AMD Radeon 660M vs NVIDIA RTX A2000 Mobile Comparison
AMD Radeon 660M
RTX A2000 Mobile
PERFORMANCE BENCHMARKS
Analysis: AMD Radeon 660M vs NVIDIA RTX A2000 Mobile
The NVIDIA RTX A2000 Mobile and AMD Radeon 660M are both integrated-class mobile GPUs, yet they deliver vastly different performance profiles. The data shows a clear overall winner: the RTX A2000 Mobile leads the Radeon 660M by a razor-thin 0.1% in average benchmark score (13821 vs 13812), but this aggregate hides a massive divide in compute and graphics workloads. In the two head-to-head tests available, the NVIDIA part wins both, with leads of 338.9% in Geekbench OpenCL and 260.4% in Geekbench Vulkan. The Radeon 660M, however, is a far more power-efficient design, drawing 40 W versus 95 W, and offers a more modern 6 nm process node. This is a classic case of raw throughput versus efficiency, and the choice depends entirely on whether the workload prioritizes compute density or sustained battery life.
FAQ
Q: Which GPU has the higher average benchmark score?
A: The NVIDIA RTX A2000 Mobile edges out the AMD Radeon 660M with an average score of 13821, compared to 13812 for the AMD part. This is a delta of just 0.1%, making them statistically tied in aggregate.
Q: How large is the performance gap in compute workloads?
A: The gap is enormous. In Geekbench OpenCL, the RTX A2000 Mobile scores 56518, which is 338.9% higher than the Radeon 660M’s 12876. In Geekbench Vulkan, the NVIDIA part scores 53146, a 260.4% advantage over the AMD’s 14748.
Q: What are the power consumption differences?
A: The RTX A2000 Mobile is rated at 95 W TDP, while the Radeon 660M is rated at 40 W. This makes the AMD part more than twice as power-efficient on paper, a critical factor for thin-and-light laptops.
Q: Do both GPUs support the same modern APIs?
A: Yes. Both the RTX A2000 Mobile and Radeon 660M support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. This means feature-level compatibility for modern games and applications is identical.
Q: Which GPU has more shading units and ray tracing cores?
A: The RTX A2000 Mobile has 2560 shading units and 20 ray tracing cores. The Radeon 660M has 384 shading units and 6 ray tracing cores. The NVIDIA part has a 6.7x advantage in shading units and a 3.3x advantage in RT cores.
Q: How do their transistor densities compare?
A: The Radeon 660M uses a 6 nm TSMC process with 13,100 million transistors on a 208 mm² die, yielding 63.0M transistors per mm². The RTX A2000 Mobile uses an 8 nm Samsung process with 8,700 million transistors on a 200 mm² die, yielding 43.5M per mm².
Where Each One Wins
The NVIDIA RTX A2000 Mobile wins decisively in raw compute and graphics throughput. It dominates both head-to-head benchmarks: Geekbench OpenCL (56518 vs 12876) and Geekbench Vulkan (53146 vs 14748). This makes it the clear choice for workloads like GPU-accelerated rendering, machine learning inference, or any task that leverages OpenCL or Vulkan compute. Its FP32 performance of 8.637 TFLOPS dwarfs the Radeon 660M’s 1,459.2 GFLOPS, and its texture rate of 135.0 GTexel/s is nearly 3x higher than the AMD’s 45.60 GTexel/s. For any scenario where the GPU is the bottleneck and power draw is not a constraint, the RTX A2000 Mobile is the superior part.
The AMD Radeon 660M wins on efficiency and integration. Its 40 W TDP is 55 W lower than the NVIDIA part’s 95 W, making it far better suited for ultraportable designs with limited cooling and battery capacity. It also uses a more advanced 6 nm process node from TSMC, versus Samsung’s 8 nm node, which contributes to its higher transistor density (63.0M/mm² vs 43.5M/mm²). The Radeon 660M’s memory is system shared and system dependent, which simplifies system design and reduces cost, but this also means its performance scales with the host system’s RAM speed. For users who prioritize battery life and quiet operation over peak frame rates, the 660M is the better fit.
Architecture Differences
The two GPUs come from fundamentally different architectural lineages. The RTX A2000 Mobile is built on NVIDIA’s Ampere architecture, using the GA107 chip. It is a discrete-class mobile GPU with dedicated GDDR6 memory, a 128-bit bus, and 192.0 GB/s of bandwidth. The architecture includes 80 tensor cores alongside 20 ray tracing cores, enabling advanced AI acceleration and hardware-accelerated ray tracing. Its FP16 performance matches FP32 at 8.637 TFLOPS (1:1 ratio), which is beneficial for workloads that can use reduced precision.
The Radeon 660M is based on AMD’s RDNA 2.0 architecture, using the Rembrandt chip. It is an integrated graphics processor (IGP) that shares system memory, with no dedicated VRAM. The architecture implements 6 ray tracing cores but no tensor cores, meaning it lacks dedicated AI acceleration hardware. Its FP16 performance is 2.918 TFLOPS (2:1 ratio), which is exactly double its FP32 rate of 1,459.2 GFLOPS. This indicates a different compute strategy: RDNA 2.0 prioritizes FP32 throughput while offering half-rate FP16, whereas Ampere offers full-rate FP16. Both support DirectX 12 Ultimate and Vulkan 1.4, but the underlying execution resources are vastly different in scale.
Specification Differences
The most significant specification gap is in compute units. The RTX A2000 Mobile has 2560 shading units, 80 TMUs, and 48 ROPs, while the Radeon 660M has 384 shading units, 24 TMUs, and 16 ROPs. This translates to a 6.7x advantage in shader count and a 3x advantage in ROPs for the NVIDIA part. Pixel rate is 80.98 GPixel/s for the RTX A2000 Mobile versus 30.40 GPixel/s for the Radeon 660M, a 2.7x difference. Texture rate is 135.0 GTexel/s versus 45.60 GTexel/s, a 3.0x difference.
Clock speeds favor AMD in raw frequency. The Radeon 660M has a base clock of 1500 MHz and a boost of 1900 MHz, while the RTX A2000 Mobile has a base of 1215 MHz and a boost of 1687 MHz. However, the NVIDIA part’s massive shader count overcomes this frequency deficit. Memory is another key differentiator: the RTX A2000 Mobile has 4 GB of GDDR6 on a 128-bit bus with 192.0 GB/s bandwidth, while the Radeon 660M uses system shared memory with system dependent bandwidth. The bus interface also differs, with the NVIDIA part using PCIe 4.0 x16 and the AMD part using PCIe 4.0 x8. TDP is 95 W for NVIDIA and 40 W for AMD. Finally, the process node differs: 8 nm Samsung for NVIDIA versus 6 nm TSMC for AMD.
Head-to-Head Benchmarks
The two available head-to-head benchmarks tell a stark story. In Geekbench OpenCL, the RTX A2000 Mobile scores 56518 against the Radeon 660M’s 12876, a delta of 338.9%. This is not a marginal win; it is a crushing defeat. The NVIDIA part delivers over 4.4x the OpenCL score of the AMD part. This benchmark likely reflects the RTX A2000 Mobile’s far higher FP32 throughput (8.637 TFLOPS) and dedicated memory bandwidth (192.0 GB/s), which are critical for compute-heavy OpenCL kernels.
In Geekbench Vulkan, the margin narrows slightly but remains enormous. The RTX A2000 Mobile scores 53146, while the Radeon 660M scores 14748, a delta of 260.4%. The NVIDIA part still achieves over 3.6x the Vulkan score. Vulkan is often more sensitive to driver optimization and CPU overhead, which may explain the smaller (but still massive) gap. The RTX A2000 Mobile’s 20 ray tracing cores and 80 tensor cores likely contribute to its strength in Vulkan workloads that leverage these features. The Radeon 660M’s 6 ray tracing cores are simply outmatched.
These results indicate that the Radeon 660M is not competitive with the RTX A2000 Mobile in any measured compute scenario. The aggregate benchmark scores (13821 vs 13812) are nearly identical only because the AMD part has no Passmark scores listed—only two Geekbench entries—which skews the average. If Passmark tests were available for the 660M, the aggregate would likely diverge further.
The Verdict
The data is unambiguous: the NVIDIA RTX A2000 Mobile is the superior GPU for performance-intensive tasks. It wins every head-to-head benchmark, offers 6.7x more shading units, 3.3x more ray tracing cores, and delivers 8.637 TFLOPS of FP32 compute versus 1,459.2 GFLOPS for the AMD Radeon 660M. Its 4 GB of GDDR6 memory with 192.0 GB/s bandwidth provides a dedicated, high-speed memory pool that the 660M’s shared system memory cannot match. Any user needing maximum graphics throughput, compute acceleration, or ray tracing performance should choose the RTX A2000 Mobile without hesitation.
However, the Radeon 660M is not without merit. Its 40 W TDP is a 58% reduction from the RTX A2000 Mobile’s 95 W, making it an ideal choice for ultra-thin laptops where thermal and power budgets are extremely tight. Its 6 nm process node from TSMC is more advanced than NVIDIA’s 8 nm Samsung node, resulting in higher transistor density (63.0M/mm² vs 43.5M/mm²). For users who rarely push the GPU to its limits and prioritize battery life, silent operation, and portability, the 660M is the more rational pick. The 0.1% aggregate score difference suggests that in mixed, non-compute workloads, the two parts may be closer than the Geekbench results imply—but the data available does not support that conclusion. The verdict is simple: pick the RTX A2000 Mobile for performance, or the Radeon 660M for efficiency. There is no middle ground.