GPU Comparison

The NVIDIA GeForce GTX 670 and the NVIDIA RTX A2000 Mobile represent two very different eras of GPU design, separated by nearly a decade of architectural evolution. The data shows a decisive generational shift, but the specifics of that shift, where the newer card wins, how it wins, and what it leaves behind, reveal a nuanced picture beyond simple age. This analysis pits the 2012 desktop Kepler card against the 2021 mobile Ampere part, using only the benchmark results and specifications provided.

Head-to-Head Benchmarks

The head-to-head data is stark and unambiguous. Across the two shared benchmark tests, the RTX A2000 Mobile wins both, and it does so by a colossal margin. In the Geekbench OpenCL test, the RTX A2000 Mobile scores 56,518 against the GTX 670's 15,341. This represents a deltaPct of -72.9% for the GTX 670, meaning the newer card is approximately 3.7 times faster in this compute-oriented workload. The Geekbench Vulkan test tells a similar story, with the RTX A2000 Mobile scoring 53,146 versus 15,553 for the GTX 670, a deltaPct of -70.7%.

These are not incremental gains; they are transformative leaps. The GTX 670's scores in both tests are nearly identical (15,341 and 15,553), suggesting it is hitting a hard architectural limit in these synthetic workloads. The RTX A2000 Mobile, in contrast, nearly quadruples the output, indicating that its Ampere architecture handles modern compute and graphics APIs with far greater efficiency. The data shows the GTX 670 wins zero of the two head-to-head tests, while the RTX A2000 Mobile takes both. This is not a close contest by any metric, but the magnitude of the victory, consistently above 70%, points to a fundamental advantage in processing throughput per clock and per core.

The Verdict

From the benchmark data alone, the verdict is clear: the NVIDIA RTX A2000 Mobile is the superior GPU in every measurable performance category. Its average benchmark score of 13,821 places it in the 55th percentile of all GPUs, while the GTX 670's average of 12,773 sits in the 52nd percentile. The RTX A2000 Mobile's lead in the head-to-head tests is so substantial that it overshadows any other consideration.

The data suggests that anyone choosing between these two for a modern workload, whether gaming, content creation, or compute, should select the RTX A2000 Mobile. It offers roughly 3.5x the raw compute performance in OpenCL and Vulkan, along with a higher overall benchmark average. The GTX 670, while historically significant, is simply outclassed. The only scenario where the GTX 670 might be considered is if a system literally cannot accommodate a mobile IGP form factor, but from a pure performance standpoint, the RTX A2000 Mobile is the definitive winner. The percentile difference, while modest at 3 points, reinforces the trend: the newer card is positioned higher in the global performance hierarchy.

Architecture Differences

The architectural chasm between these two GPUs explains the benchmark results. The GTX 670 is built on the Kepler architecture, using the GK104 chip fabricated on a 28 nm process at TSMC. It packs 3,540 million transistors into a 294 mm² die, yielding a transistor density of 12.0M per mm². In contrast, the RTX A2000 Mobile uses the Ampere architecture with the GA107 chip, manufactured by Samsung on an 8 nm process. This newer node allows for 8,700 million transistors in a smaller 200 mm² die, achieving a density of 43.5M per mm², over 3.6 times denser than the Kepler chip.

This density advantage translates directly into features. The RTX A2000 Mobile includes 20 ray tracing cores and 80 tensor cores, hardware that the GTX 670 completely lacks. These are not just additions; they represent entirely new processing paradigms for real-time ray-traced lighting and AI-accelerated tasks like DLSS. The GTX 670's Kepler architecture has no equivalent hardware, meaning it cannot accelerate these workloads at all. Furthermore, the RTX A2000 Mobile supports DirectX 12 Ultimate (12_2), while the GTX 670 maxes out at DirectX 12 (11_0). The Vulkan API support also differs significantly, with the newer card supporting version 1.4 versus 1.2.175 on the older one. The fundamental compute capabilities are also divergent: the RTX A2000 Mobile delivers 8.637 TFLOPS of FP32 performance, with a 1:1 FP16 rate of 8.637 TFLOPS, whereas the GTX 670 offers 2.634 TFLOPS of FP32 and no listed FP16 capability.

Specification Differences

The specification sheet reveals a host of differences beyond the architecture. The RTX A2000 Mobile has more shading units (2,560 vs. 1,344), more ROPs (48 vs. 32), and a higher boost clock (1,687 MHz vs. 980 MHz). However, the GTX 670 has more texture mapping units (112 vs. 80) and a wider memory bus (256-bit vs. 128-bit). Memory configuration also diverges: the GTX 670 uses 2 GB of GDDR5, while the RTX A2000 Mobile uses 4 GB of GDDR6. Interestingly, their memory bandwidth is nearly identical, 192.3 GB/s for the GTX 670 and 192.0 GB/s for the RTX A2000 Mobile, despite the different bus widths and memory types.

The power and physical profiles are starkly different. The GTX 670 has a TDP of 170 W, requires a dual-slot cooler, and needs 2x 6-pin power connectors with a suggested 450 W PSU. The RTX A2000 Mobile, by contrast, has a TDP of 95 W, is an IGP (integrated graphics processor) form factor, and requires no power connectors. The bus interface also advances from PCIe 3.0 x16 on the GTX 670 to PCIe 4.0 x16 on the RTX A2000 Mobile. Pixel and texture fill rates favor the newer card significantly: 80.98 GPixel/s vs. 27.44 GPixel/s, and 135.0 GTexel/s vs. 109.8 GTexel/s, respectively. The GTX 670 offers a fixed set of display outputs (2x DVI, 1x HDMI 1.4a, 1x DisplayPort 1.2), while the RTX A2000 Mobile's outputs are listed as "Portable Device Dependent," reflecting its mobile nature.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA RTX A2000 Mobile has a higher average benchmark score of 13,821, compared to 12,773 for the NVIDIA GeForce GTX 670.

Q: How much faster is the RTX A2000 Mobile in Geekbench OpenCL?

A: The RTX A2000 Mobile scores 56,518 in Geekbench OpenCL, while the GTX 670 scores 15,341. This results in a deltaPct of -72.9% for the GTX 670, indicating the newer card is substantially faster.

Q: Does the GTX 670 support ray tracing?

A: No. The GTX 670 has no ray tracing cores listed, while the RTX A2000 Mobile includes 20 RT cores as part of its Ampere architecture.

Q: What is the difference in memory size between the two?

A: The GTX 670 has 2 GB of GDDR5 memory, while the RTX A2000 Mobile has 4 GB of GDDR6 memory. Their memory bandwidth is nearly identical, at 192.3 GB/s and 192.0 GB/s, respectively.

Q: Which GPU has a higher pixel fill rate?

A: The RTX A2000 Mobile has a significantly higher pixel fill rate of 80.98 GPixel/s, compared to 27.44 GPixel/s for the GTX 670.

Q: What are the power requirements for each card?

A: The GTX 670 has a TDP of 170 W, requires 2x 6-pin power connectors, and a suggested 450 W PSU. The RTX A2000 Mobile has a TDP of 95 W and requires no power connectors, as it is an IGP.

Where Each One Wins

The RTX A2000 Mobile wins in virtually every performance-driven category. It dominates in compute benchmarks (OpenCL and Vulkan), offers higher FP32 and FP16 throughput, has dedicated ray tracing and tensor cores, and supports newer API versions like DirectX 12 Ultimate and Vulkan 1.4. It also wins on efficiency, delivering roughly 3.3x the FP32 performance of the GTX 670 while consuming only 95 W versus 170 W. For any modern workload, whether it's gaming with ray tracing, AI inference, or professional 3D rendering, the RTX A2000 Mobile is the clear choice.

The GTX 670's wins are narrower and more historical. It has a wider 256-bit memory bus, which in theory could offer latency advantages, but its memory bandwidth of 192.3 GB/s is effectively identical to the RTX A2000 Mobile's 192.0 GB/s. It also has more TMUs (112 vs. 80), but its texture fill rate of 109.8 GTexel/s is actually lower than the newer card's 135.0 GTexel/s. The only true advantages are its dual-slot desktop form factor with fixed display outputs (2x DVI, 1x HDMI, 1x DisplayPort), which might be preferable for legacy multi-monitor setups that rely on DVI connections. The GTX 670 also has a lower transistor density, which is irrelevant to performance. In summary, the RTX A2000 Mobile wins on compute, features, efficiency, and API support, while the GTX 670 retains a niche edge in legacy connectivity and a physical size that fits traditional desktop slots. For any practical modern use case, the data points overwhelmingly to the RTX A2000 Mobile.