3DMark, Power, Sound and Conclusions

Although we don't put a lot of weight behind the results from 3DMark, I know that a lot of readers use this as a point of reference so I have included it.

3DMark has never been as kind to the Kepler architecture as it has to AMD's Tahiti and Hawaii parts, and that shows here once again. The Radeon R9 295X2 is 16% faster than the GeForce GTX Titan Z and the GTX 780 Ti cards in SLI is 2% faster than the R9 295X2.

Power Consumption and Sound Levels

With performance testing done it's time to look at the other aspects of the graphics card experience including power consumption and noise testing. When you are looking at these dual-GPU cards you are going to see a lot of power being utilized, and that's expected. How does the new GeForce GTX Titan Z stack up?

Very well in fact! The Radeon R9 295X2 still uses the most power, drawing 676 watts as a full system including our Sandy Bridge-E processor and platform. The pair of GeForce GTX 780 Ti cards in SLI uses 624 watts, 52 watts less than the 295X2. The GeForce GTX Titan Z was actually quiet efficient, with a total system power consumption 530 watts, 146 watts less than the 295X2. 

That is not a small difference. With a TDP of 500 watts, we knew that that Radeon R9 295X2 was on the outside looking in when it comes to graphics cards, but the Titan Z is clearly keeping things more in line with the expected values. The TDP of 375 watts is within specification of the PCIe standards: 75 watts from the PCIe bus and 150 watts from each of the two 8-pin power connectors. There are no strict power supply amperage requirements like we saw on the second page of our R9 295X2 review which gives users and system builders more flexibility in chassis and design.

Another area of strength for the GTX Titan Z is the lower sound level it creates compared to the R9 295X2 and the pair of GTX 780 Ti cards in SLI. At idle the Titan Z is almost silent, measuring nearly 5 dbA lower than the Radeon R9 295X2. The Radeon card is hindered not by the fan on the card but also by the fan on the radiators and the pump noise itself. 

Under a full gaming load, after at least 20 minutes of heavy use, the GTX Titan Z continues to be the more quiet option, about 3 dbA lower than the 295X2. 

To be fair, none of the solutions seen here are obnoxiously loud, but if you are very serious about sound levels the design of the NVIDIA cooler has clearly addressed the aspect more directly than the R9 295X2.

Pricing and Availability

NVIDIA has an uphill battle to climb with the GeForce GTX Titan Z when it comes to drawing in high end gamers and our performance results on the previous pages didn't get us off to a good start. The pricing hurts it even more. 

The GeForce GTX Titan Z stands out like a sore thumb in this list, coming it at two times the price of the Radeon R9 295X2 and more than two times the price of the pair of GeForce GTX 780 Ti cards that can run in SLI. Based purely on gaming performance, the Titan Z does not present a compelling option for enthusiasts that are concerned with performance per dollar metrics. If the Titan Z were priced at $1500 then there are all kinds of arguments you might be able to make for it over the R9 295X2, but we really can't do it today.

Final Thoughts

To be fair, we kind of knew how the GeForce GTX Titan Z was going to fare before going into this review, but it is something that needed to be tested and confirmed with hardware for our readers and fans. I think the results fall in line with those expectations but the full analysis involves some further discussion.

NVIDIA sees gaming as just a part of the Titan Z market and will point to the CUDA development as another outlet for this hardware. It's true too – software developers that need a lot of CUDA processing capability will not find a more compact way to write and produce projects. There are also system builders that will sell through all the allocation of Titan Z; some of the wealthier gamers that can spend $6k+ on a gaming machine may just run with the Titan Z due to familiarity with the brand or a desire to use other features like G-Sync.

But there are obviously users that have a lot more money to invest into a gaming machine and would consider cards in this price range, and for those, the GTX Titan Z just isn't a good choice. The AMD Radeon R9 295X2 is the better multi-GPU solution offering both better gaming performance (sometimes by a significant amount) and better value for your dollar. AMD has made a lot of progress with their multi-GPU frame pacing software updates and they are really showcased in the R9 295X2 that requires it. In fact, in some places the frame pacing performance of the Hawaii GPUs is better than that found in the GK110 parts on the Titan Z. This is something that NVIDIA is working to fix and take back that particular crown.

For users that wish to stay inside the NVIDIA ecosystem but want the performance of the graphics combinations tested here, the pair of GeForce GTX 780 Ti cards are much better option. They perform better than the GTX Titan Z in many instances and only cause a problem in a handful of places when running at 4K with that pesky 3GB memory capacity per GPU. And if that is a big concern, you could also jump up to a pair of GeForce GTX Titan Black cards (each with 6GB of memory) for a cool $1000 each. In fact, with a pair of 780 Ti or Titan Black cards you'll actually be able to support two 4K DisplayPort monitors as well, something you cannot do on the Titan Z.

In some ways, NVIDIA isn't being pressured to compete on pricing in any market. Even though the AMD Radeon R9 series of cards has had advantages in performance per dollar for a long time, and that AMD has had some impressive bundle programs at the same time, NVIDIA has maintained or gained marketshare, depending on who you ask. I bet that many inside the company simply don't see the need to change the direction or pricing of products without some pressure from the gamers that are actually spending the money.

NVIDIA was put in a tight spot with the GeForce GTX Titan Z. It was announced in late March, when the company was unsure what AMD's plans were for the dual-GPU market. When the R9 295X2 launched with a water cooler and 500 watt TDP at half the price, the world was flipped for NVIDIA. More than likely the company attempted to increase TDPs and get clock speeds high enough to compete with AMD's new addition to the market (maybe a cause for the delay) but it wasn't able to pull it off. Rather than eat crow and scrap the card, NVIDIA instead is pushing forward and targeting the Titan Z to very unique and specific markets. What I would love to see from NVIDIA is a version of the Titan Z that disables double precision floating point capability (the same difference between the GTX 780 Ti and the Titan Black) and allows them to sell it for $1500 without looking bad for investors and fans. Will they though? No clue.

There are several reasons to bash on the Titan Z including performance and price, but the design of the card is still impressive. The looks and style, in addition to the sound levels and power consumption, do give the card some strengths. It just isn't enough to recommend over the Radeon R9 295X2.

« PreviousNext »