We put this “meiúca” of the R$ 2 thousand to fight!
It’s normal for people to be in doubt between some models, but few get tangled up more than the GeForce RTX 2060, the GeForce RTX 3050 and the Radeon RX 6600. On the one hand, we already have the dispute of the current generation, with the RTX 3050 coming to fight with the RX 6600 in the entry/intermediate market, but with the RTX 2060 continuing on the market (and even gaining a 12GB version) it gave a good tangle in the dispute between the current GeForce, the current Radeon and the previous GeForce. What is the best?
Let’s make a comparison with the three, seeking to find the best option. Starting with the technical specifications of the models that we will put in the dispute:
|Price at launch||U$ 349.99||U$ 249.00||U$ 329.00|
|updated price||BRL 3,599.00||BRL 3,349.00||U$ 329.00|
Among the specifications, the most relevant is the amount of VRAM. The RTX 2060 has 6GB, which is just enough in Full HD and can pose challenges in Quad HD. In many cases, texture quality adjustments already solve the problem, but the 8GB of the RTX 3050 and RX 6600 are certainly more welcome.
In ray tracing we have the advantage for the RTX 3050. It brings the new generation of Nvidia architecture with tensor and RT cores. This makes it the most efficient for technologies like DLSS and Ray Tracing, but even it needs parsimony in the use of RT. Next comes the RTX 2060, which also achieves something even with ray tracing enabled, but also with many adjustments, while the Radeon RX 6600 even does something, but it is already a card with difficulties dealing with ray tracing in some titles, even at 1080p.
In technologies the RTX 3050 and RTX 2060 combo have RTX features, such as Studio with background clipping and noise correction, as well as NVENC for video encoding, great for recording gameplay or doing live.
The advantage for the RX 6600 will appear lower in the tests: it tends to do well in games that only involve rasterization. It’s also based on RDNA 2, the same architecture as the next-gen consoles, which could help this model in the long run.
Our test platform is based on an AMD Ryzen 9 5900X processor. Several other high-performance components accompany this system, such as NVMe SSDs and 32GB of RAM clocked at 3200MHz (CL16). Below are some pictures of the card installed in our new video card test bench:
– Continues after advertising –
Before the tests, details of the machine, operating system, drivers and software/games used in the tests:
Machine used in the tests:
– AMD Ryzen 9 5900X processor
– GIGABYTE X570 AORUS Xtreme motherboard
– HyperX Predator RGB 32GB (2x16GB) 3200MHz CL16 memory kit
– Kingston KC2500 250GB + 2TB SSD
– CM MasterLiquid ML360 V2 RGB cooling system
– CM v1300W Platinum power supply
– Custom CM MasterFrame 700 Cabinet
Operating System and Drivers:
– Windows 11 Pro 64 Bit
– AMD Catalyst Adrenalin 22.5.2
– Adobe Premiere CC 2021 (GPU rendering)
– SPECviewpeft 13 (Solid Works/Maya, GPU rendering)
– 3DMark (Fire Strike Ultra / Port Royal)
– Assassin’s Creed Valhalla (DX12)
– Forza Horizon 4 (DX12)
– Grand Theft Auto 5 (DX11)
– Rainbow Six Siege (Vulkan)
– Red Dead Redemption 2 (Vulkan)
– Resident Evil Village (DX12)
– The Division 2 (DX12)
– Watch Dogs: Legion (DX12)
With the increase in applications that take advantage of the processing power of GPUs, we have updated our battery of tests with some of the most important software on the market.
Adobe Premiere CC 2022
Adobe Premiere is a world reference when it comes to video editing software, and in its latest versions it has also taken advantage of GPUs to help speed up rendering. Below is the behavior of the compared boards:
– Continues after advertising –
The professional application test suite is comprised of a comprehensive battery of hardware-intensive scenarios to render various uses ranging from architecture, mining, and medicine. We ran two tests, one focused on performance in Maya and the other in SolidWorks.
And if we talk about benchmarks, we couldn’t leave out one of the most iconic tests in the world, especially for video card performance, 3DMark. Our battery consists of three tests, but 2 of them show technologies that only newer board models bring, Ray Tracing (Port Royal) and DLSS (DLSS Feature Test).
We ran the latest version of the application from UL Benchmark (which purchased Futuremark), with all tests assuming the profile’s default configuration, with no changes. Below are the results:
Tests in games
Now let’s get to what really matters: the performance tests on some of the main games on the market.
To help you understand the following graphics: above 60fps is ideal for monitors that operate at this frequency. The closer to 30fps, the worse the fluidity becomes and, below 30, the game starts to become “unplayable”
Assassin’s Creed Valhalla
Open world game has wide scenarios and a benchmark with a good amount of characters and structures, making it a challenge for both the processor and the video card. The game uses the Ubisoft Anvil engine, an evolution of the AnvilNext 2.0 present in the series since Assassi’s Creed Unity. The version used in Valhalla on PC is based on the DirectX 12 API.
CD Proejct RED’s sales success but critical debacle highlights the widespread use of diverse technologies such as AMD’s FidelityFX and both DLSS and Nvidia RTX hardware-accelerated ray tracing. In the test we ran through the busy streets of Night City, making the work of both the processor and the graphics card challenging, and putting extra effort on the graphics card when light ray tracing is enabled.
Forza Horizon 4
The Playground Games game uses its own graphics engine and, as exclusive to Microsoft systems, it is fully developed for DirectX 12. This game stands out for its excellent graphics and the ability to deliver a good level of performance on multiple hardware, including some more limited.
Grand Theft Auto V
The game is already a classic and after years it still stands firm as one of the most played games. Based on DirectX 11, it also brings a sense of older graphics engines based on the still popular Microsoft API. It’s a very demanding test on the processor, and faster memories also have very noticeable impacts. For modern video cards, it’s no longer a big challenge.
Rainbow Six Siege
The Ubisoft game’s highlights are the use of the low-level Vulkan API in its latest implementation. This Esport demands high frame rates to be played satisfactorily, and is often one of the most efficient games in achieving this performance across multiple components.
Red Dead Redemption 2
RockStar game, with beautiful graphics is a good reference to measure the behavior of video cards. Our test considers the game running on the Vulkan API, which behaves very well on both AMD and NVIDIA cards.
Resident Evil Village
Capcom’s game uses the excellent RE Engine, a graphics engine that delivers interesting results from high-end PC hardware to more limited platforms like the Nintendo Switch. Resident Evil 8 highlights complex and richly detailed scenarios, with the use of Ray Tracing in the lighting of the scenarios and with resources such as FidelityFX available. In the tests, we take a tour of Dimitrescu Castle, one of the heaviest and most detailed locations in the game.
Tom Clancy’s The Division 2
The Division 2 uses its own graphics engine developed by Ubisoft Massive, dealing with complex scenarios and large amounts of particles on the screen.
Game supported by Nvidia is based on the Disrupt graphics engine and has a wide use of RTX technologies, such as DLSS, and also has Ray Tracing, being accelerated by both GeForce RTX and Radeon RDNA 2 hardware. Its main highlight is a futuristic London full of geometry and characters, which combined with the light ray tracing effects make it quite a challenge to run the game.
It is not difficult to understand how difficult it is for people to choose between these three models. They get tangled up in terms of performance and costs.. Especially in the duel between the GeForce options and the Radeon option, depending on the game and the technology used, who is ahead in the performance graphics varies greatly. Even in the part of professional tests we had a curious alternation, with the Radeons “patrolling” the GeForce in the test with Specviewperf, and reversing the situation with the test of Adobe Premiere.
The three models have similar costs and performance levels
In some games the Radeon RX 6600 jumps ahead, and when it’s not in the lead, it’s usually just behind the RTX 2060, something under 10%. In raw performance she is the winner of the dispute, although it is in some games that we see this difference being more noticeable. In others, it is in the same club as the 2060.
Here’s a moment when the GeForce RTX 2060’s utility kit weighs in in your favor. With several games bringing a difference of 10% or less, and not getting a bad performance in tests that it lags behind, the Nvidia card may make more sense by bringing features such as DLSS and RTX functionality, in addition to NVENC for recording and gameplay streaming.
And the RTX 3050? Our tests showed a considerable difference between it and the RTX 2060, between 10 and 15%. This means that the RTX 3050 only makes sense if it appears that same 10 to 15% cheaper, at least. Having more VRAM is welcome, but it’s not a big enough differentiator to justify a price too similar to the RTX 2060. And in the competition with the RX 6600, it lags even further behind Radeon in games than the red side of the force. stands out.
The Radeon RX 6600 has the most performance in some games, the RTX has the best balance of features and performance, and the RTX 3050 is only up for grabs if it’s priced lower.
Here it is also good to bring an additional model to the dispute. The RTX 2060 also has a 12GB version, which in addition to more VRAM, has a configuration difference with more CUDA cores.. Unfortunately we don’t have a model for testing, so our collection and our database are missing your performance results, but keep an eye out for this difference when you’re researching models.