPU die. The actual percentage is probably comparable to the percentage of car buyers that base their purchase primarily on engine displacement. Any sensible buyer focuses on performance, features and price. How the performance is achieved is irrelevant to the user experience.

That is a strawman argument and doesn't help the effort for reasonable debate rather than inflammatory remarks meant to make yourself look smarter by calling others idiots. No one is arguing that consumers look at die size as the determining factor for how much they will spend. The argument is that historically consumers paid X amount for Nvidia's mid-range chip and now that price has gone up significantly. To continue with your car analogy, it's like paying X amount for the V6 version of a car and then a few years later paying almost double that amount. A hard pill to swallow even if the today's V6 offers more performance.

That isn't to say that AMD's prices haven't gone up as well, just not to the same degree (i.e. top end 6970 that was 389mm and $369 --> top end 7970 that was 352mm for $550).
 

moonbogg

Lifer
Jan 8, 2011
10,636
3,095
136
Adjusting for inflation doesn't make sense because most people don't make any more money than they did 10 years ago. Charging $800 for a GPU is just as ridiculous as it was 10 years ago. Its not like people see $800 today and say "Oh, that's not as bad as $800 10 years ago" because its exactly as bad as $800 was 10 years ago. The only difference is that $800 ten years ago actually bought you a big die high end chip. Today it gets you a small die midranger.
 

Pariah

Elite Member
Apr 16, 2000
7,357
20
81
That is a strawman argument and doesn't help the effort for reasonable debate rather than inflammatory remarks meant to make yourself look smarter by calling others idiots. No one is arguing that consumers look at die size as the determining factor for how much they will spend. The argument is that historically consumers paid X amount for Nvidia's mid-range chip and now that price has gone up significantly. To continue with your car analogy, it's like paying X amount for the V6 version of a car and then a few years later paying almost double that amount. A hard pill to swallow even if the today's V6 offers more performance.

That isn't to say that AMD's prices haven't gone up as well, just not to the same degree (i.e. top end 6970 that was 389mm and $369 --> top end 7970 that was 352mm for $550).

I called no one an idiot. Stop twisting my words to suit your viewpoint. I said it was an idiotic line of thinking which does not imply anything about person. You can be a genius and still think up some really stupid stuff.

Over time prices go up for things. If you told someone 10 years ago, that you could buy a factory Ford Focus from a Ford dealer for over $42k they'd probably think you were smoking crack (disclaimer: no, I do not think you are a crack smoker). Still a 4 cylinder. Is it a hard pill to swallow for RS owners? Probably not. They're too busy enjoying their car to worry about what a 4 cylinder Focus cost back in 2007.
 
Last edited:

Ranulf

Platinum Member
Jul 18, 2001
2,398
1,287
136
Adjusting for inflation doesn't make sense because most people don't make any more money than they did 10 years ago. Charging $800 for a GPU is just as ridiculous as it was 10 years ago. Its not like people see $800 today and say "Oh, that's not as bad as $800 10 years ago" because its exactly as bad as $800 was 10 years ago. The only difference is that $800 ten years ago actually bought you a big die high end chip. Today it gets you a small die midranger.

Bingo.
 

Elfear

Diamond Member
May 30, 2004
7,105
676
126
I called no one an idiot. Stop twisting my words to suit your viewpoint. I said it was an idiot line of thinking which does not imply anything about person. You can be a genius and still think up some really stupid stuff.

Fair enough but "Idiotic line of thinking" sounds very similar to calling someone an idiot.

Over time prices go up for things. If you told someone 10 years ago, that you could buy a factory Ford Focus from a Ford dealer for over $42k they'd probably think you were smoking crack (disclaimer: no, I do not think you are a crack smoker). Still a 4 cylinder. Is it a hard pill to swallow for RS owners? Probably not. They're too busy enjoying their car to worry about what a 4 cylinder Focus cost back in 2007.

The RS is a significant departure from a regular Ford Focus. I was referring to something like the 2002 200hp V6 Camaro ($18k base or $24k adjusted for inflation) vs the 2016 335hp V6 Camaro ($27k base). With inflation included it's about 12% higher now vs 140% higher going from the GTX 560Ti to the GTX 1080 ($268 in today's dollars --> $600).
 
  • Like
Reactions: Ranulf

Head1985

Golden Member
Jul 8, 2014
1,866
699
136
As others have mentioned, you've created these categories from your own opinion and not from a set of universally acknowledged set of facts. I assume this thread is in response to the Nvidia Q3 revenue thread where you laid out your definition of high-end, mid-range, etc. If you go by price, then you'd be correct. For example the 680, 980, and 1080 would all be "high-end" cards. From a manufacturing (i.e. die size, memory bandwidth, etc.) and Nvidia family code names standpoint, all three of those cards are mid-range. The argument that others have made and you seem to ignore is that we are now paying significantly higher prices for Nvidia's mid-range chips than we used to, mainly due to how they release their product stack. In the past we had the following releases:

High-end (big die, 529mm, GF100) GTX 480 - Mar '10 --> mid-range (medium die, 332mm, GF104) GTX 460 - July '10
High-end (big die, 520mm, GF110) GTX 580 - Nov '10 --> mid-range (medium die, 332mm, GF114) GTX 560Ti - Jan '11

Nvidia's big die, max effort chip was released first and then the mid-range chip came later. The high end chips were $500-650 while the mid-range chips hovered around the $250 mark.

In 2012, Nvidia changed their game (partly because they could now use their mid-range chips to compete with AMD's biggest die chips and because they weren't ready with their big die max effort GPUs). The mid-range GTX 680 (294mm, GK104) was released in March of that year for $500 which moved it a full tier up the product stack. The GTX 780 (561mm, GK110) was released in May '13 for $650. This was in essence the 2nd tier high-end card since it was cut down from the full die 780Ti. Nvidia's 2nd tier cards (GTX 470, 570, etc.) typically sold for ~$350 but that got bumped up big time for the GTX 780.

We now see Nvidia continuing the same trend with the 900 series and 1000 series. The mid-range GTX 980 (398mm, GM204) was released in Sept '14 for $550 and the 2nd tier high-end GTX 980Ti (601mm, GM200) came out in June '15 for $650. The mid-range GTX 1080 (314mm, GP104) released at $600 and we're still waiting on the mainstream big-die chip.

Nvidia's chips have moved up a full tier in the price stack. The ****4 series chips we historically paid $250-350 for we now pay $500-600 for. The ****0/1/2 chips are even worse. We now get the Titan series cards that are essentially the 1st or 2nd tier high-end chips for $1000-1200. A massive jump in price. Even the mainstream cut down high end 980Ti was expensive at $650.

Nvidia is a corporation and will work for their shareholders to maximize profit, just like most other corporations. If I were a shareholder I would be ecstatic at Nvidia's revenue and profit margins. As a consumer it's the opposite and I now pay a significantly higher price for the same chip in their historical product stack.
Very good post i agree 100%.Maybe prices should be FE so 1080 is 700USD card but outside that i think everyone get the point.