Geforce gtx 780 ti gaming tests. Video cards. Tests in games

Test Methodology

Test stand configuration
CPU	Intel Core i7-3960X @ 4.6 GHz (100x46)	Intel Core i7-3970X @ 4.6 GHz (100x46)
Motherboard	ASUS P9X79 Pro
RAM	DDR3 Kingston HyperX 4x2 GB @ 1600 MHz, 9-9-9
ROM	Intel SSD 520 240 GB
Power Supply	Corsair AX1200i 1200W	Seasonic Platinum-1000, 1000 W
CPU cooling	Thermalright Archon
Frame	CoolerMaster Test Bench V1.0
Operating system	Windows 7 Ultimate X64 Service Pack 1
AMD Card Software	AMD Catalyst 13.11 Beta v8 (for Radeon R9 290); v5 (for others) + Catalyst Application Profiles 13.5 CAP1
NVIDIA card software	331.70 (for GeForce GTX 780) 331.58 WHQL (for others)

To measure the power of the system, a stand with a Corsair AX1200i power supply is used. Power-saving CPU technologies are disabled in all tests. The PCI-Express bus operates in 3.0 mode. To activate PCI-E 3.0 on GeForce 600 and 700 series video cards in a system based on the X79 chipset, a patch from NVIDIA is applied.

In settings NVIDIA drivers the CPU is always selected as the processor for the PhysX calculation. In the AMD settings, the Tesselation setting is always moved from the AMD Optimized state to Use application settings.

A set of benchmarks
Program	API	Settings	Anisotropic filtering, full-screen anti-aliasing	Permission
3D Mark 2011	DirectX 11	Extreme Profile	-	-
3DMark	DirectX 11	Fire Strike test (not Extreme)	-	-
Unigine Heaven 2	DirectX 11	DirectX 11 max. quality, tessellation in Extreme mode	AF 16x, MSAA 4x	1920x1080 / 2560x1440
Crysis Warhead + Framebuffer Crysis Warhead Benchmarking Tool	DirectX 10	DirectX 10 max. quality. Frost flythrough demo	AF 16x, MSAA 4x	1920x1080 / 2560x1440
Battlefield 3 + FRAPS	DirectX 11	Max. quality. Start of the Going Hunting mission	AF 16x, MSAA 4x + FXAA	1920x1080 / 2560x1440
Batman: Arkham City. Built-in benchmark	DirectX 11	Max. quality	AF, MSAA 4x	1920x1080 / 2560x1440
DiRT Showdown. Built-in benchmark	DirectX 11	Max. quality, Global Illumination incl. Shibuya track, 8 cars	AF, AA 4x	1920x1080 / 2560x1440
Far Cry 3 + FRAPS	DirectX 11	DirectX 11 max. quality, HDAO. Beginning of the Secure the Outpost mission	AF, MSAA 4x	1920x1080 / 2560x1440
Tomb Raider. Built-in benchmark	DirectX 11	Max. quality	AF 16x, SSAA 4x	1920x1080 / 2560x1440
Bioshock Infinite. Built-in benchmark	DirectX 11	Max. quality. postprocessing: normal	AF 16x, FXAA	1920x1080 / 2560x1440
Crysis 3 + FRAPS	DirectX 11	Max. quality. Beginning of the Post Human mission	AF 16x, MSAA 4x	1920x1080 / 2560x1440
Metro: Last Light. Built-in benchmark	DirectX 11	Max. quality	AF 16x, SSAA 4x	1920x1080 / 2560x1440

⇡ Test participants

The following video cards took part in performance testing:

⇡ Overclocking, temperature, power consumption

In normal mode, the GeForce GTX 780 Ti maintains fairly high clock speeds under load. During the test run of Crysis 3, the GPU ran at 1006-1020 MHz (not much more than the GTX 780 and GTX TITAN) with occasional drops to 993 MHz. At the same time, the maximum voltage on the GPU was 1.187 V, while the GTX 780 and GTX TITAN are content with 1.162 V.

If we compare the GTX 780 Ti with the Radeon R9 290X, then it turns out that both video cards (assuming the R9 290X is running in Uber Mode with a higher fan speed) stably keep the frequency around 1 GHz under load.

But the GK110 processor at the same time maintains a temperature within 83 ° C and the impeller of its cooling system spins up to 2336 rpm, and Hawaii warms up to 93 ° C, while the cooler rotates at a speed of 2727 rpm.

It's funny, but judging by the test in Crysis 3, the GeForce GTX 780 Ti slightly outperforms the Radeon R9 290X in peak power consumption, and the cooling systems of the rivals look the same in terms of power, but the R9 290X clearly has problems with heat dissipation. Perhaps the reason is the different crystal areas of GK110 and Hawaii: the first has 533 mm 2, and the second has 455 mm 2, hence the different heat flux density.

In addition, it should be noted that the cooling system of the GTX 780 Ti is good not only for its exquisite design and luminous inscription, but also for its excellent acoustic characteristics. Even at 4261 rpm, which is the maximum for the impeller, the noise cannot be compared with the roar of the stock Radeon R9 290X cooler at the same speed. In addition, NVIDIA's noise spectrum is softer and more pleasant to the ear.

In terms of overclocking capabilities, the GeForce GTX 780 Ti is inferior to the “vanilla” version of the GTX 780, which is quite expected, given the additional heat dissipation from the three previously dormant SMXs and the increased voltage on the GPU. In addition, the new flagship has a meager TDP margin - it can be increased only by 6%. Just in case, by running the fan at full speed so that the temperature does not affect the dynamics of the frequency, and prioritizing the temperature in EVGA Precision, we were able to increase the base frequency to 986 MHz, and the maximum recorded value increased to 1130 MHz. The frequency under load is quite stable - during the test in Crysis 3, it did not fall below 1111 MHz.

Judging by the state of the Power/Voltage/OV/Utilization Limit indicators, which NVIDIA has equipped with video cards with GPU Boost 2.0 technology, the power consumption limit and at the same time the GPU voltage become an obstacle to overclocking a video card. Attempts to turn the Overvoltage option simply did not lead to anything: the specified voltage rise by 75 mV was not registered on the graphs.

	Base clock, MHz	Max. Boost Clock, MHz	Base Clock, MHz (overclocking)	Max. registered Boost Clock, MHz (overclocking)
GeForce GTX 780 Ti	876	1020 (+144)	986	1130 (+144)
GeForce GTX TITAN	836	1006 (+145)	966	1150 (+184)
GeForce GTX 780	863	1006 (+143)	1053	1215 (+162)

Performance: synthetic benchmarks

In this test, the GeForce GTX 780 Ti is predictably faster than all single-processor video cards, and even surpasses the GeForce GTX 690 in the 3DMark Score parameter.

But in a more recent version of the benchmark, the results of the GeForce GTX 780 Ti are strikingly exactly the same as what the Radeon R9 290X achieved.
Once again, the GTX 780 Ti scored higher than the GTX 690.

Unigine Heaven 2

The GTX 780 Ti hits the heels of both dual-GPU graphics cards.
The difference between the GTX 780 Ti and the Radeon R9 290X is quite large. Thanks to fast geometry processing, which is more important than anything else in this test, rich in hardware tesslation.

⇡ Performance: gaming tests

Crysis Warhead

The superiority of the GTX 780 Ti over previous video cards with the GK110 chip is not as pronounced as in other tests.
The GTX 780 Ti is slightly, but still, slower than the Radeon R9 290X.

Battlefield 3

Here the GTX 780 Ti once again confidently outperforms the R9 290X.

Batman Arkham City

The GTX 780 Ti is the king of this benchmark. He managed to outdo not only the Radeon R9 290X, but also the “two-headed” GeForce GTX 690.

DiRT Showdown

The advantage again went to the side of the Radeon R9 290X.
However, within the NVIDIA team on the GK110 platform, the new flagship is significantly superior to the "regular" GeForce GTX 780.

Far Cry 3

Far Cry 3 is one of those games where the difference between the GeForce GTX 780 Ti and the Radeon R9 290X is the biggest (in favor of NVIDIA).
Moreover, the GTX 780 Ti performed almost as well as the Radeon HD 7990.

tomb raider

Another devastating defeat for the Radeon R9 290X.

Bioshock Infinite

Here the GTX 780 Ti is still ahead of the R9 290X, although the difference is less pronounced.

Crysis 3

Of the two graphics cards - the GeForce GTX 780 Ti and the Radeon R9 290X - the former is faster, and by a lot.
In fact, the GTX 780 Ti is approaching the positions of the Radeon HD 7990.

Metro: Last Light

The Radeon R9 290X holds up confidently, losing no more than 1 FPS to its opponent.
The GeForce GTX 780 Ti itself is almost equal to the GeForce GTX 690.

⇡ Performance: overclocking

Overclocking the GeForce GTX 780 Ti significantly affects the scores in synthetic 3DMark tests, but in games the increase in frequency of great importance does not have. Like it or not, overclocking an NVIDIA GPU with GPU Boost technology is not a very fruitful exercise. Moreover, the GTX 780 Ti is very limited by the TDP and does not provide an opportunity to increase the supply voltage, as a result, the frequency increase was not as large as in the case of the GeForce GTX 780 in the basic version.

⇡ Game tests + Unigine Heaven 2

⇡ Conclusions

In the $400 to $550 price range normally occupied by enthusiast gaming graphics cards, AMD is stronger than ever. In fact, for the first time in the history of confrontation between the two main manufacturers of discrete GPUs, AMD video adapters not only are not inferior to competitors of equal cost offered by NVIDIA, but even noticeably outperform them in performance. But NVIDIA had secret weapon- Fully unlocked GK110 GPU. While the GeForce GTX 780 can't be made cheap enough to deliver the same performance for $400 (how much does the Radeon R9 290 cost), the Greens are able to make a graphics card that's much more expensive than the Radeon R9 290X, but it's unequivocally outperforms the latter in performance.

It's just that the GK110 initially had a margin of scaling up to the heights that the GeForce GTX 780 Ti demonstrated today, and the Hawaii GPU in the Radeon R9 290X has already shown its full potential. First, of course, the GK110 includes a billion more transistors than the Hawaii. But, no less important, NVIDIA managed this budget differently. AMD has invested in a lot of ROPs, making the R9 290X more capable of 4K gaming, but NVIDIA has put more emphasis on texture filtering units. The advantage in geometry processing is again on the side of NVIDIA's top-end GPU. As the tests show, at resolutions up to 2560x1440 the last parameters are still more important than the pixel fill rate.

It is curious that the GTX 780 Ti and Radeon R9 290X are very close in terms of GPU clock speeds and, apparently, have approximately the same power consumption (at least 280-290 W), which in itself speaks in favor of NVIDIA, because the GK110 is more complex and high performance chip than Hawaii. But with all this, the NVIDIA video card also heats up less. We can't help but praise the reference cooler that the GTX 780 Ti inherited from its predecessors, but it's clearly not just about the cooling system. It looks like the GK110 itself is more efficient at dissipating heat.

However, judging from a purely practical point of view, the GeForce GTX 780 Ti and Radeon R9 290X are not direct competitors due to the price difference of $150. The new NVIDIA flagship, like the basic version of the GTX 780 earlier, turned out to be the only representative of its own price category for enthusiasts who are ready to pay extra money in excess of the usual $500-550 that make up the price of a High-End video card. As such, he is unlikely to have an alternative until the GPU production masters the 20 nm process technology.

Finally, let's say a word about the GeForce GTX TITAN, which until now has been considered the most productive single-processor video card from NVIDIA. After the release of the GeForce GTX 780 Ti, it will continue to be sold, and even there were no official statements about a price reduction. Apparently, the TITAN will still cost $1,000, but is now being touted as a "prosumer" video adapter with the ability to perform double precision non-graphics calculations at 1/3 the speed of an FP32. In fact, a kind of analogue of the Quadro K6000 for non-pros. The value of TITAN as a gaming graphics card after the release of the GeForce GTX 780 Ti has been completely destroyed.

Due to the release of the AMD Volcanic Islands family of video cards, a hard boom began in the graphics adapter market. Nvidia even lowered the prices of their top graphics cards, but that didn't help either. I had to do something and somehow change the situation in my favor. We needed a special model that would take the flagship - AMD Radeon R9 290X and removed it from the list of competitors.

In principle, nVidia has such a video card - this is the GeForce GTX TITAN. This is the fastest video card on a single chip at the time of release. But its price is too high and for most it is simply inaccessible. Even not every gamer decides to buy one. After all, for this money you can easily buy a good computer with excellent performance.

And now, nVidia has finally released a video card that is better than its competitor from AMD and more affordable - the GeForce GTX 780 Ti. She also became the head of the GTX 7xx family of video cards.

It can be assumed that this is a kind of trump card that the greens kept up their sleeves. A certain pause was simply waited for and the right moment was chosen for the release.

Specifications

Three out of fifteen large blocks from the Kepler architecture - streaming multiprocessors - were returned to this video card. It is much more powerful, especially in terms of shader instructions and texturing.

It also supports 7 GHz video memory, thereby increasing the bandwidth by as much as 14%.

Reference version dimensions

Height: 26.67 cm / 10.5 inches
Length: 11.16 cm / 4.376 inches
Width: Two slots

GPU Specifications

CUDA cores: 2880
Base clock: 875
Boost Clock: 928
Texture Fill Rate: 210 (GigaTexels/s)

Memory specifications

Memory speed (Gbps): 7.0
Memory capacity: 3072 MB
Memory interface: 384-bit GDDR5
Maximum Memory Bandwidth: 336

Capabilities

FXAA and TXAA: +
Support for NVIDIA SLI technology: +
pure video: +
3D Vision: +
PhysX: +

Software environment: CUDA

DirectX: 11
OpenGL: 4.3
Bus: PCI-E 3.0
3D Games: +
Blu Ray 3D: +

Specifications

Maximum digital resolution: 4096×2160
Maximum VGA resolution: 2048×1536
Media connection: HDMI, DisplayPort, Dual Link DVI-I, Dual Link DVI-D
Multi-Display Support: +
HDCP: +
HDMI: +
HDMI Audio Input: Internal

Power and temperature

Maximum temperature: 95
Power consumption: 250 W
Minimum system requirements by power: 600 W
Power connectors: 6-pin & 8-pin

Comparison of GTX 780 TI and GTX 780

	GeForce GTX 780	GeForce GTX 780 Ti
GPU	GK110	GK110
Number of transistors	7.1 billion	7.1 billion
Process technology, nm	28	28
GPU clock frequency, MHz: Base Clock / Boost Clock	863/900	875/928
Stream Processors	2304	2880
texture blocks	192	240
Rasterization Units (ROPs)	48	48
Video memory: type, size, MB	GDDR5, 3072	GDDR5, 3072
Memory clock frequency: real (effective), MHz	1753 (7010)	1502 (6008)
Memory bus width, bit	384	384
Interface	PCI-Express 3.0 x16
Image Output
Interfaces	1 x DL DVI-I 1 x DL DVI-D 1 x HDMI 1.4a 1 x Display Port 1.2
Maximum Resolution	VGA: 2048×1536 DVI: 2560×1600 HDMI: 4096x2160, DisplayPort: 4096×2160
Maximum power consumption, W	250	250

Manufacturers

MSI GTX 780 Lightning 3 GB

It has a core frequency: 980 MHz and a memory frequency of 1502 MHz

MSIGTX 780TwinFrozrGaming 3GB

ASUS GTX 780 DirectCU II OC 3 GB

It has a core frequency: 889 MHz and a memory frequency of 1502 MHz

EVGA GTX 780 Superclocked w/ ACX Cooler 3 GB

It has a core frequency: 967 MHz and a memory frequency of 1502 MHz

Video card Gigabyte GTX 780 WindForce OC 3 GB

It has a core clock of 954 MHz and a memory clock of 1502 MHz

PalitGTX 780SuperJetStream 3GB

It has a core frequency of 980 MHz and a memory frequency of 1550 MHz

MSI GTX 780 Gaming 6 GB

It has a core clock of 902 MHz and a memory clock of 1502 MHz

Comparison with competitors

Specifications/Model	AMD Radeon R9 290X	nVidia GeForce GTX 780	nVidia GeForce GTX 780 Ti	nVidia GeForce GTX Titan
GPU	Hawaii XT	GK110 (GK110-300-A1)	GK110 (GK110-425-B1)	GK110 (GK110-400-A1)
Process technology	28 nm	28 nm	28 nm	28 nm
Number of transistors	6.2 billion	7.1 billion	7.1 billion	7.1 billion
GPU clock speed (base frequency)	—	864 MHz	876 MHz	837 MHz
GPU clock speed (Boost frequency)	1.000 MHz	902 MHz	928 MHz	876 MHz
Memory frequency	1.250 MHz	1.502 MHz	1.750 MHz	1.502 MHz
Memory type	GDDR5	GDDR5	GDDR5	GDDR5
Memory size	4.096 MB	3.072 MB	3.072 MB	6.144 MB
Memory bus width	512 bit	384 bit	384 bit	384 bit
Memory Bandwidth	320.0 GB/s	288.4 GB/s	336 GB/s	288.4 GB/s
DirectX Version	11.2	11.1	11.1	11.1
Stream Processors	2.816	2.304	2.880	2.688
texture blocks	176	192	240	224
Raster Operations Pipelines (ROPs)	64	48	48	48
TDP	> 250W	250 W	250 W	250 W

Comparing the 780 Ti with its competitors, you can understand its pros and cons. For example, the number of streaming processes is not much different from others, but the throughput is still slightly higher. But here AMD has a much higher amount of video memory, Boost frequency and pixel fill rate. Now it is PowerTune and GPU Boost that greatly affect performance. And it is not so easy to evaluate it. So comparing all these basic characteristics is not so important.

Overclocking and tests

Overclocking the GTX 780 Ti is something. The video card has significant hidden reserves.

The company's engineers did their best. Due to a more advanced chip and an improved power system, you can overclock it by 200 MHz without raising the voltage.

Tests in benchmarks

Tests in games

Crysis 3

In Crysis 3, the video card showed excellent results, up to 36 fps at maximum settings.

Total War ROME II

This game is just crammed with various modern graphic bells and whistles.

In FullHd resolution, fps ranged from 65-99 at maximum settings.

Call of Duty Ghosts

In FullHd resolution, fps ranged from 132 to 148 at maximum settings.

conclusions

We can say that the GTX 780 Ti is an example of a high quality graphics card that is designed for serious games. Nvidia has taken a step forward in the development of the Kepler family of GPUs. It can be argued that this architecture has reached its ceiling.

Having released the video card GeForce GTX TITAN ($999) in February 2013 and supplementing it in May with the more affordable GeForce GTX 780 ($649), NVIDIA could rest on its laurels until the end of October, collecting all the "cream" in the enthusiast market - overclockers and the wealthiest gamers. Now no one can say how long this, from NVIDIA's point of view, idyll would have continued, but it was the announcement of the AMD Radeon R9 290X that prompted the company from Santa Clara, firstly, to seriously reduce the recommended prices for the GeForce GTX 780/GTX 770 and , secondly, to finally release a new flagship - a video card NVIDIA GeForce GTX 780 Ti.

Now you can safely forget about TITAN, and NVIDIA fans should urgently get rid of their GeForce GTX 780, since the new Ti-version does not leave them a single chance at once in all key characteristics. Better/faster/cheaper based on the Kepler architecture and the 28-nm process technology NVIDIA will definitely not release anything, so we have every reason to call the GeForce GTX 780 Ti the real crown of the Kepler architecture, its apogee and extravaganza. However, let's talk about everything in order.

1. NVIDIA GeForce GTX 780 Ti video card review

specifications and recommended price

The technical characteristics of the GeForce GTX 780 Ti graphics card are shown in the table in comparison with the reference versions of the GeForce GTX TITAN, GeForce GTX 780 and AMD Radeon R9 290X:

As you can see from the specifications table, the key difference between the new graphics card is the graphics processor. The GeForce GTX 780 Ti is equipped with a full-fledged Kepler GK110 architecture, containing all 15 streaming multiprocessors (SMX), like the professional Tesla K40 accelerators, 2880 unified shader processors and 240 texture units. This is 25% more than the regular GeForce GTX 780 and 7.1% more than the TITAN. In addition, the frequencies of the graphics processor have slightly increased, which are now 876/928 MHz. But the increase in the frequency of video memory is much more impressive and amounts to almost 1000 effective megahertz (7000 MHz versus 6008 MHz for TITAN or 780).

In addition, the GeForce GTX 780 Ti supports the latest Power Balancing technology, which is to organize the most balanced nutrition video cards due to the automatic distribution of power from its various sources (external/PCI-E) and should help increase the overclocking potential. Later, NVIDIA promises to implement this technology in GeForce GTX TITAN video cards with GTX 780 (apparently, with new GPU revisions). And, finally, GPU Boost version 2.0 already known to us, according to NVIDIA, makes the new video card unique in terms of achieving maximum frequencies/performance under high 3D loads.

Of the software updates, we note support for the GeForce GTX 780 Ti new technology shadowplay. Using the hardware support of the H.264 NVENC encoder by the Kepler family of GPUs, this technology allows capturing game footage. With a minimal impact on performance (less than 10% - according to NVIDIA), ShadowPlay allows you to record videos at 60 frames per second, but so far only at 1920x1080 (1080p). Possibly, in the future possibilities of video capture will be expanded. In addition to ShadowPlay, we also highlight the new hardware and software technology for vertical synchronization G-SYNC. We plan to pay special attention to both NVIDIA innovations in one of the future articles.

PCB design and features

Of the external differences between the new GeForce GTX 780 Ti and the GTX 780 and GTX TITAN, only the inscription on the front casing of the cooling system, a darker heatsink under the glass, and the absence of memory chips on the back of the printed circuit board can be noted:

The upper part of the cooler housing retains the illumination of the GEFORCE GTX symbols, which will surely please aesthetes. The dimensions of the video card are standard for this line of 267x100x39 mm, and it weighs 922 grams.

The GeForce GTX 780 Ti is equipped with four video outputs: DVI-I and DVI-D (both Dual-Link), HDMI version 1.4a and DisplayPort version 1.2:

From the reverse side of the video card, you can see an open radiator and screws securing the cooling system casing:

In general, everything looks like the GeForce GTX 780 or GTX TITAN, including additional power connectors and connectors for SLI modes:

It is noteworthy that despite the increased number of SMX multiprocessors, increased frequencies of the graphics processor and video memory, NVIDIA for the new GeForce GTX 780 Ti leads the same maximum level of heat dissipation as the GTX 780 or GTX TITAN, equal to 250 watts. This was probably done for marketing reasons, or maybe the new revision of the B1 GPU turned out to be really more economical?

The GeForce GTX 780 Ti circuit board resembles the GeForce GTX TITAN / 780 boards in design and layout:

However, the power section clearly has eight phases of the GPU power supply on DrMOS power transistors instead of six for the GTX TITAN / 780:

At the same time, the control controller has not changed - it is the well-proven NCP4206 manufactured by ON Semiconductor:

GPU die with an area of 561 sq.mm. released in Taiwan on the 34th week of 2013 (mid-August) and, most importantly, refers to a brand new B1 revision:

Marking - GK110-425-B1 (recall that TITAN crystals have the marking GK-110-400-A1). The frequencies of the GeForce GTX 780 Ti GPU in 3D mode are 876/928 MHz at a voltage of 1.175 V. When switching to 2D mode, the frequency drops to 324 MHz, and the voltage drops to 0.875 V. However, as practice shows, voltages can vary each copy of the GK110. We told you about the characteristics of the GPU in the previous subsection of the article.

Let's add that the ASIC quality of our GPU turned out to be 77.8%:

Like the GeForce GTX 780, on the front side of the GeForce GTX 780 Ti printed circuit board, 12 GDDR5 memory chips are soldered in an FCBGA package with a total volume of 3 GB. The chips are manufactured by SK Hynix (all TITANs and GTX 780s were Samsung, mind you) and are labeled H5GQ2H24AFR-R2C :

These are Hynix's fastest mass-produced chips at the moment, capable of operating at an effective frequency of 7000 MHz at a voltage of 1.5 V, on which the GeForce GTX 780 Ti memory functions. With a 384-bit memory bus, this results in a theoretical bandwidth of 336 GB/s, which is even higher than the Radeon R9 290X with its 512-bit bus and 5000 MHz memory (320 GB/s).

The latest version of the GPU-Z utility available at the time of writing this article is already familiar with the characteristics of the GeForce GTX 780 Ti:

cooling system efficiency

The cooling system of the GeForce GTX 780 Ti has not undergone any design changes. This is the same TITAN/GTX 780 cooler with an evaporative chamber at the base and an aluminum radiator, a metal substrate, and a radial fan with adaptive speed control:

However, there is still one change, even if it is not constructive - the heatsink of the GeForce GTX 780 Ti is nickel-plated, which is clearly visible both through the transparent glass on top and through the end of the video card, where heated air is ejected outside the system unit case:

According to the monitoring data, the fan speed can be adjusted in the range from 1000 to 4250 rpm.

For check temperature regime as a load, we used five cycles of the test of a very resource-intensive game Aliens vs. Predator (2010) at maximum graphics quality at a resolution of 2560x1440 pixels with 16x anisotropic filtering and MSAA4x:

MSI Afterburner version 3.0.0 beta 16 and GPU-Z utility version 0.7.4 were used to monitor temperatures and other parameters. All tests were carried out in a closed case of the system unit, the configuration of which you can see in the next section of the article, at a room temperature of 25 degrees Celsius.

So, let's look at the capabilities of the GeForce GTX 780 Ti cooler in automatic mode:

Auto mode

As a result, we got 86 degrees Celsius peak temperature at top speed fan 3090 rpm. Recall that the GeForce GTX TITAN, equipped with exactly the same cooling system, warmed up up to 87 degrees Celsius at a maximum fan speed of 3040 rpm, and the GeForce GTX 780 - up to 82 degrees Celsius at 2350 rpm. So the new GeForce GTX 780 Ti has become both hotter and, as a result, noisier than the GTX 780. In addition, you should pay attention to the fact that the frequency of the GeForce GTX 780 Ti GPU during testing was constantly equal to 1020 MHz, instead of those stated in boost mode 928 MHz. We add that the Power Limit during the tests was increased to 106%.

With the GeForce GTX 780 Ti manually set to the maximum fan speed, the peak temperature of the GPU dropped to 71 degrees Celsius:

Max Speed

True, the speed of 4250 rpm, at which the cooler operated, is too noisy for a gaming video card, but still a home video card.

overclocking potential

Do you know what is the most important thing in the GeForce GTX 780 Ti? The Kepler core is not at all full-fledged in terms of the number of streaming multiprocessors and blocks, not high-frequency memory with a record bandwidth, and not even a nickel-plated cooling system radiator, but a core revision - B1. Because if someone thinks that “all the juices” have already been squeezed out of the GK110-425-B1 GPU in the GeForce GTX 780 Ti, then this is a delusion, and from the point of view of overclocking, the delusion is almost criminal.

The fact is that, as it turned out, even with extremely limited opportunities to increase the Power Limit (only 106%), the complete absence of the possibility of increasing the voltage and using a rather mediocre standard cooling system in terms of efficiency, the graphics processor was able to function stably with an increase in frequency immediately at 240 MHz! And this, mind you, we are talking about about base frequency . For the same GeForce GTX TITAN, this would be a record value "in the air", and in the GTX 780 only the best, most successful specimens with original coolers are capable of such an impressive overclocking. Here immediately +240 MHz, and even +760 MHz for video memory:

The frequencies of the video card after overclocking were 1116/1168/7760 MHz:

At the same time, in fact, according to the monitoring data, during a high load, the frequency of the GPU reached 1260 (!) MHz, although most often it was around 1246 MHz:

For example, after overclocking the MSI GeForce GTX 780 Lightning, literally stuffed with various proprietary technologies, we managed to reach a frequency of 1227 MHz, and this is lower than today's result of the reference sample GeForce GTX 780 Ti. One can only guess what the nominal frequencies will be for serial GeForce GTX 780 Ti with high-performance original coolers, and what will be their overclocking. We add to our results that in order to achieve them, we were forced to manually set the fan speed at around 85% or 3530 rpm, which, of course, is noisy, but otherwise the standard cooling system simply could not cope with the cooling of a large and hot processor , and the frequencies decreased.

Energy consumption

Measurement of system power consumption with various video cards was carried out using the Zalman ZM-MFC3 multifunctional panel, which shows the consumption of the system "from the outlet" as a whole (excluding the monitor). The measurement was carried out in 2D mode, during normal work in Microsoft Word or Internet "surfing", as well as in 3D mode. In the latter case, the load was created using four consecutive cycles of the "Swamp" intro scene from the game Crysis 3 at 2560x1440 pixels at maximum graphics quality settings, but without using MSAA anti-aliasing.

To compare the power consumption level of NVIDIA GeForce GTX 780 Ti, we included in the test systems with NVIDIA GeForce GTX TITAN video cards at nominal frequencies and during overclocking, the reference NVIDIA GeForce GTX 690, EVGA GeForce GTX 780 Superclocked ACX at frequencies reduced to the level of the reference GeForce GTX 780, and also an AMD Radeon HD 7990 and a CrossFireX bundle of two MSI Radeon R9 280X. Here's what happened:

The power consumption of the NVIDIA GeForce GTX 780 Ti is expected to have increased compared to its predecessors, but it cannot be said that much. For example, a system with a GeForce GTX 780 Ti draws only 29 watts more than a GTX 780 and 10 watts more than a GTX TITAN. Moreover, even an overclocked GeForce GTX 780 Ti as part of a fairly fast configuration easily fits into 600 watts, and such power supply units, as we know, are not out of the ordinary today. Against this background, the level of power consumption of systems with dual-processor graphics cards NVIDIA GeForce GTX 690 and, in particular, AMD Radeon HD 7990 looks much more impressive. Well, the record holder in terms of power consumption today is the CrossFireX tandem of two MSI Radeon R9 280X, which will require a high-quality power supply power of at least 750 watts.

Let's supplement the review of the video card with a link to the BIOS, and move on to the testing methodology with a test configuration.

2. Test configuration, tools and testing methodology

Video card performance testing was carried out on a system with the following configuration:

Motherboard: Intel Siler DX79SR (Intel X79 Express, LGA 2011, BIOS 0590 from 07/17/2013);
CPU: Intel Core i7-3970X Extreme Edition 3.5/4.0 GHz(Sandy Bridge-E, C2, 1.1V, 6x256KB L2, 15MB L3);
CPU cooling system: Phanteks PH-TC14PE (2xCorsair AF140, 900 rpm);
Thermal interface: ARCTIC MX-4 ;
Video cards:

NVIDIA GeForce GTX TITAN 6 GB 837/876/6008 MHz and 972/1011/7288 MHz;
NVIDIA GeForce GTX 690 2x2 GB 915/1020/6008 MHz;
EVGA GeForce GTX 780 Superclocked ACX 3 GB (03G-P4-2784-KR) 863/916/6008 MHz;
AMD Radeon HD 7990 2x3 GB 1050/6000 MHz;
CrossFireX MSI Radeon R9 280X Gaming 2x3 GB 1050/6000 MHz;

RAM: DDR3 4x8GB G.SKILL TridentX F3-2133C9Q-32GTX(2133 MHz, 9-11-11-31, 1.6 V);
System Drive: 256 GB Crucial m4 SSD (SATA-III, CT256M4SSD2, BIOS v0009);
Drive for programs and games: Western Digital VelociRaptor (SATA-II, 300 GB, 10000 rpm, 16 MB, NCQ) in a Scythe Quiet Drive 3.5" box;
Backup disk: Samsung Ecogreen F4 HD204UI (SATA-II, 2 TB, 5400 rpm, 32 MB, NCQ);
Case: Antec Twelve Hundred (front wall - three Noiseblocker NB-Multiframe S-Series MF12-S2 at 1020 rpm; back - two Noiseblocker NB-BlackSilentPRO PL-1 at 1020 rpm; top - standard 200 mm fan at 400 rpm);
Control and monitoring panel: Zalman ZM-MFC3 ;
Power supply: Corsair AX1200i (1200W), 120mm fan;
Monitor: 27" Samsung S27A850D (DVI-I, 2560x1440, 60 Hz).

Of course, first of all, we will compare the new product with the most expensive single-processor NVIDIA video card - GeForce GTX TITAN, which was tested at nominal frequencies and during overclocking:

An even more formidable rival to the GeForce GTX 780 Ti is the dual-processor GeForce GTX 690, although it was released more than a year and a half ago, it is still one of the fastest video cards of our time. And since we included a dual-processor NVIDIA video card in testing, then AMD cannot be ignored - it is presented today by the AMD Radeon HD 7990:

Note that if the frequencies of the GeForce GTX 690 are nominal, then the Radeon HD 7990 has them slightly raised (+50 MHz for the core). This was done not for any selfish purpose, but in order to compare this video card in the next article with the performance of a pair of new MSI Radeon R9 280X Gaming in CrossFireX mode, whose GPU frequency is exactly 1050 MHz. We hope you will forgive us for taking this liberties in today's article.

To compare the GeForce GTX 780 Ti with the regular GeForce GTX 780, the tests included the EVGA GeForce GTX 780 Superclocked ACX video card, the GPU frequencies of which were brought to the reference ones for the GeForce GTX 780. In turn, the CrossFireX bundle of two MSI Radeon R9 280X Gaming functioned at frequencies 1050/6000 MHz:

Here is a rather large-scale comparison we are planning today. Unfortunately, there will be no new AMD Radeon R9 290X/R9 290 in this test, since the company did not send these video cards for testing. All questions and angry reviews about this, please direct directly to AMD.

As usual, to reduce the dependence of video card performance on platform speed, the 32nm six-core processor at a multiplier of 48, a reference frequency of 100 MHz and Load-Line Calibration activated was overclocked to 4.8 GHz by increasing the voltage in the motherboard BIOS to 1.38 AT:

Hyper-Threading Technology is enabled. At the same time, 32 GB of RAM operated at a frequency of 2.133 GHz with timings of 9-11-11-20_CR1 at a voltage of 1.6125 V.

Testing, which began on November 2, 2013, was conducted under the control of operating system Microsoft Windows 7 Ultimate x64 SP1 with all critical updates as of the specified date and with the following drivers installed:

motherboard chipset Intel Chipset Drivers - 9.4.4.1006 WHQL dated 21.09.2013;
DirectX End-User Runtimes libraries, release date November 30, 2010;
video card drivers for NVIDIA GPUs - GeForce 331.70 Beta dated 11/02/2013 for GTX 780 Ti and GeForce 331.65 WHQL dated 10/23/2013 for all other NVIDIA video cards;
graphics card drivers AMD processors – AMD Catalyst 13.11 BETA8 (13.250.18.0) dated October 29, 2013;

The performance of video cards was tested in two resolutions: 1920x1080 and 2560x1440 pixels. For tests, two graphics quality modes were used: "Quality + AF16x" - the default quality of textures in drivers with the inclusion of anisotropic filtering of the 16x level, and "Quality + AF16x + MSAA 4x (8x)" with the inclusion of anisotropic filtering of the 16x level and full-screen anti-aliasing of the degree 4x or 8x, in cases where the average fps remained high enough for a comfortable game. Anisotropic filtering and full-screen anti-aliasing were enabled directly in the game settings. If these settings were not available in games, then the parameters were changed in the control panels of the Catalyst and GeForce drivers. In the same place, vertical sync was forcibly disabled. In addition to the above, no additional changes were made to the driver settings.

Graphics cards were tested in two graphics tests and fourteen games updated to latest versions on the date of commencement of the preparation of the material. Of the new items in our test list, Batman: Arkham Origins appeared:

3DMark (2013)(DirectX 9/11) - version 1.0, tested in "Cloud Gate", "Fire Strike" and "Fire Strike Extreme" scenes;
Unigine Valley Bench(DirectX 11) - version 1.0, maximum quality settings, AF16x and (or) MSAA 4x, resolution 1920x1080;
(DirectX 11) - version 1.1.0, built-in test (Battle of Sekigahara) at maximum graphics quality settings and using one of the MSAA 8x modes;
Battlefield 3(DirectX 11) - version 1.4, all graphics quality settings on "Ultra", double sequential passage of the scripted scene from the beginning of the "Hunting" mission lasting 110 seconds;
Sniper Elite V2 Benchmark(DirectX 11) - version 1.05, used Adrenaline Sniper Elite V2 Benchmark Tool v1.0.0.2 BETA maximum graphics quality settings ("Ultra"), Advanced Shadows: HIGH, Ambient Occlusion: ON, Stereo 3D: OFF, Supersampling: OFF, double sequential test run;
Sleeping Dogs(DirectX 11) - version 1.5, used Adrenaline Action Benchmark Tool v1.0.2.1, maximum graphics quality settings for all items, Hi-Res Textures pack installed, FPS Limiter and V-Sync disabled, double sequential test run with total anti-aliasing at the "Normal" level and at the "Extreme" level;
Hitman: Absolution(DirectX 11) - version 1.0.447.0, built-in test with graphics quality set to "Ultra", tessellation, FXAA and global illumination enabled.
Crysis 3(DirectX 11) - version 1.2.0.1000, all graphics quality settings at maximum, blur level - medium, glare on, modes with FXAA and MSAA4x anti-aliasing, double sequential pass of the scripted scene from the beginning of the "Swamp" mission lasting 110 seconds;
Tomb Raider (2013)(DirectX 11) - version 1.1.743.0, used Adrenaline Action Benchmark Tool, quality settings at "Ultra", V-Sync disabled, FXAA and 2xSSAA anti-aliasing modes, TressFX technology enabled, double sequential pass of the test built into the game;
BioShock Infinite(DirectX 11) - version 1.1.21.65455, used Adrenaline Action Benchmark Tool with "Ultra" and "Ultra+DOF" quality settings, double run of the test built into the game;
Metro: Last Light(DirectX 11) - version 1.0.0.15, built-in test was used, graphics quality and tessellation settings were set to "Very High", Advanced PhysX technology was enabled, tests with and without SSAA anti-aliasing, double sequential scene pass "D6".
GRID 2(DirectX 11) - version 1.0.85.8679, built-in test was used, graphics quality settings were set to the maximum level for all positions, tests with and without MSAA8x anti-aliasing, eight cars on the Chicago track;
Company of Heroes 2 beta(DirectX 11) - version 3.0.0.10894, double sequential run of the test built into the game at maximum settings for graphics quality and physical effects;
total war: Rome II(DirectX 11) - version 1.0.0.4, quality settings at the level "Extreme", V-Sync disabled, SSAA anti-aliasing enabled, double sequential pass of the test built into the game;
ArmA III(DirectX 11) - version 1.04.111745, used ArmA3Mark test, quality settings at "Ultra", V-Sync disabled, double sequential pass of the test;
Batman: Arkham Origins(DirectX 11) - version 1.0.4, quality settings at "Ultra", V-Sync disabled, all effects enabled, all "DX11 Enhanced" features enabled, Hardware Accelerated PhysX = High, double sequential pass of the in-game test.

If the games implemented the ability to fix the minimum number of frames per second, then it was also reflected in the diagrams. Each test was carried out twice, the best of the two obtained values was taken as the final result, but only if the difference between them did not exceed 1%. If the deviations of the test runs exceeded 1%, then the testing was repeated at least once more in order to obtain a reliable result.

3. Performance test results

In the charts, generations of AMD GPUs are highlighted in traditional red, NVIDIA GPUs in green, and the new GeForce GTX 780 Ti is marked in dark teal. Let's see what happened and analyze the results.

3DMark (2013)

At its nominal frequencies, the GeForce GTX 780 Ti easily outperforms not only its predecessor GeForce GTX 780, but also the more expensive GeForce GTX TITAN in all 3DMark (2013) test scenes without exception. The dual-processor GeForce GTX 690 is slightly ahead, but the result of the GeForce GTX 780 Ti is easily overclocked. The leaders of this test are the dual-processor AMD Radeon HD 7990 and the CrossFireX bundle of two new MSI Radeon R9 280X.

Unigine Valley Bench

The situation is completely different in the Unigine Valley scene:

The "reds" don't look as confident here as in 3DMark, although they win in the anti-aliased mode (AMD was often more successful than NVIDIA in coping with image quality enhancement techniques). However, the gap between the GeForce GTX 780 Ti and the GTX 780 that is most impressive here is 16% in light graphics mode and 11% with MSAA8x. In addition, the GTX 780 Ti leaves behind the GTX TITAN, but the dual-processor power of the GeForce GTX 690 is "too tough" for the new product only with a significant increase in frequencies. An excellent video card was released by NVIDIA a year and a half ago, wasn't it?

Total War: SHOGUN 2 - Fall of the Samurai

The very first gaming test confirmed the previously obtained results in semi-synthetic benchmarks. In the nominal mode of operation, the GeForce GTX 780 Ti is 12-14% faster than the GeForce GTX 780, and 3-4% faster than the GeForce GTX TITAN. Forcing the frequencies of the video card during overclocking allows it to accelerate immediately by 14-19% and come close to the GeForce GTX 690, which is extremely effective in this game. Unfortunately, we cannot say the same about the AMD Radeon HD 7990, which is surprisingly outperformed even by a pair of MSI Radeon R9 280X in CrossFireX mode. The last bundle, by the way, is the leader in Total War: SHOGUN 2 - Fall of the Samurai.

Battlefield 3

Battlefield 3 is not a serious load on modern top-end video cards, but its results allow us to compare the performance of video cards:

Here GeForce GTX 780 Ti looks even more convincing against the background of GeForce GTX 780 (+16~19%) and at the level of GeForce GTX TITAN (+7~10%). The performance scalability of the novelty during overclocking is simply excellent - the increase was from 12 to 19%, depending on the quality mode. Nevertheless, despite the impressive overclocking, both dual-processor video cards and the CrossFireX tandem are slightly ahead, except for high-quality anti-aliasing modes.

Sniper Elite V2 Benchmark

The GeForce GTX 780 Ti wins its sisters no less confidently in the Sniper Elite V2 test:

The GeForce GTX 780 lags behind it by 18-20%, and the GeForce GTX TITAN is sadly inferior by 7-10%. The increase in frequencies brings the GeForce GTX 780 Ti an additional 17-20% and this allows the new product to reach the same performance level as the dual-processor GeForce GTX 690. Of course, the latter also overclocks. Competitors from AMD are not asleep, demonstrating a very high level of performance in this game.

Sleeping Dogs

The results of Sleeping Dogs, like Battlefield 3, are best evaluated by modes using anti-aliasing, or at least at a resolution of 2560x1440 pixels:

Video cards based on AMD GPUs are traditionally strong here, but this did not stop the heroine of today's article from demonstrating all her power. So, the GeForce GTX 780 lags behind it by up to 21%, and the GeForce GTX TITAN - up to 6%. The results of the dual-processor GeForce GTX 690, the new product cannot be beaten even with impressive overclocking of the GPU and video memory.

Hitman: Absolution

In the game Hitman: Absolution, the positions of AMD video cards are also quite strong, so NVIDIA products only compete with each other:

Apparently, due to the high processor dependence of this game in modes without anti-aliasing and an engine that is not very suitable for NVIDIA, the results of the GeForce GTX 780 Ti do not look so impressive against the background of the GeForce GTX 780 and GeForce GTX TITAN. So, for example, the first video card can be outperformed by a maximum of 10%, and the second by only 6%. The GeForce GTX 690 is the fastest from NVIDIA, but unlike previous tests, the GeForce GTX 780 easily overrides it at higher frequencies.

Crysis 3

It will soon be a year since Crysis 3 came out, but this game is still one of the most difficult tasks for modern video cards, including the most expensive ones:

Therefore, there is where to "turn around" the new GeForce GTX 780 Ti. The GeForce GTX 780, which has sharply fallen in price in recent weeks, lags behind the freshly baked flagship by 13-15%, and the previously formidable and accessible only to very wealthy units GeForce GTX TITAN is 3-7% slower. Only owners of the dual-processor GeForce GTX 690 can sleep peacefully - their video card has long and reliably been the fastest on NVIDIA GPUs. By the way, even the impressive overclocking of the GeForce GTX 780 Ti helped it outperform the nominal GeForce GTX 690 in only one of the three Crysis 3 test modes. We should also note the confident performance of the Radeon HD 7990, and we will talk about the efficiency of CrossFireX on a pair of MSI Radeon R9 280X in the next article.

Tomb Raider (2013)

The performance of the GeForce GTX 780 Ti in Tomb Raider cannot be called anything other than brilliant:

The regular GeForce GTX 780 is immediately behind the GeForce GTX 780 Ti by 19-22%, the GeForce GTX TITAN by 8-10%, and even the dual-processor GeForce GTX 690 trembles in front of the overclocked GTX 780 Ti, though only at nominal frequencies. But who definitely does not tremble before the novelty is AMD video cards: The Radeon HD 7990 is faster than the GeForce GTX 780 Ti by 20 percent or more, and CrossFireX looks just as good.

BioShock Infinite

The low minimum frames per second is typical for all the video cards of today's testing in BioShock Infinite:

As for the average FPS results, the picture is quite typical for today's testing. In the nominal mode of operation, the GeForce GTX 780 Ti is ahead of the GeForce GTX 780 by 15-18%, faster than the GeForce GTX TITAN by 7-9% and slower than the dual-processor GeForce GTX 690 by 13-18%. Overclocking the GeForce GTX 780 Ti allows it to get close to the GeForce GTX 690, but still not get ahead of it. The AMD Radeon HD 7990 holds up well, almost at the level of the GeForce GTX 690, but there are questions about the CrossFireX bundle, since, following the logic, it should not lose the Radeon HD 7990 in BioShock Infinite.

Metro: Last Light

The test results are shown with the "Advanced PhysX" technology enabled:

Apparently, the efficiency of multiprocessor technologies in the game Metro: Last Light still leaves much to be desired, since the results of the GeForce GTX 690 do not look as impressive as in all previous games, and the Radeon HD 7990 with CrossFireX-link from a pair of MSI The Radeon R9 280X was unimpressed. In the confrontation between single-processor video cards, the GeForce GTX 780 Ti wins a decisive victory, outperforming the GeForce GTX 780 by 17-23%, the GeForce GTX TITAN by 8-14%, and even more impressive with its overclocking performance.

GRID 2

The GeForce GTX 780 Ti is no less confident in dealing with its competitors in the game GRID 2:

So the GeForce GTX 780 is left behind by 15-16%, the "Titan" is 6-10% slower, and even the "hard nut" GeForce GTX 690 turned out to be too tough for the overclocked GeForce GTX 780 Ti. At the same time, the same cannot be said about the Radeon HD 7990 - the results of this video card in GRID 2 are unattainable for all other test participants, including the successfully overclocked GeForce GTX 780 Ti.

Company of Heroes 2

Radeon performs no less confidently in the game Company of Heroes 2:

True, there is one important “but”. The fact is that in order to get such high FPS for Company of Heroes 2 in the Catalyst drivers, we had to activate the “AFR” mode, which, although it increased the performance of the Radeon HD 7990 and the CrossFireX bundle, led to permanent flickering of textures during the entire test. So this AMD victory can be called "Pyrrhic". As for the results of the GeForce GTX 780 Ti against the background of the remaining test participants, everything is as usual great here: + 17-18% over the GeForce GTX 780, + 4-5% over the GeForce GTX TITAN and a slight loss of the GeForce GTX 690, easily compensated by overclocking the GPU and video memory GeForce GTX 780 Ti.

Total War: Rome II

The game Total War: Rome II was created with the support of AMD, so the high results of video cards on its GPUs do not surprise us. In turn, the new GeForce GTX 780 Ti easily defeated the GeForce GTX 780 (+15-21%), and slightly upset the owners of the thousand-dollar GeForce GTX TITAN, outperforming this video card by 6-8% at a resolution of 1920x1080 pixels and lagging behind by 3-5%. at a resolution of 2560x1440 pixels. Based on these results, Total War: Rome II was the only test game that was running out of 3GB of GeForce GTX 780 Ti video memory. Well, the dual-processor GeForce GTX 690 could not be forced to work at full capacity in Total War: Rome II.

ArmA III

The performance of video cards in the ArmA III game turned out to be limited by the speed of the platform, but we can still draw some conclusions from the most resource-intensive mode:

The GeForce GTX 780 Ti is the test leader even in the nominal mode. Only the dual-processor Radeon HD 7990 was able to surpass the result shown by the novelty, but even then only by 1 frame per second.

Batman: Arkham Origins

We are testing Batman: Arkham Origins for the first time, so here are the graphics settings that were fully used during the test:

As you can see, the line Hardware Accelerated PhysX is set to "High", which is not supported by video cards based on AMD GPUs, so they are not shown in the diagram. For reference, let's add that in the Hardware Accelerated PhysX = Normal mode, with a resolution of 2560x1440 pixels and using MSAA8x anti-aliasing radeon video card HD 7990 scores 76 average and 25 minimum frames per second.

The results of NVIDIA graphics cards are shown below:

Apparently, the SLI technology in the game Batman: Arkham Origins needs to be optimized, so the results of the GeForce GTX 690 are not as convincing as in previous tests and are barely higher than the nominal GeForce GTX 780 Ti. The latter, in turn, is faster than the GeForce GTX 780 by 11-15% and the GeForce GTX TITAN by 5-9%. Once again, we note the excellent scalability of the GeForce GTX 780 Ti during overclocking - increasing the frequencies of the video card from 876/928/7000 MHz to 1116/1168/7760 MHz leads to an increase in performance in Batman: Arkham Origins by 17-19%.

Traditionally, we will supplement the final charts and move on to the summary charts.

4. Pivot Charts

First of all, let's evaluate how much the new GeForce GTX 780 Ti is ahead of the now old GeForce GTX 780. The results are given at the nominal frequencies of both video cards, and the GeForce GTX 780 is taken as a starting point:

We note a friendly increase in performance, with the exception of games and modes with a high dependence on the speed of the platform. The superiority of the GeForce GTX 780 Ti is most pronounced in Metro: Last Light, Tomb Raider, Total War: Rome II and Sniper Elite V2, and the average for all tests is 13.4-15.2% at a resolution of 1920x1080 pixels and 15.1 -16.5% at 2560x1440 pixels. At the same time, we recall that based on the recently adjusted prices for the GeForce GTX 780, the cost of the GTX 780 Ti ($699) is $200 or 40% higher than the recommended price of the GeForce GTX 780 ($499).

The following pair of summary charts will tell us what is the difference between GeForce GTX 780 Ti and GeForce GTX TITAN at nominal frequencies:

Excluding results at 2560x1440 pixels in Game Total War: Rome II, the novelty is everywhere ahead of its more expensive sister and the average advantage of the GeForce GTX 780 Ti is 5.9-7.1% at 1920x1080 and 5.8-6.7% at 2560x1440. Clearly, with a $300 difference in favor of the GeForce GTX 780 Ti, the already short era of the GeForce GTX TITAN is nearing its end for gamers.

Unlike the GeForce GTX TITAN, the old dual-processor GeForce GTX 690 still has the right to market life, since in total it is still faster than the GeForce GTX 780 Ti:

Only in the game Total War: Rome II the SLI technology turned out to be inoperative, but this incident will most likely be fixed soon by new GeForce drivers and updates to the game itself.

Not faster than GeForce GTX 780 Ti and dual-socket AMD Radeon HD 7990:

The exception is Metro: Last Light, where CrossFireX is not effective enough, and the activated "Advanced PhysX" does not play into the hands of video cards based on AMD GPUs. On average in all tests, not taking into account the flickering Company of Heroes 2, the GeForce GTX 780 Ti is slower than the Radeon HD 7990 by 8.9-11.7% at 1920x1080 pixels and by 12.1-13.4% at 2560x1440 pixels.

Separately, I would like to note the excellent performance scalability of the GeForce GTX 780 Ti during overclocking, which we paid attention to when analyzing the results. Recall that we managed to overclock the reference video card from 876/7000 MHz to 1116/7760 MHz (+27.4/10.9%), and here are the performance gains that such overclocking led to:

The maximum overclocking effect was achieved in the resource-intensive game modes Hitman: Absolution and Crysis 3, and in general, in all gaming tests, it amounted to +12.4~15.0% in modes without anti-aliasing and +15.2~17.5% with its activation. We add that with such overclocking of the GeForce GTX 780 Ti, neither the NVIDIA GeForce GTX 690 nor the AMD Radeon HD 7990 (at their nominal frequencies, of course) are terrible.

5. Testing in CUDA and OpenCL applications

Today, in addition to traditional tests in graphics applications and games, we additionally studied the effectiveness of the new revision GK110 and GeForce GTX 780 Ti in computational and encoding programs using CUDA and OpenCL algorithms. Of course, only video cards based on NVIDIA GPUs take part in this test.

The first in line was once again the SmoothVideo Project (SVP) test - SVPMark 3.0.3a. For the result, we took the number of points obtained in the synthetic GPU test:

As you can see, the difference in performance between the video cards is negligible, and the GeForce GTX 780 is nominally the leader.

Next, we measured HD video encoding speed using Freemake Video Converter version 4.1.0. For encoding, a 1.77 GB FullHD video was taken (Disney's short cartoon "Puss in Boots: Three Little Devils", which we converted to iPad 3 format. The output was a file of about 441 MB. The smaller the time indicated on chart in seconds, the better the result:

The dual-processor GeForce GTX 690 is also the fastest here, processing video in just 79 seconds. However, the second place was taken just by the GeForce GTX 780 Ti in overclocking, and in the nominal mode its performance is exactly the same as that of the overclocked GeForce GTX TITAN. The slowest, as expected, was the GeForce GTX 780. By the way, the six-core Intel Core i7-3970X Extreme Edition processor at 4.8 GHz copes with this task in only 144 seconds, so CUDA is definitely useful here. , there is.

Another test in this section was DirectCompute Benchmark version 0.45b. In the diagrams, we present the results of DirectCompute and OpenCL:

In the nominal mode of operation, the GeForce GTX 780 Ti does not show anything special in DirectCompute Benchmark, but after overclocking, thanks to excellent scalability, it leads in the DirectCompute test, ahead of the overclocked GeForce GTX TITAN, and confidently takes second place in the OpenCL test, where there is no equal to the dual-processor GeForce GTX 690.

Next in line is the LuxMark test:

Unlike the previous three tests, the dual-processor GeForce GTX 690 does not look at all confident in Luxmark, although this test fully supports such video cards. Obviously, the GK104 core, on which the GeForce GTX 690 is based, cannot be compared with the GK110 in terms of computing power. All GeForce 7-series turned out to be faster than the GeForce GTX 690, and the heroine of today's test, the GeForce GTX 780 Ti, demonstrates the best results.

The latest test of Compute Unified Device Architecture (CUDA) was a rendering program based on Arion 2 technology and using algorithms that are significantly accelerated by NVIDIA GPUs. In testing, we used two rendering methods: hardware, when only the resources of the video card are involved, and hybrid, when both the video card and the central processor are used. Let's look at the results:

The results in both modes of the test have the same pattern: the GeForce GTX 780 Ti in the nominal mode is as fast as the dual-processor GeForce GTX 690 and the overclocked GeForce GTX TITAN. If you overclock the GeForce GTX 780 Ti itself, then it has no competitors in this test at all. And again, in the "Software" rendering mode, the six-core Intel Core i7-3970X Extreme Edition processor at 4.8 GHz in the Arion test gains only 832 points. That is, an overclocked GeForce GTX 780 Ti is four times faster than such a processor.

Conclusion

NVIDIA GeForce GTX 780 Ti without reservations and lyrical digressions is the fastest single-processor 3D video card. In terms of performance, its advantage over the regular GeForce GTX 780 is 13~17%, and over the GeForce GTX TITAN - 6~7%. Today, only dual-processor video cards NVIDIA GeForce GTX 690 and AMD Radeon HD 7990 are capable of competing with it in performance, but they also fail during overclocking. It is logical that in this case the novelty turned out to be not only hotter, but also noisier, which, however, will soon be easily solved by the appearance of serial models with original and highly efficient cooling systems. At the same time, the level of power consumption has increased slightly, yet ±20 watts in this class cannot and should not be taken into account. But the new B1 revision of the GK110 GPU impressed us today with a very high level of overclocking potential, and the video card itself with excellent overclocking performance scalability. The recommended price of the GeForce GTX 780 Ti is noticeably higher than the price of the GeForce GTX 780 (taking into account the recent decline), but lower than the price of the GeForce GTX TITAN, so among this trinity of video cards, in our opinion, the choice is obvious. Not sure about TITAN, but among the GeForce GTX 780 there is every reason to expect revision B1 GPUs, and therefore even more attractive products for overclockers. Time will tell.

We thank the Russian representative office of NVIDIA
and personally Irina Shekhovtsova
for the video card provided for testing.

19.11.2013 00:54

Race graphic arms continues. This summer, the fastest single-chip video card from NVIDIA was released under the name, but in the fall of 2013, AMD launched its generation of modern adapters on the market. Flagship red The AMD Radeon R9 290X has not yet been in our hands, but overseas colleagues let us make sure that NVIDIA should be nervous, because at the highest resolutions the card from AMD turns out to be the most productive solution.

The people at NVIDIA know many things ahead of time, and even more so than the most nimble journalists. That is why in November the release of a top-end device called the GeForce GTX 780 Ti took place, a video card that carries on its board 2880 CUDA cores. An unimaginable number to this day. It is this accelerator that is designed to become the most powerful video adapter on the market. One way or another, we will find out after the wider distribution of the AMD Radeon R9 290X product, and now we will try to study in detail the technical characteristics of the reference design GeForce GTX 780 Ti.

It is this accelerator that is designed to become the most powerful video adapter on the market.

AT green the card uses a full-fledged chip GK110, which is known from video cards and , however, in titanium versions of the kernel feature are somewhat limited, that is, in a natural way cut down(in the 780th version, the technical power is completely underestimated even more). This applies not only to the number of stream processors, texture units, but also clock frequencies. This means that we have a one-piece GK110, which boasts its original appearance and features originally designed by NVIDIA engineers. It is reasonable to assume that this is the last device on this chip, because a subsequent increase in power on this particular chip is no longer possible, for obvious physical reasons. All the more intriguing will be the expectation of a chip on an updated technical process.

So, the number of stream processors in the GeForce GTX 780 Ti - 2880 pieces, texture blocks - 240 , and ROP 48 units. RAM standard - GDDR5, and its volume is 3 GB. Nominal clock speeds for core and RAM - 928/7000 MHz respectively (taking into account automatic acceleration). The memory width remained unchanged - 384-bit.
For additional power, two connectors are used - 6-pin and 8-pin, and for stable operation of the GeForce GTX 780 Ti, you will need 600 W power unit.

The GeForce GTX 780 Ti uses the full-fledged GK110 chip, which is known from the GeForce GTX TITAN and GeForce GTX 780 graphics cards.

As a software bonus, NVIDIA has updated image smoothing technologies - FXAA and TXAA, it is on this card that such opportunities can be seen throughout the race. The maximum degree of anti-aliasing has always been a distinctive prerogative of NVIDIA video cards, most often it was GeForce that offered its user the highest possible level of anti-aliasing.

The factory cooling system of the GeForce GTX 780 Ti is no different from CO and, this statement is also true in relation to the dimensions of the adapter itself. The noise of the GeForce GTX 780 Ti is simply incredible, even in 2D mode. Cooling as many as 2880 stream processors is not so easy, which means that the maximum temperature equal to 84 degrees, is quite justified. The adapter is incredibly hot.
Not without branded LEDs from the end of the GeForce GTX 780 Ti. In the case, the product looks very chic.

For testing, we used the following system components (in addition to the GeForce GTX 780 Ti itself): CPU (3700 MHz), RAM Kingston HyperX 10th Anniversary Edition (KHX24C11X3K4/16X), motherboard and a HuntKey X7 900W power supply. Driver for 3D accelerator - ForceWare 331.65.

The noise of the GeForce GTX 780 Ti is simply incredible, even in 2D mode.

Even before testing began, there was no doubt that the GeForce GTX 780 Ti would be an incredibly powerful product, and indeed it turned out. Even the GeForce GTX TITAN is not a competitor to the new product, not to mention, although the latest device is not a direct competitor. The GeForce GTX 780 Ti is one of the first 3D accelerators to hit 100 fps in modern computer games and even more so, this adapter is no exaggeration suitable for any modern project, even such a complex one from a technical point of view as Hitman: Absolution and Company of Heroes 2.

Shoulder GeForce GTX 780 Ti and above voracious test in 3DMark, which is called Fire Strike, which means that Futuremark should think about releasing a new, even more graphically complex benchmark.

Even the GeForce GTX TITAN is not a competitor to the new product, not to mention the AMD Radeon R9 280X, although the latter device is not a direct competitor.

It makes little sense to overclock the GeForce GTX 780 Ti, especially for the average user, for whom the performance of the new product is more than enough. But professional overclockers will certainly take note of this accelerator. At least overcome psychological mark at 1 GHz on the GeForce GTX 780 Ti without any problems, even with the factory cooling system, but keep in mind that with each additional tens of megahertz, heat dissipation (and power consumption) will increase accordingly.

Together with the GeForce GTX 780 Ti video card, NVIDIA released a rather curious service called. It is designed to record game videos using the NVENC hardware H.264 encoder. This feature is only available on GeForce GTX 600 and GeForce GTX 700 series graphics cards.

ShadowPlay works continuously, saving selected moments from the gameplay (10-20 minutes each) in the buffer memory of the hard disk. If the user decides to capture any passage, even the past, just press the Alt + F10 keys, and the process recoding will start automatically.

You can conquer the psychological mark of 1 GHz on the GeForce GTX 780 Ti without any problems, even with the help of a factory cooling system.

Ready video files can be edited in programs: Sony Vegas, Adobe Premier or Windows Movie Maker. NVIDIA plans to implement full and instant synchronization of the application with Youtube and Twitch services.

Note that the maximum size of the received video file is 4 GB (for Windows 7), and with synchronous playing and recording, the performance of the graphics subsystem is reduced by only 5-10%. The number of supported games is unlimited, the program code does not require compatibility with a specific project, as is the case with GeForce Experience, which means that ShadowPlay will work without problems with any gaming application that supports DirectX 9 technology and higher.

Already, the GeForce GTX 780 Ti can be found on store shelves. Moreover, the cost is below $ 1000. Real price - 26000-27000 rubles, exactly as much as the NVIDIA GeForce GTX 780 board cost at the time of release. It's funny, but the less productive GeForce GTX TITAN does not doesn't want lower the price bar entrenched on the 1000$ with a ponytail.

NVIDIA has made its move, it remains to wait for a response from AMD, which this time is quite serious swung on the graphic crown leader. The winner, as you know, will be the adapter that will not only be the most productive, but also the most affordable for the average buyer.

NVIDIA GeForce GTX 780 Ti test results: