• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Pricing, buying strategy, and the medium size (mid range) vs big chip (high end)

C

CakeMonster

Golden Member
I see this argued here all the time, and I would love for someone to explain it.

This is how things have been looking so far for the consumer performance leader:
980 (medium) -> 980Ti (big) -> 1080 (medium) -> "1080Ti" (big)

I bought the "medium" size chip with 980. At the time it was the fastest card if you disregard the $1000+ Titan. 980Ti would only be released 9(?) months later. With 1080, NV is doing the same thing, releasing a medium sized chip but it is still in fact the fastest consumer card at the time of release.

So, people here argue that we are being "tricked" or are acting "stupid" when getting the medium size chip. But the medium size chip is (by the record so far) cheaper than the big chip at release. If this cycle is the new normal, why should I not keep buying the medium size chip? Sure, they have better margins on it since its cheaper to produce, but I get to have the performance leader every other card for a lower price. It just makes more sense to me, even though I in a sense are still being screwed over since they are "holding back" and making twice as many "top cards" than before by not releasing medium and big at the same time.

Would anyone mind putting some context on this, or explain why it is so common here to recommend buying the big chip and stay on the more expensive cycle?
 
I'll give you some advice that you should carry with you the rest of your livelong days...

Do not worry about how other people feel you should spend your money.
 
Very wise advice in this context 🙂

As for explaining the angst, I'm guessing too, but lets see! At heart I think a lot of it is basically displaced annoyance about the slowing down of the silicon die shrink cycles.

There definitely seem to be some people who feel like they should upgrade year on year but can't get big enough differences to sanely justify it and somehow end up conflicted.

Some of it is also, I think, annoyance at the way NV are doing their annual +~30% thing so smoothly and consistently. Makes it look too easy in some sense, so there's a suspicion they could be trying for more.
(It does of course take truly huge amounts of work/organisation to do what NV are doing.).

I'm not sure of concrete values of the performance deltas and time spent at the top, but doubt if the differences are large. For myself, I wouldn't touch the big chips due to their power draw/noise absent using giant coolers.
 
980 was $550
980Ti was $650

If you're telling me you prefer a 980 over the 980Ti, that's on you. Many people would prefer to wait and just get the 980Ti, as its FAR superior to the 980. I wouldn't even take a 980 out of the box if someone had given it to me for free recently.

A 980Ti? Oh yeah baby.....
 
tential said:
980 was $550
980Ti was $650

If you're telling me you prefer a 980 over the 980Ti, that's on you.
Click to expand...

^ That's missing the point. They weren't available at the same time. We are seemingly in a cycle of top cards being released with ~1yr intervals and the chips are alternating between medium/big.

If the performance delta is different then that's a valid argument against my observation. But presenting 980 and 980Ti as if they were available at the same time is not. I bought the 980 because I needed something faster for DAI release, 980Ti would need to time travel to be an alternative then.
 
Yup 🙂 The actual answer to the query at the end of your initial post is probably simply that a good number are on the 'big' cycle for historical reasons and can't see any reason to jump on to the medium one.
(Because there isn't much of one!).

The same way there's basically no reason to jump from 'medium' to 'big'.

I guess that it'd maybe be interesting to see the time line for length at top/speed increase each time
 
There's never anything wrong with buying the fastest if you want the fastest at that time.

I think it's more that the full Medium chip has been a much worse price-to-performance option. In your scenario there was still the option of buying the GTX 970 for $330 and saving $220 and only forfeiting ~15% performance. The performance gap between the 980 and 980 Ti is bigger and the price gap is smaller.

But again, if you want the fastest at a given time then you're welcome to...

But here's the thing. Getting into the big chip cycle (as long as your ignoring the Titan) of 980 Ti is actually a better cycle to be in for performance-per-dollar than the full medium 980/1080 cycle.

Not only that, getting into the cutdown medium chip cycle of 970/1070 is also better on a performance-per-dollar standpoint.

The 980/1080 cycle is beaten in raw performance and/or performance-per-dollar by the other cycles. I would never recommend this cycle as it is quite poor relative to the other 2. The 980/1080 are really only good for those that always want the best regardless of budget and quickly upgrade to Titan / Ti anyway.
 
CakeMonster said:
^ That's missing the point. They weren't available at the same time. We are seemingly in a cycle of top cards being released with ~1yr intervals and the chips are alternating between medium/big.

If the performance delta is different then that's a valid argument against my observation. But presenting 980 and 980Ti as if they were available at the same time is not. I bought the 980 because I needed something faster for DAI release, 980Ti would need to time travel to be an alternative then.
Click to expand...

You didn't even remotely address the thread topic.
"Would anyone mind putting some context on this, or explain why it is so common here to recommend buying the big chip and stay on the more expensive cycle?"

Why would anyone stay on the more expensive cycle is the point.
The 980 was $550, the 980Ti was $650. The 980Ti offers more performance per dollar at their releases, hence why the 980Ti and flagship release cycle is a better value.
 
The context is simple. You take a 560Ti level card and launch it first for $600-700. Going back in time it would be 560Ti --> relabeled 580 and price increase from $249 to $699 Founders (NV would call GTX560Ti as a GTX480 or 580 at launch then). In benchmarks, it would slaughter GTX285/HD4890. Then a year or so later, release a cut down GTX480 or a full die GTX580. Call it GTX580Ti now and raise the price to $799-899.

$699 1080 = $549 980 = $499 GTX680 => all $249 560Ti successors.

If you don't see how this is bad for gamers or why they are pissed at this marketing game, buy the $699 1080 FE and enjoy it until $599-699 Volta GV104 is released.

Fact is 1080 is an upper mid-range Pascal chip priced as a high end flagship. NV and AMD are simply splitting the generation into 2 or even 3 parts now.l and manipulating marketing names. In the past, 1080 would be 1060Ti. That's what it is. 314mm2 die, 320GB/sec, 256-but bus, no HBM2. It's a mid-range card that just happens to beat last gen's flagship, exactly like was almost always the case for 15 years.
 
Last edited:
dazelord said:
Exactly, the big chip releases have given us 30-35% or so more performance. Medium chip releases more like 5-20%.
Click to expand...

On the other hand, the medium chip does tend to gain relative to the previous big chip after launch. This makes it tricky to justify a two year cycle of big chips. Once you start accepting price/perf as a driving factor it inevitably tends towards medium crop chip though.
 
Qwertilot said:
Yup 🙂 The actual answer to the query at the end of your initial post is probably simply that a good number are on the 'big' cycle for historical reasons and can't see any reason to jump on to the medium one.
(Because there isn't much of one!).

The same way there's basically no reason to jump from 'medium' to 'big'.

I guess that it'd maybe be interesting to see the time line for length at top/speed increase each time
Click to expand...

I am afraid Nvidia is going to completely get away with charging top prices for small dies. The only thing that can stop them is AMD. Traditionally, yes, big dies are where the performance is. But when you compare an old big die to a new small die, it doesn't really matter which one you pick from a performance stand point. Maybe one is better than the other, but the choice is less obvious now with the new release schedule. Its confusing as all hell and I'm sure Nvidia knows it.
What they probably want, and what they will probably get, is that consumers will completely forget about the past and what we used to get for our money. We will completely forget what it was like to enjoy the performance crown for about 2 years, give or take a little for refresh cycles or the occasional "ultra" version product.
We will completely forget what kind of value we used to get out of a mid range product. The excitement we used to feel when a mid range card wasn't too much slower than the big die high end but cost less than half the price. That sense of getting such a killer deal is a thing of the past.
Oh well. Screw it I guess.
 
Qwertilot said:
Yup 🙂 The actual answer to the query at the end of your initial post is probably simply that a good number are on the 'big' cycle for historical reasons and can't see any reason to jump on to the medium one.
(Because there isn't much of one!).

The same way there's basically no reason to jump from 'medium' to 'big'.

I guess that it'd maybe be interesting to see the time line for length at top/speed increase each time
Click to expand...

This makes absolutely no sense. (apologize up front)

If someone just wants to buy every-other-cycle and have the 'big die' to have the 'big die' sure...but what sense does that make? That leads me to believe the performance isn't even needed.

Games and equipment don't fit nicely into every-other-year instance cycles. For someone who just upgraded to a 4K monitor, or wants to play VR, the 1080/1070 is VERY compelling. They don't care that Vega may be out in 6 months or big Pascal in 12 months. Give me a break!

However, if someone is still rocking a nice 1440P display, their 980Ti/Fury is probably a great card to keep them until Vega or P100 is released.

I am getting so sick of so many people on this forum trying to force their own reasons for buying (or not buying) a specific GPU onto others.

You buy because you NEED it. If you just buy to buy (which is fine too) then just be open about it and move on. For some, benching and upgrading is enjoyable and $$ or even usage needs doesn't really factor in.
 
This is so situational and is going to depend on what each user needs and how tight their budget is. If you are planning to get one card this year and have to keep it for a long time, and selling a card at a loss to buy the newest one is not an option for you, then you should wait because you will get more value out of the large die. However, even then if you are waiting 9 months that may be 9 months of playing games at lesser settings.

Personally I buy the card that has the price, performance, and features I want and I do not care if it is the large or medium die. Honestly, if the 1080 had FreeSync I would buy it right away. 6+ months wait and I may as well wait another 6 for the next cycle anyway.
 
CakeMonster said:
Would anyone mind putting some context on this, or explain why it is so common here to recommend buying the big chip and stay on the more expensive cycle?
Click to expand...

Is it really more common here to recommend buying the big chip? I would think that the 970 is more often recommended. And I expect the 1070 to be a repeat of the same.

For CPUs, it might be different. Buy a 6700k. Well, technically even that isnt true anymore since there is now a 6800k, but I doubt there are many would would seriously recommend it over a 6700k.
 
RussianSensation said:
The context is simple. You take a 560Ti level card and launch it first for $600-700. Going back in time it would be 560Ti --> relabeled 580 and price increase from $249 to $699 Founders (NV would call GTX560Ti as a GTX480 or 580 at launch then). In benchmarks, it would slaughter GTX285/HD4890. Then a year or so later, release a cut down GTX480 or a full die GTX580. Call it GTX580Ti now and raise the price to $799-899.

$699 1080 = $549 980 = $499 GTX680 => all $249 560Ti successors.

If you don't see how this is bad for gamers or why they are pissed at this marketing game, buy the $699 1080 FE and enjoy it until $599-699 Volta GV104 is released.

Fact is 1080 is an upper mid-range Pascal chip priced as a high end flagship. NV and AMD are simply splitting the generation into 2 or even 3 parts now.l and manipulating marketing names. In the past, 1080 would be 1060Ti. That's what it is. 314mm2 die, 320GB/sec, 256-but bus, no HBM2. It's a mid-range card that just happens to beat last gen's flagship, exactly like was almost always the case for 15 years.
Click to expand...

QFE! Its sad to see how badly Nvidia is milking customers.
 
Bacon1 said:
QFE! Its sad to see how badly Nvidia is milking customers.
Click to expand...

I guess I see it differently...

The current new GPU gen each year, with a 'big' chip every other year works fine for me. This allows more innovation for features and software to be introduced every year, which is really a big focus for NV. I would press anyone to argue that this hasn't been successful for them recently.

The alternative is what we saw a number of years ago where the big chip was every 2-2.5 years. That is great, but then you have absolutely no opportunity to tweak that arch mid-stream. Tech just moves faster these days and that is not a great strategic method to choose. Prices aside, its a good way to execute.

I do not dispute that prices are higher, but as long as folks are willing to pay and competition is lacking, it will not correct itself. 🙂
 
exar333 said:
I guess I see it differently...

The current new GPU gen each year, with a 'big' chip every other year works fine for me. This allows more innovation for features and software to be introduced every year, which is really a big focus for NV. I would press anyone to argue that this hasn't been successful for them recently.

The alternative is what we saw a number of years ago where the big chip was every 2-2.5 years. That is great, but then you have absolutely no opportunity to tweak that arch mid-stream. Tech just moves faster these days and that is not a great strategic method to choose. Prices aside, its a good way to execute.

I do not dispute that prices are higher, but as long as folks are willing to pay and competition is lacking, it will not correct itself. 🙂
Click to expand...

All of these things are only good for Nvidia's stock holders, not gamers.


Nvidia has been releasing worse and worse performance / $ cards each generation and no one seems to care because they are faster, but you are paying more to get that speed.

Yes its faster than a 980, it's also $200 more!

As a gamer and not a stock holder, I'm disappointed. If I was a stock holder I'd be gitty to see the massive profits from these cards.
 
exar333 said:
Games and equipment don't fit nicely into every-other-year instance cycles. For someone who just upgraded to a 4K monitor, or wants to play VR, the 1080/1070 is VERY compelling. They don't care that Vega may be out in 6 months or big Pascal in 12 months. Give me a break!
Click to expand...
The jury is still out on that one, as for VR I'd say multi GPU is the way to go cause there's a good chance that even big Pascal/vega won't be able to handle it on their own, in case of a single card. Also VR is just an ultra niche atm, anyone spending 600~700$ on a 1080 for VR is not thinking straight IMO, same goes for Polaris.
 
RussianSensation said:
The context is simple. You take a 560Ti level card and launch it first for $600-700. Going back in time it would be 560Ti --> relabeled 580 and price increase from $249 to $699 Founders (NV would call GTX560Ti as a GTX480 or 580 at launch then). In benchmarks, it would slaughter GTX285/HD4890. Then a year or so later, release a cut down GTX480 or a full die GTX580. Call it GTX580Ti now and raise the price to $799-899.

$699 1080 = $549 980 = $499 GTX680 => all $249 560Ti successors.

If you don't see how this is bad for gamers or why they are pissed at this marketing game, buy the $699 1080 FE and enjoy it until $599-699 Volta GV104 is released.

Fact is 1080 is an upper mid-range Pascal chip priced as a high end flagship. NV and AMD are simply splitting the generation into 2 or even 3 parts now.l and manipulating marketing names. In the past, 1080 would be 1060Ti. That's what it is. 314mm2 die, 320GB/sec, 256-but bus, no HBM2. It's a mid-range card that just happens to beat last gen's flagship, exactly like was almost always the case for 15 years.
Click to expand...

Fact is 1080 is a flagship card. It's the fastest single GPU available and therefore it can be priced however NV sees fit. You (or anybody else) have zero proof that NV is even going to release a GP100 card other than basing assumptions on historical trends. All the consumer cares about right now is that the 1080 smashes everything else, and it's actually priced fair for what it is. NV could have priced this card at $1250 (since it sits right between titan X and probably the Pro Duo) and nobody could have said boo.

Now if you want to look at the techie side of it, the reason this keeps happening has literally nothing to do with NV marketing or milking customers. It has to do with the fact the AMD has jack shit to bring to the market and keeps playing catch up with ridiculous propositions. AMD's answer to the 980ti was a card that is rated for 275W and comes bundled with water cooling, and their reply to Titan X came way too late at 375W and $1500 (lol).

You can sit and wait for the 1080ti (or Titan Y, or whatever), or buy the 1080 and enjoy it now. Let's stop with the anti-consumer conspiracy theories, this is just simple straight forward capitalism due to lack of competition. If the master race don't like it, then maybe they should switch to another hobby or just buy consoles...those are by far the best perf/$!


Bacon1 said:
All of these things are only good for Nvidia's stock holders, not gamers.


Nvidia has been releasing worse and worse performance / $ cards each generation and no one seems to care because they are faster, but you are paying more to get that speed.

Yes its faster than a 980, it's also $200 more!

As a gamer and not a stock holder, I'm disappointed. If I was a stock holder I'd be gitty to see the massive profits from these cards.
Click to expand...

And here is where you are wrong. Nobody is forcing you to buy it. If you (or your peer gamers) are buying it, then it's you who is enabling the problem. You can't eat the steak and not pay for it.
 
Last edited:
Ok, lets actually answer the original question 🙂 Percentages a bit approximate - just looked at the odd game in the AT reviews - someone has a neater list somewhere.

680 release March 2012, +30% ish vs 580, (499)
780 release May 2013, +35%(?) vs 680, (649)
980 release September 2014, seemingly ~30% vs 780, (549)
980ti release May 2015, +35% (?) vs 980, (649)
1080 release May 2016, +30% (?) vs 980ti, (599)

Clockwork! The amount of resources/organisation required to do something like this is really quite scary.

Looks like there are definitely good reasons to ask the original question - it looks very close to me. Maybe a little more performance/longetivity from the 'big' cards but definite counter balances for the medium cards too.

Odd that, you might imagine someone at NV has done the sums and made sure it was 🙂 (Of course they have!)

They definitely can't price these x80's at 400 or something at launch. If they did that there would be no reason for anyone sane to ever buy the big cards.

There is also the rogue factor of the 780ti (July 2013, 699) which came in between the 780 and 980. Looks like that might have been the last thing (so far) that AMD have 'forced' NV to do.
 
^ Thanks! That does confirm my suspicions more or less, but I'll let people question the numbers and add context as much as they desire.

Prices are rising, but more so on the medium chips... although I strongly suspect that this time around the 1080Ti will also go up.
 
Qwertilot said:
Ok, lets actually answer the original question 🙂 Percentages a bit approximate - just looked at the odd game in the AT reviews - someone has a neater list somewhere.

680 release March 2012, +30% ish vs 580, (499)
780 release May 2013, +35%(?) vs 680, (649)
980 release September 2014, seemingly ~30% vs 780, (549)
980ti release May 2015, +35% (?) vs 980, (649)
1080 release May 2016, +30% (?) vs 980ti, (599)

Clockwork! The amount of resources/organisation required to do something like this is really quite scary.

Looks like there are definitely good reasons to ask the original question - it looks very close to me. Maybe a little more performance/longetivity from the 'big' cards but definite counter balances for the medium cards too.

Odd that, you might imagine someone at NV has done the sums and made sure it was 🙂 (Of course they have!)

They definitely can't price these x80's at 400 or something at launch. If they did that there would be no reason for anyone sane to ever buy the big cards.

There is also the rogue factor of the 780ti (July 2013, 699) which came in between the 780 and 980. Looks like that might have been the last thing (so far) that AMD have 'forced' NV to do.
Click to expand...

1080 should be $699, there are no cards for $599 for sale.

Also compared to a $600 (standard price for 980 TI, probably cheaper since 1080 announcement) with custom coolers it doesn't get 30% performance increase
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top