Busting the myth of AMD longevity.

[Thinker_145]

It seems I hear entire essays on how AMD cards have better longevity on this forum every now and then. And yet nobody has actually ever proved it with an apples to apples comparison with a reasonable sample.

https://youtu.be/BfSi-Z8r12M

680 and 7970 GHz are dead equal in 2017 despite the latter having 50% more VRAM.

I would really like to see a similar comparison showing otherwise or people who keep on going about the magical AMD longevity overreacting over a single game at release should just stop. Even the reviewer remarked that the results were nothing like how people predicted them to be in the comments.







Troll threads are not allowed in tech or anywhere for that matter.


esquared
Anandtech Forum Director

 

[Headfoot]

My 290 when I bought it at release for $400 was neck to neck with a 780, now it leaves it in the dust. Maybe the word here isn't myth, but "legend" as in the 290 has proven to be legendary value. My two favorite cards of all time are the 290 and the 8800 GT. If things keep going how they are Hawaii will have the same longevity and value as G92 did

 

[Thinker_145]

My 290 when I bought it at release for $400 was neck to neck with a 780, now it leaves it in the dust. Maybe the word here isn't myth, but "legend" as in the 290 has proven to be legendary value. My two favorite cards of all time are the 290 and the 8800 GT. If things keep going how they are Hawaii will have the same longevity and value as G92 did

Any proof for this claim? We need a comparison like this to really ascertain that.

I mean it's so funny that the 7970 Ghz is hailed as a "legendary" card when in fact it is no better than the 680. Goes to show how much bias there is.

 

[kawi6rr]

Most of the recent benchmarks I've seen don't even have the 680 on it anymore but they do have the 280x. So even many of the websites seem to think the 280x is still worth comparing too.

 

[zinfamous]

wow, you found two specific cards to compare. Myth busted, everybody!

 

[crisium]

If there's multiple new game benchmark threads per year in which "people who keep on going about the magical AMD longevity overreacting over a single game at release" then it's not a single game.

 

[Det0x]

I would really like to see a similar comparison showing otherwise or people who keep on going about the magical AMD longevity overreacting over a single game at release should just stop. Even the reviewer remarked that the results were nothing like how people predicted them to be in the comments.

http://www.anandtech.com/bench/product/1722?vs=1719

Please note, this a non ghz editon of 7970

Also please note: as soon as you start to remove some of the older games and replace them with 2016-2017 (and or dx12) games, the 680 gets left in the dust..

But if you really wanna see the current champion of AMD's longevity, look no further then to Hawaii... Kepler really gets slaughtered in that comparison, as its even stepping on the toes of some of the higher up Pascal cards in certain benchmarks..

 

[Erenhardt]

It seems I hear entire essays on how AMD cards have better longevity on this forum every now and then. And yet nobody has actually ever proved it with an apples to apples comparison with a reasonable sample.

https://youtu.be/BfSi-Z8r12M

680 and 7970 GHz are dead equal in 2017 despite the latter having 50% more VRAM.

I would really like to see a similar comparison showing otherwise or people who keep on going about the magical AMD longevity overreacting over a single game at release should just stop. Even the reviewer remarked that the results were nothing like how people predicted them to be in the comments.

Sorry, but that was not 2017

That is 2016 (770vs280x = 680v7970):

And 2017 will be even uglier for 680
Also, this see this:

http://www.anandtech.com/bench/product/1722?vs=1719

Please note, this a non ghz editon of 7970

Also please note: as soon as you start to remove some of the older games and replace them with 2016-2017 (and or dx12) games, the 680 gets left in the dust..

But if you really wanna see the current champion of AMD's longevity, look no further then to Hawaii... Kepler really gets slaughtered in that comparison, as its even stepping on the toes of some of the higher up Pascal cards in certain benchmarks..

 

[crisium]

If more people do bring up comparisons, remember to also key price into the equation. People often equate the chips wrong.

Outside of the initial Nvidia launch price undercutting, you must not forget that AMD heavily price cut for most of that generation. 7950 usually competed with the 660 Ti, not the 670 which is sometimes misunderstood. And the 7870 was usually closer in price to the 660, not the 660 Ti. The 7850/265 consistently undercut the 660 and the 270X undercut the 760.

Some snapshot of prices mid-generation (scroll down):
December 2012: https://www.techpowerup.com/reviews/Club_3D/HD_7870_jokerCard_Tahiti_LE/
June 2013: https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_760/
Feb 2014: https://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_750_Ti/

So when more benchmarks start rolling, never forget prices too. For the US anyway.

 

[poofyhairguy]

Here is that graph cut down to just the competing cards from that generation. It isn't perfect but it gets across the point:

Hawaii is a monster, as the 290X and the 290 beat the 780 ti (aka the 290x competitor) today.

But look at Tahiti- it is almost equal to a 780 that at one point was a 290 competitor.

 

[nurturedhate]

At this point the OP really needs to change to title from Busting to Proving.

 

[DisarmedDespot]

Pfft, I shouldn't have read this thread. I was gonna get an R9 280X or 290, but then the crypto mining craze hit, and I ended up settling for a GTX 770 2G for a little over $300. If I had spent a little more money earlier and nabbed a custom 290, I'd still be happily running it.

 

[Thinker_145]

lol at people using game GPU charts. It was already shown how inaccurate they are since they lump the monetary FPS together instead of using the % for each game. Please come up with more credible data.

 

[zinfamous]

at first I thought this thread had run its expected course, but it looks like new life awaits!

 

[IllogicalGlory]

lol at people using game GPU charts. It was already shown how inaccurate they are since they lump the monetary FPS together instead of using the % for each game. Please come up with more credible data.

It is wrong, but the difference between using the proper method won't be more than 1-2%. If you want to, you could look at the reviews and average performance using the right method yourself.

 

[poofyhairguy]

Here is the 780 vs the 290 on Anandtech benches:

The word that comes to mind is "bloodbath."

 

[Face2Face]

I would like to see a current battle between a custom cooled R9 290 and GTX 780 - stock and overclocked. I'm tired of seeing reference cards being benchmarked where both cards don't even hit 1Ghz on the core...

 

[Mercennarius]

A 290X/390X is substantially faster on average than a 780/Ti. The Hawaii GPUs is one of the best "aging" GPUs i can think of. 3.5 years old and still very competitive with today's higher end GPUs.

 

[Det0x]

lol at people using game GPU charts. It was already shown how inaccurate they are since they lump the monetary FPS together instead of using the % for each game. Please come up with more credible data.

Is it a coincidence that you firmly ignored my previous post then ? Didn't like what it showed i take it.. Lets try this one more time then:

7970 = 280x
680 = slower then 770. ((1006core/1502mem vs 1046core/1753mem)otherwise same)

And what do we learn looking at this comparison then ?

https://www.computerbase.de/thema/grafikkarte/rangliste/

Also please note: as soon as you start to remove some of the older games and replace them with 2016-2017 (and or dx12) games, the 680 gets left in the dust..

Still stand here..

At this point the OP really needs to change to title from Busting to Proving.

Indeed, this is pretty much quote of the thread

I hope this will be my last post in this thread, so let me repeat myself on more time: if you really wanna see the current champion of AMD's longevity, look no further then to Hawaii

 

[Headfoot]

Is it a coincidence that you firmly ignored my previous post then
I hope this will be my last post in this thread, so let me repeat myself on more time: if you really wanna see the current champion of AMD's longevity, look no further then to Hawaii

You could repeat all day and it wouldnt make a difference, this is definitely a troll thread so the goal is endless posting churn, not seeking facts

 

[Thinker_145]

Is it a coincidence that you firmly ignored my previous post then ? Didn't like what it showed i take it.. Lets try this one more time then:

7970 = 280x
680 = slower then 770. ((1006core/1502mem vs 1046core/1753mem)otherwise same)

And that do we learn looking at this benchmark then ?

https://www.computerbase.de/thema/grafikkarte/rangliste/


Still stand here..


Indeed, this is pretty much quote of the thread

I hope this will be my last post in this thread, so let me repeat myself on more time: if you really wanna see the current champion of AMD's longevity, look no further then to Hawaii

Nope only a 7970 is a 7970 no substitute can come in to work for a true longevity comparison.

 

[zinfamous]

I feel like this is why Trump is our president...

 

[Bacon1]

I feel like this is why Trump is our president...

These are alternative benchmarks.

 

[Atreidin]

Nope only a 7970 is a 7970 no substitute can come in to work for a true longevity comparison.

You're either trolling or emotionally invested in showing AMD is worse than they are. Either way it's a waste of everyone's time.

 

[Thinker_145]

And btw even if 1 generation of cards show a difference doesn't really mean much in the grand scheme of things.