GeForce RTX 3090 vs Radeon RX 6950 XT

Intro

Compare all that to the Radeon RX 6950 XT, which comes with core speeds of 1925 MHz on the GPU, and 2250 MHz on the 16384 MB of GDDR6 memory. It features 5120 SPUs along with 320 Texture Address Units and 128 ROPs.

Display Graphs

Hide Graphs

Power Usage and Theoretical Benchmarks

Power Consumption (Max TDP)

Radeon RX 6950 XT		335 Watts
GeForce RTX 3090		350 Watts
	Difference: 15 Watts (4%)

Memory Bandwidth

Theoretically speaking, the GeForce RTX 3090 should perform much faster than the Radeon RX 6950 XT overall. (explain)

GeForce RTX 3090		958668 MB/sec
Radeon RX 6950 XT		589824 MB/sec
	Difference: 368844 (63%)

Texel Rate

The Radeon RX 6950 XT is quite a bit (approximately 35%) more effective at anisotropic filtering than the GeForce RTX 3090. (explain)

Radeon RX 6950 XT		616000 Mtexels/sec
GeForce RTX 3090		457560 Mtexels/sec
	Difference: 158440 (35%)

Pixel Rate

The Radeon RX 6950 XT is much (about 58%) more effective at FSAA than the GeForce RTX 3090, and will be able to handle higher screen resolutions better. (explain)

Radeon RX 6950 XT		246400 Mpixels/sec
GeForce RTX 3090		156240 Mpixels/sec
	Difference: 90160 (58%)

Please note that the above 'benchmarks' are all just theoretical - the results were calculated based on the card's specifications, and real-world performance may (and probably will) vary at least a bit.

Price Comparison

Specifications

Memory Bandwidth: Memory bandwidth is the max amount of information (in units of MB per second) that can be transferred past the external memory interface in one second. The number is worked out by multiplying the card's bus width by its memory clock speed. In the case of DDR memory, the result should be multiplied by 2 again. If DDR5, multiply by 4 instead. The better the bandwidth is, the better the card will be in general. It especially helps with AA, HDR and higher screen resolutions.

Texel Rate: Texel rate is the maximum number of texture map elements (texels) that are applied per second. This number is worked out by multiplying the total number of texture units of the card by the core speed of the chip. The better this number, the better the card will be at texture filtering (anisotropic filtering - AF). It is measured in millions of texels in one second.

Pixel Rate: Pixel rate is the maximum number of pixels that the graphics chip can possibly record to its local memory in one second - measured in millions of pixels per second. Pixel rate is calculated by multiplying the number of ROPs by the the core speed of the card. ROPs (Raster Operations Pipelines - also called Render Output Units) are responsible for filling the screen with pixels (the image). The actual pixel output rate also depends on lots of other factors, most notably the memory bandwidth - the lower the bandwidth is, the lower the potential to get to the maximum fill rate.