NVIDIA GeForce RTX 2080 SUPER vs NVIDIA GeForce RTX 3090 Comparison

Where Each One Wins

The recorded benchmark data splits decisively in favor of the newer Ampere part. The NVIDIA GeForce RTX 3090 wins 9 of the 10 head-to-head comparisons, with the RTX 2080 SUPER taking only a single victory in the Vulkan compute test. This is not a close contest in aggregate terms; the RTX 3090's average benchmark score of 27,565 sits 14% above the RTX 2080 SUPER's 24,170, and the percentile ranking reflects that gap, with the RTX 3090 at the 73rd percentile of all GPUs versus the 69th percentile for the older card.

The RTX 3090 dominates in compute-oriented workloads. The largest gap appears in 3DMark Steel Nomad DX12, where the RTX 3090 scores 5,118 against 1,882 for the RTX 2080 SUPER, a 171.9% difference. The PassMark GPU Compute test shows a similar pattern, with the RTX 3090 at 15,356 versus 8,290, an 85.2% advantage. These are generational leaps, not incremental improvements. In OpenCL compute, the RTX 3090 posts 172,758 against 99,226, a 74.1% lead, confirming that the raw throughput advantage extends beyond a single API.

The RTX 2080 SUPER's only win comes in Geekbench Vulkan, where it scores 111,284 against the RTX 3090's 53,927. That is a 51.5% deficit for the newer card, an unusual reversal that suggests driver or workload-specific behavior in that particular test. The RTX 3090 still wins the DirectX-oriented PassMark tests, including DirectX 12 (110 vs 75, a 46.7% lead), DirectX 11 (220 vs 165, a 33.3% lead), and DirectX 10 (182 vs 140, a 30% lead). Even in the older DirectX 9 workload, the RTX 3090 leads by 18.1% (268 vs 227). The 2D test is closer, with the RTX 3090 ahead by 15.5% (1,063 vs 920), while the main 3D score shows a 36.7% gap (26,645 vs 19,490).

The use-case split is clear. For modern DX12 gaming, ray-traced workloads, compute-heavy tasks, and anything that stresses raw shader throughput, the RTX 3090 is the stronger card by a wide margin. The RTX 2080 SUPER retains relevance only in the specific Vulkan scenario where it posts a surprising win, and even there, the overall benchmark suite shows it trailing badly in most other categories.

Architecture Differences

The two cards come from different architectural generations and different foundries. The RTX 3090 uses the GA102 chip on Ampere architecture, fabricated on an 8 nm process at Samsung, with 28,300 million transistors on a 628 mm² die. The RTX 2080 SUPER uses the TU104 chip on Turing architecture, built on a 12 nm process at TSMC, with 13,600 million transistors on a 545 mm² die. The transistor density difference is substantial: the Ampere part packs 45.1 million transistors per square millimeter versus 25.0 million for Turing, a direct consequence of the more advanced process node.

The core configurations diverge sharply. The RTX 3090 has 10,496 shading units, 328 texture mapping units, and 112 render output units. The RTX 2080 SUPER has 3,072 shading units, 192 TMUs, and 64 ROPs. That is more than three times the shader count on the RTX 3090, which explains much of the compute performance gap. Ray tracing cores also differ: 82 on the RTX 3090 versus 48 on the RTX 2080 SUPER. Tensor cores are present on both, but the RTX 3090 has 328 while the RTX 2080 SUPER has 384, an interesting inversion where the older card actually carries more tensor cores, though the newer implementation is likely more efficient per core.

Memory architecture is another major divergence. The RTX 3090 uses 24 GB of GDDR6X on a 384-bit bus, delivering 936.2 GB/s of bandwidth. The RTX 2080 SUPER uses 8 GB of GDDR6 on a 256-bit bus, delivering 495.9 GB/s. The RTX 3090 has both more capacity and nearly double the bandwidth. Clock speeds tell a different story: the RTX 2080 SUPER has a higher base clock at 1,650 MHz versus 1,395 MHz, and a higher boost clock at 1,815 MHz versus 1,695 MHz. The RTX 3090 compensates with sheer core count and memory throughput. Memory clock ratings also differ, with the RTX 3090's memory running at 1,219 MHz (19.5 Gbps effective) versus 1,937 MHz (15.5 Gbps effective) for the RTX 2080 SUPER, reflecting the different memory types.

The FP32 compute figures are stark: 35.58 TFLOPS for the RTX 3090 versus 11.15 TFLOPS for the RTX 2080 SUPER. FP16 performance is 35.58 TFLOPS (1:1 ratio) on the RTX 3090, while the RTX 2080 SUPER reaches 22.30 TFLOPS at a 2:1 ratio. Pixel and texture rates follow the same trend, with the RTX 3090 at 189.8 GPixel/s and 556.0 GTexel/s versus 116.2 GPixel/s and 348.5 GTexel/s for the older card.

Power and physical specifications also differ. The RTX 3090 has a 350 W TDP, is triple-slot, and uses a single 12-pin power connector with a suggested 750 W power supply. The RTX 2080 SUPER has a 250 W TDP, is dual-slot, uses one 6-pin and one 8-pin connector, and suggests a 600 W power supply. The RTX 3090 is also physically larger: 336 mm long, 140 mm tall, and 61 mm wide, versus 267 mm long, 116 mm tall, and 35 mm wide for the RTX 2080 SUPER. The bus interface differs as well, with PCIe 4.0 x16 on the RTX 3090 versus PCIe 3.0 x16 on the RTX 2080 SUPER. Display outputs are similar but not identical: the RTX 3090 has one HDMI 2.1 and three DisplayPort 1.4a, while the RTX 2080 SUPER has one HDMI 2.0, three DisplayPort 1.4a, and one USB Type-C.

Head-to-Head Benchmarks

The 3DMark Steel Nomad DX12 result is the single largest margin in the dataset. The RTX 3090 scores 5,118, which is 171.9% higher than the RTX 2080 SUPER's 1,882. This is a modern DirectX 12 workload, and the Ampere architecture's raw compute advantage shows clearly. The RTX 3090 finishes nearly three times as fast, a gap that no driver optimization could close for the Turing card.

In Geekbench OpenCL, the RTX 3090 posts 172,758 versus 99,226, a 74.1% lead. This test stresses general-purpose compute, and the 10,496 shading units on the RTX 3090 provide a massive throughput advantage over the 3,072 on the RTX 2080 SUPER. The PassMark GPU Compute test shows an even larger relative difference: 15,356 versus 8,290, an 85.2% gap. Compute-heavy workloads are clearly the RTX 3090's home turf.

The PassMark G3D test, which measures overall 3D graphics performance, shows the RTX 3090 at 26,645 versus 19,490 for the RTX 2080 SUPER, a 36.7% advantage. This is a more moderate gap than the compute tests, suggesting that the 3D test does not fully saturate the RTX 3090's compute resources. The DirectX API tests show similar patterns: DirectX 12 gives the RTX 3090 a 46.7% lead (110 vs 75), DirectX 11 gives a 33.3% lead (220 vs 165), and DirectX 10 gives a 30% lead (182 vs 140). Even DirectX 9, a legacy API, shows an 18.1% advantage for the RTX 3090 (268 vs 227).

The 2D test is the closest result. The RTX 3090 scores 1,063 against 920 for the RTX 2080 SUPER, a 15.5% lead. This test likely does not stress the GPU's compute units heavily, so the gap narrows to a more modest level. The Geekbench Vulkan result is the outlier: the RTX 2080 SUPER wins 111,284 to 53,927, a 51.5% margin in its favor. This is the only test where the older card leads, and the magnitude of the reversal is striking. It may reflect a specific Vulkan driver path or workload characteristic that favors Turing, but the database records the result as-is.

FAQ

Q: Which card has the higher average benchmark score?

A: The NVIDIA GeForce RTX 3090 has an average benchmark score of 27,565, while the NVIDIA GeForce RTX 2080 SUPER has an average of 24,170. The RTX 3090 sits at the 73rd percentile of all GPUs, compared to the 69th percentile for the RTX 2080 SUPER.

Q: Does the RTX 2080 SUPER win any benchmarks against the RTX 3090?

A: Yes, the RTX 2080 SUPER wins the Geekbench Vulkan test, scoring 111,284 against the RTX 3090's 53,927, a 51.5% margin in its favor. This is the only head-to-head win for the RTX 2080 SUPER out of 10 recorded tests.

Q: How much faster is the RTX 3090 in 3DMark Steel Nomad DX12?

A: The RTX 3090 scores 5,118 in 3DMark Steel Nomad DX12, which is 171.9% higher than the RTX 2080 SUPER's 1,882. This is the largest performance gap between the two cards in any recorded test.

Q: What are the memory specifications for each card?

A: The RTX 3090 has 24 GB of GDDR6X memory on a 384-bit bus, with 936.2 GB/s of bandwidth. The RTX 2080 SUPER has 8 GB of GDDR6 memory on a 256-bit bus, with 495.9 GB/s of bandwidth.

Q: How do the compute capabilities compare?

A: The RTX 3090 has 10,496 shading units and 35.58 TFLOPS of FP32 compute. The RTX 2080 SUPER has 3,072 shading units and 11.15 TFLOPS of FP32 compute. The RTX 3090 also has 82 ray tracing cores versus 48 on the RTX 2080 SUPER.

Q: Which card has higher clock speeds?

A: The RTX 2080 SUPER has higher clock speeds, with a base clock of 1,650 MHz and a boost clock of 1,815 MHz. The RTX 3090 has a base clock of 1,395 MHz and a boost clock of 1,695 MHz. Despite lower clocks, the RTX 3090 wins most benchmarks due to its larger core count and memory bandwidth.

Specification Differences

The two cards differ across nearly every specification category. The process node is 8 nm for the RTX 3090 (Samsung) versus 12 nm for the RTX 2080 SUPER (TSMC). Transistor count is 28,300 million versus 13,600 million, and die size is 628 mm² versus 545 mm². The RTX 3090 has a transistor density of 45.1 million per mm², while the RTX 2080 SUPER has 25.0 million per mm².

Core counts show the largest divergence: shading units are 10,496 versus 3,072, TMUs are 328 versus 192, and ROPs are 112 versus 64. Ray tracing cores are 82 versus 48, but tensor cores are 328 versus 384, with the older card having more. FP32 compute is 35.58 TFLOPS versus 11.15 TFLOPS, and FP16 compute is 35.58 TFLOPS (1:1) versus 22.30 TFLOPS (2:1). Pixel rate is 189.8 GPixel/s versus 116.2 GPixel/s, and texture rate is 556.0 GTexel/s versus 348.5 GTexel/s.

Memory differs in size, type, bus width, and bandwidth: 24 GB GDDR6X on a 384-bit bus with 936.2 GB/s versus 8 GB GDDR6 on a 256-bit bus with 495.9 GB/s. Clock speeds are higher on the RTX 2080 SUPER for both base (1,650 MHz vs 1,395 MHz) and boost (1,815 MHz vs 1,695 MHz). The memory clock is 1,219 MHz (19.5 Gbps effective) on the RTX 3090 versus 1,937 MHz (15.5 Gbps effective) on the RTX 2080 SUPER.

Power specifications differ as well: TDP is 350 W versus 250 W, slot width is triple-slot versus dual-slot, and power connectors are 1x 12-pin versus 1x 6-pin plus 1x 8-pin. The suggested power supply is 750 W versus 600 W. Bus interface is PCIe 4.0 x16 versus PCIe 3.0 x16. The RTX 3090 is larger at 336 mm by 140 mm by 61 mm, while the RTX 2080 SUPER measures 267 mm by 116 mm by 35 mm. Display outputs are 1x HDMI 2.1 plus 3x DisplayPort 1.4a on the RTX 3090, versus 1x HDMI 2.0, 3x DisplayPort 1.4a, and 1x USB Type-C on the RTX 2080 SUPER. Both cards support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The RTX 3090 was released on 2020-08-31 with a launch MSRP of 1,499 USD, while the RTX 2080 SUPER was released on 2019-07-22 with a launch MSRP of 699 USD.