What's new
  • Please do not post any links until you have 3 posts as they will automatically be rejected to prevent SPAM. Many words are also blocked due to being used in SPAM Messages. Thanks!

The NVIDIA TITAN Z Performance Review

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,841
Location
Montreal
NVIDIA’s TITAN Z is an odd GPU. On one hand it is the most powerful graphics card the company has ever launched and yet the amount of media coverage for it has been surprisingly limited. Reviews have been few and far between and despite a very public announcement, its eventual coming out party was muted at best. Considering the well-oiled PR machine NVIDIA has working behind every one of their new product introductions, the TITAN Z’s somewhat flaccid release was atypical. And yet from our perspective, after some last minute changes it went exactly the way they wanted it to.

To understand the TITAN-Z’s troubled inception and ultimate relegation to a footnote in NVIDIA’s product line, we have to go back to before NVIDIA's chief Jen Hsun Huang actually announced it back at the end of March. NVIDIA had already been working on their dual GPU follow-up to the extremely popular GTX 690 for quite some time. It included a pair of toned-down GK110 cores topped off by an air cooled heatsink and remained true to Kepler’s “efficient, quiet performance” mantra. They took a relatively conservative approach with the knowledge that AMD had no way to roll out a double-dipped R9 290X without either severely limiting its clock speeds or boosting fan speeds to laughably high levels.

Right before that fateful March 25th keynote, the first confirmations of AMD’s so-called R9 295X2 began to trickle out and it wasn’t good news for NVIDIA. Instead of going the air-cooled route, AMD’s engineers and product specialists threw their weight behind what they hoped was the market’s broader acceptance of AIO water cooling units. More importantly, by going the liquid cooled route, the R9 295X2 was able to run at high clock speeds without throttling, had a good amount of overclocking headroom and remained blissfully quiet. With that knowledge, plans were changed.

Instead of announcing a card which had a cost in-line with a pair of GTX 780 Ti’s the so-called TITAN-Z ended up being marketed to “CUDA developers” and had a suitably stratospheric price of $3000. What followed were delays as NVIDIA tried to dial in clock speeds that could speed up gaming scenarios so it could be cross marketed to a broader user base. Not to mention, higher performance would go a long way towards somewhat justifying its extreme cost.

CHART.jpg

The TITAN-Z was ultimately released with two full-enabled GK110 cores that operated at frequencies which are drastically lower than the GTX 780 Ti and TITAN Black. It does however board a mammoth 12GB of GDDR5 memory operating at 7Gbps so there should be no worries about bandwidth bottlenecks.

NVIDIA decided to go with lower clock speeds for the TITAN Z in order to hit a preset set of thermal and power consumption targets. Without those, the card would either hit temperatures that are beyond an air cooler’s capability to contain or throttle down to frequencies that are lower than its stated Base clock. Neither of those results was optimal so the core speed was reigned in. Make no mistake about it though; the two GK110s sitting on the Z’s PCB are carefully selected so they maximize performance at lower TDP levels. High leakage cores need not apply.

TITAN-Z-1.jpg

While it may have a triple slot height, the TITAN Z’s defining characteristic is its adherence to an air cooled design. Despite the use a two large vapor chambers, this speaks to the inherent efficiency of NVIDIA’s architecture and does tend to broaden the TITAN Z’s compatibility in some instances when compared against the R9 295X2.

The design language remains faithful to other high end Kepler-based cards with a glowing GeForce logo, windows that look down onto the internal heatsinks and a TITAN Z logo etched into the aluminum heatsink shroud. We should also mention that it uses a pair of 8-pin power connectors and requires a good 850W power supply.

TITAN-Z-2.jpg

The card’s back has a black anodized heatsink that covers half of the included GDDR5 memory modules and also helps with the cooling of other PCB-mounted components. What you can’t see here is a state-of-the-art 12 phase PWM design which features dynamic load balancing so both cores can expand their performance envelopes if needed while remaining within the same strict TDP constraints.

Before we get any further into this review, some transparency is in order. While the rumor that NVIDIA actively worked on preventing TITAN Z cards from getting into the heads of reviewers may be nothing more than hearsay, they had no active hand in this review. We borrowed this card from a local developer who will be using a pair of them for high level CUDA debugging and development purposes. This review may be late but the tardiness isn’t for lack of us trying to get loaner samples from board partners and NVIDIA directly. Due to the limited time we had, comprehensive GPU Compute benchmarks weren’t possible so we approached this one solely from a gaming standpoint which, according to NVIDIA’s launch video, is the TITAN Z’s primary task.
 
Last edited:

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,841
Location
Montreal
Clock Speed Stability & Thermal Imaging

Clock Speed Stability & Thermal Imaging


One of the areas we had to delve into was how the TITAN Z handled itself when faced with continual high workloads. Several other hot-running air cooled cards like AMD’s R9 290X tend to downclock after 10 minutes or less of gaming and this card is a prime candidate for the same thing. Remember, AMD needed a water cooler to bring their R9 295X2 within their required specifications while NVIDIA has taken a more traditional approach of optimizing TDP to achieve their goals.

TITAN-Z-70.jpg

Our temperature results show the TITAN Z hitting a maximum of 83°C after which it levels out without any more upwards movement as fan speeds increase to properly limit additional heat buildup. While the R9 295X2 was able to achieve much lower temperatures, we can’t forget that it does use an all in one water cooling setup which adds complexity and quite a few more points of failure (however remote they may be). This is pretty impressive considering the not-so-insignificant TDP of two GK110 cores.

TITAN-Z-69.jpg

In order to get thermals properly managed, there are a few frequency gymnastics going on behind the scenes but nothing that would adversely affect performance. When overhead allowed, the TITAN Z offered short sprints to an incredible 1005MHz while normalized clocks evened out between 980MHz and 993MHz. Both are quite a bit higher than NVIDIA’s stated Boost speed of 876MHz. We’re sure there will be instances when this card will offer lower clocks when TDP-limited situations arise but throughout testing the card didn’t exhibit any adverse behavior that would have noticeably impacted performance.

TITAN-Z-68.jpg

With stable temperatures and frequencies framerates remained extremely constant throughout the benchmarking runs. Even after an hour or more, we didn’t see any large variances.

All of these results point towards a well-balanced card that can easily level its various aspects and hit a given performance bracket. As you will see in the next few pages, the TITAN Z is capable of some incredible feats but it has been rolled into a package which costs an absolute fortune.

TITAN-Z-3.jpg
TITAN-Z-4.jpg

TITAN-Z-5.jpg

There is however one caveat here. While the TITAN Z is able to cool off the cores quite well, all the heat they produce has to go somewhere and in this case it mostly gets dumped directly into your enclosure’s interior. In addition, the backplate gets ludicrously hot so you’ll need plenty of airflow to keep things in check. This means managing thermals of other components in form factor builds will be exceedingly difficult with a TITAN Z installed.
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,841
Location
Montreal
Acoustics & Power Consumption

Acoustical Testing


What you see below are the baseline idle dB(A) results attained for a relatively quiet open-case system (specs are in the Methodology section) sans GPU along with the attained results for each individual card in idle and load scenarios. The meter we use has been calibrated and is placed at seated ear-level exactly 14” away from the GPU’s fan. For the load scenarios, Hitman Absolution is used in order to generate a constant load on the GPU(s) over the course of 15 minutes.

TITAN-Z-66.jpg

The last thing we wanted to see what NVIDIA using extremely high fan speeds to bring their flagship card under heel and somehow, they didn’t. The TITAN Z is actually one of the quietest cards we’ve ever tested when under in-game load, handily beating the R9 295X2 and all other alternatives. We can’t stress how large of an achievement this is.


System Power Consumption


For this test we hooked up our power supply to a UPM power meter that will log the power consumption of the whole system twice every second. In order to stress the GPU as much as possible we used 15 minutes of Hitman Absolution while letting the card sit at a stable Windows desktop for 15 minutes to determine the peak idle power consumption.

Please note that after extensive testing, we have found that simply plugging in a power meter to a wall outlet or UPS will NOT give you accurate power consumption numbers due to slight changes in the input voltage. Thus we use a Tripp-Lite 1800W line conditioner between the 120V outlet and the power meter.

TITAN-Z-67.jpg

Make no mistake about it: the TITAN Z is one hell of a power hungry card but it can’t hold a light to the massive figures posted by AMD’s R9 295X2 and dual R9 290X solutions. It does however consume significantly more than a GTX 780 SLI setup.
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,841
Location
Montreal
Test System & Setup

Main Test System

Processor: Intel i7 4930K @ 4.7GHz
Memory: G.Skill Trident 16GB @ 2133MHz 10-10-12-29-1T
Motherboard: ASUS P9X79-E WS
Cooling: NH-U14S
SSD: 2x Kingston HyperX 3K 480GB
Power Supply: Corsair AX1200
Monitor: Dell 2412M (1440P) / ASUS PQ321Q (4K)
OS: Windows 8.1 Professional


Drivers:
AMD 14.6 Beta
NVIDIA 337.88 Beta

*Please note that a previous version of this page erroneously listed AMD's 14.4 driver. All testing on AMD hardware was conducted with the 14.6 beta driver.

*Notes:

- All games tested have been patched to their latest version

- The OS has had all the latest hotfixes and updates installed

- All scores you see are the averages after 2 benchmark runs

All IQ settings were adjusted in-game and all GPU control panels were set to use application settings


The Methodology of Frame Testing, Distilled


How do you benchmark an onscreen experience? That question has plagued graphics card evaluations for years. While framerates give an accurate measurement of raw performance , there’s a lot more going on behind the scenes which a basic frames per second measurement by FRAPS or a similar application just can’t show. A good example of this is how “stuttering” can occur but may not be picked up by typical min/max/average benchmarking.

Before we go on, a basic explanation of FRAPS’ frames per second benchmarking method is important. FRAPS determines FPS rates by simply logging and averaging out how many frames are rendered within a single second. The average framerate measurement is taken by dividing the total number of rendered frames by the length of the benchmark being run. For example, if a 60 second sequence is used and the GPU renders 4,000 frames over the course of that time, the average result will be 66.67FPS. The minimum and maximum values meanwhile are simply two data points representing single second intervals which took the longest and shortest amount of time to render. Combining these values together gives an accurate, albeit very narrow snapshot of graphics subsystem performance and it isn’t quite representative of what you’ll actually see on the screen.

FCAT on the other hand has the capability to log onscreen average framerates for each second of a benchmark sequence, resulting in the “FPS over time” graphs. It does this by simply logging the reported framerate result once per second. However, in real world applications, a single second is actually a long period of time, meaning the human eye can pick up on onscreen deviations much quicker than this method can actually report them. So what can actually happens within each second of time? A whole lot since each second of gameplay time can consist of dozens or even hundreds (if your graphics card is fast enough) of frames. This brings us to frame time testing and where the Frame Time Analysis Tool gets factored into this equation.

Frame times simply represent the length of time (in milliseconds) it takes the graphics card to render and display each individual frame. Measuring the interval between frames allows for a detailed millisecond by millisecond evaluation of frame times rather than averaging things out over a full second. The larger the amount of time, the longer each frame takes to render. This detailed reporting just isn’t possible with standard benchmark methods.

We are now using FCAT for ALL benchmark results, other than 4K.
 
Last edited:

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,841
Location
Montreal
1440P: Assassin’s Creed IV: Black Flag / Battlefield 4

Assassin’s Creed IV: Black Flag


<iframe width="640" height="360" src="//www.youtube.com/embed/YFgGnFoRAXU?rel=0" frameborder="0" allowfullscreen></iframe>​

The fourth iteration of the Assassin’s Creed franchise is the first to make extensive use of DX11 graphics technology. In this benchmark sequence, we proceed through a run-through of the Havana area which features plenty of NPCs, distant views and high levels of detail.


2560 x 1440

TITAN-Z-48.jpg

TITAN-Z-30.jpg


Battlefield 4


<iframe width="640" height="360" src="//www.youtube.com/embed/y9nwvLwltqk?rel=0" frameborder="0" allowfullscreen></iframe>​

Amidst its teething problems since its release, BF4 has been a bone of contention among gamers. In this sequence, we use the Singapore level which combines three of the game’s major elements: a decayed urban environment, a water-inundated city and finally a forested area. We chose not to include multiplayer results simply due to their randomness injecting results that make apples to apples comparisons impossible.

2560 x 1440

TITAN-Z-49.jpg

TITAN-Z-31.jpg
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,841
Location
Montreal
1440P: Call of Duty: Ghosts / Far Cry 3

Call of Duty: Ghosts


<iframe width="640" height="360" src="//www.youtube.com/embed/gzIdSAktyf4?rel=0" frameborder="0" allowfullscreen></iframe>​

The latest Call of Duty game may have been ridiculed for its lackluster gameplay but it remains one of the best-looking games out there. Unfortunately due to mid-level loads, getting a “clean” runthrough without random slowdowns is nearly impossible, even with a dual SSD system like ours. Hence why you should ignore any massive framerate dips as they are anomalies of poor loading optimizations. For this benchmark we used the first sequence of the 5th Chapter entitled Homecoming as every event is scripted so runthroughs will be nearly identical.

2560 x 1440

TITAN-Z-50.jpg

TITAN-Z-32.jpg


Far Cry 3


<iframe width="560" height="315" src="http://www.youtube.com/embed/mGvwWHzn6qY?rel=0" frameborder="0" allowfullscreen></iframe>​

One of the best looking games in recent memory, Far Cry 3 has the capability to bring even the fastest systems to their knees. Its use of nearly the entire repertoire of DX11’s tricks may come at a high cost but with the proper GPU, the visuals will be absolutely stunning.

To benchmark Far Cry 3, we used a typical run-through which includes several in-game environments such as a jungle, in-vehicle and in-town areas.



2560 x 1440

TITAN-Z-52.jpg

TITAN-Z-34.jpg
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,841
Location
Montreal
1440P: Hitman Absolution / Metro: Last Light

Hitman Absolution


<iframe width="560" height="315" src="http://www.youtube.com/embed/8UXx0gbkUl0?rel=0" frameborder="0" allowfullscreen></iframe>​

Hitman is arguably one of the most popular FPS (first person “sneaking”) franchises around and this time around Agent 47 goes rogue so mayhem soon follows. Our benchmark sequence is taken from the beginning of the Terminus level which is one of the most graphically-intensive areas of the entire game. It features an environment virtually bathed in rain and puddles making for numerous reflections and complicated lighting effects.


2560 x 1440

TITAN-Z-53.jpg

TITAN-Z-35.jpg


Metro: Last Light


<iframe width="640" height="360" src="http://www.youtube.com/embed/40Rip9szroU" frameborder="0" allowfullscreen></iframe>​

The latest iteration of the Metro franchise once again sets high water marks for graphics fidelity and making use of advanced DX11 features. In this benchmark, we use the Torchling level which represents a scene you’ll be intimately familiar with after playing this game: a murky sewer underground.


2560 x 1440

TITAN-Z-54.jpg

TITAN-Z-36.jpg
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,841
Location
Montreal
1440P: Thief / Tomb Raider

Thief


<iframe width="640" height="360" src="//www.youtube.com/embed/p-a-8mr00rY?rel=0" frameborder="0" allowfullscreen></iframe>​

When it was released, Thief was arguably one of the most anticipated games around. From a graphics standpoint, it is something of a tour de force. Not only does it look great but the engine combines several advanced lighting and shading techniques that are among the best we’ve seen. One of the most demanding sections is actually within the first level where you must scale rooftops amidst a thunder storm. The rain and lightning flashes add to the graphics load, though the lightning flashes occur randomly so you will likely see interspersed dips in the charts below due to this.


2560 x 1440

TITAN-Z-55.jpg

TITAN-Z-37.jpg



Tomb Raider


<iframe width="560" height="315" src="http://www.youtube.com/embed/okFRgtsbPWE" frameborder="0" allowfullscreen></iframe>​

Tomb Raider is one of the most iconic brands in PC gaming and this iteration brings Lara Croft back in DX11 glory. This happens to not only be one of the most popular games around but it is also one of the best looking by using the entire bag of DX11 tricks to properly deliver an atmospheric gaming experience.

In this run-through we use a section of the Shanty Town level. While it may not represent the caves, tunnels and tombs of many other levels, it is one of the most demanding sequences in Tomb Raider.


2560 x 1440

TITAN-Z-56.jpg

TITAN-Z-38.jpg
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,841
Location
Montreal
4K: Assassin’s Creed IV: Black Flag / Battlefield 4

Assassin’s Creed IV: Black Flag


3840 x 2160

TITAN-Z-57.jpg

TITAN-Z-39.jpg


Battlefield 4


3840 x 2160

TITAN-Z-58.jpg

TITAN-Z-40.jpg
 

SKYMTL

HardwareCanuck Review Editor
Staff member
Joined
Feb 26, 2007
Messages
12,841
Location
Montreal
4K: Call of Duty: Ghosts / Far Cry 3

Call of Duty: Ghosts


3840 x 2160

TITAN-Z-59.jpg

TITAN-Z-41.jpg


Far Cry 3


3840 x 2160

TITAN-Z-61.jpg

TITAN-Z-43.jpg
 
Top