Is there any reason to upgrade from a 780Ti?

[CPA]

Checking out 970 and 980 cards and from my rudimentary research (GPUBoss primarily), I just don't see a strikingly large enough difference (and the 780TI seems to outperform some 970's) to justify paying $300-$600 for an upgrade. Am I missing something? Too early to upgrade?

 

[IEC]

Checking out 970 and 980 cards and from my rudimentary research (GPUBoss primarily), I just don't see a strikingly large enough difference (and the 780TI seems to outperform some 970's) to justify paying $300-$600 for an upgrade. Am I missing something? Too early to upgrade?

Not worth the upgrade cost. Wait for 16nm or whatever the next process shrink will be. Unless Radeon 390X blows Titan X out of the water and is <$600 won't be worth upgrading either (unless you have a specific display setup that really needs more horsepower)

AnandTech bench comparison of 780 Ti and 970:
http://anandtech.com/bench/product/1072?vs=1355

May be worth the sidegrade if you can find some chump to sell the 780 Ti to for the cost of a 970 or better.

 

[Skurge]

Probably better off waiting for 980Ti

 

[LTC8K6]

That may depend on whether you think NV is crippling Kepler in the latest drivers/games...

 

[IEC]

That may depend on whether you think NV is crippling Kepler in the latest drivers/games...

Why would it depend on whether you *think* they are or not? It's clear that whether it is something in the drivers or in the game engines themselves, Kepler is not performing as it "should" be in certain games. This could be resolved by future software updates, however, and may not even matter to OP if he doesn't play those games.

Several options for OP:
1) Wait to pick up a "cheap" 780 Ti and SLI
2) Sell 780 Ti and pick up a 970/980/980 Ti
3) Sell 780 Ti and pick up dual 970s and SLI
4) Sell 780 Ti and pick up a R390X (if good)
5) None of the above. Wait for 16nm FinFET or whatever the next process shrink is to upgrade

 

[CPA]

Not worth the upgrade cost. Wait for 16nm or whatever the next process shrink will be. Unless Radeon 390X blows Titan X out of the water and is <$600 won't be worth upgrading either (unless you have a specific display setup that really needs more horsepower)

AnandTech bench comparison of 780 Ti and 970:
http://anandtech.com/bench/product/1072?vs=1355

May be worth the sidegrade if you can find some chump to sell the 780 Ti to for the cost of a 970 or better.

Holy crap, that amazes me. Is that a driver issue (I can't imagine it is) or a specific hardware change that the next generation performs so poor comparatively.

Please keep in mind that I don't nearly keep up with this stuff as I used to.

 

[CPA]

Why would it depend on whether you *think* they are or not? It's clear that whether it is something in the drivers or in the game engines themselves, Kepler is not performing as it "should" be in certain games. This could be resolved by future software updates, however, and may not even matter to OP if he doesn't play those games.

Several options for OP:
1) Wait to pick up a "cheap" 780 Ti and SLI
2) Sell 780 Ti and pick up a 970/980/980 Ti
3) Sell 780 Ti and pick up dual 970s and SLI
4) Sell 780 Ti and pick up a R390X (if good)
5) None of the above. Wait for 16nm FinFET or whatever the next process shrink is to upgrade

Yes, I will be playing most of those games even IF my Steam library is overflowing . Option 1 may be the best route, if needed.

 

[IEC]

Holy crap, that amazes me. Is that a driver issue (I can't imagine it is) or a specific hardware change that the next generation performs so poor comparatively.

Please keep in mind that I don't nearly keep up with this stuff as I used to.

It's an apples to oranges comparison since the 780 Ti was top tier at its release versus the 970 which is basically a cut-down 980.

Even comparing a 780 Ti to a 980 though (still not quite apples to apples - except maybe on price $$$, that'll have to wait for 980 Ti), you won't see HUGE performance gains in most titles:
http://anandtech.com/bench/product/1442?vs=1441

It's the nature of a refresh product. Both are manufactured on a 28nm process so it's not like there was a die shrink involved.

So yeah, you can probably afford to wait

 

[happy medium]

Simple answer No

 

[LTC8K6]

Why would it depend on whether you *think* they are or not? It's clear that whether it is something in the drivers or in the game engines themselves, Kepler is not performing as it "should" be in certain games. This could be resolved by future software updates, however, and may not even matter to OP if he doesn't play those games.

Several options for OP:
1) Wait to pick up a "cheap" 780 Ti and SLI
2) Sell 780 Ti and pick up a 970/980/980 Ti
3) Sell 780 Ti and pick up dual 970s and SLI
4) Sell 780 Ti and pick up a R390X (if good)
5) None of the above. Wait for 16nm FinFET or whatever the next process shrink is to upgrade

If I thought the bad performance of the 780ti and other high end Kepler cards in the latest games was due to NV crippling Kepler, I don't know that I'd buy an expensive NV card again.

If I didn't think so, it wouldn't matter.

However, I'd still have an expensive card that seems to perform relatively poorly in the latest games, and I'd need to decide what to do.

It's easy to look at older games and say that the 780ti is keeping up.

Anyway, I am not the one complaining about poor Kepler performance. I read about it here. I hadn't heard about it before.

The fastest Kepler I have ever owned was a 660.

 

[LTC8K6]

I'd be pissed if a 960 beat my 780 in a new game.

NV is investigating:

https://forums.geforce.com/default/...he-witcher-3-wild-hunt-/post/4533205/#4533205

 

[DaveSimmons]

I still have a 680 and haven't bothered to upgrade it yet.

I'd have to get a Titan X to see a 100% jump in power, and there aren't any games that I'm playing yet where I find my 680 isn't good enough.

But my monitor is only 19x12 and I don't care much about extra-wavy grass patches or uber-textures for slightly grainier dirt.

I'll probably upgrade after the next die shrink but that keeps getting pushed back.

 

[CPA]

It's an apples to oranges comparison since the 780 Ti was top tier at its release versus the 970 which is basically a cut-down 980.

Even comparing a 780 Ti to a 980 though (still not quite apples to apples - except maybe on price $$$, that'll have to wait for 980 Ti), you won't see HUGE performance gains in most titles:
http://anandtech.com/bench/product/1442?vs=1441

It's the nature of a refresh product. Both are manufactured on a 28nm process so it's not like there was a die shrink involved.

So yeah, you can probably afford to wait

Thanks IEC!

 

[imaheadcase]

Depending on performance increase, i will be switching to 980TI from 780TI.

 

[swilli89]

If I thought the bad performance of the 780ti and other high end Kepler cards in the latest games was due to NV crippling Kepler, I don't know that I'd buy an expensive NV card again.

If I didn't think so, it wouldn't matter.

However, I'd still have an expensive card that seems to perform relatively poorly in the latest games, and I'd need to decide what to do.

It's easy to look at older games and say that the 780ti is keeping up.

Anyway, I am not the one complaining about poor Kepler performance. I read about it here. I hadn't heard about it before.

The fastest Kepler I have ever owned was a 660.

Exactly what this guy said. If nvidia indrrd crippled performance of Kepler, would you seriously let their plan work and fork out for a new nvidia card?

If 390x is around the same at titan x for $500-$600 you've got a huge win right there.

 

[guskline]

CPA, what resolution monitor are you using?

 

[biostud]

Do you feel your games have choppy gameplay? If not, then don't upgrade.