Following the launch of its consumer cards based on the Maxwell Architecture, NVIDIA updated its Titan line-up of Workstation grade Gaming GPUs with its new GM200 Maxwell chip in the form of the new GTX Titan X. This GPU really takes things to the extreme with its configuration. We find out if it can output the performance to do justice to the Titan moniker.
[section label=”Specification” anchor=”Specification”]Specification of the Titan X
The card has a ROP count of 96 and comes with a whopping 12GB of Hynix GDDR5 VRAM. Core clock is set to 1002 MHz and boost clock is 1089MHz. The VRAM is clocked at an effective 7012MHz. The VRAM should be able to overclock decently as should the chip itself if earlier experience with Maxwell parts is anything to go by.
The rear panel of the card sports a DVI-D port, a HDMI port and three DisplayPort slots. Of these only four slots can be used at a time in a multi-monitor setup. There are two SLI connectors as well.
The launch price for the GTX Titan X is ₹85,099 including taxes.
[section label=”Build Quality” anchor=”bq”]Build Quality and Packaging
Our review sample came in a special Black Package that showcased the card much like the one we had seen with the GTX 980. Though this box has no accessories, retail versions should ship with a Driver CD/DVD, Manual and Power Cable adapters.
The card comes in an all-black casing adapting the heatsink and cooler design adapted from NVIDIA’s existing cards. There is no heat sheath on the back however, leaving the PCB exposed to the air. In terms of bling, there is the green backlit GEFORCE GTX logo on the side that can be controlled through the GEForce Experience application. The choice of Black colour means that the body tends to get really hot when the card is under load. Given the trend with the Titan series, we are unlikely to see any custom cooler designs from the manufacturers, which means the flaws with the reference design will be seen in all cards.
The cooling vents at the back are cut in the triangular honeycomb pattern, which makes for an interesting looking design. The GDDR5 RAM uses Hynix Chips, which are known for being quality overclockers. The card has dimensions of 27cmx11.1cmx4cm, which can make for a tight fit in some small cases.
[section label=”Testing Methodology” anchor=”TM”]Testing Methodology
We tested the GTX Titan X not only for the average Frames per Second(FPS) but also for the 99th Percentile Frame time which tells us about the performance of the GPU within the second. Within the Second testing is useful to understand micro-stutter which can render a game unplayable despite FPS being high. Fraps 3.5.99 allowed us to calculate both.
Since the card is a Top End GPU, we decided to use highest possible settings in our benchmark games and compare with other Top End GPUs. We tried to disable CPU dependent settings or minimise their impact where possible. VSync and frame buffering were disabled for testing. All tests were run at 1920×1080 on a single monitor configuration.
Test System:
CPU: Intel i7-4790k @4GHZ (4.4GHz Turbo)
Motherboard: Asus Maximus VII Hero
RAM: 2x8GB G.Skill TridentX 2666MHz 12-13-13-36
PSU: Corsair TX650 650W
SSD: Samsung 840 Evo 1TB (For OS and Benchmarks)
HDD: Seagate Barracuda 7200.12 1TB, Seagate Barracuda 2TB, 2xWesten Digital Red 3TB (Storage)
OS: Windows 8.1 x64
NVIDIA Driver: 347.09 for G1 Gaming GTX 980
NVIDIA Driver: 347.88 for GTX Titan X
While this system may not look like a purpose built test rig, we decided to use a normal usage PC so as to better reflect real world scores of the card. The games were also tested with a few applications like Antivirus, Browser, VOIP tool and others running in the background to get a realistic usage scenario. All HDDs were thoroughly defragmented prior to usage and SSDs were optimised for maximum performance. Due to lack of equipment, we were unable to conduct acoustic and power testing. Since this is a new test rig, we are lacking a comparative benchmark database for it, as it will be built up as we get more cards for testing. For now, we will compare the Titan X against our best performing GPU to date, the Gigabyte G1 Gaming GTX 980.
[section label=”3DMark” anchor=”3DM”]3DMark Firestrike
3Dmark is an artificial Benchmarking tool whose Firestrike Test is very thorough on DirectX 11cards powering High-end PCs. The full test run for Firestrike includes 2 GPU only tests, a CPU dependent Physics Test and a Combined Graphics and Physics Test. The Tool is also useful for stress testing a GPU when run on loop.
Given that we are looking for the Performance of the Card itself, one should look at the Graphics score and the FPS for Graphics tests 1 and 2. We also tested the card for the Firestrike Ultra setting that tests for 4K performance. The Physics and Combined tests are CPU dependent, which is the limiting factor of our test rig.
[section label=”Game Benchmarks” anchor=”GB”]Game Benchmarks
Battlefield 4
Battlefield 4 uses the Frostbite 3 engine to push the visual processing boundaries of current hardware. Since Mantle is only for AMD cards, we ran the DX11 version with the highest possible settings. The game offers no benchmark tool, so we used areas from the first single player campaign mission for the FRAPS run.
As expected, the Titan X easily beats the G1 Gaming GTX 980. The margin isn’t that high though.
The FPS spikes a lot for the duration of the run and the frametime graph is also densely packed with a bit of periodic spiking that might be seen as microstutter.
Company of Heroes 2
Relic’s Company of Heroes 2 is a tough nut to crack for quite a few GPUs, though it’s dependent on CPUs to a great degree as well. We used highest possible Settings with Physics turned off and Low AA for the short 45s benchmark run the game offers.
This is the first time a GPU has managed to tame this game in our testing. The Titan X maintains a similar performance margin over the G1 Gaming GTX 980 here too.
The frametimes are quite tight with the stray spikes. The FPS also remained easily playable and visually provided quite a smooth run.
Far Cry 4
Far Cry 4 adds even more visual effects to the DUNIA engine of Far Cry 3 and it has some NVIDIA specific effects like God Rays and realistic Fur as well. This makes for an absolute visual treat when all settings are cranked up to the max like in our 2 minute benchmark run.
The Titan X has a marginal advantage in this game over the G1 Gaming GTX 980 that is curious given the higher Minimum FPS value.
There are a lot of spikes and the stuttering was noticeable as the game warped in objects in the distance. The FPS was fluctuating quite a bit. Seems Far Cry 4 needs more optimisation, rather than the GPU needing better drivers if it performs so poorly even on its showcase GPUs.
GRID Autosport
GRID Autosport is the next game in the GRID series from Codemasters that has specific effects reserved for Intel GPUs, which makes it great to judge the performance of NVIDIA and AMD cards as it provides an even playing field. We used the highest settings and used game’s benchmark in a 2 minute run.
Once more, the Titan X takes the crown over the G1 Gaming GTX 980. However, the margin is a bit lower than what we saw earlier.
The graph is tightly packed in the frametimes with no major spikes. Even the fps looks quite stable.
Metro Last Light Redux
4A Games have really cranked up the eye candy in Metro Last Light, and the Redux version comes with even more improvements to visuals. The game looks beautiful in its cramped corridors as well as its open outdoor environments and cranking up the settings can easily bring a GPU to its knees in the menu screen itself. We used the highest settings available except for SSAA that was set at 2x to accommodate a few other cards we tested. The game has no inbuilt benchmark, so we chose a particular area to conduct out tests run in for a time of 2 minutes.
The Titan X easily beats the G1 Gaming GTX 980 by a handsome margin across all criteria.
The game stays consistent for the most part and the spikes are not significant. There was no noticeable stuttering where the spikes occurred. The FPS stays pretty consistent too.
Middle-earth: Shadow of Mordor
Monolith have used some fancy effects to add some visual appeal to the dreary land of Mordor. Shadow of Mordor also uses PhysX for particle effects. The rain though doesn’t quite look natural and the game requires up to 6GB VRAM for using its Ultra HD textures. Curiously, it scales the game based on the screen’s natural resolution instead of offering resolution options. We used the highest available settings for the inbuilt benchmark, which runs for less than the standard 2 minutes of our other benchmarks.
The G1 Gaming GTX 980 was much closer to the Titan X in this game, even with the Ultra textures on.
Other than the large spike at the start and end, which were caused by activating and ending the benchmarking in FRAPS, there are no noticeable spikes and the game did not stutter a bit. The FPS curve is relatively smooth.
Ryse: Son of Rome
Ryse pushes the newest iteration of the CryEngine to its limits as claimed by Crytek on the PC. It doesn’t have a benchmark mode, so we ran a fixed scenario for 2 minutes. The settings were set to the highest possible. This game scales the game as per the resolution of the screen instead of offering resolution options.
There was no comparative benchmark for this game since we used different settings for the cards. These are the values we got by turning everything to the maximum in this game.
There are a few spikes but the 99th percentile graph, which is pretty tight otherwise. The FPS graph remains pretty constant giving a good idea of the card’s limits.
Thief (2014)
Thief (2014) is a game that boasts support for AMD’s Mantle and its TrueAudio tech. It is a graphical showcase and has a built in benchmark, though the run time is less than our standard 2 minutes. We set all settings to the highest possible.
The Titan X almost breaks the 100 fps mark in this game soundly beating the G1 Gaming GTX 980.
The Frametime graph is tight for the most part. The FPS graph doesn’t show huge fluctuations either.
Tomb Raider (2013)
Tomb Raider (2013) introduced us to a new Lara with fabled TressFX hair that behaves a lot more realistically than the pre-rendered mop we were used to. However, the card does not seem to cope well with this technology and we had to drop it in our test run. We used highest possible settings with TressFX on. The Test run was shorter since the benchmark tool offered by the game runs for less than out 120s target time.
The Titan X stamps it authority all over this contest, despite the challenge posed by the TressFX technology easily exceeding the G1 Gaming GTX 980.
The frametime graph is pretty densely packed with a few spikes here and there. The FPS graph is also very smooth.
Total War: Rome II
Total War: Rome II is another CPU heavy game that offers significant visual goodness. It offers a benchmark tool that focuses more on GPU power though and that is what we used for our test run of 120s. We set the game to Extreme and Unit Size to Small so as to reduce the impact of the CPU. Unlimited Video Memory was off so the game could scale down visual settings if it reached a bottleneck.
This CPU heavy game is no threat to the Titan X that retains its performance advantage over the G1 Gaming GTX 980.
The frametimes are densely distributed with few variations. The FPS graph shows a large amount of fluctuations.
[section label=”Acoustics” anchor=”acoustics”]Acoustics
Our acoustics testing consisted of trying to determine how noticeable the noise output from the card was, when kept in a case at 1m distance with the side panel closed, as it might be in a real world scenario. Noise is a very relative characteristic that depends not only on the person hearing, but also on the background noise of their surroundings. During the course of our testing, we found that the card barely went over 50% fan speed even at full load, which was silent to our ears. However, setting the fans at 100% rpm makes a noticeable din that sounds something like a toy R/C Car.
[section label=”Temperatures” anchor=”Temps”]Temperatures
At idle the card hovered in the 45°-47°C range with an ambient temperature of 33°C, which is quite warm. The card easily went to 85°C under load. While overclocked it would plateau out at 85°C as the fan kicked up its RPM, which is the same as the stock peak. Owners of this card would do well to look at aftermarket cooling solutions to extract even more performance from it.
[section label=”Power” anchor=”Pwr”]Power
The card comes in a 6pin+8pin power connector configuration and is rated at 250W TDP though it barely reaches that when overclocked. The card requires a good PSU for overclocking. The extra power as compared to the stock design increases the overclocking headroom of the card. However, this does mean that the card becomes slightly less power efficient, something that has been a key selling point for the Maxwell chips.
[section label=”Overclocking” anchor=”oc”]Overclocking the Titan X
Using MSI Afterburner, we were able to push the core clock to 1242MHz with boost clock at 1316 MHz, ( in boost mode, it even reached frequencies of 1400MHz) and the Hynix VRAM chips to 1953 MHz (7812 MHz effective). This is a total overclock of about 24% over stock for the core clock and about 11.5% for the VRAM. These are impressive figures for overclocking without any alterations to the card or without increasing the voltage, but increasing Power limit to 110%. The Voltage can be increased to improve stability of the overclock. To keep the temperatures in check, it is a good idea to set the fan to 60-70% from the outset, especially when the ambient temperatures are higher. All our tests however, were conducted in the stock configuration.
We ran 3DMark to check improvements in the overclocked performance and the graphics score increased from 14474 to 16261.
We also ran the benchmark offered by the game Thief to get a real world idea of the performance gains. We used the Very High settings for our runs that concluded in less time than our standard 120 second runs. One must note that Thief is slightly unreliable as a benchmark due to inconsistency in results across various runs. That said we got an improvement of about 12-13 fps on average with a very slightly higher 99th percentile frametime, which is significant.
[section label=”Feature Set” anchor=”FS”]Feature Set
With NVIDIA’s GeForce Experience, the card can support features like Shadowplay and game streaming. The card’s LED strip can also be controlled to display some cool effects. The new card also lets one use a technique called Dynamic Super Resolution that can be used to enhance the visual quality of shoddy console ports. The 12GB of VRAM shows its worth when using 3D Rendering applications that require such large amounts of Video Memory. Games like GTA V only needed about half this capacity to max out all settings at 4K resolutions.
[section label=”Conclusion” anchor=”End”]Conclusion
While the Titan X does set the bar higher for single GPU performance, its target market isn’t purely gamers, but rather workstation PCs. The large amounts of RAM it packs is overkill for most modern games but should come in handy for 3D Rendering.
The all Black case design comes with its share of flaws and can mean a really hot card running inside the case in the absence of a custom cooling solution. The lack of a heat sheath on top also exposes the PCB to the elements. The stock cooler is a passable design and has a lot of scope for improvement.
The overclocking performance of the card is stellar and provides a tangible performance boost. The most touted point about the Maxwell parts is the power efficiency, though this card doesn’t quite match its other Maxwell brethren in this area. Given that the target market segment of the card is enthusiasts and workstations, we feel the relatively higher price isn’t quite justified by the performance gap to the GTX 980, which is why this GPU loses out on the Value for Money front. We award this card the iLL Gaming Gold simply based on its performance and overclocking potential.
We are extremely grateful to NVIDIA for providing us with a test sample for reviewing.
