AMD Instinct MI350P vs AMD Radeon RX 7900 GRE Comparison

Head-to-Head Benchmarks

The database records no direct head-to-head benchmark results between the AMD Instinct MI350P and the AMD Radeon RX 7900 GRE. Instead, the comparison relies on their individual recorded scores and percentiles. The RX 7900 GRE has an extensive benchmark suite, while the MI350P has no recorded benchmarks in the database, placing it at the 50th percentile of all GPUs with an average benchmark score of 0. The RX 7900 GRE sits at the 77th percentile with an average benchmark score of 32,456.

Looking at the RX 7900 GRE's recorded results, its strongest showing is in Geekbench OpenCL with a score of 175,758. Its Vulkan score in Geekbench reaches 99,850. In Passmark tests, the G3D score is 27,089, and the GPU compute score is 15,016. The DirectX 11 test records 300, DirectX 9 records 310, and DirectX 10 records 139. The DirectX 12 test shows 107, and the G2D score is 1,180. The Steel Nomad DX12 test in 3DMark records 4,814.

The MI350P's zero benchmark scores mean no direct numerical wins can be attributed to it. However, the architecture data indicates capabilities that the RX 7900 GRE cannot match in specific workloads. The MI350P's FP32 throughput is 36.04 TFLOPS, while the RX 7900 GRE delivers 45.98 TFLOPS. In FP16, the RX 7900 GRE reaches 91.96 TFLOPS (2:1 ratio), whereas the MI350P achieves 36.04 TFLOPS (1:1 ratio). The MI350P's texture rate is 1,126.4 GTexel/s versus 718.4 GTexel/s for the RX 7900 GRE.

The nearest rivals for the RX 7900 GRE provide context for its performance tier. The AMD FirePro S10000 scores 32,388, a delta of 0.2 percent. The FirePro S9300 X2 scores 32,540, a negative 0.3 percent delta. The Radeon RX 590 GME scores 32,601, negative 0.4 percent. The Radeon Pro 570X scores 32,176, a positive 0.9 percent delta. These margins are all within 1 percent, indicating the RX 7900 GRE's average benchmark score sits in a tightly clustered performance band.

FAQ

Q: Does the AMD Instinct MI350P have any recorded benchmark scores?

A: No. The database shows zero benchmarks for the MI350P, with an average benchmark score of 0 and a percentile of 50 among all GPUs.

Q: What is the RX 7900 GRE's highest recorded benchmark score?

A: Its Geekbench OpenCL score of 175,758 is the highest single result. The Geekbench Vulkan score of 99,850 follows, and the Passmark G3D score of 27,089 is the third highest.

Q: How does the MI350P's FP32 throughput compare to the RX 7900 GRE?

A: The RX 7900 GRE has higher FP32 performance at 45.98 TFLOPS compared to the MI350P's 36.04 TFLOPS. In FP16, the RX 7900 GRE reaches 91.96 TFLOPS with a 2:1 ratio, while the MI350P offers 36.04 TFLOPS with a 1:1 ratio.

Q: What memory configurations do the two cards use?

A: The MI350P uses 144 GB of HBM3e memory with an 8192-bit bus and 8.19 TB/s bandwidth. The RX 7900 GRE uses 16 GB of GDDR6 memory on a 256-bit bus with 576.0 GB/s bandwidth.

Q: What is the RX 7900 GRE's percentile ranking?

A: The RX 7900 GRE ranks at the 77th percentile of all GPUs with an average benchmark score of 32,456. Its nearest rivals are all within 1 percent of this score.

Q: Does the MI350P have any display outputs?

A: No. The MI350P has no display outputs. The RX 7900 GRE has 1x HDMI 2.1a, 2x DisplayPort 2.1, and 1x USB Type-C.

Architecture Differences

The two AMD cards use fundamentally different architectures built for different purposes. The MI350P uses CDNA 4.0 architecture on a 3 nm process from TSMC. The RX 7900 GRE uses RDNA 3.0 architecture on a 5 nm process, also from TSMC. The MI350P's chip is designated MI350 128CU, while the RX 7900 GRE uses the Navi 31 chip with the codename Plum Bonito.

Transistor counts and die sizes differ sharply. The MI350P packs 73,000 million transistors on a 1190 mm² die, yielding a density of 61.3M transistors per mm². The RX 7900 GRE has 57,700 million transistors on a 529 mm² die, achieving a higher density of 109.1M per mm². The MI350P is the larger physical chip despite its lower transistor density.

Compute resources vary. The MI350P has 8,192 shading units and 512 texture mapping units but no ROPs, giving it a pixel rate of 0 MPixel/s and a texture rate of 1,126.4 GTexel/s. The RX 7900 GRE has 5,120 shading units, 320 TMUs, and 160 ROPs, producing a pixel rate of 359.2 GPixel/s and a texture rate of 718.4 GTexel/s. The MI350P doubles the shading unit count, but the RX 7900 GRE brings rasterization capability that the MI350P lacks entirely.

The RX 7900 GRE includes 80 ray tracing cores. The MI350P records no ray tracing cores. Neither card lists tensor cores. API support diverges completely: the RX 7900 GRE supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The MI350P lists N/A for DirectX, OpenGL, and Vulkan, reflecting its compute-oriented design without graphics APIs.

Memory architecture reinforces the divide. The MI350P uses HBM3e with 144 GB capacity, an 8192-bit bus, and 8.19 TB/s bandwidth. The RX 7900 GRE uses GDDR6 with 16 GB, a 256-bit bus, and 576.0 GB/s bandwidth. The MI350P delivers over 14 times the memory bandwidth.

Power and connectivity also differ. The MI350P has a TDP of 600 W with a single 16-pin connector and a suggested PSU of 1000 W. The RX 7900 GRE has a TDP of 260 W with two 8-pin connectors and a suggested PSU of 600 W. The MI350P uses PCIe 5.0 x16, while the RX 7900 GRE uses PCIe 4.0 x16. Both are dual-slot cards, but the MI350P is shorter at 267 mm versus 276 mm for the RX 7900 GRE. The MI350P is also thinner at 40 mm versus 51 mm.

The Verdict

The data presents two distinct products. The RX 7900 GRE is a fully featured graphics card with display outputs, graphics API support, and a complete benchmark profile. It ranks at the 77th percentile with an average score of 32,456. Its nearest rivals in the database are the FirePro S10000, FirePro S9300 X2, RX 590 GME, and Radeon Pro 570X, all within 0.9 percent of its score.

The MI350P is an accelerator without graphics outputs or API support. It has no recorded benchmarks, placing it at the 50th percentile. Its strengths are theoretical: massive memory capacity, enormous bandwidth, and high texture throughput. The MI350P's 8.19 TB/s bandwidth and 144 GB capacity dwarf the RX 7900 GRE's 576.0 GB/s and 16 GB. Its 1,126.4 GTexel/s texture rate exceeds the RX 7900 GRE's 718.4 GTexel/s.

For graphics workloads, the RX 7900 GRE is the only viable choice based on its API support and display connectivity. For compute workloads that require large memory pools and high bandwidth, the MI350P's design indicates capability that the RX 7900 GRE cannot match. The RX 7900 GRE has a production status of Active and a launch MSRP of 549 USD. The MI350P's production status and launch price are not recorded.

The RX 7900 GRE's release date is recorded as 2023-07-26, with a predecessor of Navi II and successor of Navi IV. The MI350P's release date is 2026-05-06, with a predecessor of Radeon Instinct. No successor is listed for the MI350P.

Specification Differences

The following fields differ between the two cards:

• Architecture: CDNA 4.0 (MI350P) versus RDNA 3.0 (RX 7900 GRE)
• Process Node: 3 nm (MI350P) versus 5 nm (RX 7900 GRE)
• Transistors: 73,000 million (MI350P) versus 57,700 million (RX 7900 GRE)
• Die Size: 1190 mm² (MI350P) versus 529 mm² (RX 7900 GRE)
• Transistor Density: 61.3M / mm² (MI350P) versus 109.1M / mm² (RX 7900 GRE)
• Base Clock: 1000 MHz (MI350P) versus 1287 MHz (RX 7900 GRE)
• Boost Clock: 2200 MHz (MI350P) versus 2245 MHz (RX 7900 GRE)
• Game Clock: Not recorded (MI350P) versus 1880 MHz (RX 7900 GRE)
• Memory Clock: 2000 MHz, 8 Gbps effective (MI350P) versus 2250 MHz, 18 Gbps effective (RX 7900 GRE)
• Memory Size: 144 GB (MI350P) versus 16 GB (RX 7900 GRE)
• Memory Type: HBM3e (MI350P) versus GDDR6 (RX 7900 GRE)
• Memory Bus Width: 8192 bit (MI350P) versus 256 bit (RX 7900 GRE)
• Memory Bandwidth: 8.19 TB/s (MI350P) versus 576.0 GB/s (RX 7900 GRE)
• Shading Units: 8192 (MI350P) versus 5120 (RX 7900 GRE)
• TMUs: 512 (MI350P) versus 320 (RX 7900 GRE)
• ROPs: 0 (MI350P) versus 160 (RX 7900 GRE)
• RT Cores: Not recorded (MI350P) versus 80 (RX 7900 GRE)
• Pixel Rate: 0 MPixel/s (MI350P) versus 359.2 GPixel/s (RX 7900 GRE)
• Texture Rate: 1,126.4 GTexel/s (MI350P) versus 718.4 GTexel/s (RX 7900 GRE)
• FP32: 36.04 TFLOPS (MI350P) versus 45.98 TFLOPS (RX 7900 GRE)
• FP16: 36.04 TFLOPS, 1:1 (MI350P) versus 91.96 TFLOPS, 2:1 (RX 7900 GRE)
• TDP: 600 W (MI350P) versus 260 W (RX 7900 GRE)
• Power Connectors: 1x 16-pin (MI350P) versus 2x 8-pin (RX 7900 GRE)
• Suggested PSU: 1000 W (MI350P) versus 600 W (RX 7900 GRE)
• Bus Interface: PCIe 5.0 x16 (MI350P) versus PCIe 4.0 x16 (RX 7900 GRE)
• Display Outputs: No outputs (MI350P) versus 1x HDMI 2.1a, 2x DisplayPort 2.1, 1x USB Type-C (RX 7900 GRE)
• DirectX: N/A (MI350P) versus 12 Ultimate (12_2) (RX 7900 GRE)
• OpenGL: N/A (MI350P) versus 4.6 (RX 7900 GRE)
• Vulkan: N/A (MI350P) versus 1.4 (RX 7900 GRE)
• Dimensions: 267 mm length, 111 mm height, 40 mm width (MI350P) versus 276 mm length, 110 mm height, 51 mm width (RX 7900 GRE)
• Release Date: 2026-05-06 (MI350P) versus 2023-07-26 (RX 7900 GRE)

Where Each One Wins

The RX 7900 GRE wins in every measured benchmark category because it has recorded scores and the MI350P has none. Its average benchmark score of 32,456 places it at the 77th percentile, well above the MI350P's 50th percentile with a score of 0.

The RX 7900 GRE wins on rasterization. Its 160 ROPs and 359.2 GPixel/s pixel rate give it real graphics output capability. The MI350P has no ROPs and a 0 MPixel/s pixel rate, meaning it cannot produce a rendered frame for display.

The RX 7900 GRE wins on graphics API support. DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4 are all present. The MI350P lists N/A for all three, confirming it does not handle graphics APIs.

The RX 7900 GRE wins on FP32 throughput with 45.98 TFLOPS versus 36.04 TFLOPS. It also wins on FP16 with 91.96 TFLOPS versus 36.04 TFLOPS, though the MI350P's 1:1 ratio may indicate different precision handling.

The MI350P wins on memory capacity and bandwidth. Its 144 GB of HBM3e and 8.19 TB/s bandwidth exceed the RX 7900 GRE's 16 GB and 576.0 GB/s by wide margins. For workloads that fit in memory, this difference is decisive.

The MI350P wins on texture throughput. Its 1,126.4 GTexel/s texture rate surpasses the RX 7900 GRE's 718.4 GTexel/s, driven by its 512 TMUs versus 320.

The MI350P wins on shading unit count with 8,192 versus 5,120. It also uses a newer PCIe 5.0 x16 interface versus PCIe 4.0 x16, and a smaller physical footprint at 267 mm length and 40 mm width.

The MI350P wins on process technology with a 3 nm node versus 5 nm, though the RX 7900 GRE achieves higher transistor density at 109.1M per mm² versus 61.3M per mm².

Power consumption favors the RX 7900 GRE. Its 260 W TDP and 600 W suggested PSU are far below the MI350P's 600 W TDP and 1000 W suggested PSU. The MI350P uses a single 16-pin connector while the RX 7900 GRE uses two 8-pin connectors.