AMD Radeon 890M vs NVIDIA Rubin GPU Comparison

Head-to-Head Benchmarks

The recorded database contains no direct head-to-head benchmark comparisons between the AMD Radeon 890M and the NVIDIA Rubin GPU. The 890M has a full suite of ten benchmark results, while the Rubin GPU currently has zero recorded benchmark scores. This absence of data is itself informative: the Rubin GPU is a server-class accelerator with no graphics API support, which explains the lack of DirectX, OpenCL, or Vulkan test results in the database.

Looking at the 890M's standalone scores provides a baseline for understanding what the comparison would look like if direct measurements existed. The 890M achieves a 3DMark Steel Nomad DX12 score of 590, a Geekbench OpenCL score of 37,254, and a Geekbench Vulkan score of 40,808. In Passmark tests, it scores 33 in DirectX 10, 73 in DirectX 11, 37 in DirectX 12, and 97 in DirectX 9. Its Passmark G2D score is 979, G3D score is 8,076, and GPU compute score is 4,157.

The 890M's average benchmark score across all tests is 9,210, placing it in the 45th percentile of all GPUs in the database. Its nearest rivals include the AMD Radeon Vega 8 with an average score of 9,221 (0.1% higher), the NVIDIA GeForce GTX 960 at 9,273 (0.7% higher), the NVIDIA GeForce GTX 465 at 9,294 (0.9% higher), and the NVIDIA GeForce GTX 850M at 9,302 (1.0% higher). These margins are tight, indicating the 890M sits in a competitive cluster of mid-range performance parts.

The Rubin GPU, by contrast, has an avgBenchmarkScore of 0 and a 50th percentile ranking, but this ranking appears to be a placeholder given the absence of any actual test data. Without benchmark entries, the database cannot generate meaningful win counts; both winsA and winsB are recorded as zero.

Architecture Differences

The architectural gap between these two processors is substantial and reflects their entirely different design goals. The AMD Radeon 890M uses the Strix Point chip with RDNA 3.5 architecture, part of the Navi III IGP generation for mobile platforms. It is manufactured on a 4 nm process at TSMC, with 34,000 million transistors on a 233 mm² die, yielding a transistor density of 145.9 million per square millimeter.

The NVIDIA Rubin GPU uses the GR100 chip with Rubin architecture, belonging to the Server Rubin (Rxx) generation for data center deployment. It is built on a more advanced 3 nm process, also at TSMC, with 336,000 million transistors on a 1,456 mm² die, resulting in a transistor density of 230.8 million per square millimeter. The Rubin GPU packs nearly ten times the transistor count into a die roughly six times larger, with significantly higher packing density.

Clock speeds diverge sharply. The 890M runs at a base clock of 400 MHz and boosts to 2,900 MHz. The Rubin GPU has a base clock of 700 MHz and a boost clock of 2,267 MHz. Despite the lower boost clock, the Rubin GPU's massive shader array more than compensates in raw throughput.

Memory architecture could not be more different. The 890M uses system shared memory with a bus width and type also listed as system shared, making bandwidth system dependent. The Rubin GPU uses 288 GB of HBM4 memory on a 16,384-bit bus, delivering 22.1 TB/s of bandwidth with a memory clock of 2,695 MHz (10.8 Gbps effective). This is a server-class memory subsystem designed for massive data movement, not for graphics output.

Compute resources differ by an order of magnitude. The 890M has 1,024 shading units, 64 texture mapping units, 32 raster operation pipelines, and 16 ray tracing cores. The Rubin GPU has 28,672 shading units, 896 texture mapping units, 24 ROPs, and 896 tensor cores. The Rubin GPU lists no ray tracing cores, while the 890M has dedicated RT hardware.

The 890M delivers 92.80 GPixel/s pixel rate, 185.6 GTexel/s texture rate, 5.939 TFLOPS FP32, and 5.939 TFLOPS FP16 at a 1:1 ratio. The Rubin GPU produces 54.41 GPixel/s pixel rate (lower due to fewer ROPs), 2,031.2 GTexel/s texture rate, 130.0 TFLOPS FP32, and 260.0 TFLOPS FP16 at a 2:1 ratio. The Rubin GPU's FP16 throughput is double its FP32, indicating a design optimized for tensor and AI workloads.

Where Each One Wins

The AMD Radeon 890M wins in scenarios requiring integrated graphics with low power consumption. Its TDP is 15 W, it uses no power connectors, and it is an IGP (integrated graphics processor) with PCIe 4.0 x8 interface. It supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, making it suitable for gaming and general graphics workloads on mobile platforms. Its display outputs are portable device dependent, meaning it can drive laptop and handheld screens.

The 890M's benchmark results show balanced performance across graphics APIs. Its highest Passmark scores come from DirectX 9 (97) and DirectX 11 (73), while DirectX 10 (33) and DirectX 12 (37) are lower. The Geekbench Vulkan score of 40,808 exceeds the OpenCL score of 37,254, suggesting Vulkan is the stronger compute path for this IGP.

The NVIDIA Rubin GPU wins in raw compute density and memory bandwidth, but it is not a graphics product at all. It has no display outputs, no DirectX support, no OpenGL support, and no Vulkan support. Its API list is entirely N/A. This is a server accelerator designed for data center workloads where rendering output is irrelevant. Its 22.1 TB/s memory bandwidth, 288 GB HBM4 capacity, and 260 TFLOPS FP16 performance target AI training, scientific simulation, and large-scale data processing.

The Rubin GPU's SXM Module slot width and 2,300 W TDP with a suggested PSU of 2,700 W confirm it is a rack-mounted component, not a desktop or mobile part. The 890M, with its 15 W TDP and IGP form factor, is the opposite: a power-sipping component integrated into a mobile processor.

The Verdict

The data supports a clear separation of purpose. The AMD Radeon 890M is an integrated graphics solution for portable devices, evidenced by its 15 W TDP, IGP slot width, system shared memory, and portable device dependent display outputs. Its benchmark scores, while modest in absolute terms, place it in the 45th percentile of all GPUs, with nearest rivals within 1% of its average score. This indicates it performs competitively within its class of integrated and older discrete GPUs.

The NVIDIA Rubin GPU is a server accelerator with no graphics functionality. The absence of any benchmark scores in the database, combined with its N/A API support and lack of display outputs, means it cannot be evaluated through the same graphics-oriented tests as the 890M. Its specifications target a completely different workload profile: 28672 shading units, 896 tensor cores, 288 GB HBM4, and 260 TFLOPS FP16. These figures suggest massive parallel compute capability, but the database contains no measurements to confirm real-world performance.

Who should pick which depends entirely on the use case as defined by the recorded data. A mobile device needing integrated graphics with DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4 support would use the 890M. A server deployment requiring high-bandwidth memory and tensor core compute, with no display output requirements, would use the Rubin GPU. There is no overlap in their target applications, and the benchmark data reflects this complete divergence.

FAQ

Q: Which GPU has higher FP32 performance?

A: The NVIDIA Rubin GPU delivers 130.0 TFLOPS FP32, while the AMD Radeon 890M delivers 5.939 TFLOPS FP32.

Q: What memory does each GPU use?

A: The Radeon 890M uses system shared memory with bandwidth that is system dependent. The Rubin GPU uses 288 GB of HBM4 memory on a 16,384-bit bus with 22.1 TB/s bandwidth.

Q: Does the NVIDIA Rubin GPU support DirectX?

A: No. The Rubin GPU lists DirectX as N/A, along with OpenGL and Vulkan, and has no display outputs.

Q: What is the power consumption difference?

A: The Radeon 890M has a TDP of 15 W with no power connectors. The Rubin GPU has a TDP of 2,300 W and a suggested PSU of 2,700 W.

Q: How does the Radeon 890M compare to its nearest rivals?

A: The 890M has an average benchmark score of 9,210, which is 0.1% lower than the AMD Radeon Vega 8, 0.7% lower than the NVIDIA GeForce GTX 960, 0.9% lower than the GTX 465, and 1.0% lower than the GTX 850M.

Q: What is the transistor count for each chip?

A: The Radeon 890M uses 34,000 million transistors on a 233 mm² die. The Rubin GPU uses 336,000 million transistors on a 1,456 mm² die.

Specification Differences

| Specification | AMD Radeon 890M | NVIDIA Rubin GPU |

|---|---|---|

| Architecture | RDNA 3.5 | Rubin |

| Process Node | 4 nm | 3 nm |

| Transistors | 34,000 million | 336,000 million |

| Die Size | 233 mm² | 1,456 mm² |

| Base Clock | 400 MHz | 700 MHz |

| Boost Clock | 2,900 MHz | 2,267 MHz |

| Memory Size | System Shared | 288 GB |

| Memory Type | System Shared | HBM4 |

| Memory Bus Width | System Shared | 16,384 bit |

| Memory Bandwidth | System Dependent | 22.1 TB/s |

| Shading Units | 1,024 | 28,672 |

| TMUs | 64 | 896 |

| ROPs | 32 | 24 |

| RT Cores | 16 | N/A |

| Tensor Cores | N/A | 896 |

| Pixel Rate | 92.80 GPixel/s | 54.41 GPixel/s |

| Texture Rate | 185.6 GTexel/s | 2,031.2 GTexel/s |

| FP32 | 5.939 TFLOPS | 130.0 TFLOPS |

| FP16 | 5.939 TFLOPS (1:1) | 260.0 TFLOPS (2:1) |

| TDP | 15 W | 2,300 W |

| Slot Width | IGP | SXM Module |

| Power Connectors | None | N/A |

| Suggested PSU | N/A | 2,700 W |

| Bus Interface | PCIe 4.0 x8 | PCIe 6.0 x16 |

| Display Outputs | Portable Device Dependent | No outputs |

| DirectX | 12 Ultimate (12_2) | N/A |

| OpenGL | 4.6 | N/A |

| Vulkan | 1.4 | N/A |

| Release Date | 2024-07-14 | 2025-12-31 |

| Predecessor | Navi II IGP | Server Blackwell |