AMD Instinct MI350P vs NVIDIA GeForce RTX 5090 Mobile Comparison

AMD Instinct MI350P and NVIDIA GeForce RTX 5090 Mobile occupy opposite ends of the hardware spectrum, one being a dual-slot accelerator with no display outputs and the other an integrated graphics processor for laptops. The database shows a stark contrast in design goals, with the MI350P targeting throughput and the RTX 5090 Mobile focusing on versatility and efficiency. While direct head-to-head benchmark comparisons are absent from the recorded data, the specification sheets and available benchmark scores for the RTX 5090 Mobile allow for a detailed analysis of their respective positions.

Where Each One Wins

The AMD Instinct MI350P is built for raw compute density. Its strengths lie in massive memory capacity and bandwidth, with 144 GB of HBM3e memory and 8.19 TB/s of bandwidth, figures that dwarf the mobile NVIDIA part. The MI350P also leads in raw shading throughput, delivering 36.04 TFLOPS of FP32 performance compared to 31.80 TFLOPS for the RTX 5090 Mobile. This makes the MI350P the clear choice for workloads that depend on large datasets residing in fast memory, such as training or inference scenarios where the entire model fits within the accelerator's local memory.

The NVIDIA GeForce RTX 5090 Mobile wins in every practical deployment metric that matters for client devices. It is an IGP (integrated graphics processor) with a 95 W TDP, whereas the MI350P is a dual-slot card requiring a 600 W TDP and a 1000 W suggested PSU. The RTX 5090 Mobile also has full API support, including DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, while the MI350P reports N/A for all graphics APIs. The RTX 5090 Mobile has display outputs (Portable Device Dependent) and a production status of Active, while the MI350P has no display outputs. The NVIDIA part is also the only one with recorded benchmark scores, achieving an 84th percentile ranking among all GPUs in the database.

Architecture Differences

The two processors are built on different architectures and manufacturing processes. The AMD Instinct MI350P uses CDNA 4.0 architecture, fabricated on a 3 nm process at TSMC. The NVIDIA GeForce RTX 5090 Mobile uses Blackwell 2.0 architecture, fabricated on a 5 nm process, also at TSMC. The transistor counts reflect the different design philosophies: the MI350P integrates 73,000 million transistors on a 1190 mm² die, resulting in a transistor density of 61.3M per mm². The RTX 5090 Mobile contains 45,600 million transistors on a 378 mm² die, achieving a much higher transistor density of 120.6M per mm².

The compute units differ significantly. The MI350P has 8192 shading units, 512 TMUs, and 0 ROPs, with no dedicated RT or tensor cores listed. Its pixel rate is 0 MPixel/s, confirming it is not designed for rasterization. The RTX 5090 Mobile has 10496 shading units, 328 TMUs, and 112 ROPs, plus 82 RT cores and 328 tensor cores. Its pixel rate is 169.7 GPixel/s. The MI350P does feature a higher texture rate at 1,126.4 GTexel/s compared to 496.9 GTexel/s for the NVIDIA part.

Clock behavior also differs. The MI350P has a base clock of 1000 MHz and a boost clock of 2200 MHz. The RTX 5090 Mobile has a lower base clock of 990 MHz and a much lower boost clock of 1515 MHz. This suggests the MI350P is designed to run at sustained high clocks under a 600 W power envelope, while the RTX 5090 Mobile operates within a 95 W mobile constraint.

Head-to-Head Benchmarks

No direct head-to-head benchmark results exist in the database between the AMD Instinct MI350P and the NVIDIA GeForce RTX 5090 Mobile. The MI350P has no recorded benchmark scores, an average benchmark score of 0, and a 50th percentile ranking. The RTX 5090 Mobile, by contrast, has a substantial set of recorded results across multiple test suites.

The RTX 5090 Mobile achieves an average benchmark score of 45152. Its nearest rivals in the database provide context for this figure. The AMD Radeon Pro 5500 XT scores 45384, which is 0.5% higher than the RTX 5090 Mobile. The NVIDIA GeForce RTX 4070 Ti scores 44795, which is 0.8% lower. The Intel Arc A730M scores 45592, 1% higher. The NVIDIA RTX 5880 Ada Generation scores 45972, 1.8% higher. This places the RTX 5090 Mobile in a tight competitive band around the 45,000 score mark.

Specific benchmark results for the RTX 5090 Mobile show its strongest performance in compute-oriented tests. It scores 201834 in Geekbench OpenCL and 198405 in Geekbench Vulkan. In 3DMark Steel Nomad DX12, it scores 5871. PassMark results show 30034 in G3D, 13401 in GPU Compute, and lower scores in legacy API tests: 324 in DirectX 9, 269 in DirectX 11, 183 in DirectX 10, and 138 in DirectX 12. The G2D score is 1057. These numbers indicate the RTX 5090 Mobile's compute performance is strong relative to its graphics API performance in older tests.

Given the absence of MI350P benchmark data, the only quantitative comparison available is the specification-level differences in FP32 throughput, memory bandwidth, and power consumption.

Specification Differences

The two parts differ in nearly every measurable specification. The AMD Instinct MI350P uses the MI350 128CU chip, while the NVIDIA GeForce RTX 5090 Mobile uses GB203. The MI350P is on a 3 nm process, the RTX 5090 Mobile on a 5 nm process. Transistor counts are 73,000 million versus 45,600 million. Die size is 1190 mm² versus 378 mm². Transistor density is 61.3M per mm² versus 120.6M per mm².

Memory configurations are vastly different. The MI350P has 144 GB of HBM3e on a 8192-bit bus with 8.19 TB/s bandwidth. The RTX 5090 Mobile has 24 GB of GDDR7 on a 256-bit bus with 896.0 GB/s bandwidth. Memory clocks are 2000 MHz (8 Gbps effective) for the MI350P and 1750 MHz (28 Gbps effective) for the RTX 5090 Mobile.

Compute resources differ. The MI350P has 8192 shading units, 512 TMUs, and 0 ROPs. The RTX 5090 Mobile has 10496 shading units, 328 TMUs, 112 ROPs, 82 RT cores, and 328 tensor cores. The MI350P's FP32 output is 36.04 TFLOPS, its FP16 output is 36.04 TFLOPS (1:1), and its texture rate is 1,126.4 GTexel/s. The RTX 5090 Mobile's FP32 output is 31.80 TFLOPS, FP16 is 31.80 TFLOPS (1:1), texture rate is 496.9 GTexel/s, and pixel rate is 169.7 GPixel/s.

Power and physical specifications diverge sharply. The MI350P has a 600 W TDP, is dual-slot, uses a single 16-pin power connector, and requires a 1000 W suggested PSU. It measures 267 mm in length, 111 mm in height, and 40 mm in width. The RTX 5090 Mobile has a 95 W TDP, is an IGP, uses no power connectors, and has no listed dimensions. Both use a PCIe 5.0 x16 bus interface.

API support is another major divider. The MI350P reports N/A for DirectX, OpenGL, and Vulkan. The RTX 5090 Mobile supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The MI350P has no display outputs; the RTX 5090 Mobile has Portable Device Dependent outputs. The MI350P's release date is 2026-05-06, while the RTX 5090 Mobile's release date is 2025-03-26. The MI350P's predecessor is Radeon Instinct, and the RTX 5090 Mobile's predecessor is GeForce 40 Mobile.

FAQ

Q: Which processor has higher FP32 compute throughput?

A: The AMD Instinct MI350P delivers 36.04 TFLOPS of FP32 performance, which is higher than the NVIDIA GeForce RTX 5090 Mobile's 31.80 TFLOPS.

Q: What is the memory capacity and bandwidth difference?

A: The MI350P has 144 GB of HBM3e memory with 8.19 TB/s bandwidth on a 8192-bit bus. The RTX 5090 Mobile has 24 GB of GDDR7 memory with 896.0 GB/s bandwidth on a 256-bit bus.

Q: Does the AMD Instinct MI350P support graphics APIs?

A: No. The MI350P reports N/A for DirectX, OpenGL, and Vulkan. The NVIDIA GeForce RTX 5090 Mobile supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Q: How does the RTX 5090 Mobile compare to its nearest rivals in average benchmark score?

A: The RTX 5090 Mobile has an average benchmark score of 45152. The AMD Radeon Pro 5500 XT scores 45384 (0.5% higher), the NVIDIA GeForce RTX 4070 Ti scores 44795 (0.8% lower), the Intel Arc A730M scores 45592 (1% higher), and the NVIDIA RTX 5880 Ada Generation scores 45972 (1.8% higher).

Q: What are the power requirements for each processor?

A: The AMD Instinct MI350P has a 600 W TDP and requires a 1000 W suggested PSU. The NVIDIA GeForce RTX 5090 Mobile has a 95 W TDP and no power connectors.

Q: Which processor has ray tracing and tensor cores?

A: The NVIDIA GeForce RTX 5090 Mobile has 82 RT cores and 328 tensor cores. The AMD Instinct MI350P does not list RT cores or tensor cores in its specifications.