Introduction

We finally have with us the NVIDIA GeForce RTX 4060 Ti 16 GB graphics card. This model, which doubles the VRAM size over the "standard" RTX 4060 Ti 8 GB, was announced to much fanfare back in May 2023, but NVIDIA, for reasons we'll find out in this review, decided against sampling the card to reviewers ahead of its official July 18 launch. There is no first-party Founders Edition card, and NVIDIA's board partners weren't too keen on sampling their custom-design cards, either. So we spent north of $500, (€562.50 to be precise) to buy one of these cards from the market after launch, and it finally made its way to our labs. Technically, this is a Gainward GeForce RTX 4060 Ti 16 GB Panther OC, but we are treating it as a de facto NVIDIA review, because there is no FE available and we have no other reviews planned of this 16 GB model.



The GeForce RTX 4060 Ti Ada is designed to consolidate the mid-range for NVIDIA this generation. It offers maxed-out gaming at 1080p, including ray tracing, and 1440p gaming is very much possible at high-thru-ultra settings. For 1440p with ray tracing, you might need to dial down some settings, or better yet, let GeForce Experience find the right settings for you. The best option would be to take advantage of features such as DLSS. Since it's based on the latest Ada Lovelace graphics architecture, the RTX 4060 Ti gets DLSS 3 Frame Generation, a path-breaking feature to increase frame-rates with minimal impact on image quality.

The GeForce Ada Lovelace graphics architecture debuts the third generation of NVIDIA RTX, the technology that mainstreamed ray tracing for real-time graphics and gaming applications. By itself, ray tracing is a the holy grail of 3D graphics, as it's the most computationally intensive way to generate photorealistic graphics. NVIDIA figured out a way to combine conventional raster 3D graphics with certain real-time ray traced elements such as lighting, shadows, reflections, global illuminations, and motion-blur; to significantly improve realism in games. Even this much of ray tracing requires tremendous compute power, and so NVIDIA innovated the RT core, a fixed function hardware unit to calculate ray intersections. NVIDIA leverages AI for de-noising, and hence the GPU has Tensor cores that accelerate neural network building and training, besides features such as DLSS. Ada introduces NVIDIA's highest IPC CUDA cores that support shader execution re-ordering; 3rd generation RT cores that support displaced micro-meshes, and the new optical flow accelerator that enables DLSS 3 Frame Generation.

The GeForce RTX 4060 Ti 16 GB is based on the 5 nm AD106 silicon, which it nearly maxes out, enabling 34 out of 36 streaming multiprocessors (SM) physically present. This works out to 4,352 CUDA cores, 136 Tensor cores, 34 RT cores, 136 TMUs, and 48 ROPs. In a bit of a controversial decision, NVIDIA reduced the memory bus width of the RTX 4060 Ti to 128-bit, compared to the 256-bit that the RTX 3060 Ti enjoys. The memory speed has generationally increased to 18 Gbps, but the memory bandwidth still falls a bit short. This shouldn't be a problem, NVIDIA claims, as the company re-architected the memory sub-system to place greater importance on 8-12 times larger on-die caches, which should reduce the round-trips to the video memory. The 16 GB version of the RTX 4060 Ti is simply a doubling in memory size, using the same 128-bit GDDR6 memory interface, with a clamshell memory configuration, which puts the chips on both sides of the PCB. The memory speed is unchanged at 18 Gbps, as is the GPU's core-configuration. You get the same number of shaders, and the same GPU clock speeds. Only the power limit has been increased by 5 W, to account for the higher power draw of the additional 8 GB memory.

NVIDIA is asking a steep $499 for the 16 GB GeForce RTX 4060 Ti, a 25% premium over the RTX 4060 Ti 8 GB. From what little we know about GDDR6 memory spot-pricing, the 16 GB variant is essentially a $10-20 increase in the bill of materials, but for whatever reason, it lugs that huge premium. At $499, this card is $100 shy of the GeForce RTX 4070, which is about 30% faster than the RTX 4060 Ti 8 GB. In this review, we'll find out by just how much NVIDIA has managed to slim that gap with the 16 GB RTX 4060 Ti.

Packaging

The Card

The Gainward Panther OC comes with a clean color theme that uses black as the main color, with some white highlights. On the back you'll find a metal backplate with cutouts that let some air flow through the card.


Dimensions of the card are 25.0 x 11.5 cm, and it weighs 753 g.


Installation requires three slots in your system. The card's width is 45 mm.


Display connectivity includes three standard DisplayPort 1.4a ports and one HDMI 2.1a (same as Ampere).

The new 8th Gen NVENC now accelerates AV1 encoding, besides HEVC. You also get an "optical flow accelerator" unit that is able to calculate intermediate frames for videos, to smooth playback. The same hardware unit is used for frame generation in DLSS 3.


Gainward opted for a single 8-pin PCIe power connector, this configuration is good for up to 225 W.

Teardown

The thermal solution on the Gainward Panther uses two heatpipes. The main heatsink also provides cooling for the memory chips and VRM circuitry.


The backplate is made of metal and protects the card against damage during installation and handling. Please note the two thermal pads, which provide cooling for the memory chips on this side of the PCB.

High-resolution PCB Pictures

These pictures are for the convenience of volt modders and people who would like to see all the finer details on the PCB. Feel free to link back to us and use these in your articles, videos or forum posts.


High-resolution versions are also available (front, back).

Circuit Board (PCB) Analysis

GPU voltage is a six-phase design, managed by a uPI uP9512R controller.


Monolithic Power Systems MP87990 DrMOS components are used for GPU voltage.


Memory voltage is a dual-phase design, managed by a uPI uP1666Q controller.


For memory, Sinopower SM7342EKKP are used.


The memory chips are Samsung K4ZAF325BC-SC20, these are 20 Gbps-rated GDDR6 memory chips. These are the same chips as on the 8 GB version, the only difference is that the 16 GB models have eight memory chips, four on each side.


NVIDIA's AD106 graphics processor is the company's fourth Ada Lovelace GPU. It is built using a 5 nanometer process at TSMC Taiwan, with a transistor count of 22.9 billion and a die size of 190 mm². Please note that the 16 GB version gets its own GPU variant, called AD106-351, whereas the 8 GB version has AD106-350 onboard.

Test System

Benchmark scores in other reviews are only comparable when this exact same configuration is used.

• All games and cards are tested with the drivers listed above—no performance results were recycled between test systems. Only this exact system with exactly the same configuration is used for all results in this review.
• All graphics cards are tested using the same game version.
• All games are set to their highest quality setting unless indicated otherwise.
• AA and AF are applied via in-game settings, not via the driver's control panel.
• Before starting measurements, we heat up the card for each test to ensure a steady state is tested. This ensures that the card won't boost to unrealistically high clocks for only a few seconds until it heats up, as that doesn't represent prolonged gameplay.
• For better real-life applicability, all game tests use custom in-game test scenes, not the integrated benchmarks
• All cards used for comparison are reference designs. When a reference design does not exist, we go the extra mile and buy the closest possible match, using reference clocks and default power limit.

Each game is tested at these screen resolutions:

• 1920x1080: Most popular monitor resolution.
• 2560x1440: Intermediary resolution between Full HD and 4K, with reasonable performance requirements.
• 3840x2160: 4K Ultra HD resolution, available on high-end monitors.