AMD Instinct MI308X vs NVIDIA GeForce RTX 5080 Comparison

Head-to-Head Benchmarks

The database contains no overlapping benchmark scores for the AMD Instinct MI308X and the NVIDIA GeForce RTX 5080, so a direct head-to-head comparison on identical workloads cannot be made from recorded measurements. What the data does show is a stark contrast in profile: the MI308X holds a 50th percentile ranking across all GPUs with an average benchmark score of 0, while the RTX 5080 sits at the 87th percentile with an average benchmark score of 56,083. That percentile gap is large, but it reflects different testing ecosystems, not identical tests.

The RTX 5080 delivers measurable results across a range of workloads. In 3DMark Steel Nomad DX12, its score is 8,637. Geekbench OpenCL returns 235,901, and Geekbench Vulkan returns 255,450. Passmark GPU compute scores 21,789, while Passmark G3D scores 36,565. The RTX 5080 also records legacy DirectX results: Passmark DX9 at 389, DX10 at 208, DX11 at 324, and DX12 at 151. Its 2D score is 1,415. These numbers confirm a wide application footprint.

The MI308X has no recorded benchmark entries in the database, meaning its 0 average score and 50th percentile are placeholders rather than performance verdicts. Its capabilities must be assessed from architectural specifications, not from measured results.

Relative to its nearest rivals, the RTX 5080 holds a narrow edge. It leads the AMD Radeon 8060S by 0.6 percent, the AMD Radeon RX 6750 GRE 12 GB by 0.7 percent, and the AMD Radeon Pro W5700X by 2.3 percent. The AMD Radeon RX 9070 GRE sits ahead by 2.2 percent. These deltas indicate that the RTX 5080 is tightly clustered with its competition, not dominant by a wide margin.

The MI308X has no nearest rivals listed, so no comparative deltas exist. That absence means the database provides no measured context for how it stacks against other accelerators, only its raw specification sheet.

FAQ

Q: Which GPU has the higher percentile ranking in the database?

A: The NVIDIA GeForce RTX 5080 ranks at the 87th percentile across all GPUs, while the AMD Instinct MI308X ranks at the 50th percentile.

Q: What is the average benchmark score for each card?

A: The RTX 5080 has an average benchmark score of 56,083. The MI308X has an average benchmark score of 0, with no individual benchmark entries recorded.

Q: How does the RTX 5080 compare to its closest rival, the AMD Radeon 8060S?

A: The RTX 5080 scores 0.6 percent higher than the AMD Radeon 8060S, with average scores of 56,083 and 55,757 respectively.

Q: Which card has more shading units?

A: The AMD Instinct MI308X has 19,456 shading units, compared to 10,752 on the NVIDIA GeForce RTX 5080.

Q: What memory configurations do the two cards use?

A: The MI308X uses 192 GB of HBM3 on an 8192-bit bus, delivering 5.32 TB/s of bandwidth. The RTX 5080 uses 16 GB of GDDR7 on a 256-bit bus, delivering 960.0 GB/s of bandwidth.

Q: Which card has a higher boost clock?

A: The NVIDIA GeForce RTX 5080 has a boost clock of 2617 MHz, while the AMD Instinct MI308X has a boost clock of 2100 MHz.

Where Each One Wins

The NVIDIA GeForce RTX 5080 wins in every measured benchmark category, because it is the only one of the two with recorded test scores. Its strengths show up across compute and graphics workloads. In Geekbench Vulkan, it scores 255,450, and in Geekbench OpenCL, 235,901, indicating solid general-purpose compute performance. Its 3DMark Steel Nomad DX12 score of 8,637 confirms modern gaming-workload capability. Passmark G3D at 36,565 and Passmark GPU compute at 21,789 reinforce a balanced profile between rasterization and compute.

The RTX 5080 also holds a measurable lead over its nearest rivals, which strengthens its position in competitive comparisons. It is 0.6 percent ahead of the AMD Radeon 8060S and 0.7 percent ahead of the AMD Radeon RX 6750 GRE 12 GB. Those margins are slim, but they are consistent across the entire benchmark set.

The AMD Instinct MI308X wins on raw specification breadth. It has nearly double the shading units (19,456 versus 10,752), more than triple the texture mapping units (1,216 versus 336), and a memory subsystem that dwarfs the RTX 5080 in capacity and bandwidth. Its 192 GB of HBM3 on an 8192-bit bus with 5.32 TB/s of bandwidth targets large-scale data workloads, not client graphics. Its FP32 throughput of 81.72 TFLOPS exceeds the RTX 5080's 56.28 TFLOPS by a wide margin.

The MI308X also wins on power envelope headroom, with a 750 W TDP and a suggested PSU of 1150 W, though that is not a benchmark win, it is a design choice for sustained compute. The RTX 5080 fits a dual-slot form factor with a 360 W TDP and a 750 W suggested PSU, making it far more suitable for desktop integration.

Specification Differences

The two cards diverge on nearly every measurable specification. The AMD Instinct MI308X uses a 5 nm process from TSMC, as does the NVIDIA GeForce RTX 5080, but the die sizes and transistor counts are far apart. The MI308X has 153,000 million transistors on a 1017 mm² die, for a density of 150.4M per mm². The RTX 5080 has 45,600 million transistors on a 378 mm² die, for a density of 120.6M per mm².

Memory differs completely. The MI308X offers 192 GB of HBM3 with a 8192-bit bus and 5.32 TB/s bandwidth. The RTX 5080 offers 16 GB of GDDR7 with a 256-bit bus and 960.0 GB/s bandwidth. The MI308X memory clock is 1300 MHz (5.2 Gbps effective), while the RTX 5080 runs at 1875 MHz (30 Gbps effective).

Compute resources also differ sharply. The MI308X has 19,456 shading units, 1,216 TMUs, and no ROPs, no ray tracing cores, and no tensor cores listed. Its pixel rate is 0 MPixel/s and its texture rate is 2,553.6 GTexel/s. The RTX 5080 has 10,752 shading units, 336 TMUs, 112 ROPs, 84 ray tracing cores, and 336 tensor cores. Its pixel rate is 293.1 GPixel/s and its texture rate is 879.3 GTexel/s.

Clock speeds are lower on the MI308X, with a 1000 MHz base and 2100 MHz boost, compared to the RTX 5080's 2295 MHz base and 2617 MHz boost. TDP differs by 390 W: 750 W for the MI308X versus 360 W for the RTX 5080. The MI308X uses an OAM module slot width with no power connectors, while the RTX 5080 is dual-slot with a single 16-pin connector. Suggested PSUs are 1150 W and 750 W respectively.

The RTX 5080 has a full display output suite (1x HDMI 2.1b and 3x DisplayPort 2.1b) and supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The MI308X has no display outputs and no API support listed. The RTX 5080 has physical dimensions of 304 mm length, 137 mm height, and 40 mm width. The MI308X lists no dimensions.

Release dates differ by about 14 months. The MI308X launched on 2023-12-05, and the RTX 5080 launched on 2025-01-29. The RTX 5080 has a launch MSRP of 999 USD. The MI308X has no launch MSRP listed.

Architecture Differences

The AMD Instinct MI308X is built on the CDNA 3.0 architecture, using the Aqua Vanjaram chip. It belongs to the Instinct (MIx) generation and follows the Radeon Instinct line. Its design is compute-focused, with no graphics pipeline: no ROPs, no display outputs, and no consumer API support. The architecture prioritizes memory capacity and bandwidth above all else, with HBM3 memory and an 8192-bit interface that is unmatched in the database. The 5 nm process from TSMC yields a 1017 mm² die, which is one of the largest recorded.

The NVIDIA GeForce RTX 5080 uses the Blackwell 2.0 architecture on the GB203 chip. It belongs to the GeForce 50-series and the GeForce 50 generation, succeeding the GeForce 40 line and preceding GeForce 60. The architecture is a full client GPU design, with 84 ray tracing cores and 336 tensor cores, plus a complete ROP and TMU setup. It supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, and it has display outputs for consumer use. The 5 nm process yields a 378 mm² die, much smaller than the MI308X, with a lower transistor density of 120.6M per mm².

The MI308X targets accelerators with no rasterization path. Its FP16 and FP32 throughput are identical at 81.72 TFLOPS (1:1), which suits mixed-precision compute workloads. The RTX 5080 also runs FP16 and FP32 at the same rate, 56.28 TFLOPS (1:1), but adds dedicated ray tracing and tensor hardware. That feature set is absent from the MI308X entirely.

The API support gap is total. The MI308X lists N/A for DirectX, OpenGL, and Vulkan. The RTX 5080 supports all three with current versions. This reflects different product categories: the MI308X is a server accelerator, the RTX 5080 is a desktop graphics card.

Power delivery also differs architecturally. The MI308X uses an OAM module with no external power connectors, indicating a board-level power solution. The RTX 5080 uses a standard 16-pin connector and a dual-slot cooler, fitting conventional desktop power supplies. The suggested PSU difference of 400 W (1150 W versus 750 W) reinforces the divergent thermal and electrical designs.

The production status is listed as null for the MI308X and Active for the RTX 5080. The RTX 5080 also has a defined successor, the GeForce 60, while the MI308X has no successor listed. These differences place the two cards in separate lifecycle stages and separate market segments, despite sharing the same process node.