This article will examine the most powerful graphics card available for purchase, the newly released Nvidia GeForce GTX 1080 Ti, which is built on the Pascal architecture. Featuring 11GB of GDDR5X graphics memory and the innovative GP102-350 GPU, we are poised to achieve remarkable benchmarks today. It has been eight months since Nvidia introduced the first product utilizing the GP102 architecture, the Titan X. To this day, the Titan X remains an exceptionally impressive graphics card, sharing many similarities with the model we are reviewing today. Essentially, the 1080 Ti can be viewed as a variant of the Titan X, albeit with one gigabyte less of GDDR5X memory and one ROP partition less.
Allow me to present a concise overview of the new high-performance GTX 1080 Ti. This graphics card is equipped with 3,584 CUDA Cores, 224 Texture Units, and a 352-bit memory controller, along with 11 GB of high-speed GDDR5X memory operating at 11 Gbps. It utilizes the same “GP102” GPU found in the TITAN X Pascal, albeit with slight modifications. Notably, the 352-bit GDDR5X memory interface was an unexpected feature, comprising 11 memory chips that function at an effective speed of 11 GHz, resulting in a memory bandwidth of 484 GB/s. Consequently, there has been a reduction in the ROP count to 88, down from 96 on the TITAN X Pascal, while the TMU count remains at 224. A standard reference Ti card can achieve a boost frequency of approximately 1,600 MHz, with overclocking capabilities similar to other GeForce Pascal cards, reaching around the 2 GHz mark.
| Nvidia GTX 1080 Ti | Nvidia Titan X | GeForce GTX 1080 | Geforce GTX Titan X | |
|---|---|---|---|---|
| Architecture | Pascal | Pascal | Pascal | Maxwell |
| GPU | GP102-350 | GP102-400 | GP104-400 | GM200 |
| Fab | 16nm Finfet | 16nm Finfet | 16nm Finfet | 28nm |
| Shader procs | 3584 | 3584 | 2560 | 3072 |
| Base | 1480 MHz | 1417 MHz | 1607 MHz | 1000 MHz |
| Boost | 1582 MHz | 1531 MHz | 1733 MHz | 1075 MHz |
| Perf | 11.5 TFLOPS | 11 TFLOPS | 8,87 TFLOPS | 6,6 TFLOPS |
| Mem | 11GB GDDR5X | 12GB GDDR5X | 8 GB GDDR5X | 12 GB GDDR5 |
| Mem freq | 11000 MHz | 10000 MHz | 10000 MHz | 7000 MHz |
| Mem bus | 352-bit | 384-bit | 256-bit | 384-bit |
| Mem bandw | 484 GB/s | 480 GB/s | 320 GB/s | 336,5 GB/s |
| TDP | 250W | 250W | 180W | 250 W |
Consequently, for those who primarily engage in gaming at 1080p, this graphics card may present a challenge; it is unlikely to provide significant benefits at a resolution of 1920×1080, as performance will be constrained by the capabilities of your processor. Even the most advanced 8 and 10 core processors may struggle to deliver optimal performance. Gamers who are willing to invest in this product should aim for at least Wide Quad HD (2560×1440) or Ultra HD gaming experiences. For resolutions below this threshold, it would be advisable to consider options such as the GTX 1070 or 1080. The Pascal GP102, manufactured using a 16nm process with FinFET technology, has proven to be highly effective for Nvidia. The 1050, 1060, 1070, and 1080 have consistently demonstrated impressive performance since their introduction.
The 16nm FinFET node proves to be advantageous for Nvidia. Similar to the 1080 architecture, this iteration features high clock speeds and an impressive memory configuration (11 GHz effective!). This product line is poised to attract significant interest, albeit at a premium price. The Nvidia GeForce GTX 1080 Ti will showcase a design reminiscent of the 1070/1080 Founder Edition coolers, albeit with some aesthetic modifications. The Pascal-based unit is indeed a formidable contender. It is supported by a 7+2 phase power delivery system (2x dual-FET power), and the GPU die measures 471 square millimeters. In the broader product lineup, the GeForce GTX 1080 is equipped with 2,560 shader processors, while the GeForce GTX 1070 has 1,920, and the GeForce GTX 1060 contains 1,280. The Nvidia GeForce 1080 Ti boasts 3,584 active shader processors within the GP102 GPU, with the term “active” used intentionally, as it is not a fully enabled GPU. This configuration includes 28 active streaming multiprocessors (28 SMs x 128 shader cores, or 2×64). Additionally, this 11 GB model will feature fast GDDR5X memory, which is connected to a 352-bit wide bus operating at 11 GHz (effective GDDR5X). The synergy of this memory type and clock speed results in an effective memory bandwidth of 484 GB/s. To better understand the product’s capabilities, let us examine some comparative figures.
The specifications may initially seem overwhelming; however, from a gaming perspective, it is evident that this graphics card will perform exceptionally well in Ultra HD settings. It is capable of handling popular titles such as Grand Theft Auto V, Resident Evil 7, Battlefield 1, The Division, and numerous other contemporary games. Additionally, we will examine its performance with Gears of War 4, Dishonored 2, Doom, Watch Dogs 2, and Sniper Elite 4.
In the following sections, we will explore the technology and architecture in greater detail, though we will maintain a balanced approach given the extensive material we need to address. The graphics card is equipped with a robust dual-slot, single-fan cooling system. Under full load, the temperature of this card stabilizes around 80 degrees Celsius, and it operates with a commendable level of quietness. The new GTX 1080 Ti boasts a boost clock of 1582 MHz, with a base clock of 1480 MHz. The memory is set at a reference speed of 11 Gbps. This card includes both a 6-pin and an 8-pin power connector, and we will also engage in overclocking today. Let us commence this review, but not before you have the opportunity to examine the card itself.
The Nvidia GeForce GTX 1080 Ti Founders Edition is equipped with the Nvidia GP102-350-A1 graphics processing unit, featuring a core clock speed of 1,480 MHz, a boost clock speed of 1,582 MHz, and memory operating at 11,010 MHz.
No account yet?
Create an Account