AMD Radeon RX 7900M vs NVIDIA GeForce RTX 5090 D Comparison

# AMD Radeon RX 7900M vs NVIDIA GeForce RTX 5090 D: A Generational Gap in Mobile and Desktop GPUs

The AMD Radeon RX 7900M and NVIDIA GeForce RTX 5090 D represent two vastly different segments of the GPU market, separated by architecture generation, form factor, and performance class. The RX 7900M is an integrated mobile GPU from AMD's RDNA 3.0 lineup, while the RTX 5090 D is a flagship desktop card built on NVIDIA's Blackwell 2.0 architecture. Benchmark data reveals a decisive performance advantage for the NVIDIA card, but the comparison also highlights the distinct design philosophies and target use cases of each product.

FAQ

Q: Which GPU wins in the 3DMark Steel Nomad DX12 benchmark?

A: The NVIDIA GeForce RTX 5090 D wins decisively, scoring 14,326 points versus the AMD Radeon RX 7900M's 4,201 points. This represents a 70.7% advantage for the RTX 5090 D.

Q: How do the two cards compare in OpenCL performance?

A: The RTX 5090 D scores 310,674 in Geekbench OpenCL, which is 58.3% higher than the RX 7900M's score of 129,499. The NVIDIA card's compute advantage is substantial.

Q: What is the memory configuration difference?

A: The RX 7900M features 16 GB of GDDR6 memory on a 256-bit bus with 576.0 GB/s bandwidth, while the RTX 5090 D has 32 GB of GDDR7 memory on a 512-bit bus with 1.79 TB/s bandwidth. The NVIDIA card has double the capacity and over three times the bandwidth.

Q: Which GPU has more shading units?

A: The RTX 5090 D has 21,760 shading units, compared to 4,608 on the RX 7900M. The NVIDIA card also has 680 tensor cores, which the AMD card lacks entirely.

Q: How does the power consumption compare?

A: The RX 7900M has a TDP of 180 W and uses no power connectors as an integrated GPU. The RTX 5090 D has a TDP of 575 W, requires a 1x 16-pin power connector, and has a suggested PSU rating of 950 W.

Q: What are the release timelines for these GPUs?

A: The AMD Radeon RX 7900M was released on October 18, 2023, while the NVIDIA GeForce RTX 5090 D launched on January 29, 2025.

The Verdict

The data paints an unambiguous picture: the RTX 5090 D outperforms the RX 7900M by a massive margin across every benchmark tested. In 3DMark Steel Nomad DX12, the NVIDIA card is 70.7% faster. In OpenCL and Vulkan workloads, it leads by 58.3% and 57.9%, respectively. The wins total 3-0 in favor of the RTX 5090 D.

For users seeking maximum desktop performance, particularly in DX12 gaming and compute-heavy tasks, the RTX 5090 D is the clear choice. Its 104.8 TFLOPS FP32 performance and 32 GB of GDDR7 memory position it as a top-tier workstation and enthusiast GPU. The RX 7900M, by contrast, targets portable devices where the 180 W TDP and integrated form factor are advantages. Its 38.52 TFLOPS FP32 and 16 GB GDDR6 are respectable for a mobile solution, but they cannot compete with a dual-slot desktop card.

The percentile rankings are telling: the RX 7900M sits at the 94th percentile among all GPUs, while the RTX 5090 D ranks at the 92nd percentile. This counterintuitive result reflects the RX 7900M's strong standing within the mobile GPU space, while the RTX 5090 D's average benchmark score of 77,712 is dragged down by its PassMark legacy tests (DX9, DX10, DX11), which are not representative of modern workloads. The RTX 5090 D's launch MSRP is 2,299 USD.

Head-to-Head Benchmarks

The 3DMark Steel Nomad DX12 test shows the most dramatic gap between the two GPUs. The RTX 5090 D's score of 14,326 dwarfs the RX 7900M's 4,201, a 70.7% deficit for AMD. This benchmark is a heavy DX12 load that stresses modern rendering features, and the NVIDIA card's 170 ray tracing cores and 21,760 shading units provide a clear architectural advantage.

In Geekbench OpenCL, the RTX 5090 D achieves 310,674 points versus the RX 7900M's 129,499. The 58.3% delta reflects the NVIDIA card's raw compute throughput: 104.8 TFLOPS FP32 versus 38.52 TFLOPS for AMD. The RTX 5090 D also offers 680 tensor cores, which accelerate AI and machine learning workloads absent from the AMD card's feature set.

The Vulkan benchmark follows a similar pattern. The RTX 5090 D scores 376,915, while the RX 7900M manages 158,760. The 57.9% gap here is consistent with the OpenCL result, indicating that the NVIDIA card's advantage is not API-specific but rather a fundamental performance difference. Both cards support DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, so software compatibility does not explain the disparity.

The RX 7900M's nearest rivals in average benchmark score include the AMD Radeon Pro VII (0.4% higher), NVIDIA Quadro RTX 6000 (4.3% higher), and AMD Radeon Instinct MI60 (5.4% lower). Its average score of 97,487 places it in a professional-grade workstation tier. The RTX 5090 D, with an average score of 77,712, is closest to the AMD Radeon RX 6650M XT (1.1% higher) and RX 6850M XT (1.6% lower), suggesting that its average is heavily influenced by legacy PassMark tests where it scores poorly (e.g., 219 in DX12, 231 in DX10).

Specification Differences

The two GPUs differ in nearly every measurable specification. The RX 7900M uses 5 nm process technology from TSMC with 57,700 million transistors on a 529 mm² die, giving a transistor density of 109.1M per mm². The RTX 5090 D also uses TSMC 5 nm but packs 92,200 million transistors on a larger 750 mm² die, achieving 122.9M per mm² density.

Memory is a major differentiator. The AMD card has 16 GB GDDR6 with a 256-bit bus and 576.0 GB/s bandwidth. The NVIDIA card doubles capacity to 32 GB GDDR7, doubles the bus to 512-bit, and more than triples bandwidth to 1.79 TB/s. Memory clocks also differ: 2250 MHz (18 Gbps effective) for AMD versus 1750 MHz (28 Gbps effective) for NVIDIA.

Compute resources show the NVIDIA card's scale advantage. The RX 7900M has 4,608 shading units, 288 TMUs, and 192 ROPs. The RTX 5090 D has 21,760 shading units, 680 TMUs, and 176 ROPs. While the NVIDIA card has fewer ROPs, its pixel rate of 423.6 GPixel/s still exceeds AMD's 401.3 GPixel/s. Texture rate tells a different story: 1,636.8 GTexel/s for NVIDIA versus 601.9 GTexel/s for AMD.

Clock speeds favor NVIDIA as well. The RTX 5090 D has a base clock of 2017 MHz and boost of 2407 MHz, while the RX 7900M operates at 1825 MHz base and 2090 MHz boost. The RX 7900M's FP16 performance of 77.05 TFLOPS (2:1 ratio) is half its FP32 rate, while the RTX 5090 D achieves 104.8 TFLOPS in both FP16 and FP32 (1:1 ratio).

Form factor and power delivery diverge sharply. The RX 7900M is an integrated GPU (IGP) with no slot width, no power connectors, and a 180 W TDP. The RTX 5090 D is a dual-slot card measuring 304 mm in length, 137 mm in height, and 48 mm in width, requiring a 1x 16-pin connector and a 950 W suggested PSU. The RX 7900M uses PCIe 4.0 x16, while the RTX 5090 D uses PCIe 5.0 x16. Display outputs are portable-device dependent for AMD, whereas NVIDIA provides 1x HDMI 2.1b and 3x DisplayPort 2.1b.

Architecture Differences

The RX 7900M is built on AMD's RDNA 3.0 architecture with the Navi 31 chip, codenamed Plum Bonito. It belongs to the Navi Mobile generation within the RX 7000M series. The RTX 5090 D uses NVIDIA's Blackwell 2.0 architecture with the GB202 chip and is part of the GeForce 50-series. This generation gap is fundamental: RDNA 3.0 launched in late 2023, while Blackwell 2.0 arrived in early 2025.

Ray tracing capabilities differ in scale. The RX 7900M has 72 ray tracing cores, while the RTX 5090 D has 170. The NVIDIA card also includes 680 tensor cores, which AMD's card does not list. This makes the RTX 5090 D significantly more capable for ray-traced workloads and AI-accelerated compute tasks.

Transistor architecture reflects the design goals. The RX 7900M uses 57,700 million transistors with a density of 109.1M per mm², optimized for power efficiency within a 180 W mobile envelope. The RTX 5090 D uses 92,200 million transistors at 122.9M per mm², prioritizing raw performance at the cost of 575 W power consumption.

The predecessor relationships also differ. The RX 7900M succeeds Polaris Mobile, while the RTX 5090 D succeeds GeForce 40 and has a listed successor in GeForce 60. This suggests NVIDIA's product line is positioned for continued generational iteration, whereas AMD's mobile offering represents a more distinctive architectural shift from its Polaris-era predecessor.

FP16 compute reveals a key architectural choice. The RX 7900M delivers 77.05 TFLOPS at a 2:1 ratio to FP32, indicating separate FP16 paths. The RTX 5090 D achieves 104.8 TFLOPS at a 1:1 ratio, meaning unified FP32/FP16 execution. This makes the NVIDIA card more efficient for mixed-precision workloads common in AI inference and scientific computing.