NVIDIA GeForce GTX 670MX vs NVIDIA RTX A400 Comparison

The NVIDIA RTX A400 and the NVIDIA GeForce GTX 670MX represent two distinct eras of graphics hardware, separated by over a decade of architectural evolution. The data reveals a massive performance gulf, but the comparison is more nuanced than a simple list of specifications. The A400 is a modern, power-efficient workstation part built on Ampere, while the GTX 670MX is a legacy mobile gaming chip from the Kepler generation. Benchmark results show the A400 dominating in compute and graphics workloads, but the GTX 670MX's higher shading unit count and texture fillrate point to a different design philosophy that was common in its era.

FAQ

Q: How much faster is the NVIDIA RTX A400 than the GTX 670MX in OpenCL?

A: In the Geekbench OpenCL test, the RTX A400 scores 22844, which is 273% higher than the GTX 670MX's score of 6125. This indicates a massive advantage for the newer card in general-purpose compute tasks.

Q: What about the Vulkan performance gap between these two GPUs?

A: The RTX A400's Geekbench Vulkan score of 22237 is 318.3% higher than the GTX 670MX's 5316. This is the largest delta in the head-to-head data, showing the A400's modern architecture excels in this low-level graphics API.

Q: Which GPU has a higher transistor density?

A: The RTX A400, built on an 8 nm Samsung process, has a transistor density of 43.5M per mm². This is significantly denser than the GTX 670MX, which has a density of 12.0M per mm² on TSMC's 28 nm node.

Q: Does the RTX A400 support ray tracing or tensor cores?

A: Yes, the RTX A400 includes 6 RT cores and 24 tensor cores, which are absent from the GTX 670MX's Kepler architecture. This gives the A400 features for ray-traced workloads and AI acceleration that the older card cannot perform.

Q: How do the average benchmark scores compare?

A: The RTX A400 has an average benchmark score of 6078, which is 6.2% higher than the GTX 670MX's average score of 5721. This places the A400 in the 35th percentile of all GPUs, while the 670MX sits in the 33rd percentile.

Q: Which GPU has a wider memory bus?

A: The GTX 670MX has a 192-bit memory bus, which is wider than the RTX A400's 64-bit bus. However, the A400's GDDR6 memory at 12 Gbps effective provides higher total bandwidth of 96.00 GB/s compared to the 670MX's 67.20 GB/s.

Architecture Differences

The RTX A400 and GTX 670MX are built on fundamentally different architectures. The A400 uses the Ampere architecture on a chip designated GA107, while the GTX 670MX uses the older Kepler architecture on a GK104 chip. The process technology shows a major generational leap: the A400 is fabricated on an 8 nm Samsung process, whereas the GTX 670MX uses a 28 nm TSMC process. This explains the stark difference in transistor density — 43.5M per mm² for the A400 versus 12.0M per mm² for the 670MX.

The A400's GA107 chip packs 8,700 million transistors into a 200 mm² die, while the GTX 670MX's GK104 contains 3,540 million transistors on a larger 294 mm² die. Despite having fewer shading units (768 vs 960), the A400 compensates with much higher clock speeds. The A400 runs at a base clock of 1417 MHz and a boost of 1762 MHz, compared to the 670MX's fixed 601 MHz. This clock advantage, combined with architectural improvements, drives the A400's superior FP32 performance of 2.706 TFLOPS versus the 670MX's 1,153.9 GFLOPS.

Feature support is another major differentiator. The A400 includes 6 RT cores and 24 tensor cores, enabling hardware-accelerated ray tracing and AI workloads. The GTX 670MX has no such capabilities. API support also differs: the A400 supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the 670MX is limited to DirectX 12 (11_0) and Vulkan 1.2.175. The A400 also has a newer PCIe interface (4.0 x8) compared to the 670MX's PCIe 3.0 x16, though the older card's x16 link offers more lanes.

Head-to-Head Benchmarks

The head-to-head data includes only two benchmark tests, but both show a decisive victory for the RTX A400. In Geekbench OpenCL, the A400 scores 22844 against the 670MX's 6125, a delta of 273%. This is a massive margin that reflects the A400's modern compute capabilities, including its tensor cores and higher memory bandwidth. The OpenCL result suggests the A400 is not just a gaming card but a competent compute device for workstation tasks.

The Geekbench Vulkan test shows an even larger gap. The A400's score of 22237 is 318.3% higher than the 670MX's 5316. Vulkan is a low-overhead API that benefits from efficient hardware design, and the A400's Ampere architecture clearly handles it far better than the Kepler-based 670MX. This result indicates the A400 can leverage modern graphics APIs to their full potential, while the 670MX is bottlenecked by its older design and limited driver support for newer features.

It is notably the GTX 670MX has no wins in the head-to-head benchmarks. The RTX A400 wins both tests, with a total of 2 wins and 0 losses. This aligns with the average benchmark scores, where the A400's 6078 edges out the 670MX's 5721. However, the average scores are much closer than the individual head-to-head results, suggesting that the 670MX performs relatively better in some legacy tests that are not part of the head-to-head comparison. The A400's nearest rivals in the benchmark database include the NVIDIA GeForce MX230 (deltaPct 0) and Quadro P2000 (deltaPct 0.5), while the 670MX sits near the Intel Iris Pro Graphics P6300 (deltaPct 0.2) and GeForce GTX 550 Ti (deltaPct -0.2).

Specification Differences

The two GPUs differ across nearly every specification field. The RTX A400 has 4 GB of GDDR6 memory, while the GTX 670MX has 3 GB of GDDR5. The memory bus width is a notable difference: the A400 uses a 64-bit bus, but the 670MX uses a wider 192-bit bus. Despite the narrower bus, the A400 achieves higher bandwidth at 96.00 GB/s versus 67.20 GB/s, thanks to its faster memory clock of 1500 MHz (12 Gbps effective) compared to 700 MHz (2.8 Gbps effective).

The shading unit count favors the 670MX, which has 960 units to the A400's 768. The 670MX also has more texture mapping units (80 vs 24) and more raster output units (24 vs 16). However, the A400's higher clock speeds give it a higher pixel rate of 28.19 GPixel/s versus 12.02 GPixel/s, and a texture rate of 42.29 GTexel/s versus 48.08 GTexel/s. The 670MX actually wins in texture rate due to its higher TMU count, despite its lower clocks.

Power characteristics differ significantly. The A400 has a TDP of 50 W with no power connectors required and a suggested PSU of 250 W. The GTX 670MX has a TDP of 75 W and also uses no power connectors, but no suggested PSU is listed. The A400 is a single-slot card measuring 163 mm in length, while the 670MX's dimensions are not specified, as it is a portable device dependent component. Display outputs also differ: the A400 offers 4x mini-DisplayPort 1.4a, while the 670MX's outputs are listed as portable device dependent. The production status shows the A400 is active, while the 670MX is end-of-life.

Where Each One Wins

The RTX A400 is the clear winner in modern compute and graphics workloads. Its 273% lead in OpenCL and 318.3% lead in Vulkan demonstrate that it is far more capable of handling current applications, including those that use ray tracing or AI acceleration via its 6 RT cores and 24 tensor cores. The A400 also wins in memory bandwidth (96.00 GB/s vs 67.20 GB/s), pixel rate (28.19 GPixel/s vs 12.02 GPixel/s), and FP32 performance (2.706 TFLOPS vs 1,153.9 GFLOPS). Its lower TDP of 50 W makes it more power-efficient, and its active production status means it remains a viable purchase for workstation use.

The GTX 670MX retains a few technical advantages, though they do not translate to benchmark victories in the data provided. It has more shading units (960 vs 768), more TMUs (80 vs 24), and more ROPs (24 vs 16). Its texture rate of 48.08 GTexel/s is higher than the A400's 42.29 GTexel/s, suggesting it might handle certain texture-heavy workloads better, though this is not reflected in the available benchmark scores. The wider 192-bit memory bus could theoretically benefit some memory-intensive tasks, but the A400's higher bandwidth negates this in practice. The 670MX's portable device dependent design suggests it was intended for mobile systems, where its lower pixel rate and older architecture were acceptable at the time.

The Verdict

The data unequivocally favors the NVIDIA RTX A400 for anyone prioritizing modern performance. The A400's 273% and 318.3% leads in the head-to-head benchmarks make it the obvious choice for compute-heavy or Vulkan-based applications. Its support for DirectX 12 Ultimate, Vulkan 1.4, and hardware ray tracing ensures it can handle future software, while the GTX 670MX is locked to older API levels. The A400's active production status and lower TDP also make it a more practical option for system builders today.

The GTX 670MX appeals only to specific use cases that rely on its unique specification profile. Its higher shading unit count and texture rate could theoretically benefit older games or applications that are not sensitive to clock speed, but the lack of benchmark data supporting this makes it a speculative choice. Given that the 670MX is end-of-life and its successor is the GeForce 700M series, it offers no future-proofing. The average benchmark scores, while closer than the head-to-head results, still place the A400 6.2% higher overall. For most users, the RTX A400 is the only rational pick based on the facts. The GTX 670MX is a relic of a bygone era, and the data shows it cannot compete with the modern Ampere architecture.