980Ti was released in 2015 with 6gb, 250W TDP for $649

Please correct, I know wasn't formal but it looks fine to me.

>2015 node and tech is cheaper in 2019
who would've thunk.
It took 4 years for nvidia to add 2 more GB of vram form 2007.
PROGRESS.

Attached: turkroach_medicine.jpg (474x474, 17K)

>Isn't Pascal most revolutionary GPU architecture our time?
would be, Isn't Pascal the most revolutionary GPU architecture of our time?, other than that it's fine

>know wasn't
know it* wasn't

>comparing top of the line xx80ti product with mid-high range xx70 product
I seriously hope you don't do that

Except the 1070 matched the 980ti for performance and for far less. It was a good 40% jump compared to the last gen equivalent, which was great and put them far ahead of what AMD gained from fury or vega.

A 40% gen to gen improvement was fairly common during the node shrinking days with a tick tock arch, but now that it has slowed down we have to wait a bit longer. Here's hoping navi gives a 40% jump over polaris, even if it's only for the mid range.

>A 40% gen to gen improvement was fairly common during the node shrinking days with a tick tock arch, but now that it has slowed down we have to wait a bit longer. Here's hoping navi gives a 40% jump over polaris, even if it's only for the mid range.
huge bumps in gpu performance is very common when you do a new μarch.
Why?
becaus by changing a trivial things in the pipeline can get you performance.
e.g. when gaymes wanted more fillrate and more texels/sec, gpus just added it to the pipeline and the older gpus got crippled when running new games or greater resolutions.
nvidia had a shitload of rops in their gpus and started the jewish tricks with the subpixel AA years ago and they just re-iterate some AA algorithm and fight amd because they have plenty of ROPs.
In the end, you can fix every gayme with the proper settings and you can make any card look good or bad.
amd had more generic approach with gcn and relied on pure horsepower to render anything, that's why it aged so well.
novidia cards on the other hand, lacked support for newer api, lacked support for features like DC5.0 for a couple generations and their design was always tailored to the current gen of gaymes.
New games came out and the performance went south.

>xx80ti xx70

It makes zero sense to write this like you did, particularly when the talk is about the 980 Ti, but it would have been retarded regardless either way.

Also the 980 Ti was the end of the 9-series, so it makes perfect sense to compare them as the casual cards are never meant to compete with top of the line while the 70 cards are.

>Isn't Pascal most revolutionary GPU architecture our time?
No. It's a programming language I learned back in the 80s.

Attached: 1531547533818.jpg (500x842, 155K)

>780Ti was released Nov 2013 for $699
>970 was released Sept 2014 for $329
Isn't Maxwell the most revolutionary architecture of our time?

Performance gains between Kepler and maxwell are not as great as Pascal and maxwell.

>Got a 1080Ti a year and a half ago
>Paid MSRP (~$1000 CAD after import tax from the US)
>Went up in value over time thanks to miners
>Now
>Could probably easily sell it for $800 CAD but there's literally no reason to upgrade anyways
FEELS GOOD

The 970 is greater than or equal to the 780Ti, re-re

>thunk

bring dich um du dumer schluchtenscheißer lol.