AMD Instinct MI350P vs NVIDIA GeForce RTX 5080 SUPER Comparison

Where Each One Wins

The recorded data for the AMD Instinct MI350P contains no benchmark entries, while the NVIDIA GeForce RTX 5080 SUPER has one benchmark result: a 3DMark Steel Nomad DX12 score of 3075. This single measurement places the RTX 5080 SUPER at the 19th percentile among all GPUs in the database, meaning 81 percent of recorded graphics cards score higher in this particular test. The MI350P has no comparable score in the database, so no direct performance comparison is possible for this workload.

The NVIDIA card shows a clear edge in the one measured metric. Its 3075 score sits slightly behind the NVIDIA Quadro P1000, which averages 3163, a delta of -2.8 percent. It also trails the Intel Arc Pro B60, which averages 3182, a delta of -3.4 percent. Conversely, the RTX 5080 SUPER leads the NVIDIA GeForce 820A by 3.1 percent (2983 average) and the NVIDIA GeForce GTX 860M by 3.6 percent (2967 average). These rival deltas are modest, indicating the RTX 5080 SUPER clusters closely with lower-tier cards in this specific benchmark, despite its high-end positioning.

The MI350P, by contrast, reports zero benchmark wins and zero losses in the head-to-head data. Its average benchmark score is 0, and its percentile rank is 50. This suggests the database holds no validated performance measurements for the accelerator. The absence of data does not indicate poor performance; it simply means the MI350P has not been subjected to the same standardized tests as the RTX 5080 SUPER. The MI350P targets a different workload profile, as its specifications reveal, so the lack of gaming-oriented benchmark data is consistent with its intended role.

Architecture Differences

The two processors diverge fundamentally in architecture. The MI350P uses AMD's CDNA 4.0 architecture on a 3 nm TSMC process, while the RTX 5080 SUPER employs NVIDIA's Blackwell 2.0 architecture on a 5 nm TSMC process. The MI350P is built around the MI350 128CU chip, a 1190 mm² die containing 73,000 million transistors. The RTX 5080 SUPER uses the GB203 chip, a 378 mm² die with 45,600 million transistors. Transistor density tells a different story: the MI350P packs 61.3 million transistors per square millimeter, while the RTX 5080 SUPER achieves 120.6 million per square millimeter, indicating a tighter packing on the smaller NVIDIA die.

Core configurations differ sharply. The MI350P has 8192 shading units and 512 texture mapping units, but reports zero ROPs. The RTX 5080 SUPER has 10752 shading units, 336 TMUs, and 112 ROPs. The NVIDIA card also includes 84 ray tracing cores and 336 tensor cores, features entirely absent from the MI350P specification listing. The MI350P's pixel rate is recorded as 0 MPixel/s, consistent with its lack of display outputs and ROPs, whereas the RTX 5080 SUPER delivers 293.1 GPixel/s.

Memory architecture represents the largest gap. The MI350P carries 144 GB of HBM3e across an 8192-bit bus, yielding 8.19 TB/s of bandwidth. The RTX 5080 SUPER has 24 GB of GDDR7 on a 256-bit bus, providing 1.02 TB/s. The MI350P's memory bandwidth is roughly eight times higher, a decisive advantage for data-intensive compute workloads. Clock speeds also differ: the MI350P runs at 1000 MHz base and 2200 MHz boost, while the RTX 5080 SUPER runs at 2295 MHz base and 2617 MHz boost. The NVIDIA card's higher clocks contribute to its higher FP32 throughput of 56.28 TFLOPS versus the MI350P's 36.04 TFLOPS, with both reporting 1:1 FP16.

The Verdict

The data supports a clear workload split. The RTX 5080 SUPER is a graphics-oriented card with display outputs, full API support, and a measured 3DMark score. The MI350P is an accelerator with no display outputs, no DirectX/OpenGL/Vulkan support, and no benchmark entries in the database. For any task involving rasterization, ray tracing, or graphics APIs, the RTX 5080 SUPER is the only viable option between the two. Its 10752 shading units, 112 ROPs, and 84 RT cores give it the hardware foundation for rendering workloads.

For memory-intensive compute tasks, the MI350P offers 144 GB of HBM3e, an 8192-bit bus, and 8.19 TB/s of bandwidth. The RTX 5080 SUPER cannot match these figures: 24 GB, 256-bit, 1.02 TB/s. The MI350P also draws 600 W versus 415 W for the RTX 5080 SUPER, and its dual-slot 267 mm length is shorter than the RTX 5080 SUPER's 304 mm. Both use PCIe 5.0 x16 and a single 16-pin power connector.

The RTX 5080 SUPER has a launch MSRP of 999 USD. The MI350P has no launch MSRP in the database. The RTX 5080 SUPER's production status is Active, while the MI350P's status is not recorded. Release dates differ by roughly four months, with the RTX 5080 SUPER dated 2025-12-31 and the MI350P dated 2026-05-06. Users needing a working graphics card with measured performance should select the RTX 5080 SUPER. Users needing massive memory capacity and bandwidth for server-side compute should evaluate the MI350P, bearing in mind that no benchmark scores exist for it in the database.

FAQ

Q: Which card has a higher FP32 compute throughput?

A: The RTX 5080 SUPER delivers 56.28 TFLOPS FP32, while the MI350P delivers 36.04 TFLOPS FP32. Both report 1:1 FP16 ratios.

Q: How much memory does each card have?

A: The MI350P has 144 GB of HBM3e on an 8192-bit bus with 8.19 TB/s bandwidth. The RTX 5080 SUPER has 24 GB of GDDR7 on a 256-bit bus with 1.02 TB/s bandwidth.

Q: Does the MI350P support DirectX or Vulkan?

A: No. The MI350P lists DirectX as N/A, OpenGL as N/A, and Vulkan as N/A. The RTX 5080 SUPER supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: What is the RTX 5080 SUPER's benchmark score?

A: Its only recorded score is 3075 in 3DMark Steel Nomad DX12, placing it at the 19th percentile among all GPUs. It trails the Quadro P1000 by 2.8 percent and the Arc Pro B60 by 3.4 percent, while leading the GeForce 820A by 3.1 percent and the GTX 860M by 3.6 percent.

Q: Do both cards use the same power connector?

A: Yes, both use a single 16-pin power connector. The MI350P has a 600 W TDP and a suggested PSU of 1000 W, while the RTX 5080 SUPER has a 415 W TDP and no suggested PSU listed.

Q: Which card has more shading units?

A: The RTX 5080 SUPER has 10752 shading units. The MI350P has 8192 shading units.

Head-to-Head Benchmarks

The head-to-head benchmark array is empty, so no direct comparative scores exist between these two products. The only measured score belongs to the RTX 5080 SUPER: 3075 in 3DMark Steel Nomad DX12. That score indicates the card performs near the bottom quarter of the database, ranking at the 19th percentile. Its nearest rivals in the database are the Quadro P1000 at 3163 (-2.8 percent), the Arc Pro B60 at 3182 (-3.4 percent), the GeForce 820A at 2983 (+3.1 percent), and the GTX 860M at 2967 (+3.6 percent). The margin between the RTX 5080 SUPER and its nearest rivals is narrow, under 4 percent in every case.

The MI350P records no benchmarks, no average score, and no rivals. Its percentile of 50 is a neutral placeholder rather than a measured rank. The database contains zero wins for either card in head-to-head comparisons, confirming that no standardized cross-testing has been logged. The practical implication is that any performance comparison must rely on specification analysis rather than direct measurements. The RTX 5080 SUPER's FP32 figure of 56.28 TFLOPS exceeds the MI350P's 36.04 TFLOPS by about 56 percent. The MI350P's memory bandwidth of 8.19 TB/s exceeds the RTX 5080 SUPER's 1.02 TB/s by a factor of roughly eight. These are the largest quantitative gaps between the two, and they align with each card's intended function.

Specification Differences

The two cards differ in nearly every measured category. Process node: 3 nm for the MI350P, 5 nm for the RTX 5080 SUPER. Die size: 1190 mm² versus 378 mm². Transistor count: 73,000 million versus 45,600 million. Transistor density: 61.3M per mm² versus 120.6M per mm². Base clock: 1000 MHz versus 2295 MHz. Boost clock: 2200 MHz versus 2617 MHz. Memory clock: both run at 2000 MHz, but effective rates differ: 8 Gbps for the MI350P versus 32 Gbps for the RTX 5080 SUPER.

Memory size, type, bus width, and bandwidth all favor the MI350P: 144 GB HBM3e, 8192-bit, 8.19 TB/s versus 24 GB GDDR7, 256-bit, 1.02 TB/s. Shading units favor the RTX 5080 SUPER: 10752 versus 8192. TMUs favor the MI350P: 512 versus 336. ROPs: 0 for the MI350P, 112 for the RTX 5080 SUPER. The RTX 5080 SUPER has 84 RT cores and 336 tensor cores; the MI350P lists none. Pixel rate: 0 MPixel/s versus 293.1 GPixel/s. Texture rate: 1126.4 GTexel/s for the MI350P versus 879.3 GTexel/s for the RTX 5080 SUPER. FP32: 36.04 TFLOPS versus 56.28 TFLOPS. FP16: 36.04 TFLOPS versus 56.28 TFLOPS.

TDP: 600 W versus 415 W. Slot width: both dual-slot. Power connector: both 1x 16-pin. Suggested PSU: 1000 W for the MI350P, none listed for the RTX 5080 SUPER. Bus interface: both PCIe 5.0 x16. Display outputs: none for the MI350P, 1x HDMI 2.1b and 3x DisplayPort 2.1b for the RTX 5080 SUPER. API support: N/A for the MI350P across DirectX, OpenGL, and Vulkan; the RTX 5080 SUPER has DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Dimensions: 267 mm length for the MI350P versus 304 mm for the RTX 5080 SUPER; 111 mm height versus 137 mm; 40 mm width for both. Release dates: 2026-05-06 for the MI350P, 2025-12-31 for the RTX 5080 SUPER.