Titan X Announced

[Head1985]

Silverforce11:

GTX980-165w TDP
http://www.techpowerup.com/gpudb/2621/geforce-gtx-980.html
TITAN X with:
50% more SP
50% more TMU
50% more Rops
50%more memory bus@bandwidth
4 vs 12GB

There is no way TITAN X is only 35% faster than GTX980 IF it will have + 50% more TDP witch is 250W.
I presume TitanX will have 250w TDP and it will be 50% faster than GTX980 in 1080P and even more faster in 4k.

IF TITAN X will have only 225W TDP then yes it will be only +-35% faster than GTX980.
But there is no room for cutdown gm200 if full GM200 is only 35% faster.

I think Full GM200 will be 50% faster than GTX980 and cutdown(2560SP, 160TMU) will be 25% above GTX980.

 

[Cloudfire777]

Many pictures of the card here
http://www.legitreviews.com/hands-nvidia-geforce-gtx-titan-x-12gb-video-card_159519

It also looks like Nvidia is giving away GTX Titan X to many reviewers.

 

[2is]

The real question is, will they be releasing a 980Ti in the next month or two that outperforms it for half the cost?

 

[DarkKnightDude]

A card that's likely over 1K USD is hardly going to change the market.

 

[Stuka87]

The real question is, will they be releasing a 980Ti in the next month or two that outperforms it for half the cost?

980Ti would be this summer at the soonest. Most likely later than that. And like the 780Ti to Titan, the Titan X will probably be faster in many non-gaming related areas, but the 980Ti will probably be faster at gaming.

 

[Black Octagon]

980Ti would be this summer at the soonest. Most likely later than that. And like the 780Ti to Titan, the Titan X will probably be faster in many non-gaming related areas, but the 980Ti will probably be faster at gaming.

The question was whether there will be a 980 Ti AT ALL...

 

[SPBHM]

if it's 1.5K it's like half the launch price of a Titan Z, and maybe not much slower! it can be a bargain!

 

[Silverforce11]

Then why did you pick 1080p when you know the 980 is the fastest GPU on the market and thus making comparisons at 1080p is worthless?

I see now. You arbitrarily pick a resolution and stick with it. When talking about how fast AMD is, you stick with 4K. Because that is where AMD looks best. When comparing Nvidia products, you do 1080p. Because that is what shows the smallest gains between Nvidia chips. And since you're worried about the lack of recent driver optimizations for older products, lets go back a few months then!

http://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_980/26.html

3x 1080p: 770 - 65, 780 TI - 94......44.6% diference
1440p: 770 - 65, 780 TI - 94..........43.9% difference
1080p 770 - 66, 780 TI - 93...........40.9% difference

OMG Where is the massive drop off in 770 performance????? There isn't. Now do you see why I picked 1440p? Because 4k isn't adopted yet, 1080p isn't what GM200 buyers are going use, and the 770 was quite capable of 1440p gaming without massive performance degradation beyond typical pixel scaling.

How many times do you I have to prove you wrong?

You didn't prove jack.

My original statement was regarding similar perf/w for GM200 compared to GM204 if we go by history of NV's bigger chip.

You didn't even bother to discuss the point but went off on a crusade to prove me wrong based on 1440p for 770 vs 780ti. The point is AT LAUNCH (in case in your worthless crusade you forgot the entire point, go here: http://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_780_Ti/27.html) the 780ti was ~35% faster than the 770 overall. When you go up to 1440p, the gap increases. I even said as much.

How about you GO BACK to the point at hand: Re: GM200 vs GM204, launch time, launch performance comparisons.

Do you think efficiency will be the same? Yes? Then assuming 35% perf gains overall (I am not gonna cherry pick a particular resolution), it should consume ~35% more power.

If it consumes 50% more power, it should be 50% faster.

Unless you somehow feel NV will pull a miracle and improve perf/w with GM200, I stand by this logical extrapolation.

 

[Silverforce11]

Silverforce11:

GTX980-165w TDP
http://www.techpowerup.com/gpudb/2621/geforce-gtx-980.html
TITAN X with:
50% more SP
50% more TMU
50% more Rops
50%more memory bus@bandwidth
4 vs 12GB

There is no way TITAN X is only 35% faster than GTX980 IF it will have + 50% more TDP witch is 250W.
I presume TitanX will have 250w TDP and it will be 50% faster than GTX980 in 1080P and even more faster in 4k.

IF TITAN X will have only 225W TDP then yes it will be only +-35% faster than GTX980.
But there is no room for cutdown gm200 if full GM200 is only 35% faster.

I think Full GM200 will be 50% faster than GTX980 and cutdown(2560SP, 160TMU) will be 25% above GTX980.

I don't know the TDP of Titan X. I merely said I am expecting it to be ~35% faster while consuming ~35% more power, or similar to the first Titan & 780ti because the 980 is 160-180W, 35% more puts it in that range.

How fast it is will depend on how aggressive NV targets the power use, that much should be obvious. If NV is targeting Titan X at 250W TDP, then definitely it should be 50% faster than the 980 overall.

 

[Enigmoid]

As far a perf/W and scaling goes with regards to die size just look at GM107, GM206, and GM204. GM107 and GM204 have pretty much identical efficiency. GM206 trails.

I would expect GM200 to perform at a similar efficiency level.

You also need to take note of just how much 12 GB GDDR5 costs. At $15 per GB (high speed binned chips) that would be a premium of $120 solely in GDDR5 costs over an equivalent 4 GB card.

 

[Silverforce11]

As far a perf/W and scaling goes with regards to die size just look at GM107, GM206, and GM204. GM107 and GM204 have pretty much identical efficiency. GM206 trails.

I would expect GM200 to perform at a similar efficiency level.

You also need to take note of just how much 12 GB GDDR5 costs. At $15 per GB (high speed binned chips) that would be a premium of $120 solely in GDDR5 costs over an equivalent 4 GB card.

Exactly, GM200 should be as efficient as the other Maxwell unless there's some major issues or improvements to change the status quo.

But your point regarding the GDDR5, that adds to power usage as well, 12GB is no small amount.

 

[madwolfa]

if it's 1.5K it's like half the launch price of a Titan Z, and maybe not much slower! it can be a bargain!

Interesting times we live in... when 1.5K card is considered a bargain.

 

[Cloudfire777]

Im only buying this one for $999.
GTX 980 SLI for $1200 will be a better buy if GTX Titan X cost $1.3K

 

[parvadomus]

GK104: 3.54billion Transistors. 1536 cores. 23046875/core
GM204: 5.2billion Transistors. 2048 cores. 25390625/core

GK110: 7.1billion Transistors. 2880 cores. 24652777/core
GM200: 8.0billion Transistors. x cores

80000000000/x = 26600000 (roughly due to higher density with Maxwell like shown with GM204)
80000000000 = 26600000x
x = 80000000000/26600000
GM200 Transistor Count: 3007
Nearest SMM: 23, Rounding up to 24

GM200: 24SMM, 3072 Cores. 384bit, 12GB VRAM
Thats my estimation :thumbsup:

GM204: 5.2billion Transistors. 398mm2
GM200: 8.0billion Transistors. -> 612mm2 <-

This if density holds. I expect it to be 45-50% faster than 980.
I also think the 390X will be more than 50% faster than 290X (IF the rumour about the 640Gb/s bandwidth is real), they might be close, specially at 4K.

 

[antihelten]

Take GTX Titan and GTX 680 from the 1440p.
+37%.

So 35% is the minimum we will see I think from GTX Titan X over 980 most likely.
With 3072 cores like I calculated above, it should be 50% faster. But Titan X is probably not as agressive clocked, so 40% ish above GTX 980 is my guess

The important thing to note though, is that GTX Titan had 75% more shaders than GTX 680, which resulted in 37% more performance.

Titan X is rumored to have 50% more shaders than 980. As such, assuming the same scaling, that would equate to about 25% better performance. So 35% isn't really the theoretical minimum.

Now obviously the poor scaling of GTX Titan, is largely a result of the lower frequency, and if we normalize for this we get that Titan X should be about 41% faster than 980. But this of course assumes that GM200 can be clocked as high as GM204.

 

[Deders]

The important thing to note though, is that GTX Titan had 75% more shaders than GTX 680, which resulted in 37% more performance.

Titan X is rumored to have 50% more shaders than 980. As such, assuming the same scaling, that would equate to about 25% better performance. So 35% isn't really the theoretical minimum.

Now obviously the poor scaling of GTX Titan, is largely a result of the lower frequency, and if we normalize for this we get that Titan X should be about 41% faster than 980. But this of course assumes that GM200 can be clocked as high as GM204.

There is also the CPU limit factor, there are games that have upwards of a 60% increase in performance (680 to titan) but this usually occurs at 2560x1440 or 1600, not so much at 1080p.

 

[antihelten]

There is also the CPU limit factor, there are games that have upwards of a 60% increase in performance (680 to titan) but this usually occurs at 2560x1440 or 1600, not so much at 1080p.

The 37% performance increase is taken from TPUs latest 1440p numbers, so the CPU bottleneck issue can probably be more or less ruled out. That's not to say that there won't be games where the Titan is significantly more than 37% faster than the 680, but likewise there will also be games where it is less than 37% faster.

 

[soccerballtux]

it is going to be an overpriced 980 x2 with 4gb of real ram + 2gb of slow ram.

nothing new..

does the 980 have 4+2???

 

[AdamK47]

Im only buying this one for $999.
GTX 980 SLI for $1200 will be a better buy if GTX Titan X cost $1.3K

You could probably find some slightly used GTX 980s once the Titan X is released.

 

[cmdrdredd]

The rumor I saw said ~$1250, but retail pricing will probably be higher at first like the other Titan cards if I remember correctly.

 

[Deders]

The 37% performance increase is taken from TPUs latest 1440p numbers, so the CPU bottleneck issue can probably be more or less ruled out. That's not to say that there won't be games where the Titan is significantly more than 37% faster than the 680, but likewise there will also be games where it is less than 37% faster.

Would be interested to see a link.

For reference I've got an Excel table of all the results in the Anandtech GPU benchmark page and the average % increase for the games is 46%.

I would still think that some games are going to be able to use 1 CPU core's worth of processing in 1 main thread, that would limit what the GPU could do.

 

[antihelten]

Would be interested to see a link.

For reference I've got an Excel table of all the results in the Anandtech GPU benchmark page and the average % increase for the games is 46%.

I would still think that some games are going to be able to use 1 CPU core's worth of processing in 1 main thread, that would limit what the GPU could do.

The numbers were posted earlier in this thread (link), and came from this review

 

[Deders]

The numbers were posted earlier in this thread (link), and came from this review

ahh cool. If you are interested Anandtech's numbers are slightly different.

The games are:
Metro Last Light
Company of heroes 2
Bioshock Infinite
Crysis 3
Crysis Warhead
Rome 2 Total War
Thief
Grid 4

At 1080p the average increase is 39%
At 1440p the average increase is 48%

For minimum framerates with CoH2 Crysis Warhead and Thief are:

1080p 46% increase
1440p 65% increase.

As we know the minimums are an important factor.

 

[Enigmoid]

The important thing to note though, is that GTX Titan had 75% more shaders than GTX 680, which resulted in 37% more performance.

Titan X is rumored to have 50% more shaders than 980. As such, assuming the same scaling, that would equate to about 25% better performance. So 35% isn't really the theoretical minimum.

Now obviously the poor scaling of GTX Titan, is largely a result of the lower frequency, and if we normalize for this we get that Titan X should be about 41% faster than 980. But this of course assumes that GM200 can be clocked as high as GM204.

Couple things.

-Lower clocks
-much of the chip scaled less. Bandwidth was 50% more, for example, the number of gpc only increased from 4 to 5 (25%)

 

[Deders]

Kitguru have some more info apparently:
http://www.kitguru.net/components/g...itan-x-12gb-what-you-need-to-know-and-expect/

Not sure why they are putting 12GB on a consumer card without Double precision.