AMD Radeon Pro WX 4100 vs NVIDIA GeForce GTX 670MX Comparison

Head-to-Head Benchmarks

The recorded database contains two direct comparison points between the AMD Radeon Pro WX 4100 and the NVIDIA GeForce GTX 670MX, and in both cases the AMD part delivers a decisive victory. In Geekbench OpenCL, the WX 4100 scores 17,642 against 6,125 for the GTX 670MX, a delta of 188% in favor of AMD. That is not a marginal gap; it is a near-tripling of raw compute throughput in a cross-platform API workload. The Vulkan result is even more lopsided: the WX 4100 posts 18,703, while the GTX 670MX manages only 5,316, producing a delta of 251.8%. Across the two head-to-head tests, the AMD card wins both, with a 2-0 record and no benchmark where the NVIDIA part takes the lead.

Looking at the broader benchmark database, the WX 4100’s average score sits at 6,330, which places it in the 37th percentile of all GPUs tracked. Its closest rivals are the AMD Radeon R7 M350 (6,327, a 0.1% delta), the NVIDIA Quadro K620 (6,282, a 0.8% delta), and the NVIDIA GeForce GTX 580M (6,389, a -0.9% delta). Those are essentially statistical ties, meaning the WX 4100 operates in a dense cluster of mid-range parts where small percentage shifts separate competitors. The GTX 670MX, by contrast, averages 5,721 and sits in the 33rd percentile. Its nearest rivals include the Intel Iris Pro Graphics P6300 (5,712, a 0.2% delta), the NVIDIA GeForce GTX 550 Ti (5,731, a -0.2% delta), and the AMD Radeon HD 8790M (5,691, a 0.5% delta). The GTX 670MX is therefore about 9.6% behind the WX 4100 in average score, a meaningful but not overwhelming margin when viewed against the full field.

The individual benchmark suite for the WX 4100 shows a wide range of performance depending on the API and workload type. Its strongest single result is in Geekbench Metal at 21,018, followed by Vulkan at 18,703 and OpenCL at 17,642. In Passmark tests, the card scores 3,699 in G3D, 1,475 in GPU compute, and lower figures in legacy DirectX paths: 56 in DirectX 9, 24 in DirectX 11, 22 in DirectX 12, and 16 in DirectX 10. The GTX 670MX only has two recorded benchmarks in the database, both of which are far lower than the AMD part’s corresponding scores. This disparity is consistent with the architectural differences between the two chips, which will be examined in the next section.

Architecture Differences

The AMD Radeon Pro WX 4100 is built on the Baffin chip using the GCN 4.0 architecture, fabricated on a 14 nm process at GlobalFoundries. It packs 3,000 million transistors into a die size of 123 mm², yielding a transistor density of 24.4 million per square millimeter. The NVIDIA GeForce GTX 670MX uses the GK104 chip with the older Kepler architecture, manufactured by TSMC on a 28 nm process. It contains 3,540 million transistors on a much larger 294 mm² die, resulting in a lower density of 12.0 million per square millimeter. The process node difference is significant: 14 nm versus 28 nm, which directly affects power efficiency and thermal characteristics.

Clock speeds differ substantially. The WX 4100 runs at a base clock of 1125 MHz with a boost up to 1201 MHz, while the GTX 670MX operates at a flat 601 MHz with no boost. Memory speeds also diverge: the AMD card uses 1500 MHz memory with 6 Gbps effective transfer, while the NVIDIA part runs at 700 MHz with 2.8 Gbps effective. The WX 4100 has 4 GB of GDDR5 on a 128-bit bus, providing 96.00 GB/s of bandwidth. The GTX 670MX has 3 GB of GDDR5 on a wider 192-bit bus, but its slower memory clock yields only 67.20 GB/s. Despite having a narrower bus, the AMD card achieves 43% more memory bandwidth.

Compute resources show a mixed picture. The WX 4100 has 1,024 shading units, 64 texture mapping units, and 16 render output units. The GTX 670MX has 960 shading units, 80 TMUs, and 24 ROPs. The NVIDIA part has more TMUs and ROPs, which typically benefits geometry and pixel throughput, but the AMD part counters with a higher pixel rate of 19.22 GPixel/s versus 12.02 GPixel/s, and a higher texture rate of 76.86 GTexel/s versus 48.08 GTexel/s. In floating-point performance, the WX 4100 delivers 2.460 TFLOPS for both FP32 and FP16 (at a 1:1 ratio), while the GTX 670MX manages 1,153.9 GFLOPS for FP32 and offers no FP16 capability.

Power consumption favors the AMD card: it is rated at 50 W TDP, while the GTX 670MX draws 75 W. The WX 4100 is a single-slot card with no power connectors and a suggested PSU of 250 W, whereas the GTX 670MX has no slot width specified, no power connectors listed, and no suggested PSU data. The bus interface also differs: the WX 4100 uses PCIe 3.0 x8, while the GTX 670MX uses PCIe 3.0 x16. Display outputs are another differentiator: the AMD card offers four mini-DisplayPort 1.4a connections, while the NVIDIA part is listed as “Portable Device Dependent,” reflecting its mobile-oriented design.

API support varies, particularly for modern graphics APIs. The WX 4100 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.3. The GTX 670MX supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The DirectX feature level difference (12_0 versus 11_0) is notable, as it indicates the AMD card can access more advanced DirectX 12 features. The Vulkan version gap is smaller but still present. The WX 4100 was released on 2016-11-09, while the GTX 670MX came out on 2012-09-30, a gap of over four years. Both are end-of-life products, with the AMD card succeeding the Radeon Pro GCN and preceding the Radeon Pro Vega, while the NVIDIA part follows the GeForce 500M and precedes the GeForce 700M.

Where Each One Wins

The AMD Radeon Pro WX 4100 wins every recorded head-to-head benchmark, so its advantage is broad rather than narrow. The largest margin comes in Vulkan, where it leads by 251.8%, suggesting that its GCN 4.0 architecture is particularly well-suited to modern, low-level graphics APIs. The OpenCL gap of 188% indicates strong general-purpose compute performance as well. For workloads that rely on compute shaders, OpenCL, or Vulkan, the WX 4100 is the clear choice based on the recorded data.

The GTX 670MX does not win any head-to-head test in the database, but its architectural strengths are still visible in its specification profile. With more TMUs and ROPs, it could theoretically handle certain rasterization-heavy tasks better, but the benchmark data does not support that conclusion. The only two tests recorded for the GTX 670MX are Geekbench OpenCL and Vulkan, both of which it loses decisively. Without recorded results in DirectX or Passmark suites, any claim of NVIDIA superiority in those areas would be speculation, not data.

For the WX 4100, the Passmark results show a different performance profile. Its G3D score of 3,699 is strong relative to its GPU compute score of 1,475, indicating that traditional graphics rendering outperforms compute workloads in Passmark’s specific tests. The legacy DirectX scores are low in absolute terms (56 in DirectX 9, 24 in DirectX 11, 22 in DirectX 12, 16 in DirectX 10), which may reflect driver overhead or the architecture’s focus on modern APIs. The Metal score of 21,018 is the highest single result for the card, suggesting excellent Apple ecosystem performance.

FAQ

Q: Which GPU has higher raw compute performance?

A: The AMD Radeon Pro WX 4100 delivers 2.460 TFLOPS of FP32 performance, while the NVIDIA GeForce GTX 670MX delivers 1,153.9 GFLOPS. The AMD card also supports FP16 at 2.460 TFLOPS, while the NVIDIA part has no FP16 capability.

Q: How do the two cards compare in memory bandwidth?

A: The WX 4100 has 4 GB of GDDR5 on a 128-bit bus, achieving 96.00 GB/s. The GTX 670MX has 3 GB of GDDR5 on a 192-bit bus, but its slower clock speed limits bandwidth to 67.20 GB/s. The AMD card has a 43% bandwidth advantage.

Q: Which card is more power-efficient?

A: The WX 4100 has a TDP of 50 W, while the GTX 670MX has a TDP of 75 W. The AMD card achieves higher performance with 33% lower power consumption.

Q: What API features differ between the two?

A: The WX 4100 supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.3. The GTX 670MX supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The AMD card has a higher DirectX feature level and a newer Vulkan version.

Q: Why is the GTX 670MX missing from most benchmark suites?

A: The database records only Geekbench OpenCL and Vulkan scores for the GTX 670MX (6,125 and 5,316 respectively). No Passmark or Metal results exist for this card, while the WX 4100 has ten recorded benchmarks across multiple suites.

Q: Which card has a higher average benchmark score?

A: The WX 4100 averages 6,330 across all recorded tests, placing it in the 37th percentile of all GPUs. The GTX 670MX averages 5,721, placing it in the 33rd percentile. The AMD card is approximately 9.6% higher.

The Verdict

The data is unambiguous: the AMD Radeon Pro WX 4100 outperforms the NVIDIA GeForce GTX 670MX in every recorded benchmark. The average score gap of 9.6% is compounded by the much larger margins in individual tests, particularly Vulkan where the AMD card is 251.8% ahead. For users choosing between these two end-of-life parts, the WX 4100 is the superior option on every measurable axis: compute throughput, memory bandwidth, power efficiency, API support, and display connectivity.

The GTX 670MX’s only advantages are architectural in nature, such as its wider memory bus (192-bit versus 128-bit), higher TMU and ROP counts, and larger transistor count. However, these specifications do not translate into benchmark wins. The Kepler architecture’s age, combined with its 28 nm process node and lower clock speeds, leaves it trailing in both synthetic and real-world workloads. Its PCIe 3.0 x16 interface is wider than the WX 4100’s x8, but the bandwidth difference does not compensate for the compute deficit.

The WX 4100 is the recommended pick for any workload involving OpenCL, Vulkan, or Metal, based on its scoring. Its 14 nm process and GCN 4.0 architecture provide a modern foundation that scales better with contemporary APIs. The GTX 670MX, with its limited benchmark presence and lower percentile ranking, is best suited for legacy use cases where its mobile form factor and lower absolute power draw (75 W) are acceptable, but only if the workload does not require modern API features or high compute throughput. In a direct comparison, the AMD part wins on merit, not just on paper.