AMD Radeon 890M vs NVIDIA GeForce RTX 5070 Mobile Comparison
AMD Radeon 890M
GeForce RTX 5070 Mobile
PERFORMANCE BENCHMARKS
Analysis: AMD Radeon 890M vs NVIDIA GeForce RTX 5070 Mobile
The AMD Radeon 890M and NVIDIA GeForce RTX 5070 Mobile occupy entirely different tiers of mobile graphics performance. The recorded data shows a decisive overall victory for the NVIDIA part, which wins 8 of the 9 head-to-head benchmark comparisons. The AMD Radeon 890M secures a single win in the PassMark G2D test, which measures 2D graphics throughput. The RTX 5070 Mobile delivers compute and 3D rendering scores that are multiple times higher, positioning it as a discrete-class solution, while the 890M functions as an integrated graphics processor (IGP) with a significantly lower power envelope.
Where Each One Wins
The RTX 5070 Mobile dominates every compute and 3D graphics workload recorded in the database. Its largest margin comes in the Geekbench OpenCL test, where it scores 122,238 against the 890M's 37,254, a difference of 69.5%. The Geekbench Vulkan result shows a similar pattern, with the NVIDIA part scoring 116,960 versus 40,808, a 65.1% advantage. In DirectX 12 (PassMark), the RTX 5070 Mobile scores 93 compared to 37, a 60.2% lead. These results indicate that the RTX 5070 Mobile is built for high-end gaming and GPU-accelerated workloads, while the 890M is positioned for basic graphics output and light gaming.
The AMD Radeon 890M's only victory is in the PassMark G2D benchmark, scoring 979 against the RTX 5070 Mobile's 896, a 9.3% advantage. This test focuses on 2D graphics operations, such as window rendering and desktop compositing. The 890M's win here suggests that its integrated design, with memory bandwidth dependent on the system's shared memory configuration, handles 2D tasks efficiently. However, this single win does little to offset the NVIDIA part's overwhelming lead in all other categories, including DirectX 9, 10, and 11, where the RTX 5070 Mobile leads by margins ranging from 54.7% to 74.4%.
The use-case split is clear. The RTX 5070 Mobile is the choice for demanding applications: modern DirectX 12 titles, Vulkan-based games, and GPU compute tasks like OpenCL workloads. The 890M is suited for everyday productivity, 2D applications, and older or less demanding games. The PassMark G3D score reinforces this, with the RTX 5070 Mobile at 20,355 versus 8,076, a 60.3% gap. The data shows no overlap in their intended performance envelopes.
Architecture Differences
The two processors are built on different architectures and process nodes. The AMD Radeon 890M uses the RDNA 3.5 architecture, fabricated on a 4 nm process at TSMC. The chip, codenamed Strix Point, contains 34,000 million transistors on a 233 mm² die, yielding a transistor density of 145.9 million transistors per square millimeter. The NVIDIA GeForce RTX 5070 Mobile uses the Blackwell 2.0 architecture on a 5 nm process, also from TSMC. Its GB206 chip contains 21,900 million transistors on a 181 mm² die, with a density of 121.0 million transistors per square millimeter.
The core configurations differ substantially. The 890M has 1024 shading units, 64 texture mapping units (TMUs), and 32 raster operation units (ROPs). It includes 16 ray tracing cores and no tensor cores. The RTX 5070 Mobile has 4608 shading units, 144 TMUs, and 48 ROPs, along with 36 ray tracing cores and 144 tensor cores. This means the NVIDIA part has 4.5 times the shading units, 2.25 times the TMUs, and 1.5 times the ROPs. The presence of 144 tensor cores gives the RTX 5070 Mobile dedicated hardware for AI and deep learning tasks, a feature the AMD part lacks entirely.
Memory architecture is another major divergence. The 890M uses system-shared memory, with size, type, bus width, and bandwidth all dependent on the host system's configuration. The RTX 5070 Mobile has dedicated 8 GB of GDDR7 memory on a 128-bit bus, delivering 384.0 GB/s of bandwidth. This dedicated memory ensures consistent performance regardless of system RAM, while the 890M's performance can vary based on system memory speed and capacity. The clock speeds also differ: the 890M boosts to 2900 MHz from a 400 MHz base, while the RTX 5070 Mobile boosts to 1425 MHz from a 907 MHz base. Despite lower clocks, the NVIDIA part achieves higher throughput due to its larger core count and dedicated memory bandwidth.
FAQ
Q: Which GPU has a higher average benchmark score?
A: The NVIDIA GeForce RTX 5070 Mobile has an average benchmark score of 29,928, while the AMD Radeon 890M averages 9,210. The RTX 5070 Mobile's score places it in the 75th percentile of all GPUs, compared to the 890M's 45th percentile.
Q: How do the two compare in terms of power consumption?
A: The AMD Radeon 890M has a TDP of 15 W, while the NVIDIA GeForce RTX 5070 Mobile has a TDP of 50 W. The 890M's lower power draw reflects its integrated design, whereas the RTX 5070 Mobile consumes more power for its higher performance.
Q: What are the memory specifications for each?
A: The RTX 5070 Mobile has 8 GB of GDDR7 memory on a 128-bit bus with 384.0 GB/s bandwidth. The 890M uses system-shared memory, meaning its size, type, and bandwidth are dependent on the host system's configuration.
Q: Which GPU is better for DirectX 11 workloads?
A: The RTX 5070 Mobile scores 192 in PassMark's DirectX 11 test, compared to the 890M's 73, a 62% lead. The NVIDIA part is substantially faster in this legacy API.
Q: Does the AMD Radeon 890M support ray tracing?
A: Yes, the 890M includes 16 ray tracing cores. However, the RTX 5070 Mobile has 36 ray tracing cores, and its overall compute advantage makes it far more capable for ray-traced workloads.
Q: What is the transistor count difference?
A: The AMD Radeon 890M has 34,000 million transistors on a 233 mm² die, while the NVIDIA GeForce RTX 5070 Mobile has 21,900 million transistors on a 181 mm² die. The AMD chip has a higher transistor density at 145.9M per mm² versus 121.0M per mm².
Specification Differences
The two GPUs differ across nearly every specification field. The process node favors AMD, with a 4 nm process versus NVIDIA's 5 nm. The 890M has a larger die at 233 mm² and more transistors at 34,000 million, while the RTX 5070 Mobile uses a smaller 181 mm² die with 21,900 million transistors. Clock speeds show the 890M boosting to 2900 MHz versus 1425 MHz for the RTX 5070 Mobile, but the NVIDIA part compensates with more cores and dedicated memory.
The shading unit count is 1024 for AMD and 4608 for NVIDIA. TMUs are 64 versus 144, and ROPs are 32 versus 48. The RTX 5070 Mobile has 36 ray tracing cores and 144 tensor cores, while the 890M has 16 ray tracing cores and no tensor cores. Pixel rates differ: 92.80 GPixel/s for the 890M versus 68.40 GPixel/s for the RTX 5070 Mobile, though the AMD part's higher pixel rate does not translate to better 3D performance. Texture rates are 185.6 GTexel/s for AMD and 205.2 GTexel/s for NVIDIA. FP32 and FP16 compute are 5.939 TFLOPS for the 890M and 13.13 TFLOPS for the RTX 5070 Mobile.
Memory is a major differentiator: the 890M uses system-shared memory, while the RTX 5070 Mobile has 8 GB of GDDR7 on a 128-bit bus with 384.0 GB/s bandwidth. The bus interface also differs, with the 890M using PCIe 4.0 x8 and the RTX 5070 Mobile using PCIe 5.0 x16. TDP is 15 W for AMD and 50 W for NVIDIA. Release dates are July 14, 2024 for the 890M and April 14, 2025 for the RTX 5070 Mobile. Both use the same API set: DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.
Head-to-Head Benchmarks
The Geekbench OpenCL test shows the largest performance gap. The RTX 5070 Mobile scores 122,238 against the 890M's 37,254, a 69.5% deficit for the AMD part. This test stresses general-purpose GPU compute, and the results confirm that the NVIDIA part's 4608 shading units and dedicated memory deliver massive compute throughput. The Geekbench Vulkan test follows a similar trend, with the RTX 5070 Mobile at 116,960 versus 40,808, a 65.1% lead, indicating strong performance in modern graphics APIs.
In PassMark's DirectX tests, the RTX 5070 Mobile leads across the board. DirectX 10 shows 129 versus 33, a 74.4% margin, the largest percentage gap in any test. DirectX 11 shows 192 versus 73, a 62% lead. DirectX 12 shows 93 versus 37, a 60.2% advantage. Even DirectX 9, the oldest API tested, shows a 54.7% lead for the NVIDIA part with 214 versus 97. These results demonstrate consistent superiority across both legacy and modern rendering paths.
The PassMark G3D test, which measures overall 3D graphics performance, shows the RTX 5070 Mobile at 20,355 versus 8,076, a 60.3% difference. This is the most representative test for gaming performance, and the gap is substantial. The PassMark GPU Compute test shows 8,279 versus 4,157, a 49.8% lead for the RTX 5070 Mobile, the smallest percentage margin in the NVIDIA part's wins. The only AMD win is PassMark G2D, where the 890M scores 979 versus 896, a 9.3% edge. This test measures 2D rendering and desktop operations, areas where the integrated 890M appears to excel.
The Verdict
The benchmark data indicates that the NVIDIA GeForce RTX 5070 Mobile is a fundamentally more powerful GPU than the AMD Radeon 890M. With an average benchmark score of 29,928 versus 9,210, the RTX 5070 Mobile delivers roughly three times the overall performance. Its percentile ranking of 75 versus 45 for the 890M places it in a higher performance class entirely. The nearest rivals for the RTX 5070 Mobile include the NVIDIA GeForce RTX 3070 Ti and RTX 2080 Ti, with scores of 29,945 and 29,783 respectively, showing it trades blows with previous-generation desktop high-end cards. The 890M's nearest rivals are the AMD Radeon Vega 8 and NVIDIA GeForce GTX 960, with scores of 9,221 and 9,273, indicating it sits in the entry-level integrated graphics tier.
Users seeking a GPU for modern gaming, 3D rendering, or compute workloads should select the RTX 5070 Mobile. Its 8 GB of GDDR7 memory, 144 tensor cores, and 13.13 TFLOPS of FP32 performance make it capable of handling demanding tasks. The 50 W TDP is higher than the 890M's 15 W, but the performance gains justify the power draw in systems designed for discrete graphics. The 890M, with its 15 W TDP and system-shared memory, is appropriate for ultraportable laptops where battery life and thermals are priorities. Its single win in PassMark G2D suggests it handles 2D tasks well, but its 5.939 TFLOPS of compute and 8,076 G3D score limit it to light gaming and productivity. The data is unambiguous: these are different classes of hardware, and the choice depends entirely on whether the system needs integrated efficiency or discrete-level performance.