The emergence of a large GPU based on the Maxwell architecture was inevitable, the only question was when it would happen and in what form. As a result, the assumption was justified that the GM200 will follow the path of its analogue from the Kepler family - GK110, debuting as part of an accelerator under the TITAN brand.
NVIDIA GeForce GTX TITAN X
There was very little time to test the new video card this time, so the review will be condensed. Having discarded unnecessary discussions, let's get straight to the point. The Maxwell architecture, compared to Kepler, is characterized by a simplified and optimized structure of streaming multiprocessors (SMM), which has made it possible to radically reduce the SMM area while maintaining 90% of the previous performance. In addition, the GM200 belongs to the second iteration of the Maxwell architecture, like the previously released GM204 (GeForce GTX 970/980) and GM206 (GeForce GTX 960) chips. And therefore, it has a more productive geometry engine, PolyMorph Engine version 3.0, and supports at the hardware level some computing functions that are likely to be included in the new feature level Direct3D 12, and are also necessary for hardware acceleration of the VXGI global illumination technology developed by NVIDIA. For a more detailed description of the first and second generation Maxwell architecture, we refer readers to the reviews of GeForce GTX 750 Ti and GeForce GTX 980.
Block diagram of NVIDIA GM200 GPU
Qualitatively, the GM200 graphics processor and lower-end GPUs in the line do not differ from each other, with the exception that only the GM206 has a dedicated H.265 (HEVC) compressed video decoding unit. The differences are purely quantitative. The GM200 includes an unprecedented number of transistors - 8 billion, so there are one and a half to two times more computing units in it than in the GM204 (depending on which ones you count). In addition, the 384-bit memory bus has returned to service. Compared to the GK110 chip, the new flagship GPU is not as intimidatingly powerful, but, for example, the number of ROPs here is double, which makes the GM200 perfectly prepared for 4K resolution.
In terms of support for double precision calculations, the GM200 is no different from the GM204. Each SMX contains only four FP64-compatible CUDA cores, so the combined performance under this workload is 1/32 of FP32.
|
|
|
|
⇡ Technical characteristics, price
TITAN X uses the most powerful version of the GM200 core with a full set of active computing units. The base GPU frequency is 1000 MHz, Boost Clock is 1076 MHz. The memory operates at the standard 7012 MHz frequency for Maxwell-based products. But the volume is unprecedented for gaming video cards - 12 GB (and TITAN X is primarily a gaming video card, at least until the appearance of the GM200 in the main, “numbered” GeForce line).
Suggested retail prices for the TITAN X were announced in the final hours before the review was published. For the US market, the price is set at $999 - the same as the first TITAN based on GK110 cost at one time.
Note: prices in the table for GeForce GTX 780 Ti and TITAN Black are given at the time the latter were discontinued.
|
Model |
GPU |
Video memory |
TDP, W |
RRP* for the US market (excluding taxes), $ |
||||||||
|
Code name |
Number of transistors, million |
Clock frequency, MHz: Base Clock / Boost Clock |
Number of CUDA cores |
Number of texture units |
Bus width, bits |
Chip type |
Clock frequency: real (effective), MHz |
Volume, MB |
||||
|
GeForce GTX 780 Ti |
||||||||||||
|
GeForce GTX TITAN Black |
||||||||||||
|
GeForce GTX 980 |
||||||||||||
|
GeForce GTX TITAN X |
||||||||||||
⇡ Design
Since the very first Titan, NVIDIA has been using the same cooling system in top-end video cards, with some variations. TITAN X stands out from its predecessors only in its completely black body (only two inserts on the sides remain unpainted).
NVIDIA GeForce GTX TITAN X
The back plate, which was equipped with the GeForce GTX 980 as an experiment, is again missing from the TITAN X, despite the fact that some memory chips are soldered on the back side of the board. Although GDDR5 chips, in general, do not require additional cooling.
NVIDIA GeForce GTX TITAN X rear view
But the radiator with an evaporation chamber has returned, which in the GTX 980 was replaced by a simpler option.
NVIDIA GeForce GTX TITAN X, cooling system
NVIDIA GeForce GTX TITAN X, cooling system
NVIDIA GeForce GTX TITAN X, cooling system
The video card has three DisplayPort connectors and one each for HDMI and Dual-Link DVI-I.
⇡ Fee
The design of the printed circuit board, which is not at all surprising, evokes associations with a series of video adapters on the GK110 chip. The voltage converter is built according to a 6+2 scheme (the number of phases for powering the GPU and memory chips, respectively). Power is supplied through one eight-pin and one six-pin connector. But this is the first time we see the power controller for the ON Semiconductor NCP81174 graphics processor here.
24 SK hynix H5GQ4H24MFR-R2C memory chips with a nominal frequency of 7 GHz are located on both sides of the board.
NVIDIA GeForce GTX TITAN X, PCB, front side
NVIDIA GeForce GTX TITAN X, PCB, back side
Test stand, testing methodology
Energy-saving CPU technologies are disabled in all tests. In the NVIDIA driver settings, the CPU is selected as the processor for PhysX calculations. In AMD drivers, the Tesselation parameter is transferred from the AMD Optimized state to Use application settings.
| Benchmarks: synthetic | |||
|---|---|---|---|
| Program | Settings | Permission | |
| 3DMark 2011 | Test Extreme | - | - |
| 3DMark | Fire Strike test (not Extreme) | - | - |
| Unigine Heaven 4 | DirectX 11, max. quality, tessellation in Extreme mode | AF 16x, MSAA 4x | 1920×1080 / 2560×1440 |
| Benchmarks: games | |||
|---|---|---|---|
| Program | Settings | Anisotropic filtering, full screen anti-aliasing | Permission |
| Far Cry 3 + FRAPS | DirectX 11, max. quality, HDAO. Start of the Secure the Outpost mission | AF, MSAA 4x | 2560×1440/3840×2160 |
| Tomb Raider. Built-in benchmark | Max. quality | AF 16x, SSAA 4x | 2560×1440/3840×2160 |
| Bioshock Infinite. Built-in benchmark | Max. quality. Postprocessing: Normal | AF 16x, FXAA | 2560×1440/3840×2160 |
| Crysis 3 + FRAPS | Max. quality. Start of the Post Human mission | AF 16x, MSAA 4x | 2560×1440/3840×2160 |
| Metro: Last Light. Built-in benchmark | Max. quality | AF 16x, SSAA 4x | 2560×1440/3840×2160 |
| Company of Heroes 2. Built-in benchmark | Max. quality | AF,SSAA 4x | 2560×1440/3840×2160 |
| Battlefield 4 + FRAPS | Max. quality. Beginning of the Tashgar mission | AF 16x, MSAA 4x + FXAA | 2560×1440/3840×2160 |
| Thief. Built-in benchmark | Max. quality | AF 16x, SSAA 4x + FXAA | 2560×1440/3840×2160 |
| Alien: Isolation | Max. quality | AF 16x, SMAA T2X | 2560×1440/3840×2160 |
Test participants
The following video cards took part in performance testing:
- NVIDIA GeForce GTX TITAN X (1000/7012 MHz, 12 GB);
⇡ Clock speeds, power consumption, temperature, overclocking
The GM110 operates at a base frequency that the GK110 never reached at reference specifications. In addition, GPU Boost acts very aggressively, raising the frequency up to 1177 MHz. At the same time, the processor is content with a voltage of 1.174 V, which is lower compared to top-end products based on GK110.
BIOS settings allow you to increase the power limit to 110% and add 83 mV to the maximum voltage on the GPU. In fact, the voltage only rises to 1.23 V, but at the same time several additional frequency/VID steps are opened: the difference between the base frequency and the maximum frequency recorded in the speaker increases to 203 MHz.
Overclocking the video card allowed us to reach a base frequency of 1252 MHz, and frequencies up to 1455 MHz were observed in dynamics. The video memory was able to add 1.2 GHz, successfully operating at an effective frequency of 8,212 MHz.
| Base Clock, MHz | Max. Boost Clock, MHz | Base Clock, MHz (overclocking) | Max. registered Boost Clock, MHz (overclocking) | |
| GeForce GTX TITAN X | 1000 | 1177 (+177) | 1252 | 1455 (+203) |
| GeForce GTX 980 | 1127 | 1253 (+126) | 1387 | 1526 (+139) |
| GeForce GTX TITAN Black | 889 | 1032 (+143) | 1100 | 1262 (+162) |
| GeForce GTX TITAN | 836 | 1006 (+145) | 966 | 1150 (+184) |
| GeForce GTX 780 Ti | 876 | 1020 (+144) | 986 | 1130 (+144) |
| GeForce GTX 780 | 863 | 1006 (+143) | 1053 | 1215 (+162) |
| GeForce GTX 770 | 1046 | 1176 (+130) | 1190 | 1333 (+143) |
In terms of power consumption, the TITAN X is close to the GTX 780 Ti and far exceeds the GTX 980. Contrary to expectations, in Crysis 3 there is no significant difference between the TITAN X and the Radeon R9 290X, but in FurMark the R9 290X (like the R9 280X) heats up more and is noticeably superior to the TITAN X.
Overclocking the TITAN X increases power by 5-25 W, depending on which test results you rely on - FurMark or Crysis 3.
The maximum temperature that is allowed for the GPU is determined by the BIOS settings, so TITAN X does not go beyond the set 83 ° C. At the same time, the cooling system turbine spins up to 49% of the maximum speed - up to 2339 rpm. At first glance, this is quite a lot, but in fact the noise from the cooler is quite acceptable.
⇡ Performance: synthetic tests
- TITAN X impresses from the very first test. Compared to the GTX 780 Ti and Radeon R9 290X, the video card is one and a half times more powerful.
- With the Radeon R9 280X and GeForce GTX 770 - adapters based on once top-end GPUs - the difference is more than twofold.
- All of the above is true for 3DMark 2013.
Unigine Heaven 4
- TITAN X maintains an advantage of about 50% over the GTX 780 Ti and Radeon R9 290X at WQHD resolution. By the way, unlike 3DMark, the GTX 980 is no better than the GTX 780 Ti in this test.
- At Ultra HD resolution, previous video adapters have reduced the distance, and yet TITAN X is head and shoulders above all rivals.
⇡ Performance: games
This time we will deviate from the standard form of describing game tests. In the case of TITAN X, it is completely pointless to describe which video card is faster for each game. In all games, the new “Titan” is ahead of its rivals by a colossal margin. Quantitative indicators tend to the formula: TITAN X is 30-50% faster than the GeForce GTX 780 Ti and Radeon R9 290X, and often twice as fast compared to the Radeon R9 280X and GeForce GTX 770. The only intrigue is to look for fluctuations within this corridor in that or the other side. In addition, there is a unique case: the TITAN X enjoys a frame rate of 24 FPS in Far Cry 4 at Ultra HD resolution and MSAA 4x anti-aliasing, while competitors can not get out of the 5-7 FPS hole (and the GeForce GTX 770 even less). Here, apparently, the Titan needed a memory capacity of 12 GB, and even 4 GB, which the Radeon R9 290X is equipped with, is not enough for such settings in FC4.
Tomb Raider
Bioshock Infinite
Crysis 3
⇡ Performance: Computing
Video decoding (DXVA Checker, Decode Benchmark)
- The dedicated H.264 decoder in the GM200 is the same as in other chips in the Maxwell family. Its performance is more than enough to play video with resolutions up to Ultra HD and frame rates of 60 Hz and higher.
- Among AMD's discrete video adapters, only the Radeon R9 285 can boast of this. The GeForce GTX 780 Ti is capable of delivering up to 35 FPS at a resolution of 3840 × 2160.
- CPUs with 6-8 x86 cores are better suited for fast decoding for video conversion, but the fixed functionality unit does this job with less power consumption, and, finally, it is simply fed to the most powerful GPU.
- The only GPU with full hardware H.265 decoding is the GM206 in the GeForce GTX 960. The remaining representatives of the Maxwell architecture, as well as Kepler, perform part of the operations on the H.264 decoder pipeline. The rest falls on the central processor.
- The performance of all these adapters with a good CPU is quite enough to play video at any reasonable resolution and frame rate. For high-speed work, a GTX 960 or a powerful CPU is better suited.
Luxmark: Room (Complex Benchmark)
- The Maxwell architecture demonstrates a surprising performance increase over Kepler in this task, thanks to which the TITAN X doubled the modest result of the GeForce GTX 780 Ti and left the Radeon R9 290X far behind. However, this does not mean that the LuxMark results are representative of any ray tracing tasks.
- The difference between TITAN X and GeForce GTX 980 is not as huge as in gaming tests.
Sony Vegas Pro 13
- AMD video adapters maintain leadership in video rendering. And TITAN X does not stand out in any way in the group of the most productive NVIDIA devices.
CompuBench CL: Ocean Surface Simulation
- TITAN X takes away the palm from the Radeon R9 290X and compensates for the failure of the GeForce GTX 980, which finds this test surprisingly difficult.
CompuBench CL: Particle Simulation
- Here, on the contrary, the GTX 980 took a big step forward from the GTX 780 Ti, and the TITAN X consolidated its success. The Radeon R9 290X cannot be compared to NVIDIA's flagship.
SiSoftware Sandra 2015: Scientific Analysis
- In double precision conditions (FP64), AMD accelerators still have no equal, and even the Radeon R9 280X based on a far from new GPU can give TITAN X a run for its money.
- Among the “green” ones, TITAN X predictably leads in FP64 performance, especially compared to the frankly weak GTX 980.
- In FP32 computing, the TITAN X stands out from all NVIDIA graphics cards. Only it provides a level of performance comparable to that of the Radeon R9 290X.
⇡ Conclusions
Considering that discrete GPU production is still within the 28nm process, the results from the GeForce GTX TITAN X look fantastic. With the same TDP as GK110-based video adapters, TITAN X reaches 130-150% of the performance of accelerators such as the GTX 780 Ti and Radeon R9 290X. If we take the first 28nm GPUs - GK104 (GTX 680, GTX 770) and Radeon R9 280X, then TITAN X is often twice as good as them.
TITAN X, like its predecessors in this position, is extremely expensive for a single GPU graphics card. The positioning has not changed compared to the previous Titans. Firstly, this is an alternative to SLI configurations of two discrete GeForce GTX 980: although the potential performance of a tandem is higher, a single GPU has more predictable performance. Secondly, compact PCs that do not have room for two video cards. And finally, non-graphical computing (GP-GPU). While GM200's FP64 performance is limited to 1/32 that of FP32, TITAN X makes up for some of this limitation with GPU brute force. In addition, FP32 calculations are dominant in "prosumer" load (the same Ray Tracing, accelerated video rendering), and in this discipline the GM200 is at least as good as the best AMD products, and often surpasses them in the same way as in gaming tests.
Spring is a time not only for nature to awaken, but also for the traditional announcement of the flagship single-chip video card of the Titan line. And although the first demonstration of NVIDIA GeForce GTX Titan X was unexpected, this announcement was surrounded by a number of rumors. A couple of days ago the official presentation of this video card took place, and over the course of several days we had the opportunity to study it in detail. Let's see what she can boast of next.
NVIDIA GeForce GTX Titan X became the fourth in the line and the third “titanium” on a single-chip system. Let me remind you that the GeForce GTX Titan Z with two GPUs stands apart. Of course, we understand that such systems cannot be called “folk”, but even without 1000-1300 dollars, this review may be interesting to get acquainted with the Maxwell architecture in its maximum implementation today. As expected, it is based on the GM200 core, the second generation of this architecture. It comes with the PolyMorph Engine 3.0 with support for feature level Direct3D 12 functions and hardware acceleration of global illumination technology. Maybe this will be the breakthrough in realism and graphics that the gaming industry has been waiting for for so long?
The GM200 contains 8 billion transistors, 256 ROPs, 3072 CUDA and a 384-bit memory bus. All these weapons are aimed at supporting 4K resolution and improving performance in 3D mode. The base core frequency is: 1000 MHz, Boost Clock - 1076 MHz. The memory operates at a frequency of 7012 MHz. There is 12 GB of graphics memory on board, which was unavailable for video cards aimed at gamers before the release of the Titan X.
Video review of NVIDIA GeForce GTX Titan X
Appearance
NVIDIA GeForce GTX Titan X did not become a revolution in the design of top-end video cards; the appearance has changed little. An air cooling system is used, which externally repeats what we saw earlier.
The changes here are minor, limited to changing the color of the body. Now it is almost completely painted black.
The video card does not have an amplification board on the back of the printed circuit board. Let me remind you that the GeForce GTX 980 had it in the reference design.
The rear panel has three DisplayPorts, HDMI and DVI. In shared workspace mode, three connectors can work, but you can connect all 5 connectors at the same time. The same approach is applied throughout the 900 line.
On the side there is a shining NVIDIA GeForce GTX logo. Unfortunately, the spectacular images of the glowing fan remain only photographs.
Cooling
The design of the installed cooling system is the same as that used in the GeForce GTX 780 Ti.
An evaporation chamber is used, which has shown its strengths in removing a large volume of heat to the radiator.
The system is collapsible, so you can completely remove the heat dissipator; this can be useful for installing a water cooling system.
Filling
The power supply system has also migrated, despite the mention of changes, upon inspection it is still the same capacitors and chokes. Even the PWM controller is familiar to us - NCP4206.
But I won’t dramatize, we ourselves, even under long-term loads, were unable to find the noise and squeaking that was mentioned in the comments to a number of video cards.
The level for increasing voltage has also been preserved. In NVIDIA GeForce GTX Titan X it can be increased within 25 W (TDP 250 W/275 W).
12 GB of SKhynix memory chips with a frequency of 1750 MHz are soldered, there are 24 of them in total.
Testing NVIDIA GeForce GTX Titan X
A test bench was used.
| Model | Data |
|---|---|
| Frame | Aerocool Strike-X Air |
| Motherboard | Biostar Hi-Fi Z87X 3D |
| CPU | Intel Core i5-4670K Haswell |
| CPU cooler | DeepCool Ice Blade Pro v2.0 |
| Video card | Inno3D iChill GeForce GTX 780Ti HerculeZ X3 Ultra |
| RAM | Corsair CMX16GX3M2A1600C11 DDR3-1600 16 GB Kit CL11 |
| HDD | ADATA XPG SX900 256 GB |
| Hard drive 2 | WD Red WD20EFRX |
| power unit | Aerocool Templarius 750W |
| Wi-Fi adapter | TP-LINK TL-WDN4800 |
| Audio | Creative Sound Blaster EVO Wireless |
| Monitor | iiyama ProLite E2773HDS |
| Monitor 2 | Philips 242G5DJEB |
| Mouse | ROCCAT Kone XTD |
| Keyboard | Razer BlackWidow Chroma |
| Stabilizer | Sven AVR PRO LCD 10000 |
| operating system | Microsoft Windows Ultimate 8 64-bit |
In all tables below, the data is provided using factory settings; proprietary software from manufacturers is not installed. Memory and core frequencies are also not affected in order to exclude the influence of extraneous factors.
1. Temperature conditions of the video card
- NVIDIA GeForce GTX Titan X - 31/83
- Inno3D iChill GeForce GTX 960 Ultra - 29/44
- GeForce GTX 980 - 34/79
- GeForce GTX 770 - 35/80
- GeForce GTX 780 - 35/77
- GeForce GTX 760 - 35/84
2. Noise
- NVIDIA GeForce GTX Titan X - 36/42
- GeForce GTX 980 - 34/79
3. Power consumption
- NVIDIA GeForce GTX Titan X-405
- Inno3D iChill GeForce GTX 960 Ultra - 260
- GeForce GTX 980 - 295
- Inno3D iChill GeForce GTX 780Ti HerculeZ X3 Ultra - 340
- NVIDIA GeForce GTX Titan X - 7133
- Inno3D iChill GeForce GTX 960 Ultra - 3522
- GeForce GTX 980 - 6050
- Inno3D iChill GeForce GTX 780Ti HerculeZ X3 Ultra - 6190
The time has come for the most spectacular tests, FPS measurements in resource-intensive games. We accompany a number of tables with visual videos with recordings during the game. Data is captured with Full HD resolution at Ultra settings. It is worth considering that in a number of moments in the videos, the actual FPS data is lower than that obtained during test runs, this is due to the expenditure of resources on video recording. For this video card, we separately tested operation at a resolution of 3840x2160 on Ultra settings.
6. Crysis 3
Crysis 3 - 3840x2160 - Very High 4x AA - 22.
- Inno3D iChill GeForce GTX 960 Ultra - 45
- GeForce GTX 980 - 69
- Inno3D iChill GeForce GTX 780Ti HerculeZ X3 Ultra - 61
- GeForce GTX 770 - 43
- GeForce GTX 780 - 47
7. Battlefield 4
Battlefield 4 - 3840x2160 - Ultra - 39
- NVIDIA GeForce GTX Titan X - 75
- Inno3D iChill GeForce GTX 960 Ultra - 52
- GeForce GTX 980 - 91
- Inno3D iChill GeForce GTX 780Ti HerculeZ X3 Ultra - 82
8. Hitman: Absolution
A very demanding game, powered by the Glacier 2 engine. The game's appetite is the envy of other new releases of the year.
Hitman: Absolution - 3840x2160 - High, 2x MSAA, 16x AF - 46
- NVIDIA GeForce GTX Titan X - 95
- Inno3D iChill GeForce GTX 960 Ultra - 44/li>
- GeForce GTX 980 - 70
- Inno3D iChill GeForce GTX 780Ti HerculeZ X3 Ultra - 62
- GeForce GTX 770 - 43
- GeForce GTX 780 - 55
- GeForce GTX 760 - 41
9. Metro Last Light
Another hardware-intensive game that uses DirectX 11 and tessellation.
Metro Last Light - 3840x2160 - Very high - 35
10. Middle Earth: Shadow Mordor
- Inno3D iChill GeForce GTX 960 Ultra - 51
- GeForce GTX 980 - 111
11. Tomb Raider
- NVIDIA GeForce GTX Titan X - 156
- Palit GeForce GTX 960 Super JetStream - 64
- Inno3D iChill GeForce GTX 960 Ultra - 68
- GeForce GTX 980 - 100
12. Watch Dogs Ultra 4x AA
- NVIDIA GeForce GTX Titan X - 80
- Inno3D iChill GeForce GTX 960 Ultra - 49
- GeForce GTX 980 - 62
13. Total War: Rome II Extreme
- NVIDIA GeForce GTX Titan X - 79
- Inno3D iChill GeForce GTX 960 Ultra - 41
- GeForce GTX 980 - 70
14. GRID Autosport Ultra 4x MSAA
- NVIDIA GeForce GTX Titan X - 154
- Inno3D iChill GeForce GTX 960 Ultra - 80
- GeForce GTX 980 - 128
15. World of Tanks
- NVIDIA GeForce GTX Titan X - 124
- Palit GeForce GTX 960 Super JetStream - 71
- Inno3D iChill GeForce GTX 960 Ultra - 75
- GeForce GTX 980 - 116
16. World of Warships
This is a new section of our tests; while the number of tested video cards is limited, comprehensive material will be presented by the end of March. The World of Warships game is difficult to evaluate in terms of graphics efficiency, but in general this data can be useful when assembling a system specifically for Wargaming games.
- NVIDIA GeForce GTX Titan X - 72
- Inno3D iChill GeForce GTX 780Ti - 72
- Palit GeForce GTX 960 Super JetStream - 59
- Radeon R9 280x - 70
- Radeon R9 280x - 70
Overclocking
By tradition, we do not limit ourselves to testing at standard frequencies. For overclocking, the MSI Afterburner program is used, the latest version at the time of testing. For NVIDIA GeForce GTX Titan X, we were able to achieve the following results without increasing the core voltage:
To compare the performance increase, the synthetic 3D Mark FireStrike test is used:
There is potential for further overclocking at maximum voltage. The core frequency can be raised to 1202 MHz and 7806 MHz from memory. Here the maximum temperature level rises to 88 degrees.
Results for NVIDIA GeForce GTX Titan X
NVIDIA GeForce GTX Titan X demonstrated an increase in performance while reducing power consumption. Given the current balance of power, this is the maximum performance on a single-chip system. AMD Radeon has not yet provided adequate responses. As an alternative, you can consider the still relevant GTX 780 Ti, GTX 980 in SLI mode, Radeon R9 290X. It will also be interesting for video rendering. NVIDIA GeForce GTX Titan X receives the well-deserved Gold Award..
The main conclusion: GeForce GTX TITAN X is the fastest single-chip gaming video card of our time. The performance of the new NVIDIA flagship is enough to effortlessly enjoy modern 3D entertainment in Full HD and WQHD at the highest graphics settings. True, the GeForce GTX 980 can do this too. “Wait, what about 4K?” - the reader will ask. Yes, although I called the article “The first for Ultra HD”, in terms of modern games at maximum graphics quality settings, the GeForce GTX TITAN X demonstrates only a conditionally playable FPS level. However, this is the best indicator among single-chip video cards. Therefore, for me personally, the GeForce GTX TITAN X is the first video card that is really capable of satisfying the requirements of a gamer who wants to conquer virtual spaces at such a high resolution. Even if in some cases you will have to delve into the settings. But a couple of these “titans” are capable of reining in any emerging next-gen. If only the drivers and optimization did not fail. However, this is a topic for a separate article.
It is 12 GB of video memory that provides a large margin of safety for the GeForce GTX TITAN X. Of course, someone will rightly note that such a volume is excessive. However, the same Assassin’s Creed Unity in 4K resolution, not at the very highest graphics quality settings, already “eats” 5-6 GB of video memory from the video card. That is almost half. That is why (and we have clearly seen this) even ultra-expensive combinations of several 3D accelerators can have a bottleneck in the form of 4 GB GDDR5. So for gaming in Ultra HD you already need to have some reserve.
As always, the NVIDIA reference showed itself to be good. The video card has decent overclocking potential. A cooler based on an evaporation chamber cools a 250-watt chip quite effectively. It is a little noisy, but quite tolerable.
Of course, if NVIDIA had released this video card in September last year (not for ~1000 bucks - author's note), then the wow effect would have been, in my opinion, stronger. However, should we be surprised by NVIDIA’s scheme, which has been proven over the years? Price is the strongest limiting factor for purchasing a GeForce GTX TITAN X. In our country, which is experiencing another economic crisis, this is even more so.
Finally, I’ll just note that the “greens” have raised the performance bar quite high, which the future AMD flagship (Radeon R9 390X?) will have to at least reach in order to restore the status quo. Or do something similar in performance, but noticeably more budget-friendly. Agree, this will be very interesting to follow.
NVIDIA GeForce GTX TITAN X receives our Editors' Choice award.
NVIDIA does not often deviate from traditions that have been formed over the years. So in 2015, according to the established spring tradition, the “greens” present a new single-chip flagship in the person of GeForceGTXTITANX. At the beginning of summer 2015, this is the world's most powerful video card based on a single graphics processor.
The new graphics adapter is the fourth in the Titans lineup and, logically, replaces . The new product is based on a graphics core labeled GM200, created on a second-generation microarchitecture. In terms of basic characteristics, the GM200 is a one and a half times “expanded” GM204 core, on which the recent single-chip flagship is based. More precisely, the number of CUDA cores, ROP and TMU units, as well as the cache size have been increased by one and a half times. Let's take a closer look at the characteristics of these two video cards.
The energy consumption of the new flagship turned out to be noticeably more than the energy consumption of the GTX 980. Naturally, this is due to the higher performance of TITAN X, which can reach 30% in comparison with the 980 model. According to the manufacturer's recommendation, the power supply for the system should be at least 600 W.
Here, perhaps, you need to pay attention to the cooling system of the new product. Namely, the GeForce GTX TITAN X will officially be supplied exclusively with a reference cooler, which should provide high performance with low noise levels.
*Highest possible graphics quality
It is quite obvious that the new product supports all currently existing NVIDIA technologies - SLI®, G-Sync™, GameStream™, ShadowPlay™, 2.0 GPU Boost™, Dynamic Super Resolution, MFAA, GameWorks™, OpenGL 4.5. Microsoft DirectX 12 API is also supported with subsequent upgrade to 12.1.
The price for the model in question at the time of the start of sales was announced by the manufacturer in the amount of $999. Actually, this is the same amount as the “black Titan”. But if you pay attention to the colossally increased performance of the new video card in comparison with its predecessor, then here NVIDIA has again made a big and timely step forward.
Here's a basic, in-depth look at the Nvidia Geforce GTX Titan X.
Object of study: 3D graphics accelerator (video card) Nvidia Geforce GTX Titan X 12288 MB 384-bit GDDR5 PCI-E
Developer Information: Nvidia Corporation (Nvidia trademark) was founded in 1993 in the USA. Headquarters in Santa Clara (California). Develops graphics processors and technologies. Until 1999, the main brand was Riva (Riva 128/TNT/TNT2), from 1999 to the present - Geforce. In 2000, the assets of 3dfx Interactive were acquired, after which the 3dfx/Voodoo trademarks were transferred to Nvidia. We don't have our own production. The total number of employees (including regional offices) is about 5,000 people.
Part 1: Theory and architecture
As you already know, back in the middle of last month, Nvidia released a new top-end video card called Geforce GTX Titan X, which became the most powerful on the market. We immediately published a detailed review of this new product, but it contained only practical research, without a theoretical part and synthetic tests. This happened due to various circumstances, including those beyond our control. But today we are correcting this shortcoming and will take a very detailed look at the March new product - nothing happened in a month to make it lose its relevance.
Back in 2013, Nvidia released the first solution of the new Geforce GTX Titan brand of video cards, named after the supercomputer at Oak Ridge National Laboratory. The first model of the new line set new records, both in terms of performance and price - the recommended price for the US market was set at $999. This was the first elite video card of the Titan series, which then continued with the not-so-popular dual-chip Titan Z and the accelerated Titan Black, which received a fully unlocked GK110 revision B graphics processor.
And in the spring of 2015, the time has come for another new Nvidia product from the “titanium” premium series. The GTX Titan X was first shown by company president Jensen Huang at the GDC 2015 gaming developer conference at an event on the Epic Unreal Engine game engine. In fact, this video card was invisibly present at the show anyway, being installed in many demo stands, but Jensen presented it officially.
Before the release of the Geforce GTX Titan X, the fastest single-chip video card was the Geforce GTX 980, based on the GM204 chip of the same Maxwell graphics architecture, introduced in September last year. This model is very energy efficient, delivering decent processing power while consuming just 165 watts of power - meaning it's twice as energy efficient as the previous generation Geforce.
At the same time, Maxwell architecture GPUs support the upcoming DirectX 12 (including Feature Level 12.1) and other latest graphics technologies of the company: Nvidia Voxel Global Illumination simulation (VXGI, we wrote about it in the article on GTX 980), a new Multi-Frame sampled anti-aliasing method AA (MFAA), Dynamic Super Resolution (DSR), etc. The combination of performance, energy efficiency and capabilities made the GM204 chip the best advanced GPU at the time of its release.
But everything changes someday, and the GPU with 2048 computational cores and 128 texture units was replaced by a new graphics processor based on the same second-generation Maxwell architecture (we remember the first from the GM107 chip, on which the Geforce GTX 750 Ti video card is based) and those the same capabilities, but with 3072 CUDA computing cores and 192 texture units - all this is already packed into 8 billion transistors. It’s clear that Geforce GTX Titan X immediately became the most powerful solution.
In fact, the top-end second-generation Maxwell chip, which we now know under the code name GM200, had been ready by Nvidia for some time before its announcement. There simply wasn’t much point in releasing another top-end video card model when even the GM204-based GeForce GTX 980 did an excellent job of being the fastest single-chip video card in the world. Nvidia has been waiting for some time for the release of a more powerful GPU-based solution from AMD, produced using the same 28 nm process technology, but it never came.
Probably, so that the product would not turn sour at all in the absence of real competition, they still decided to release it, securing the title of the company that produces the most powerful GPUs. And indeed, there was no point in waiting for the opponent’s decision, because it was postponed at least until June - it was simply unprofitable to wait that long. Well, if something happens, you can always release an even more powerful video card based on the same graphics processor, but operating at a higher frequency.
But why do we need such powerful solutions in the era of the proliferation of multi-platform games with fairly average GPU power requirements? Firstly, very soon the first gaming applications should appear that use the capabilities of DirectX 12, even if they are multi-platform - after all, the PC versions of such applications almost always offer higher-quality graphics, additional effects and higher-resolution textures. Secondly, DirectX 11 games have already been released that can use all the capabilities of the most powerful GPUs - like Grand Theft Auto V, which we will talk about in more detail below.
It is important that Nvidia's Maxwell architecture graphics solutions fully support the so-called Feature Level 12.1 from DirectX 12 - the maximum currently known. Nvidia has long provided game developers with drivers that support the upcoming version of DirectX, and now they have become available to users who installed Microsoft Windows 10 Technical Preview. It is not surprising that it was Geforce GTX Titan X video cards that were used to demonstrate the capabilities of DirectX 12 at the Game Developers Conference, where the model was first shown.
Since the video card model in question from Nvidia is based on the top-end second-generation Maxwell architecture graphics processor, which we have already reviewed and which is similar in detail to the previous Kepler architecture, before reading this material it is useful to familiarize yourself with earlier articles about the company’s video cards Nvidia:
- Nvidia Geforce GTX 970 - A good replacement for the GTX 770
- Nvidia Geforce GTX 980 - Follower of the Geforce GTX 680, outperforming even the GTX 780 Ti
- Nvidia Geforce GTX 750 Ti - Maxwell starts small... despite Maxwell
- Nvidia Geforce GTX 680 - the new single-processor leader in 3D graphics
So, let's look at the detailed characteristics of the Geforce GTX Titan X video card, based on the GM200 graphics processor.
| Geforce GTX Titan X graphics accelerator | |
|---|---|
| Parameter | Meaning |
| Chip code name | GM200 |
| Production technology | 28 nm |
| Number of transistors | about 8 billion |
| Core area | about 600 mm 2 |
| Architecture | Unified, with an array of common processors for stream processing of numerous types of data: vertices, pixels, etc. |
| DirectX hardware support | DirectX 12, with support for Feature Level 12.1 |
| Memory bus | 384-bit: six independent 64-bit wide memory controllers with support for GDDR5 memory |
| GPU frequency | 1000 (1075) MHz |
| Computing blocks | 24 streaming multiprocessors, including 3072 scalar ALUs for single and double precision floating point calculations (at 1/32 of FP32) within the framework of the IEEE 754-2008 standard; |
| Texturing blocks | 192 texture addressing and filtering units with support for FP16 and FP32 components in textures and support for trilinear and anisotropic filtering for all texture formats |
| Rasterization Units (ROPs) | 6 wide ROP blocks (96 pixels) with support for various anti-aliasing modes, including programmable and with FP16 or FP32 frame buffer format. The blocks consist of an array of configurable ALUs and are responsible for depth generation and comparison, multisampling and blending |
| Monitor support | Integrated support for up to four monitors connected via Dual Link DVI, HDMI 2.0 and DisplayPort 1.2 interfaces |
| Geforce GTX Titan X Reference Graphics Card Specifications | |
|---|---|
| Parameter | Meaning |
| Core frequency | 1000 (1075) MHz |
| Number of universal processors | 3072 |
| Number of texture blocks | 192 |
| Number of blending blocks | 96 |
| Effective memory frequency | 7000 (4×1750) MHz |
| Memory type | GDDR5 |
| Memory bus | 384-bit |
| Memory | 12 GB |
| Memory Bandwidth | 336.5 GB/s |
| Compute Performance (FP32) | up to 7 teraflops |
| Theoretical maximum fill rate | 96 gigapixels/s |
| Theoretical texture sampling rate | 192 gigatexels/s |
| Tire | PCI Express 3.0 |
| Connectors | One Dual Link DVI, one HDMI 2.0 and three DisplayPort 1.2 |
| Energy consumption | up to 250 W |
| Additional food | One 8-pin and one 6-pin connectors |
| Number of slots occupied in the system case | 2 |
| Recommended price | $999 (USA), RUB 74,990 (Russia) |
The new Geforce GTX Titan X model received a name that continues Nvidia’s line of premium solutions with a specific positioning - the letter X was simply added to it. The new product replaced the Geforce GTX Titan Black model, and is located at the very top of the company’s current product line. The only thing that remains above it is the dual-chip Geforce GTX Titan Z (although it can no longer be mentioned), and below it are the single-chip GTX 980 and GTX 970 models. The recommended price for the new board is $999, and this is expected for a Titan line board, since it is the best performance solution on the single-chip video card market.
The Nvidia model in question is based on the GM200 chip, which has a 384-bit memory bus, and the memory operates at 7 GHz, which gives a peak bandwidth of 336.5 GB / s - one and a half times more than in the GTX 980. This is quite an impressive value, especially if you remember the new on-chip information compression methods used in the second generation Maxwell, which help to use the existing memory bandwidth much more efficiently than a competitor’s GPU.
With such a memory bus, the amount of video memory installed on the video card could be 6 or 12 GB, but in the case of the elite model, it was decided to install 12 GB in order to continue the trend set by the first GTX Titan models. This is more than enough to run any 3D applications without regard to quality parameters - this amount of video memory is enough for absolutely any existing game at any screen resolution and at any quality settings, which makes the Geforce GTX Titan X video card especially tempting in the future - its the owner will never face a lack of video memory.
The official power consumption figure for the Geforce GTX Titan X is 250 W - the same as other single-chip solutions in the elite Titan series. Interestingly, 250 W is about 50% more than the GTX 980, and the number of main functional units has also increased by the same amount. Quite high consumption does not bring any problems; the reference cooler does an excellent job of dissipating such an amount of heat, and enthusiast systems after the GTX Titan and GTX 780 Ti have long been ready for a similar level of power consumption.
Architecture
The Geforce GTX Titan X video card model is based on the new GM200 graphics processor, which includes all the architectural capabilities of the GM204 chip, so everything said in the article on the GTX 980 applies in full to the premium new product - we advise you to first read the material in which they are more fully discussed namely the architectural features of Maxwell.
The GM200 GPU can be called an extreme version of the GM204, possible within the 28 nm process technology. The new chip is larger, significantly faster and more power hungry. According to Nvidia, the "big Maxwell" includes 8 billion transistors that occupy an area of about 600 mm 2 - that is, it is the company's largest graphics processor. “Big Maxwell” has 50% more stream processors, 50% more ROP units and 50% more memory bandwidth, which is why it has almost one and a half times the area.
Architecturally, the GM200 video chip is fully consistent with the younger GM204 model; it also consists of GPC clusters, which contain several SM multiprocessors. The top-end GPU contains six GPC clusters consisting of 24 multiprocessors, in total it has 3072 CUDA cores, and texture operations (sampling and filtering) are performed using 192 texture units. And with a base frequency of 1 GHz, the performance of texture modules is 192 gigatexels/sec, which is more than a third higher than the same performance of the company’s previous most powerful video card - Geforce GTX 980.
The second-generation Maxwell multiprocessor is divided into four blocks of 32 CUDA cores (a total of 128 cores per SMM), each of which has its own resources for distributing instructions, scheduling processing and buffering the instruction stream. Due to the fact that each computing unit has its own dispatcher blocks, CUDA computing cores are used more efficiently than in Kepler, which also reduces GPU power consumption. The multiprocessor itself has not changed compared to the GM204:
To improve the efficiency of GPU caches, numerous changes have been made to the memory subsystem. Each of the multiprocessors in the GM200 has a dedicated 96 KB of shared memory, and the L1 and texture caches are combined into 24 KB blocks - two blocks per multiprocessor (48 KB total per SMM). Previous generation Kepler GPUs had only 64 KB of shared memory, which also served as L1 cache. As a result of all the changes, the efficiency of Maxwell CUDA cores is approximately 1.4 times higher than in the similar Kepler chip, and the energy efficiency of the new chips is approximately twice as high.
In general, everything in the GM200 graphics processor is arranged exactly the same as in the GM204 chip we reviewed in 2014. Even the computing cores, which can perform double-precision floating-point operations at a rate of only 1/32 of the speed of single-precision calculations, were not touched - just like the Geforce GTX 980. It seems that Nvidia admitted that the release of specialized solutions for the professional market (GK210) and for gaming (GM200) is quite justified.
The memory subsystem of the GM200 is strengthened compared to the GM204 - it is based on six 64-bit memory controllers, which totals a 384-bit bus. The memory chips operate at an effective frequency of 7 GHz, which gives a peak throughput of 336.5 GB/s, which is one and a half times higher than that of the Geforce GTX 980. Don’t forget about new data compression methods from Nvidia, which allow you to achieve more efficient memory bandwidth, compared to previous products - on the same 384-bit bus. In our review of the Geforce GTX 980, we carefully examined this innovation of the second generation of Maxwell chips, which provides them with a quarter more efficient use of video memory compared to Kepler.
Like all the latest Geforce video cards, the GTX Titan X model has a base frequency - the minimum for GPU operation in 3D mode, as well as a Boost Clock turbo frequency. The base frequency for the new product is 1000 MHz, and the Boost Clock frequency is 1075 MHz. As before, turbo frequency only means the average GPU operating frequency for a certain set of gaming applications and other 3D tasks used by Nvidia, and the actual operating frequency may be higher - it depends on the 3D load and conditions (temperature, power consumption etc.)
It turns out that the GPU frequency of the new product is approximately 10% higher than that of the GTX Titan Black, but lower than that of the GTX 980, since large graphics processors always have to be clocked at a lower frequency (and the GM200 is noticeably larger in area than the GM204) . Therefore, the overall 3D performance of the new product will be approximately 33% higher than that of the GTX 980, especially when comparing Turbo Boost frequencies.
In all other respects, the GM200 chip is exactly the same as the GM204 - the solutions are identical in their capabilities and supported technologies. Even the modules for working with displays and video data were left exactly the same as those of the GM204, on which the Geforce GTX 980 is based. Accordingly, everything that we wrote about the GTX 980 and GTX 970 fully applies to the Titan X.
Therefore, for all other questions about the functional subtleties of the new product, you can refer to the reviews of Geforce GTX 980 and GTX 750 Ti, in which we wrote in detail about the Maxwell architecture, the design of streaming multiprocessors (Streaming Multiprocessor - SMM), the organization of the memory subsystem and some other architectural differences. There you can also familiarize yourself with functionality such as hardware support for accelerating VXGI global illumination calculations, new full-screen anti-aliasing methods and improved DirectX 12 graphics API capabilities.
Solving problems with the development of new technological processes
We can confidently say that everyone in the video card market has long been tired of the 28 nm technical process - we are already in the fourth year of using it, and at first TSMC was unable to take a step forward at all, and then it seemed like they managed to start 20 nm production, but there’s no point in There was no such option for large GPUs - the yield of suitable ones was quite low, and no advantages were found compared to the used 28 nm. Therefore, Nvidia and AMD had to squeeze as much as possible out of existing capabilities, and in the case of Maxwell architecture chips, Nvidia clearly succeeded in this. In terms of power and energy efficiency, GPUs of this architecture have become a clear step forward, to which AMD simply has not responded - at least not yet.
Thus, from the GM204, Nvidia engineers were able to squeeze out much more performance compared to the GK104, at the same level of power consumption, although the chip increased by a third, and the greater density of transistors allowed them to increase their number even more - from 3.5 billion to 5.2 billion. It is clear that under such conditions the GM204 contained much more execution units, which resulted in greater 3D performance.
But in the case of the largest chip of the Maxwell architecture, Nvidia designers could not increase the size of the chip too much, compared to the GK110; it already has an area of about 550 mm 2, and it was not possible to increase its area by a third or at least a quarter - such a GPU would become too complex and expensive to produce. We had to sacrifice something (compared to the older Kepler), and this something was the performance of double precision calculations - its pace in the GM200 is exactly the same as in other Maxwell solutions, although the older Kepler was more universal, suitable for graphics and for any non-graphical calculations.
This decision was not easy for Kepler - too much of the area of this chip was occupied by FP64 CUDA cores and other specialized computing units. In the case of the large Maxwell, it was decided to make do with graphical tasks and it was made simply as an enlarged version of the GM204. The new GM200 chip has become purely graphical, it does not have special blocks for FP64 calculations, and their speed remains the same - only 1/32 of FP32. But most of the GK110 area occupied by FP64 ALUs was freed up and a larger number of FP32 ALUs important for graphics were placed in their place.
This move made it possible to significantly increase graphics (and computing, if we take FP32 calculations) performance compared to the GK110 without increasing power consumption and with a slight increase in the chip area - by less than 10%. It's interesting that Nvidia deliberately chose to separate graphics and computing chips this time. Although the GM200 remains very productive in FP32 calculations, and it is quite possible for Tesla to release specialized solutions for single-precision calculations sufficient for many scientific tasks, the Tesla K40 remains the most productive for FP64 calculations.
This is the difference from the original Titan, by the way - the first solution in the line could also be used for professional purposes for double-precision calculations, since it also has a 1/3 rate for FP64 calculations. And many researchers have used the GTX Titan as a starting card for their CUDA applications and workloads, moving on to Tesla solutions with success. The GTX Titan X is no longer suitable for this; you’ll have to wait for the next generation of GPUs. Unless they are initially separated into graphics and computing chips, of course.
In expansion cards, such a separation already exists - the Tesla K80 model contains a pair of GK210 chips, which are not used in video cards and differ from the GK110 in a doubled register file and shared memory for greater performance in computing tasks. It turns out that the GK210 can be considered exclusively a “computational” processor, and the GM200 can be considered a purely “graphics” processor (with a certain degree of convention, because both GPUs have the same capabilities, just different specializations).
Let's see what happens in the next generations of Nvidia's graphic architectures, produced using a more “fine” technical process - perhaps such a separation will not be needed in them, at least at first. Or vice versa, we will immediately see a strict division into GPU models with different specializations (computing models will have more computing capabilities, and graphics models will have TMU and ROP units, for example), although the architecture will remain the same.
Video card design features
But let's return to the Geforce GTX Titan X. This is a powerful video card designed for PC gaming enthusiasts, so it should have a corresponding appearance - an original and solid design of the board and cooler. Like previous solutions from the Titan line, the Geforce GTX Titan X model is covered with an aluminum case, which gives the video card that premium look - it really looks solid.
The cooling system also looks very impressive - the Titan X cooler uses a copper alloy vapor chamber to cool the GM200 graphics processor. The vapor chamber is connected to a large dual-slot aluminum alloy heatsink, which dissipates the heat transferred from the video chip. The fan removes heated air outside the PC case, which has a positive effect on the overall temperature in the system. The fan is very quiet even during overclocking and long periods of work under load, and as a result, the GTX Titan X, with a power consumption of 250 W, is one of the quietest graphics cards in its class.
Unlike the Geforce GTX 980 reference board, the new product does not contain a special removable plate that covers the back surface of the board - this is done to ensure maximum air flow to the PCB for its cooling. To power the board, a set of one 8-pin and one 6-pin PCI Express additional power connector is used.
Since Geforce GTX Titan X is intended for enthusiasts who prefer solutions with maximum performance, all components of the new video card were selected with this in mind and even with some reserve in capabilities and characteristics.
For example, to provide the graphics processor in the Geforce GTX Titan X with energy, a 6-phase power system with the possibility of additional amplification is used. To ensure the operation of GDDR5 memory, another two-phase power system is additionally used. The 6+2-phase power supply system of the video card provides the model in question with more than enough energy, even taking into account overclocking. Thus, the Titan X reference board is capable of supplying up to 275 W of power to the GPU, provided that the maximum power target value is set to 110%.
Also, to further improve the overclocking potential, the cooling of all components of the new product was improved, compared to the original Geforce GTX Titan video card - the changed design of the board and cooler led to improved overclocking capabilities. As a result, almost all Titan X samples are capable of operating at frequencies up to 1.4 GHz or more - with the same reference air cooler.
The length of the Geforce GTX Titan X reference board is 267 mm, it has the following connectors for image output: one Dual-Link DVI, one HDMI 2.0 and three DisplayPort. Geforce GTX Titan X supports image output to displays with resolutions up to 5K, and is another video card with support for HDMI 2.0, which its competitor still does not have - this allows you to connect the new product to 4K TVs, providing maximum picture quality with a high refresh rate of 60 Hz
Game developer support
Nvidia has always been a company that is known for working very closely with software manufacturers, and especially game developers. Just look at PhysX - the most popular gaming physics effects engine, which has been used for more than 10 years in more than 500 games. The widespread use of PhysX is also due to the fact that it is integrated into some of the most popular game engines: Unreal Engine 3 and Unreal Engine 4. Thus, at the Game Developers Conference 2015, Nvidia announced free access to the source codes of the CPU- oriented part of PhysX 3.3.3 for C++ developers in versions for Windows, Linux, OS X and Android.
Developers will now be able to modify the PhysX engine code however they wish, and the modifications can even then be incorporated into the main Nvidia PhysX code. By opening the source code of PhysX to everyone, Nvidia has given access to its physics engine to an even wider range of game application developers who can use this advanced physics engine in their games.
Nvidia continues to promote another of its technologies - a fairly new algorithm for simulating dynamic global illumination VXGI, which includes support for special hardware acceleration on video cards with second-generation Maxwell GPUs, such as the Geforce GTX Titan X.
The introduction of VXGI into the game will allow developers to provide very high-quality calculation of dynamic global illumination in real time, using all the capabilities of modern GPUs and providing the highest performance. To understand the importance of calculating global illumination (rendering taking into account not only direct illumination from light sources, but also its reflection from all objects in the scene), just look at a couple of pictures - with and without GI enabled:
It is clear that this example is artificial, and in reality game designers use special methods to simulate global shading, placing additional light sources or using pre-calculated lighting - but before VXGI they were either not fully dynamic (pre-calculated for static geometry) or did not have sufficient realism and/or performance. In future games, it is quite possible to use VXGI, and not only on top GPUs.
Game developers really liked the VXGI technique. At the very least, many of them have tried the method in test scenes, are very excited about the results and are considering adding it to their games. And here is another scene with a high-quality calculation of global illumination - it also shows how important it is to take into account the rays of light reflected from all surfaces of the scene:
Until developers implement VXGI into their own engines, you can use a special version of the Unreal Engine 4 VXGI GitHub engine, which is provided to all interested developers - this makes it possible to quickly integrate VXGI into their gaming (and not only!) projects that use this popular game engine - however , this will require some modifications, VXGI cannot just be "turned on".
Let's consider another Nvidia technology - full-screen anti-aliasing using the MFAA method, which provides excellent performance and at the same time acceptable anti-aliasing quality. We have already written about this method and will only briefly repeat the essence and prospects. MFAA support is one of the key capabilities of Maxwell GPUs compared to previous generations of GPUs. Using the ability to program positions for anti-aliasing samples in the MSAA method, these samples change every frame in such a way that MFAA is almost full-fledged MSAA, but with less load on the GPU.
As a result, a picture with MFAA enabled looks almost the same as with MSAA, but the performance loss is significantly lower. For example, MFAA 4x provides speed at the level of MSAA 2x, and the anti-aliasing quality is close to MSAA 4x. Therefore, in those games where the performance is not enough to achieve high frame rates, the use of MFAA will be fully justified and can improve quality. Here is an example of the resulting performance with MSAA and MFAA on the Titan X graphics card compared to the regular Titan (at 4K resolution):
The MFAA anti-aliasing method is compatible with all games that use DirectX 10 and DirectX 11 and have MSAA support (with the exception of rare projects like Dead Rising 3, Dragon Age 2 and Max Payne 3). MFAA can be enabled manually in the Nvidia Control Panel. MFAA is also integrated into Geforce Experience, and this method will automatically be enabled for different games if optimized using Geforce Experience. The only problem is that at the moment MFAA is still not compatible with Nvidia SLI technology, which they promise to fix in future versions of video drivers.
Modern games on Geforce GTX Titan X
With all its power and capabilities, the Geforce GTX Titan X is capable of handling not only current games, but also future projects with support for the upcoming version of DirectX 12. But you can “bring the new product to your knees” even now - with the help of the most demanding modern games at maximum settings quality, with full-screen anti-aliasing enabled and high rendering resolution - like 4K.
At high resolutions and with anti-aliasing enabled, a powerful memory subsystem becomes especially important, and the Geforce GTX Titan X is absolutely fine with it - a 384-bit memory interface and chips operating at an effective frequency of 7 GHz provide a bandwidth of 336.5 GB/ s - although this is not a record, it is pretty decent.
It is also very important that all data fits into video memory, since when MSAA is enabled in 4K resolution, in many games there is simply not enough video memory - more than 4 GB of memory is needed. And Titan X has not just 6 GB, but as much as 12 GB of video memory, because this line is created for those enthusiasts who do not tolerate compromises. It is clear that with such an amount of onboard memory, the player does not need to think about whether the performance of the game in high resolution will decrease when multisampling is enabled - in all games at any settings, 12 GB will be more than enough.
At the moment, in absolutely any game you can set any settings and select any resolutions - Titan X will provide sufficient frame rates under (almost) any conditions. Here are the games Nvidia chose to demonstrate the performance of its solution:
As you can see, frame rates of 40 FPS or more are achieved in most of the “heavy” modern games, with full-screen anti-aliasing enabled, including projects such as Far Cry 4 - in this game, with Ultra settings and anti-aliasing in 4K resolution, you can achieve Acceptable rendering speeds are only possible on Titan X or multi-chip configurations.
And with the release of future games that will support DirectX 12, we can expect an even greater increase in the requirements for GPU and video memory performance - improving the quality of rendering is not “free”. By the way, at that time Nvidia had not yet tested its Titan X video card in the newest game that had recently been released - the PC version of Grand Theft Auto V. This series of games is the most popular among modern projects, in which you act as various criminal elements in the scenery the city of Los Santos, suspiciously similar to the real Los Angeles. The PC version of GTAV was highly anticipated and it finally came out in mid-April - a month after Titan X.
Even the console versions (we're talking about the current generation consoles, of course) of Grand Theft Auto V were quite good in terms of picture quality, and the PC version of the game offers several more opportunities to improve it: a significantly increased drawing distance (objects, effects, shadows), the ability to play at 60 FPS or more, including resolutions up to 4K. In addition, they promise rich and dense traffic, many dynamic objects in the scene, improved weather effects, shadows, lighting, etc.
The use of a couple of Nvidia GameWorks technologies has made it possible to further improve the picture quality in GTAV. Let us remind you that GameWorks is a special platform for game and graphics developers, providing them with 3D technologies and utilities designed for Nvidia video cards. Adding GameWorks technologies to games makes it relatively easy to achieve high-quality simulation of realistic smoke, fur and hair, waves, as well as global illumination and other effects. GameWorks makes it much easier for developers by providing examples, libraries, and SDKs ready to be used in game code.
Grand Theft Auto V uses a couple of these technologies from Nvidia: ShadowWorks Percentage-Closer Soft Shadows (PCSS) and Temporal Anti-Aliasing (TXAA), which improve the already good graphics in the game. PCSS is a special shadow rendering technique that has better quality than typical soft shadow methods. PCSS has three advantages: the degree of softness of shadow edges depends on the distance between the object casting the shadow and the surface on which it is drawn, it also provides better filtering, reducing the number of artifacts in the form of jagged edges of shadows, and the use of a shadow buffer allows you to competently handle intersections of shadows from different objects and prevent the appearance of “double” shadows.
As a result, when PCSS is enabled, the game provides soft, realistic dynamic shadows that are much better quality than what we have seen on game consoles. And for a game like Grand Theft Auto V with the bright sun constantly moving across the horizon, the quality of the shadows is very important, they are always in sight. From the following screenshots you can evaluate the difference between the two highest quality methods used in the game (AMD algorithm versus Nvidia method):
It is clearly visible that the PCSS method allows you to obtain soft edges of shadows, which become progressively blurrier the further the distance between the object from which the shadow is and the surface “receiving” the shadow. At the same time, enabling PCSS has almost no effect on the final performance of the game. While this method provides better shadow quality and realism, there is virtually a performance cost to enabling this option.
Another important addition to the PC version of GTAV is the Nvidia TXAA anti-aliasing method. Temporal Anti-Aliasing is a new anti-aliasing algorithm designed specifically to eliminate the problems with conventional anti-aliasing methods seen in motion - when individual pixels flicker. To filter pixels on the screen using this method, samples are used not only inside the pixel, but also outside it, also together with samples from previous frames, which allows you to obtain “cinematic” filtering quality.
The advantage of the method over MSAA is especially noticeable on objects with translucent surfaces such as grass, tree leaves and mesh fences. TXAA also helps smooth out pixel-by-pixel effects. In general, the method is very high quality and approaches the quality of professional methods used in 3D graphics, but the result after TXAA is slightly blurrier than MSAA, which not all users like.
The performance hit from enabling TXAA depends on the game and conditions, and correlates mainly with the speed of MSAA, which is also used in this method. But compared to pure post-processing anti-aliasing methods like FXAA, which provide maximum speed at lower quality, TXAA is designed to maximize quality at some additional performance penalty. But with a world as rich and detailed as we see in Grand Theft Auto V, turning on high-quality anti-aliasing will be very useful.
The PC version of the game has rich graphics settings that allow you to get the required picture quality with the required performance. Thus, GTAV on a PC provides acceptable rendering speed and quality on all Nvidia solutions, starting with approximately the Geforce GTX 660. Well, to fully enjoy all the graphic effects of the game, it is recommended to use something like a Geforce GTX 970/980 or even Titan X.
To check the settings, a performance test is built into the game - this benchmark contains five scenes close to real gameplay, which will allow you to evaluate the rendering speed of the game on a PC with different hardware configurations. But owners of Nvidia video cards can do it easier by optimizing the game for their own PC using Geforce Experience. This software will select and configure the optimal settings while maintaining playable rendering speeds - all with the click of a button. Geforce Experience will find the best combination of characteristics for both the Geforce GTX 660 with a FullHD monitor and the Titan X with a 4K TV, providing the best settings for a specific system.
Full support for the GTAV game has appeared in the new Geforce driver build version 350.12 WHQL, which has a special optimized profile for this application. This driver version will provide optimal gaming performance, including using other Nvidia technologies: 3D Vision, 4K Surround, Dynamic Super Resolution (DSR), GameStream, G-SYNC (Surround), Multi Frame Sampled Anti-Aliasing (MFAA) , Percentage Closer Soft Shadows (PCSS), SLI and others.
Also, the special driver version 350.12 WHQL contains updated SLI profiles for several games, including a new profile for Grand Theft Auto V. In addition to the SLI profiles, the driver updates and adds profiles for both 3D Vision technology and a profile for GTAV received a rating “Excellent”, which means excellent stereo image quality in this game - owners of appropriate glasses and monitors should try it!
Support for virtual reality technologies
The topic of virtual reality (VR) is now one of the loudest in the gaming industry. Oculus, which was then acquired by Facebook, is largely to blame for the revival of interest in VR. Until some time they showed only prototypes or SDK kits, but they also have plans to release a commercial version of the Oculus Rift helmet at the end of this year. Other companies are also not standing aside. For example, the well-known company Valve announced plans to partner with HTC to release its own virtual reality helmet also by the end of 2015.
Naturally, GPU manufacturers also see promise in VR, and Nvidia works closely with suppliers of software and hardware solutions for virtual reality in order to ensure the most comfortable work with Geforce video cards (or even Tegra, who knows?). And these are not just marketing slogans, because in order for the use of VR to be comfortable, several problems need to be solved, including reducing the delay between the player’s action (head movement) and the resulting display of this movement on the display - too much lag does not just spoil the impression of virtual reality , but can cause so-called motion sickness.
To reduce this latency, Nvidia's VR Direct software supports asynchronous time warp. Using asynchronous time warping, a scene rendered some time ago can shift based on later movements of the player's head, which were captured by the headset sensors. This reduces the delay between the action and the image output, since the GPU does not have to recalculate the entire frame before shifting. Nvidia already provides driver support for VR app developers, and they can implement asynchronous time warping in their software.
In addition to the output delay, it is very important to achieve comfortable gameplay in a virtual reality helmet not just to ensure a high frame rate, but to output frames for each eye with the smoothest possible transition. Accordingly, after the popularization of future generation VR headsets, many players will want to try them out in modern games that are very demanding on GPU power. And in some cases, you will have to create a dual-chip SLI configuration from a pair of powerful video cards like Geforce GTX Titan X.
To ensure maximum comfort in such cases, Nvidia offers VR SLI technology, which allows game developers to assign a specific pair of GPUs to each eye to reduce latency and improve performance. In this case, the image for the left eye will be rendered by one GPU, and for the right eye - by the second GPU. This obvious solution reduces latency and is ideal for virtual reality applications.
So far, VR SLI and asynchronous time warp are not available in public Nvidia drivers, but this is not particularly necessary, since using them requires changing the executable code of the games. And preview versions of GeForce video drivers with support for VR SLI and Async Time Warp are available to select Nvidia partners such as Epic, Crytek, Valve and Oculus. Well, the public driver will be released closer to the release of the final VR products for sale.
Let's add that such a powerful video card as the Geforce GTX Titan X was used in many virtual reality demonstrations at this year's gaming conference for developers - Game Developers Conference 2015. Here are just a few examples: “Thief in the Shadows” - a joint development of Nvidia, Epic, Oculus and WETA Digital - the studio that created visual effects in the film trilogy "The Hobbit", "Back to Dinosaur Island" - a reboot of the famous 14-year-old demo program "X-Isle: Dinosaur Island" from Crytek, as well as demo from Valve: "Portal ", "Job Simulator", "TheBluVR" and "Gallery". In general, it’s up to the release of VR helmets for sale, and Nvidia will be ready for this.
Conclusions on the theoretical part
From an architectural point of view, the new top-end GPU of the second generation of Maxwell architecture turned out to be very interesting. Like its “relatives,” the GM200 takes all the best from the company’s past architectures, gaining additional functionality and all the improvements of the second generation Maxwell. Therefore, functionally, the new product looks just great, corresponding to the models of the Geforce GTX 900 line. With the help of a serious modernization of the execution units, Nvidia engineers achieved in Maxwell a doubling of the ratio of performance to energy consumption, while adding functionality - we recall hardware support for VXGI global illumination acceleration and graphics API DirectX 12.
The top-of-the-line GeForce GTX Titan X is designed for ultra-enthusiast gamers who want maximum quality and performance from the latest PC games, running at the highest resolutions, with maximum quality settings, with full-screen anti-aliasing, all at acceptable frame rates. On the one hand, such a powerful GPU is required by few games, and you can install a couple of less expensive video cards. On the other hand, due to the problems of multi-chip solutions with increased latencies and unsmooth frame changes, many players will prefer one powerful GPU to a pair of less powerful ones. Not to mention the fact that a single-chip card will also provide less power consumption and noise from the cooling system.
Naturally, in such conditions, the main issue with the Geforce GTX Titan X is the price of the solution. But the fact is that it is sold in a niche where the concepts of price justification and price-quality ratio are simply not needed - solutions with maximum performance always cost noticeably more than those close to them, but still not as productive. And the Titan X is an extremely powerful and expensive graphics card designed for those who are willing to pay for maximum speed in 3D applications.
Geforce GTX Titan X is positioned as a premium (luxury, elite - whatever you want to call it) video card, and there should be no complaints about the recommended price - especially since the previous solutions in the line (GTX Titan and GTX Titan Black) initially cost exactly the same - $999 . This is the solution for those who want the fastest GPU available, regardless of its price. Moreover, for the richest enthusiasts and record holders in 3D benchmarks, systems of three or even four Titan X video cards are available - these are simply the fastest video systems in the world.
These are the requests Titan X fully justifies and provides - the top-end new product, even alone, shows the highest frame rates in all gaming applications and in almost all conditions (resolution and settings), and the amount of fast GDDR5 video memory of 12 GB allows you not to think about the lack of local memory several years ahead - even games of future generations, with support for DirectX 12, etc., simply will not be able to fill this memory so much that it will not be enough.
As with the first GTX Titan in 2013, the new GTX Titan X sets a new benchmark for performance and functionality in the premium graphics card segment. At one time, the GTX Titan became a fairly successful product for Nvidia, and there is no doubt that the GTX Titan X will repeat the success of its predecessor. Moreover, based on the largest video chip of the Maxwell architecture, the model has become the most productive on the market without any reservations. Since video cards like the GTX Titan X are produced by Nvidia itself and sell reference samples to its partners, there are no problems with availability in stores from the very moment of its announcement.
GTX Titan X meets its highest level in all respects: the most powerful GPU of the Maxwell family, excellent graphics card design in the style of previous Titan models, as well as an excellent cooling system - efficient and quiet. In terms of 3D rendering speed, this is the best video card of our time, offering more than a third more performance compared to the best models released before the Titan X - like the Geforce GTX 980. And if we don’t consider dual-chip video systems (like a pair of the same GTX 980 or one Radeon R9 295X2 from a competitor that has problems inherent in multi-chip configurations), then Titan X can be called the best solution for wealthy enthusiasts.
In the next part of our material, we will examine the rendering speed of the new Nvidia Geforce GTX Titan X video card in practice, comparing its speed with the performance of the most powerful video systems from AMD and with the speed indicators of predecessors produced by Nvidia, first in our usual set of synthetic tests, and then in games.
