Anyone buying a Titan?

[bononos]

To be fair, I have a hard time justifying $200. I've been all over the map over the last two years with video cards in terms of price/performance, and frankly the "high end" cards are really a ripoff in my eyes.

I'm currently rocking a GTX660, and it does plenty fine on every single title I play (current gen titles too). Then again, I'm not rocking 3 screens, nor 1440p, nor 4x MSAA, nor do I care about the difference between 60fps and 65fps.

There days, it's become all about separating people from their money. There is no justifiable reason that a video card should cost more than the rest of your system - combined, including the software, desk, chair, and house you live in.

A $200 card would give very good performance for a gamer at reasonable resolutions on a single monitor but we are talking about a titan of a card here, the high end of high end where the sky's the limit for pricing.

Thats the advantage that the people who put out bleeding edge products have. Its not about what you need, its about what you want to have.

 

[RussianSensation]

That's even worse than Canada. The 680 is more expensive here too but it is not that far out of whack (for instance most 7970s retail for similar value, especially HIS cards for some reason, but 680s are 60 or so more than in the US these days).

I knew that cards could be far more expensive other places around the world but did not realize the hierarchy was so warped some places.My apologies.

In Ontario you could buy 3 1Ghz 7970s after selling the games. $380 CDN 1Ghz Gigabyte 7970 with 5 free games x 3 with 13% tax = $1,288 CDN.

$1000 CDN Titan with 13% tax = $1,130 CDN

I bet you can sell those 5 games for $50 and you get 3 sets! $1,288 - $150 = $1,138 CDN. :biggrin:

I am not buying the Titan. Most games in 2013 are console ports. Titan can't max out Crysis 3 either. Waiting for 20nm GPUs.

 

[BallaTheFeared]

In Ontario you could buy 3 1Ghz 7970s after selling the games.

$380 CDN 1Ghz Gigabyte 7970 with 5 free games x 3 with 13% tax = $1,288 CDN

$1000 CDN Titan with 13% tax = $1,130 CDN

I bet you can sell those 5 games for $50 and you get 3 sets!

$1,288 - $150 = $1,138 CDN. :biggrin:

Why not just gouge your eyes out instead?

 

[Elfear]

Why not just gouge your eyes out instead?

I haven't felt the need to gouge my eyes out yet. Must be something wrong with my vision... ()

 

[bleucharm28]

Im happy with my 2 GTX 680SC+

 

[BallaTheFeared]

Have you seen an optometrist? I thought you said you rarely play games anyways?

 

[Pariah]

Exactly this. If they weren't going to be absolute useless garbage in two years, then I might consider sucking it up and throwing down for a couple of them in the near future. But I can't possibly stomach the loss of value that they will incur. This is just way too much.

NO BUY.

This is an outdated viewpoint. Maybe 5 years ago you could have said this with confidence, not any more.

The pace of improvement for the highend has almost ground to a halt by most measures the past few years. I'm currently running a 2600k oc'd to 4.8GHz using a decent air tower cooler, nothing exotic. This CPU is over 2 years old now and there is still nothing worth upgrading to. It has certainly not lost 80% of its value in 18-24 months.

Video cards aren't that bad yet, but they are getting close. The days of 6 months between refreshes and 12 months between architectures are gone. The 7970 is over a year old now and AMD has already announced, that they will not be releasing anything to beat it this year. That means AMD will have gone 2 years without a new highend card. NVidia just released Titan, and is likely in the same boat as AMD, with nothing new at the top until the end of the year as well. So, how exactly is the Titan going depreciate significantly when we already know that nothing new is going to be released at the highend for another year? Where it stands now against the competition is where it is going to be 12 months from now.

 

[RussianSensation]

Why not just gouge your eyes out instead?

No thanks. At 1200mhz a single HD7970 is going to be just 30% slower than the Titan max OC on air. Even if CF doesn't work in some games, when it works you get > 2x the performance from 3 7970s and at the current rate they are making > $200 a month in bitcoin mining. I was simply commenting that in Canada you can get 3 1Ghz 7970s for the price of 1 Titan and Titan is just 35-36% faster than a single such card. :whiste: In my case, I'll just wait until games get more demanding in 2014+ and consider picking up Maxwell instead. I have no interest in paying early adopter price premium to future-proof for next gen games that won't be out until PS4/720 drop.

So, how exactly is the Titan going depreciate significantly when we already know that nothing new is going to be released at the highend for another year? Where it stands now against the competition is where it is going to be 12 months from now.

Actually not much has changed in GPUs. People have short memory or something. GTX480 came out March 2010 and it wasn't until March 2012 that we got a 35% performance increase with the 680. GTX480 OC to 580 speeds easily and 580 was barely 15-17% faster. That means in 2010 performance grew less than 20% from NV but jumped a lot more once we moved down to 28nm. Same thing happened with AMD. We went from HD5870 Sept 2009 to HD6970 Dec 2010. That was also just a 15% bump. The large increase we got of 45% happened by January 2012 when AMD also went to 28nm.

2013 was always expected to be a refresh year (15-20% increases). There were never expectations for some mythical 40-50% faster HD8970 or GK114 with 40% more performance at $500. 2014 is different since it's another node change down to 20nm. We should expect at least a 35-45% increase from HD7970GE/GTX680 cards at the same $500-550 prices. That means in about 15-18 months, there should be a GPU with a performance of a Titan for half the price.

 

[Jaydip]

Probably getting one.

 

[lopri]

lol, no. Two jokes in a day are a bit too much. ($1K Titan, $1.3K Chromebook)

 

[BallaTheFeared]

No thanks. At 1200mhz a single HD7970 is going to be just 30% slower than the Titan max OC on air. Even if CF doesn't work in some games, when it works you get > 2x the performance from 3 7970s and at the current rate they are making > $200 a month in bitcoin mining. I was simply commenting that in Canada you can get 3 1Ghz 7970s for the price of 1 Titan and Titan is just 35-36% faster than a single such card. :whiste: In my case, I'll just wait until games get more demanding in 2014+ and consider picking up Maxwell instead. I have no interest in paying early adopter price premium to future-proof for next gen games that won't be out until PS4/720 drop.

Your chip isn't at 1200MHz though?

Really you found out the average across all samples of ACTUAL user results? That's impressive RS! Me I'll wait until overclockers get the cards personally, never cared for reviewer results they always seemed to get painfully low OC's like 770MHz 470s.

Since CF support is so touch and go if it scales with the third card or not, the inflated frame reports for broken rendering don't lend too much support for any ideology that they would actually be any faster than Titan, not to mention the ms issue. Funny how we've gone to MG is awful when the 7970 was faster, to 3 way CF in less than a week ^_^

I'm just pointing out that CF reports frames that aren't actually rendered, inflating actual performance, suffers from MS issues, and has a hit or miss chance of seeing any actual scaling from the second card, let alone the third. :whiste:


Just 40-50% faster on release drivers, just? That's faster than the 7970 was over the 580 at release, lol.

Me either, going to keep waiting until mid-range 7970s hit $250 or Nvidia brings out some more reasonably priced GK110 cards.

 

[RussianSensation]

Your chip isn't at 1200MHz though?

I was being generous with a 30% delta at 1200mhz. ~1Ghz base Titan is only 35-38% faster at computerbase.de than a 1050mhz HD7970. That means an 1150mhz HD7970 would not be more than 30% slower. I can overclock to 1200mhz but I prefer the nice balance of noise levels and power consumption at stock voltage on my chip.

Just 40-50% faster on release drivers, just? That's faster than the 7970 was over the 580 at release, lol.

40-50% faster on average? What are you smoking now.

28% faster at 1080P, 32% faster at 1600P at TPU
25% faster at 1080P, 22% faster at 1600P at Computerbase
16% faster at 1080P, 12% faster at 1600P at Hardware.fr

$3000 Titan SLI vs. $1200 HD7970GE x 3 on a single 30 inch monitor. Not looking so hot.

3x the price, not even 30% faster on average. Ouch!

 

[BallaTheFeared]

lol 1600p, what joke post.

"Ouch"

CF cheating "Ouch"
http://www.pcper.com/reviews/Graphi...ance-Review-and-Frame-Rating-Update/Frame-Rat

I was going to link some graphs of triple screen where AMD often gets no scaling or negative scaling with Tri and it gets beat pretty handily by Titan, but then I realized my first two points nullified any need for FPS charts when comparing NVidia to AMD.

 

[Pariah]

2013 was always expected to be a refresh year (15-20% increases). There were never expectations for some mythical 40-50% faster HD8970 or GK114 with 40% more performance at $500. 2014 is different since it's another node change down to 20nm. We should expect at least a 35-45% increase from HD7970GE/GTX680 cards at the same $500-550 prices. That means in about 15-18 months, there should be a GPU with a performance of a Titan for half the price.

You're right, 2013 was expected to be a refresh year. However, everyone who thought that was wrong according to AMD, because there is not going to be a refresh this year. The only performance increase we are going to see from AMD is in driver improvements for the 7970.

The only way we are getting a Titan level card in 15-18 months as you speculate is if the refresh expected early next year is skipped, which is not likely. We're not getting a new architecture 3-6 months after a refresh.

 

[Final8ty]

lol 1600p, what joke post.

"Ouch"

CF cheating "Ouch"
http://www.pcper.com/reviews/Graphi...ance-Review-and-Frame-Rating-Update/Frame-Rat

I was going to link some graphs of triple screen where AMD often gets no scaling or negative scaling with Tri and it gets beat pretty handily by Titan, but then I realized my first two points nullified any need for FPS charts when comparing NVidia to AMD.

RadeonPro, job done.

 

[BallaTheFeared]

RadeonPro, job done.

So 40 fps vs whatever titan gets?

 

[boxleitnerb]

RadeonPro, job done.

Yeah, with heavy fps losses. You realize that you should cap your fps at or around the minimum value to get rid of all microstutter, right?

After seeing the PCPER preliminary report, any comparison between CF and any Nvidia offering is completely silly. CF offers terrible value if you actually want to play, not count Fraps-fps. Nothing to sugarcoat it anymore.

 

[VulgarDisplay]

Yeah, with heavy fps losses. You realize that you should cap your fps at or around the minimum value to get rid of all microstutter, right?

After seeing the PCPER preliminary report, any comparison between CF and any Nvidia offering is completely silly. CF offers terrible value if you actually want to play, not count Fraps-fps. Nothing to sugarcoat it anymore.

Honestly all the PCper article shows is that not using Vsync is stupid. The frames in the Nvidia pics are also chopped.

I also wish that these review sites would get a 120hz monitor in there to test as well.

 

[wand3r3r]

Yeah, with heavy fps losses. You realize that you should cap your fps at or around the minimum value to get rid of all microstutter, right?

After seeing the PCPER preliminary report, any comparison between CF and any Nvidia offering is completely silly. CF offers terrible value if you actually want to play, not count Fraps-fps. Nothing to sugarcoat it anymore.

Do you have crossfire? Have you had it lately, or tried radeonpro?

I guess this thread is taking a turn from the OP.

 

[Annisman*]

Yeah, with heavy fps losses. You realize that you should cap your fps at or around the minimum value to get rid of all microstutter, right?

After seeing the PCPER preliminary report, any comparison between CF and any Nvidia offering is completely silly. CF offers terrible value if you actually want to play, not count Fraps-fps. Nothing to sugarcoat it anymore.

This is why multiple 7970s look like great value compared to Titan but totally blow it in terms of the end results. AMD needs to fix their drivers just as much as nvidia needs to lower it's prices.

 

[wand3r3r]

This is why multiple 7970s look like great value compared to Titan but totally blow it in terms of the end results. AMD needs to fix their drivers just as much as nvidia needs to lower it's prices.

I agree that they had better take it seriously and improve it.

@NVIDIOTS
It sounds like they're desperately starting to beat on the sales drum.

Now that crossfire beats titanic (whether it's worse experience or not idk, radeonpro etc) suddenly it's time to quit 'sugarcoating' it and attacking crossfire? Sounds like the focus from NV to try smear crossfire. These attempts appear to be desperately trying to justify the extremely high price of the titan. Are you shareholders or employees with all of this new desperation to smear crossfire. (I don't use crossfire myself but suddenly this is the new drumbeat from team green?)


Now a very bad example to complete the post.

If you look briefly at the numbers posted above.

3 way crossfire is nearly around (trading blows) the level of 3 way titan in those four games.

3 x $400 = ~1200
3 x $1000 = $3000

They are pretty close in BF3, Anno, Sleeping dogs, Hitman.

Say you have a 30% penalty for Radeonpro (guessing, don't see minimums).
Now you have 3 crossfire running around 70% of 3 sli in those four games.

70% performance = 30% of the price

$1200 vs. $3000!!!!

Or 140% of the performance for 300% of the price.

Sure these numbers are only showing a small part of the picture, but the premium is so bad that the only way you can defend it is if you have commercial interest in NV, or are extremely wealthy, or are a fool.

I don't have 2/3 7970's to try see the stutter myself, but those who are so vehemently attacking crossfire suddenly should be attacking NV's outrageous rip off unless they really have ties to NV.

 

[boxleitnerb]

They have done much more than 15% since release and the next big jump is realistically expected at 10-15% anyway. So even if it turns out 10%, it would be so close to the Titan once heavily oced that in most cases you would only be able to bench the difference, say sub 15% avg.

On average "much more" than 15%? Show me, because I don't believe it. I couldn't find any review than spans more Catalyst releases than this:
http://www.computerbase.de/artikel/grafikkarten/2012/bericht-amd-catalyst-12.11-beta/2/

Before you mention HardOCP and the 7950 - I think someone should verify these results, before we call them valid. Aside from that, they only tested 3-4 games - hardly what I would call "average".

And finally - how do you know that AMD hasn't pushed GCN (almost) to its full potential with the latest drivers? Just because you increased performance in the past doesn't mean you can do it again...and again...and again. So this is pure speculation

Do you have crossfire? Have you had it lately, or tried radeonpro?

I guess this thread is taking a turn from the OP.

No I haven't, but I understand the PCPER results and I understand what RadeonPro actually does (namely capping fps). Do you?

I mean if you have 70fps avg, 45fps min and have to cap fps at 45-50fps to get smooth gameplay, you can do the math yourself what scaling will be. Hint: Not good

Honestly all the PCper article shows is that not using Vsync is stupid. The frames in the Nvidia pics are also chopped.

I also wish that these review sites would get a 120hz monitor in there to test as well.

Vsync can mitigate AFR stuttering...but only as long as your fps are above the refresh rate. In my experience, Vsync can actually make things worse, it depends on the game. And remember, not everyone wants to use Vsync due to increased input lag. But I concur that PCper should test with Vsync on as well since that is relevant for at least a part of their readership.

 

[wand3r3r]

On average "much more" than 15%? Show me, because I don't believe it. I couldn't find any review than spans more Catalyst releases than this:
http://www.computerbase.de/artikel/grafikkarten/2012/bericht-amd-catalyst-12.11-beta/2/

Before you mention HardOCP and the 7950 - I think someone should verify these results, before we call them valid. Aside from that, they only tested 3-4 games - hardly what I would call "average".

And finally - how do you know that AMD hasn't pushed GCN (almost) to its full potential with the latest drivers? Just because you increased performance in the past doesn't mean you can do it again...and again...and again. So this is pure speculation



No I haven't, but I understand the PCPER results and I understand what RadeonPro actually does (namely capping fps). Do you?

I mean if you have 70fps avg, 45fps min and have to cap fps at 45-50fps to get smooth gameplay, you can do the math yourself what scaling will be. Hint: Not good

I think someone should verify these results too, before we call them valid.

You're trying pretty hard, any ties to NV? (stocks, employer, sponser)

(Nothing personal, but I mean "suddenly" this is an issue, really? the card is over a year old, cheap, and 70%+ of the 200%+ flag-turd priced NV card.)

I think crossfire 'issues' are a whole topic that should be investigated and fixed, but it still doesn't compensate for NV's shock pricing.

 

[boxleitnerb]

What's with the accusation? No ties to NV here...that is uncalled for.

As for the validity, CF stuttering has been under investigation from many sites for years now, so this is nothing new. The 40+% gains on the other hand I have never seen before in case of the 7950. Maybe I missed it, that certainly is a possibility.

As for pricing: I just cancelled my two Titans that I ordered in a weak moment (580 SLI getting old...). What does that tell you?
This was just in response to those people who think CF is an alternative with such good value. It isn't. A single 7970@OC on the other hand very much is.

Personally I'm left without any options currently which I quite hate:
580 3GB SLI too slow
670/680 4GB (not downgrading to 2GB) SLI too expensive as a short term solution, especially with watercooling in mind. I'll never go back to air cooling.
And any single GPU solution doesn't provide me with any performance benefit over my current solution
CF is out of the question as you can easily guess.

Where does that leave me? I hate Nvidia as much as everyone else right now for their pricing. Now I have to wait until GTX780 or better Maxwell while games like Tomb Raider and Crysis 3 are out. This sucks!

 

[VulgarDisplay]

Vsync can mitigate AFR stuttering...but only as long as your fps are above the refresh rate. In my experience, Vsync can actually make things worse, it depends on the game. And remember, not everyone wants to use Vsync due to increased input lag. But I concur that PCper should test with Vsync on as well since that is relevant for at least a part of their readership.

That's why I really want someone to capture 120hz and see what it does for frametime graphs. I would imagine the increased speed would result in fewer chopped frames. Also, in the few games I get 120fps on I do not really notice any input lag on my monitor.