NVIDIA GTX TITAN reviews

[Olivon]

889MHz limited, worst case turbo / 1006MHz, with additional airflow to maintain good thermal condition
Ambient : 25/26 Celsius
Benchmarks have been lauched after minutes of gaming in order to simulate real life conditions

3DMark Fire Strike : 889 / 1006 MHz
Unigne Heaven 4.0 : 876 / 954 MHz
Unigne Valley 1.0 : 876 / 954 MHz
Alan Wake 1920 Max : 876 / 993 MHz
Alan Wake 2560 Very High : 837 / 928 MHz
Alan Wake 2560 Max : 876 / 954 MHz
Anno 2070 : 837 / 902 MHz
Assassin's Creed 3 1920 MSAA4x : 876 / 967 MHz
Assassin's Creed 3 2560 FXAA HQ : 850 / 941 MHz
Assassin's Creed 3 2560 MSAA4x : 850 / 928 MHz
Batman Arkham City AA8x : 876 / 993 MHz
Batman Arkham City 2560 AA4x : 876 / 954 MHz
Battlefield 3 1920 MSAA 4x : 876 / 993 MHz
Battlefield 3 2560: 850 / 928 MHz
Battlefield 3 2560 MSAA 4x : 876 / 954 MHz
Civilization V 1920 AA8x : 850 / 928 MHz
Civilization V 2560 AA4x : 837 / 902 MHz
Civilization V 2560 AA8x : 863 / 954 MHz
Crysis 2 1920 Ultra : 876 / 993 MHz
Crysis 2 2560 Extreme : 876 / 954 MHz
Crysis 2 2560 Ultra : 863 / 967 MHz
DiRT Showdown 1920 Ultra : 876 / 954 MHz
DiRT Showdown 2560 Ultra AL OFF : 837 / 928 MHz
DiRT Showdown 2560 Ultra : 863 / 941 MHz
Far Cry 3 1920 AA4x : 876 / 954 MHz
Far Cry 3 2560 : 850 / 928 MHz
Hitman Absolution : 876 / 993 MHz
Max Payne 3 AA 4x : 876 / 954 MHz
Max Payne 3 2560 NoAA : 850 / 941 MHz
Sleeping Dogs : 837 / 902 MHz
The Witcher 2 Enhanced Edition : 876 / 954 MHz
Total War Shogun 2 1920 MSAA4x : 837 / 928 MHz
Total War Shogun 2 2560 MLAA : 837 / 928 MHz
Total War Shogun 2 2560 MSAA 4x : 837 / 901 MHz

http://www.hardware.fr/articles/887-...ost-tests.html

Frequencies are quite low compared to others reviews I have read, I believe that's the key.
12% represents the worst worst case (maybe none existing) and that's why the results are not so good.
If you read the review entirely, you'll see that Kepler/Titan seems to be quite an headache to review thanks to loony turbo add.

A complicated test ...
Some products are more complex to properly test than others. 75 fps at first try, 63 fps 3 minutes later, it quickly became apparent that the GeForce GTX Titan would complicate our lives and require special attention in the organization of tests. You will understand why ...

Here are 2 examples with Anno 2070 and Battlefield 3 with a quick test (1), test temperature stabilized after 5 minutes (2) and the same test but with the latter two 120mm fans positioned around the card (3) :
(1) (2) (3)
Anno 2070: 75 fps -> 63 fps -> 68 fps
Battlefield 3: 115 fps -> 107 fps -> 114 fps

http://www.hardware.fr/articles/887-...cole-test.html
http://www.hardware.fr/articles/887-...ergetique.html

[Russian]

My Review:http://www.brightsideofnews.com/news...z-edition.aspx

I'm also going to TRY to do 4K testing in Crysis 3. If not, then BF3 it is.

[Hornet331]

after looking at all this benchmarks it seems the titan is quite similar to the 8800 Ultra. While the avg. performance is higher then back then when you compare the 680 to the titan as opposed to the 8800 gtx to the 8800ultra, you have basically the same pricing scheme.

Wonder if the prices will come down n the next month or so. .but then again it would be pointless to buy, since new gpus would be basically right around the corner.

[Russian]

after looking at all this benchmarks it seems the titan is quite similar to the 8800 Ultra. While the avg. performance is higher then back then when you compare the 680 to the titan as opposed to the 8800 gtx to the 8800ultra, you have basically the same pricing scheme.

Wonder if the prices will come down n the next month or so. .but then again it would be pointless to buy, since new gpus would be basically right around the corner.

I'm almost 100% positive that prices won't come down any time this year on these unless AMD comes out with something faster this year.

[Stevethegreat]

after looking at all this benchmarks it seems the titan is quite similar to the 8800 Ultra. While the avg. performance is higher then back then when you compare the 680 to the titan as opposed to the 8800 gtx to the 8800ultra, you have basically the same pricing scheme.

Except the g80 was a true high end chip and gk104 ain't. So the gtx 680 is the equivalent of 8600 gts of back in the day and titan is closer to 8800 gtx/ultra.

[tajoh111]

Titan isn't even an ultra in my eyes aside from the cooler and price. Its isn't even fully enabled and the clocks are conservative. Ultra to me means all out to the point where it starts to lose its performance/watt. For a monster cards with a monster price, it is way too efficient and it loses some of its potential performance. Who at this price range cares about performance per watt?

Titan at the very least should be fully enabled for the price. Its will be slap in the nuts to earlier titan owners when Nvidia releases a Titan ultra on A2 stepping with all shaders enabled. Buying a 1000 dollar cards should atleast mean you have the fastest Single GPU card this year.

[Mungri]

The fully enabled GK110 would probably cost 10x as much. The yield rate on it is too low, even the Tesla card has one SMX unit disabled (... Well I think that's what I read somewhere).

[tajoh111]

The fully enabled GK110 would probably cost 10x as much. The yield rate on it is too low, even the Tesla card has one SMX unit disabled (... Well I think that's what I read somewhere).

The professional line doesn't necessarily have the most parts enabled or even come out first. Look at the Quadro 6000 series and Tesla 2090 series. The Top of the line Quadro 6000 only had 448 shaders out of 512 enabled, and the tesla 2090 came out 6 months after the fully enabled gtx 580 came out. With another stepping, I can easily see fully enabled Titans with higher clocks. For a first time run of gk110, Titan came out fantastic compared to fermi. A respin can only improve the yields and clocks as it did for Fermi.

As a result, I don't think after the respin, Nvidia is going to have trouble producing a fully enabled titan.

[matose]

Finally we have the review up, you can read it here: NVIDIA GTX Titan Review @ lab501.ro | English translation

[aussie-revhead]

No BF3?

[SKYMTL]

Sorry for the thread necro guys but I didn't want to make a new thread and be accused of "spamming".

NVIDIA GTX TITAN vs. SLI & Crossfire w/ Frame Time Analysis

[paulbagz]

Sorry for the thread necro guys but I didn't want to make a new thread and be accused of "spamming".

NVIDIA GTX TITAN vs. SLI & Crossfire w/ Frame Time Analysis

-PB

[mmorpgfocus]

Sorry for the thread necro guys but I didn't want to make a new thread and be accused of "spamming".

NVIDIA GTX TITAN vs. SLI & Crossfire w/ Frame Time Analysis

I appreciate the in depth frame time analysis on the TITAN - but why leave TITAN SLI off the table? My scaling experience with TITAN in SLI is disappointing.

Yes TITAN SLI is the fastest thing out there in most cases - but it does not seem to be scaling over a single card like GTX 680s do, and past SLI setups I have seem performed over the single card - do you have thoughts on this?

[solofly]

I'm guessing your CPU is holding back your Titans...

[cobra_kai]

I'm guessing your CPU is holding back your Titans...

With a 3930k at 4.5 GHz?

I'm glad to see the frame time information, hopefully it will become a key point of all future GPU reviews.

[solofly]

With a 3930k at 4.5 GHz?

All I'm seeing is 980X at 4.2 unless he's talking about something else...

[jimba86]

Finally we have the review up, you can read it here: NVIDIA GTX Titan Review @ lab501.ro | English translation

Nice review Matose! Great work!

[Mungri]

All I'm seeing is 980X at 4.2 unless he's talking about something else...

Main Test System

Processor: Intel i7 3930K @ 4.5GHz
Memory: Corsair Vengeance 32GB @ 1866MHz
Motherboard: ASUS P9X79 WS
Cooling: Corsair H80
SSD: 2x Corsair Performance Pro 256GB
Power Supply: Corsair AX1200
Monitor: Samsung 305T / 3x Acer 235Hz
OS: Windows 7 Ultimate N x64 SP1

Dont see how you can get much less bottleneck than that CPU.

[SKYMTL]

I appreciate the in depth frame time analysis on the TITAN - but why leave TITAN SLI off the table? My scaling experience with TITAN in SLI is disappointing.

Yes TITAN SLI is the fastest thing out there in most cases - but it does not seem to be scaling over a single card like GTX 680s do, and past SLI setups I have seem performed over the single card - do you have thoughts on this?

My thoughts on this are simple: I need another TITAN to conduct the testing, something I don't have right now.

[mmorpgfocus]

All I'm seeing is 980X at 4.2 unless he's talking about something else...

Does anyone have a link that shows where a 980x/990x EE @ 4.0Ghz would slow SLI on TITAN but not 680s?

From the reviews I have seen a 980x/990x at 4.0ghz++ is still very competitive. I would be cautious about buying a new CPU and MB for 2-3 frames/sec extra. I am seeing poor SLI scaling of a much greater scale.

Before the TITANS I had 680s and had no problem with them scaling very well when going from one card to two - why would my CPU bottleneck TITAN SLI but not GTX 680 SLI??

[Rollo]

Sorry for the thread necro guys but I didn't want to make a new thread and be accused of "spamming".

NVIDIA GTX TITAN vs. SLI & Crossfire w/ Frame Time Analysis

I dunno Sky. I read the review and I remain unconvinced on the high latency issue with CF. Just played a bunch of Crysis 3 at 25X16, 2X16X, all high but blur/shadows/water, and it was very smooth with 7970 CF.

If the gaming sucked on CF I'd put my GTX680s back in my main rig, but so far I'm enjoying the Crossfire. I'm sure the Titan is smoother than either the SLi or the CF, but I question how much of a factor is it?

I paid $670AMIR for two 7970 Double Ds, which have a nice HSF, and one is clocked at a GHz. Got 5 kick ass games with them, three of which were on my buy list. Just taking the cost of those three (FC3/Bioshock/Crysis 3) off the $670 gets me down around $500. acquisition cost on the cards.

Like I said, I'm guessing a Titan would be smoother, better. But is it $330 in my pocket, five games in my box better? Not in my eyes.

[BababooeyHTJ]

I'm surprised that you haven't noticed it. I haven't played Crysis 3 so I can't comment but there are a few games where I found microstutter to be very noticeable in crossfire.

Stalker was probably the worst offender. Serious Sam 3 was bad at times. Borderlands 2 needs a cap bad, that could be a driver issue though. If you force SSAA in almost any game you'll see microstutter.

It wasn't much of an issue for me with a 60hz display and a framerate cap but with a 120hz catlap that cap isn't an option.

I do agree that tahiti is a very good value for most users though.

[Rollo]

I'm surprised that you haven't noticed it. I haven't played Crysis 3 so I can't comment but there are a few games where I found microstutter to be very noticeable in crossfire.

Stalker was probably the worst offender. Serious Sam 3 was bad at times. Borderlands 2 needs a cap bad, that could be a driver issue though. If you force SSAA in almost any game you'll see microstutter.

It wasn't much of an issue for me with a 60hz display and a framerate cap but with a 120hz catlap that cap isn't an option.

I do agree that tahiti is a very good value for most users though.

I've noticed some bad gameplay when forcing AA on various games, and in FC3.

I haven't played Borderlands 2 and I haven't played Stalker in a long while. By definition Titan has to be better than any multi GPU, but NVIDIA has done very well with SLi MS, and AMD is at least working on it.

I may not be as sensitive to it as some also, I've been using 2-4 GPUs whenever possible since Voodoo 2 days. Also, you use 7950s which test higher at the MS than 7970s for some reason.

[jimba86]

Having seen the GPGPU results I only have to find out when will it be on sale.. Will need one at least.

care to link me to some?

[mmorpgfocus]

Does anyone have a link that shows where a 980x/990x EE @ 4.0Ghz would slow SLI on TITAN but not 680s?

From the reviews I have seen a 980x/990x at 4.0ghz++ is still very competitive. I would be cautious about buying a new CPU and MB for 2-3 frames/sec extra. I am seeing poor SLI scaling of a much greater scale.

Before the TITANS I had 680s and had no problem with them scaling very well when going from one card to two - why would my CPU bottleneck TITAN SLI but not GTX 680 SLI??

I knew it!!

The new Tomb Raider now runs in Ultimate mode in 5760x1200 in buttery smooth 60+ FPS~!

This driver update provides HUGE performance bump for TITAN SLI across the board, more than just the games listed.

That being said Tomb Raider Ultimate Mode - is possibly the most beautiful game I have ever seen in motion. If you have a system like mine you owe it to yourself to buy Tomb Raider just as a tech demo - wow!

Thank you NVIDIA for finally unlocking the power in these cards! BTW new Bioshock is a total letdown graphically IMHO - it looks like a 4 year old console game no matter how high I turn every option.

Best looking games after this patch - Tomb Raider 2013, Crysis 3, Sleeping Dogs, Witcher 2 Enhanced. DX11 Depth of Field is big deal when done right. So are the super high resolution texture packs (like Sleeping Dogs add-on). I can't decide whether to stare at Tomb Raider or actually play the game.

Anyone have a recommendation for a gorgeous next generation game I am missing? DX 11 used right is a big deal !!

I knew these TITANS had a ton more muscle!

All the TITAN reviews, ESPECIALLY SLI reviews should be re-done with 314.22 drivers, the cards just stomp anything else out there, especially with the EASY 1GHZ overclocks. I feel like I am already playing beyond PS4/XBOX Next quality.

So happy!! Please do NOT base your decision on buying these on the OLD reviews in this thread. They underestimate TITAN SLI hugely, in many cases 20+ FPS.