NVIDIA GeForce GTX 650 Ti Review

Author: Roger Cantwell
Editor: Howard Ha
Publish Date: Monday, November 5th, 2012
Originally Published on Neoseeker (http://www.neoseeker.com)
Article Link: http://www.neoseeker.com/Articles/Hardware/Reviews/nvidia_650_ti/
Copyright Neo Era Media, Inc. - please do not redistribute or use for commercial purposes.


Its been a busy year for NVIDIA, and it all started back in March with the release of the Kepler GPUs. Here we are several months later and NVIDIA is still at it. Hot on the heals of the GTX 660, NVIDIA unveiled the GTX 650 Ti built with the new GK106 GPU. NVIDIA's GTX 650 Ti rounds out the GTX Kepler lineup, so it will be interesting to see how well this little card can perform as we all know it has a lot to live up to.

NVIDIA's GTX 650 Ti brings the legendary GTX performance to the masses.  The GTX 650 Ti shares both hardware and software features with its predecessors. Some of these features are exclusive to the Kepler series, others have been introduced through driver updates.




Base Clock
925 MHz
Memory Clock (Data rate)
5400 MHz
L2 Cache Size
Graphics Processing Clusters
2 or 3
CUDA Cores
Texture Units
ROP Units
Total Video Memory
1024MB GDDR5
Memory Interface
Total Memory Bandwidth
86.4 GB/s
Texture Filtering Rate (Bilinear)
59.2 GigaTexels/sec
Fabrication Process
28 nm
Transistor Count
2.54 Billion
2 x Dual-Link DVI
1 x Mini-HDMI
Form Factor
Single Slot
Power Connectors
1 x 6-pin
Recommended Power Supply
400 Watts
Thermal Threshold
98° C


Let's take a closer look at NVIDIA's GTX 650 Ti. The card ships with 768 Cuda cores and four SMX units, along with all the features of previous Kepler-based cards including GPU-based PhysX, Adaptive VSync as well as FXAA made available via the control panel.

The GTX 650 Ti is equipped with a considerably larger heatsink given the overall size of the card (5.65" inches in length). As you can see, most of the PCB is covered by the fan and its shroud. Looking at the back of the GTX 650 Ti, we can see that NVIDIA did an outstanding job with the available space on the PCB.

NVIDIA's reference GTX 650 Ti is equipped with two dual-link DVI outputs as well as a mini-HDMI port. The GTX 650 Ti has been designed for best-in-class gaming at resolutions of 1920x1200 and 1920x1080.

Notice that the GTX 650 Ti's backplate is constructed without any vents, therefore all of the heated air will be vented back into the case.

With the fan assembly removed, we can get a better look at the placement of the power and memory components on the GTX 650 Ti. NVIDIA once again moved the power to the rear of the board while placing the memory under the fan assembly. By utilizing this layout, the memory receives cooling from the fan. Due to the lower power requirements of the GTX 650 Ti, NVIDIA was able to equip the reference card with a single 6-pin connector.

NVIDIA has made a few changes under the hood to both the core speed and memory sub-system on the GTX 650 Ti. The GPU core speed is clocked at 925MHz, and GPU Boost has been removed. On the memory side of things, the GTX 650 Ti is equipped with two 64-bit memory controllers giving it a total memory bandwidth of 128-bits and a overall memory capacity of 1GB of GDDR5 clocked at a 5400MHz data rate.

With a maximum TDP of 110 watts and a thermal threshold of 98 degrees Celcius. the GTX 650 Ti doesn't require a large heatsink that can be seen in the second photo below.

Overclocking the GTX 650 Ti was easy to say the least. I started in core clock increments of 10 until the GPU failed to pass our stability testing, at which point I began to increase the voltage by 10 until I had a stable overclock. My final stable overclock on the reference board was 1005MHz on the core. The memory proved to be unstable at anything above stock settings.

Hardware Configuration:

Test Setup:





Power consumption

All benchmarks will be performed at a resolution of 1920x1080 when using a single display. A resolution of 5760x1080 will be used for benchmarks in Eyefinity & Surround. Vsync will be disabled in the control panel, while AA will be set to x4 and AF set to x16. All in-game settings will be set to High, or an equivalent to Very High where possible.

Battlefield 3 is an intense FPS with massive battles that utilize the Frostbite 2 engine for features like Dynamic Global illumination along with deferred shading, adding unmatched realism to the game and enabling it to accurately render destroyed buildings, burnt-out cars and highly populated areas.

Crysis 2 is a first-person shooter developed by Crytek and is built on the CryEngine 3 engine. While the game was lacking in graphical fidelity upon its release, Crytek has since added features such as DX11 and high quality textures. This improved the in-game visuals substantially, which in turn pushes even high-end hardware to the max.

DiRT 3 takes full advantage of D11's latest features, making it one of the most visually impressive additions to the series yet thanks to tessellation and SM5.

Futuremark's latest 3DMark 2011 is designed for testing DirectX 11 hardware running on Windows 7 and Windows Vista. The benchmark includes six new tests that make extensive use of all the new DirectX 11 features including tessellation, compute shaders and multi-threading.

Total War: Shogun 2 combines turn-based and real time strategy to create a unique experience. The game is set in 16th-century feudal Japan, placing you in control of a warlord battling various rival factions. Total War: Shogun 2 is the first in the series to feature DX11 technologies to enhance the look of the game and bring even the highest-end graphics cards to their knees through massive in-game battles.

To measure the GPU temperatures, we ran two game benchmarks and recorded the idle and load temperature according to the minimum and maximum temperatures posted by MSI Afterburner. The software we tested were Crysis 2 and Unigine Heaven 3.0, running both benchmarks for 15 minutes each. This way we can give the included thermal solution and GPU time to reach equilibrium.

To measure power usage, a Kill A Watt power meter was used. Note that the numbers represent the power drain for the entire benchmarking system, not just the video cards themselves. For the idle readings we measured the power drain from the desktop, with no applications running. For load testing, we took the sustained peak power drain readings after running the system through the same in-game benchmarks we used for the temperature testing. This way we are recording real world power usage as opposed to pushing a product to its thermal threshold.

Unigine Heaven 3.0 is one of the most popular DirectX 11 benchmarks used today. Its extreme tessellation tests create quite the spectacle.

NVIDIA continues to suprise us with thier ability to produce excellent products at every price point. NVIDIA's GTX 650 Ti, built from the all new GK106, is no exception as it shares both hardware and software features of earlier Kepler-based GPUs including Adaptive VSync, FXAA, along with GPU based PhysX. NVIDIA trimmed some of the core hardware specs starting with the Cuda cores by reducing them to 768, while cutting the SMX units to four. The memory sub-system has also seen some changes as the GTX 650 Ti has a total of two memory controllers, creating a 128-bit interface.

When it comes to performance, the GTX 650 Ti is capable of delivering best in class performance even in the most demanding games and applications. With manfacturers offering their own cards at attractive prices starting around $139, you will be hard pressed to find better performance in this price range. In the performance department the GTX 650 Ti easily out performed the AMD Radeon HD 7770 in all but two tests, allowing NVIDIA to retake the throne as the fastest single graphics card solution manufacturer.

I can sum up my final thoughts on the GTX 650 Ti in two words: awesome performance.


Copyright Neo Era Media, Inc., 1999-2015.
All Rights Reserved.

Please do not redistribute or use this article in whole, or in part, for commercial purposes.