Titan X for 500$

[RussianSensation]

If it was 6 months ago it would be in my case already but with new cards coming it makes you think.

I'm more worried about the Titan X not having the proper features for a full direct X 12 game.
Mabe some 275$ midrange card will mop the floor with it in newer games late 2016 early 2017.
The value will go down and i'll be pissed.

Mabe buy it and resell it, that's sounds like a plan

I doubt a ~$279 card will be considered mid-range next gen, not new @ msrp. I guess it depends on what one defines as mid-range. Based on where R9 290 landed and then where 970/390 landed, I now consider mid-range pricing to be about $330-399. The old days of mid-range $199-229 cards at MSRP? Forget it. Those are now low-end products. The difference in performance between modern $200 GPUs and $300 GPUs is huge.

Now we have a solid amount of data with modern games to compare 970 vs. 980 and 970 SLI.

@ 1080P 980 is only about 14% faster but 970 SLI is 45% faster
http://www.techpowerup.com/reviews/ASUS/R9_380X_Strix/23.html

That means those who bought 970 SLI in Sept 2014 are getting the best of both worlds: when SLI scaling doesn't work, you get most of 980's performance but when SLI works, 980 gets killed. This is one of those cases where SLI/CF makes more sense than a single card, subject to your PSU. 970 also lost barely any value which means 970 SLI owners could have easily sold 2 of those cards and upgraded to 980Ti if they wanted to.

What am I getting at? You can keep gaming on 960, buy TX for $500, sell for $750-800. Then, even if 970's successor is $399, you can buy 2 of those for $800 - ($250-300 profit) = $500-550 out of pocket. You will essentially get next gen's GP100 performance right away in 2016 and if history repeats itself, you'll get more VRAM than 980Ti's, more features too. By that time you can get a Core i7.

From a practical point of view, if the 960 satisfies you now, why spend $500 on a flagship card when you can profit $250-300 and just upgrade when you actually need the extra GPU horsepower. Look how fast GPU prices drop over time:

Nov 2013
$699 780Ti
$549 290X

Now you can buy a used 780Ti for $200, GTX970/390 for $240-250 on timed sales. Unless you actually need the horsepower now, put aside the profits and go from there.

 

[Mondozei]

SLI is a broken hack at this stage, NV used to be good at releasing them with the game-ready drivers but these days it's often missing, sometimes for months. That's especially been true of games this year.

There's been a long-term decline in SLI/Crossfire more generally. AMD kind of bet on it big a few years ago but complexity has risen dramatically and as such SLI/Crossfire is a dysfunctional mess as it stands today.
It will surely get a lot better with DX12, but it will take some time.

As for happy medium, if you're happy with a 960 today, I doubt you'll really need the Titan X or similar performance. Honestly, such an upgrade only really makes sense if you're moving to 1440p@144 Hz or similar.

 

[Silverforce11]

I don't know, I used to think like that, but if we get a GP104 which is not exceptionally overwhelming then we might have a $300 GPU which performs around the same as a 980 Ti. That wouldn't make the 980 Ti - or the Titan X - "utterly obsolete". It would merely place them in the mainstream of performance.

Compare the custom 970 for ~$339 vs Titan.

That was on the SAME node.

Now imagine GP104 vs Titan X, 16nm ff vs 28nm, new uarch. Think about that for a second and let it sink in.

Would not be out of the ordinary to have a "Pascal 970" that's 20% faster than Titan X at ~150W.

To me, Titan X, 980Ti, Fury X (lol Fury X2) is obsolete within a few months.

Anyone saw the SIGGRAPH presentation? Oxide's Dan Baker had this to say: "Next-gen GPUs will be 200% faster at DX12 than current stuff". Not twice as fast, that's 100% faster. I'm sure that's best case scenario, but even twice as fast is already massive.

 

[MrTeal]

Compare the custom 970 for ~$339 vs Titan.

That was on the SAME node.

Now imagine GP104 vs Titan X, 16nm ff vs 28nm, new uarch. Think about that for a second and let it sink in.

Would not be out of the ordinary to have a "Pascal 970" that's 20% faster than Titan X at ~150W.

To me, Titan X, 980Ti, Fury X (lol Fury X2) is obsolete within a few months.

Anyone saw the SIGGRAPH presentation? Oxide's Dan Baker had this to say: "Next-gen GPUs will be 200% faster at DX12 than current stuff". Not twice as fast, that's 100% faster. I'm sure that's best case scenario, but even twice as fast is already massive.

The 970 was about as fast as the 780Ti when it launched, while the 670 was about 20% faster than the 580 when it launched, while also getting about 2x the density moving from 40nm to 28nm. 28nm to 16FF is supposed to also be about a 2x increase in density, so even though 28 to 16/14 is supposed to be two full nodes it will probably be a lot closer to the actual results going from 40 to 28.

The difference here is that going from GF110 to GK104 and from GK110 to GM104 was a transition from a compute heavy chip to a lean gaming focused chip. Since GM200 is already a really lean gaming chip, they won't get that extra benefit moving to GP104.

You might be able to buy a $350-400 GTX 1070 next year that is as fast as a Titan X, but I'd be shocked if it beat it by much. Considering that buying it would give him an massive upgrade from his 960 for the next 6 months or whatever it ends up being, it seems like a good deal even outside of reselling it.

 

[tential]

This thread is hurting my brain. This was such a no-brainer that I feel the op is bragging or trolling.





This is the third threadcrapping post you have made in the thread.

Now stay out of his thread.


esquared
Anandtech Forum Director

 

[Majcric]

Compare the custom 970 for ~$339 vs Titan.

That was on the SAME node.

Now imagine GP104 vs Titan X, 16nm ff vs 28nm, new uarch. Think about that for a second and let it sink in.

Would not be out of the ordinary to have a "Pascal 970" that's 20% faster than Titan X at ~150W.

To me, Titan X, 980Ti, Fury X (lol Fury X2) is obsolete within a few months.

Anyone saw the SIGGRAPH presentation? Oxide's Dan Baker had this to say: "Next-gen GPUs will be 200% faster at DX12 than current stuff". Not twice as fast, that's 100% faster. I'm sure that's best case scenario, but even twice as fast is already massive.

There is quite a big difference in being obsolete and just not being the top card anymore. Anyway, unless the card has some kind of malfunction I don't see how the OP can go wrong i.e sell the card, or keep it to use.

 

[Keysplayr]

Isn't that basically a custom 980Ti? Since they have a major boost clock gap over Titan X.

So the question is, would you grab a custom 980Ti for $500?

Edit: I would not, because I believe we're quite close to the real next-gen. Folking out $500 that's gonna be utterly obsolete in 6 months is sad panda. The end of a node with a major node jump about to start is always a bad time for major GPU buys.

There is quite a big difference in being obsolete and just not being the top card anymore. Anyway, unless the card has some kind of malfunction I don't see how the OP can go wrong i.e sell the card, or keep it to use.

Yeah I read that post and was like "WTF" ?

OP, yes. Buy it at the speed of thought.

 

[happy medium]

bought it and sold it that fast.... thanks guys.

gonna just put the extra 250$ away for a Kabylake upgrade and a buy myself a surround sound gaming headset.

Now I see my cpu is overclockable. It sure would have been cool to see how much my 6100 overclocked would have pushed a Titan.
Thanks again.

 

[lavaheadache]

I would have kept it. Where is my X for $500?

 

[Azix]

shouldn't even be a question. Nvidia people are loose with money so someone will buy that off you for decent profit. If you have the cash to buy a titan x at $500 and it works, buy it and resell it.

Its not even worth thinking about because its a simple change cash to more cash. You're basically asking if you should turn $500 into $700 - $1000

 

[happy medium]

Its not even worth thinking about because its a simple change cash to more cash. You're basically asking if you should turn $500 into $700 - $1000

SO you think it was easy to hold a Titan X in your hand and then sell it?

 

[Azix]

SO you think it was easy to hold a Titan X in your hand and then sell it?

I would do the same with a 980ti if there was good profit. I've learnt to look at things in that way (plus I expect a titan x to perform around fury non-x or worse next year).

If you think about it its really easy. Worst case you are without a card for a bit.

 

[RussianSensation]

bought it and sold it that fast.... thanks guys.

gonna just put the extra 250$ away for a Kabylake upgrade and a buy myself a surround sound gaming headset.
Thanks again.

:thumbsup: Brilliant move. I am glad you followed our advice. By the time you get i7 6700K/Kaby-Lake, say summer/fall 2016, there may be a card for $550-600 that's 20-30% faster than the Titan X, have a newer architecture (the key part is NV/AMD driver support is best for newest cards), be better suitable for DX12 games. More importantly, that $550-600 card will cost you $300-350 out of pocket after $250 profit, not $500. That means you would have $150-200 left over towards that i7 CPU upgrade that otherwise would have been eaten up by the $500 TX purchase. That's win-win on both the CPU and GPU side upgrade paths!

As a bonus, next gen AMD/NV cards should also have superior video encoding/decoding functionality, HDMI2.0a/DP1.3 connection, and with a new card you'll get 2-3 year warranty.

 

[Silverforce11]

Yeah I read that post and was like "WTF" ?

OP, yes. Buy it at the speed of thought.

I thought he was buying it to use, which is where I come from. If he buy to sell profit $250, no brainer man.

Unless you like seeing such a big sum to go obsolete so fast, buying top GPU now is silly.