GPU Comparison

The AMD FirePro W4100 and NVIDIA GeForce GTX 670MX are both end-of-life mobile or professional graphics solutions built on a 28 nm TSMC process, but they target different workloads and deliver contrasting performance profiles. The benchmark data from Geekbench shows a clear split: the NVIDIA card dominates in OpenCL compute, while the AMD card takes a decisive lead in Vulkan graphics workloads. The following sections break down the head-to-head results, architectural differences, and use-case implications based solely on the provided specifications and scores.

Head-to-Head Benchmarks

The two GPUs split their two benchmark results exactly one win apiece, but the margins are far from symmetrical. In the Geekbench OpenCL test, the NVIDIA GeForce GTX 670MX scores 6125, which is 10.6% higher than the AMD FirePro W4100’s 5478. This is the larger absolute gap of the two tests, and it aligns with the GTX 670MX’s raw compute specifications, which show a significantly higher FP32 throughput. The AMD card trails by 647 points, a sizeable margin that places it behind its rival in this particular compute API.

The Vulkan test tells the opposite story, and it is a far more lopsided result. The AMD FirePro W4100 scores 6496, which is 22.2% ahead of the NVIDIA GeForce GTX 670MX’s 5316. This 1180-point delta is nearly double the absolute difference seen in OpenCL, indicating that the FirePro W4100 is substantially stronger in Vulkan-based workloads. The GTX 670MX’s Vulkan score is not only lower than its own OpenCL result, but it also falls below the FirePro’s OpenCL score, which underscores the NVIDIA card’s relative weakness in this newer API.

Looking at the average benchmark scores, the AMD FirePro W4100 averages 5987 across its two tests, while the NVIDIA GeForce GTX 670MX averages 5721. This gives the AMD card a 4.6% advantage in overall average performance, despite losing the OpenCL test. The FirePro W4100’s higher Vulkan score pulls its average up, while the GTX 670MX’s Vulkan result drags its average down. In terms of percentile ranking, the FirePro W4100 sits at the 34th percentile of all GPUs, while the GTX 670MX is just below at the 33rd percentile, confirming that the AMD card holds a marginal edge in the broader performance distribution.

The nearest rival data reinforces the positioning of each card. The FirePro W4100’s average score of 5987 is nearly identical to the NVIDIA Quadro K4000M (5986) and Quadro K4000 (5982), with deltas of 0% and 0.1% respectively. It also sits close to the NVIDIA RTX PRO 6000 Blackwell Server (5996) and GeForce GTX 770M (6000), both of which are within 0.2% of the FirePro’s score. The GTX 670MX’s average of 5721 places it near the Intel Iris Pro Graphics P6300 (5712, 0.2% delta) and the GeForce GTX 550 Ti (5731, -0.2% delta), with the AMD Radeon HD 8790M (5691) and NVIDIA Quadro M500M (5604) trailing by 0.5% and 2.1% respectively.

Where Each One Wins

The NVIDIA GeForce GTX 670MX is the clear winner in OpenCL compute tasks, based on its 10.6% score advantage. This makes it the better choice for applications that rely heavily on OpenCL acceleration, such as certain physics simulations, video encoding, or general-purpose GPU computing that uses this API. The GTX 670MX’s higher shading unit count (960 versus 512) and FP32 throughput (1,153.9 GFLOPS versus 645.1 GFLOPS) provide the theoretical foundation for this OpenCL dominance, even though the benchmark result is the only direct evidence of real-world performance.

The AMD FirePro W4100 wins decisively in Vulkan, with a 22.2% lead that is impossible to ignore. This makes it the superior option for Vulkan-based graphics workloads, including modern game engines that use Vulkan for rendering, as well as compute applications that leverage this API. The FirePro’s 512 shading units and 16 ROPs are fewer than the GTX 670MX’s 960 shading units and 24 ROPs, yet the AMD card still manages a massive Vulkan advantage, suggesting that architecture-level optimizations or driver support play a significant role.

When considering the average benchmark score, the AMD FirePro W4100 wins overall with 5987 versus 5721, a 4.6% advantage. This means that in a mixed workload scenario where both OpenCL and Vulkan are used, the FirePro W4100 is likely to come out ahead on average. However, the GTX 670MX’s OpenCL strength cannot be dismissed, as it may be the deciding factor in environments that are exclusively OpenCL-based.

For users prioritizing a single API, the choice is straightforward: the GTX 670MX for OpenCL-heavy tasks, and the FirePro W4100 for Vulkan-heavy tasks. For general-purpose use, the FirePro W4100’s higher average score and better Vulkan performance make it the more balanced option, though the GTX 670MX’s compute capabilities should not be overlooked.

Architecture Differences

The two GPUs come from different architectural families and have distinct die configurations. The AMD FirePro W4100 uses the Cape Verde chip based on GCN 1.0 architecture, which is an older design from AMD’s Graphics Core Next line. The NVIDIA GeForce GTX 670MX uses the GK104 chip based on Kepler architecture, NVIDIA’s 2012-era design. Both are fabricated by TSMC on a 28 nm process, but the similarities end there.

The die sizes and transistor counts differ substantially. The GTX 670MX’s GK104 die measures 294 mm² and contains 3,540 million transistors, while the FirePro W4100’s Cape Verde die is only 123 mm² with 1,500 million transistors. This makes the NVIDIA chip nearly 2.4 times larger in die area and 2.36 times larger in transistor count. However, the transistor density is nearly identical: 12.2 million transistors per mm² for the AMD chip versus 12.0 million for the NVIDIA chip, reflecting the shared 28 nm process node.

The FirePro W4100’s memory configuration is more modest in capacity but faster in clock speed. It has 2 GB of GDDR5 memory on a 128-bit bus, with a memory clock of 1000 MHz (4 Gbps effective) and a bandwidth of 64.00 GB/s. The GTX 670MX has 3 GB of GDDR5 memory on a 192-bit bus, with a lower memory clock of 700 MHz (2.8 Gbps effective) but a slightly higher bandwidth of 67.20 GB/s due to the wider bus. This gives the NVIDIA card an 3.2 GB/s bandwidth advantage, despite the AMD card’s higher memory clock.

The compute resources are heavily skewed toward the NVIDIA card. The GTX 670MX has 960 shading units, 80 texture mapping units (TMUs), and 24 ROPs, compared to the FirePro W4100’s 512 shading units, 32 TMUs, and 16 ROPs. This translates to a pixel rate of 12.02 GPixel/s and a texture rate of 48.08 GTexel/s for the NVIDIA card, versus 10.08 GPixel/s and 20.16 GTexel/s for the AMD card. The FP32 performance is nearly double for the GTX 670MX at 1,153.9 GFLOPS, compared to 645.1 GFLOPS for the FirePro W4100.

In terms of API support, both cards support DirectX 12, but at different feature levels. The FirePro W4100 supports DirectX 12 (11_1), while the GTX 670MX supports DirectX 12 (11_0). Both support OpenGL 4.6, but the Vulkan versions differ slightly: the FirePro W4100 supports Vulkan 1.2.170, while the GTX 670MX supports Vulkan 1.2.175. The GTX 670MX has a marginally newer Vulkan version, yet the benchmark results show the FirePro W4100 outperforming it significantly in Vulkan, suggesting that driver maturity and architecture efficiency matter more than the API version number.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD FirePro W4100 has an average benchmark score of 5987, while the NVIDIA GeForce GTX 670MX has an average of 5721. The FirePro W4100 is 4.6% higher on average.

Q: How does the GTX 670MX compare to the FirePro W4100 in OpenCL?

A: The GTX 670MX scores 6125 in Geekbench OpenCL, which is 10.6% higher than the FirePro W4100’s 5478. This is the NVIDIA card’s single biggest win.

Q: What is the Vulkan performance difference between the two?

A: The FirePro W4100 scores 6496 in Geekbench Vulkan, which is 22.2% higher than the GTX 670MX’s 5316. This is the AMD card’s dominant win.

Q: Which GPU has more shading units and higher FP32 performance?

A: The NVIDIA GeForce GTX 670MX has 960 shading units and 1,153.9 GFLOPS FP32 performance, compared to the AMD FirePro W4100’s 512 shading units and 645.1 GFLOPS.

Q: What are the memory sizes and bandwidths of each card?

A: The FirePro W4100 has 2 GB of GDDR5 on a 128-bit bus with 64.00 GB/s bandwidth. The GTX 670MX has 3 GB of GDDR5 on a 192-bit bus with 67.20 GB/s bandwidth.

Q: Do both cards support DirectX 12 and OpenGL 4.6?

A: Yes, both support DirectX 12 and OpenGL 4.6, but the FirePro W4100 supports DirectX 12 (11_1) while the GTX 670MX supports DirectX 12 (11_0). Both support Vulkan, with the GTX 670MX having version 1.2.175 and the FirePro W4100 having 1.2.170.

Specification Differences

The following specifications differ between the AMD FirePro W4100 and the NVIDIA GeForce GTX 670MX:

• Chip: Cape Verde (AMD) vs GK104 (NVIDIA)
• Architecture: GCN 1.0 (AMD) vs Kepler (NVIDIA)
• Generation: FirePro GCN (Wx100) vs GeForce 600M
• Transistors: 1,500 million vs 3,540 million
• Die Size: 123 mm² vs 294 mm²
• Transistor Density: 12.2M / mm² vs 12.0M / mm²
• Base Clock: N/A vs 601 MHz
• Boost Clock: N/A vs 601 MHz
• Memory Clock: 1000 MHz (4 Gbps effective) vs 700 MHz (2.8 Gbps effective)
• Memory Size: 2 GB vs 3 GB
• Memory Bus Width: 128 bit vs 192 bit
• Memory Bandwidth: 64.00 GB/s vs 67.20 GB/s
• Shading Units: 512 vs 960
• TMUs: 32 vs 80
• ROPs: 16 vs 24
• Pixel Rate: 10.08 GPixel/s vs 12.02 GPixel/s
• Texture Rate: 20.16 GTexel/s vs 48.08 GTexel/s
• FP32: 645.1 GFLOPS vs 1,153.9 GFLOPS
• TDP: 50 W vs 75 W
• Slot Width: Single-slot vs N/A
• Suggested PSU: 250 W vs N/A
• Display Outputs: 4x mini-DisplayPort 1.2 vs Portable Device Dependent
• DirectX Support: 12 (11_1) vs 12 (11_0)
• Vulkan Support: 1.2.170 vs 1.2.175
• Dimensions: 171 mm (6.7 inches) length, 69 mm (2.7 inches) height vs N/A
• Release Date: 2014-08-12 vs 2012-09-30
• Predecessor: FirePro Terascale vs GeForce 500M
• Successor: Radeon Pro Polaris vs GeForce 700M
• Geekbench OpenCL Score: 5478 vs 6125
• Geekbench Vulkan Score: 6496 vs 5316
• Percentile vs All GPUs: 34 vs 33
• Average Benchmark Score: 5987 vs 5721