NVIDIA GeForce RTX 2080 SUPER

NVIDIA’s GeForce RTX 2080 SUPER, built on the 12 nm Turing architecture with the TU104 chip, posts an average benchmark score of 25378 across the database suite. This places it at the 69th percentile of all GPUs tracked, a mid-to-high standing that reflects its position as a capable enthusiast-class card from the GeForce 20-series generation. The data reveals a peculiar clustering around its average score: its four nearest rivals all land within a 0.8% delta, making the RTX 2080 SUPER’s performance envelope remarkably crowded despite the architectural differences at play. The card’s individual benchmark results vary widely by API, with a PassMark G3D score of 19490 and a Geekbench OpenCL score of 114190, while its 3DMark Steel Nomad DX12 result of 1882 offers a modern workload comparison point.

Benchmark Performance

The RTX 2080 SUPER’s 25378 average score is derived from a broad set of tests, but the individual figures tell a more nuanced story. In synthetic 3DMark Steel Nomad DX12, the card scores 1882, a result that reflects its Turing architecture’s handling of DirectX 12 Ultimate workloads. Meanwhile, Geekbench results show a noticeable split: OpenCL scores 114190, while Vulkan trails at 108401, a difference of 5789 points that suggests the card’s driver and hardware combination favors OpenCL compute paths in that benchmark. PassMark results paint a mixed picture across legacy APIs — DirectX 9 scores 227, DirectX 10 drops to 140, DirectX 11 climbs to 165, and DirectX 12 falls sharply to 75. This pattern implies the RTX 2080 SUPER is strongest in older, lighter API loads, with the DirectX 12 score being a notable outlier that may indicate driver overhead or benchmark-specific limitations rather than raw hardware capability.

The PassMark G3D score of 19490 and G2D score of 920 provide additional context, with the 2D result being a minor component of overall system performance. Compute workloads, measured via PassMark GPU Compute at 8290, show the card’s parallel processing throughput is solid but not exceptional. The pixel rate of 116.2 GPixel/s and texture rate of 348.5 GTexel/s, derived from the 64 ROPs and 192 TMUs respectively, translate into fill-rate headroom that supports high-resolution rendering. The FP32 throughput of 11.15 TFLOPS, paired with FP16 at 22.30 TFLOPS (2:1 ratio), indicates the card can leverage half-precision for compatible workloads, though the 2:1 ratio is not the full-rate advantage seen in some competitors. The 69th percentile ranking means roughly a third of all GPUs in the database score higher, positioning the RTX 2080 SUPER as a solid mid-to-upper tier option rather than a flagship.

How It Compares

Against the AMD Radeon RX 580 2048SP, the RTX 2080 SUPER holds a razor-thin 0.4% lead in average score (25378 vs 25287). This is statistically negligible, suggesting that in aggregate benchmarks, the two cards perform nearly identically despite the architectural gap — the RX 580 2048SP is an older, lower-tier product, yet the data shows no meaningful performance separation. This implies the RTX 2080 SUPER’s advantage lies elsewhere, likely in feature support rather than raw framerate.

The AMD Radeon 890M is even closer, trailing by just 0.5% (25246 vs 25378). The 890M is an integrated graphics solution, which makes this comparison striking: a discrete Turing GPU with 3072 shading units is barely ahead of an iGPU in average benchmark scores. This suggests the 890M benefits from modern memory bandwidth or driver optimizations in specific tests, but the RTX 2080 SUPER’s dedicated VRAM and higher power envelope should give it an edge in sustained workloads.

The AMD Radeon RX 6600 sits 0.6% behind (25235 vs 25378), another near-parity result. The RX 6600 is a newer discrete card, and its close score indicates that generational improvements in AMD’s architecture have closed the gap with the older RTX 2080 SUPER. The delta is small enough that game-specific optimizations would likely decide any head-to-head match, rather than raw compute capability.

Finally, the AMD Radeon RX 6700M trails by 0.8% (25180 vs 25378), the largest gap among the nearest rivals but still under one percentage point. The RX 6700M is a mobile GPU, so its proximity in score is notable — the RTX 2080 SUPER’s desktop power budget yields only a marginal aggregate advantage over a laptop part. This suggests that the 2080 SUPER’s performance is not substantially ahead of modern mid-range options, and its value proposition hinges on features like ray tracing rather than raw benchmark dominance.

Ray Tracing and Feature Set

The RTX 2080 SUPER is equipped with 48 RT cores and 384 tensor cores, a combination that defines its Turing feature set. These dedicated RT cores enable hardware-accelerated ray tracing, while the tensor cores provide AI-accelerated capabilities such as deep learning super sampling (DLSS) — though the FACT PACK does not specify DLSS versions or performance metrics for these features. The API support includes DirectX 12 Ultimate (12_2), which is significant because it ensures compatibility with the latest DirectX 12 features, including ray tracing and mesh shaders. OpenGL 4.6 and Vulkan 1.4 round out the API portfolio, meaning the card supports both legacy and modern graphics interfaces across a wide range of applications.

The presence of 48 RT cores on the TU104 chip is a defining characteristic of the GeForce 20-series, distinguishing it from the predecessor GeForce 10 series, which lacked dedicated ray tracing hardware. The tensor cores, numbered at 384, are equally critical for AI-driven workloads, though the FACT PACK does not provide specific TFLOPS for tensor operations. The 12 nm process node from TSMC, housing 13,600 million transistors on a 545 mm² die, results in a transistor density of 25.0M per mm² — a figure that reflects the manufacturing technology of its 2019 release. The display outputs include 1x HDMI 2.0, 3x DisplayPort 1.4a, and 1x USB Type-C, offering modern connectivity for VR headsets and high-refresh monitors, though the HDMI 2.0 port may limit 4K high-refresh output compared to newer standards.

FAQ

Q: What is the RTX 2080 SUPER’s average benchmark score?

A: The average benchmark score across the database is 25378, which places it at the 69th percentile of all GPUs tracked.

Q: How does the RTX 2080 SUPER compare to the AMD Radeon RX 6600?

A: The RTX 2080 SUPER is 0.6% ahead of the RX 6600 in average score (25378 vs 25235), a negligible margin that indicates near-identical aggregate performance.

Q: Does the RTX 2080 SUPER support DirectX 12 Ultimate?

A: Yes, it supports DirectX 12 Ultimate (12_2), along with OpenGL 4.6 and Vulkan 1.4.

Q: What is the memory configuration of the RTX 2080 SUPER?

A: It features 8 GB of GDDR6 memory on a 256-bit bus, providing 495.9 GB/s of bandwidth.

Q: How many RT and tensor cores does it have?

A: The card has 48 RT cores and 384 tensor cores, enabling ray tracing and AI-enhanced features.

Q: What is the launch MSRP of the RTX 2080 SUPER?

A: The launch MSRP is 699 USD.

Who Should Consider It

Benchmark results indicate the RTX 2080 SUPER is suited for 1440p gaming and entry-level 4K, though the data does not include specific frame rates per resolution. The 69th percentile ranking and the near-parity with rivals like the RX 6600 and RX 6700M suggest that users should not expect flagship-level performance. The 116.2 GPixel/s pixel rate and 495.9 GB/s memory bandwidth provide enough throughput for high-resolution textures and moderate anti-aliasing, but the DirectX 12 PassMark score of 75 is a red flag for modern API workloads — users prioritizing DX12 titles may see inconsistent results. The 3072 shading units and 192 TMUs deliver solid rasterization performance, as evidenced by the 19490 PassMark G3D score, making it a reasonable choice for 1080p and 1440p at high settings.

For ray tracing, the 48 RT cores offer dedicated hardware, but the FACT PACK provides no ray tracing benchmark scores, so the actual impact on frame rates is unquantified. Gamers who want to enable ray tracing in supported titles should consider the 2080 SUPER’s position relative to its 0.4-0.8% rivals — since those rivals lack RT cores, the 2080 SUPER’s feature set gives it a qualitative advantage in RT-enabled games, even if aggregate scores are similar. Users targeting 4K should note the 8 GB VRAM and 256-bit bus, which may become limiting in texture-heavy titles, though the bandwidth is adequate for most current releases. The card’s production status is end-of-life, so buyers should focus on used or remaining stock, and its 2019 release date means it predates the GeForce 30-series successor.

Power and Cooling

The RTX 2080 SUPER has a TDP of 250 W, requiring a suggested power supply of 600 W. Power is delivered via a 1x 6-pin and 1x 8-pin connector configuration, which is standard for enthusiast cards of its era. The dual-slot design, measuring 267 mm in length, 116 mm in height, and 35 mm in width, means it fits most mid-tower cases but may require checking clearance for longer GPUs. The 12 nm process node, while not the most efficient by modern standards, allows the 250 W TDP to deliver the 11.15 TFLOPS FP32 performance noted in benchmarks. Cooling is handled by a dual-slot solution, but the FACT PACK does not specify the cooler type (e.g., blower vs open-air), so thermal performance under load is unquantified. The power requirements are consistent with the card’s performance class — the 600 W PSU recommendation provides headroom for a typical system with a mid-range CPU, though users with high-end processors should consider additional margin. The end-of-life status suggests that driver optimizations may have stabilized, so power draw should be predictable.

Memory Subsystem

The RTX 2080 SUPER is equipped with 8 GB of GDDR6 memory, operating at 1937 MHz with a 15.5 Gbps effective data rate. This is spread across a 256-bit bus, yielding a total bandwidth of 495.9 GB/s. The 256-bit bus width is a key factor: it provides a balanced memory interface that supports high-resolution textures without the complexity of a wider bus, though it is narrower than some higher-tier cards. The 495.9 GB/s bandwidth is sufficient for 1440p gaming at high settings and can handle 4K in many titles, but the 8 GB capacity may become a bottleneck in future games that require more than 8 GB of VRAM at ultra settings. The GDDR6 type offers better efficiency than older GDDR5, and the 15.5 Gbps effective speed is competitive for the 2019 release window. For compute workloads, the memory subsystem’s bandwidth directly impacts the 8290 PassMark GPU Compute score, as memory-bound tasks will scale with available bandwidth. The FP16 2:1 ratio of 22.30 TFLOPS suggests that half-precision workloads could benefit from the memory bandwidth, but the practical impact depends on software support. Overall, the memory configuration is adequate for the card’s performance tier, but users pushing 4K with high-resolution texture packs should monitor VRAM usage closely.