AMD Radeon Pro WX 3100 vs NVIDIA GeForce GTX 680M Comparison

AMD Radeon Pro WX 3100 and NVIDIA GeForce GTX 680M represent two very different eras of mobile and workstation graphics, separated by five years of architectural evolution. The data shows a clear split: the older NVIDIA part wins the single available compute benchmark outright, while the AMD part brings a newer feature set and far better efficiency. Below, the numbers from the FACT PACK are broken down to show exactly where each GPU stands.

Head-to-Head Benchmarks

The only direct benchmark comparison in the data is the Geekbench OpenCL test, and the result is unambiguous. The NVIDIA GeForce GTX 680M scores 9230, while the AMD Radeon Pro WX 3100 scores 7333. That is a delta of -20.6% for the AMD part, meaning NVIDIA leads by roughly a fifth in raw compute throughput. This is a decisive margin, not a narrow one. The GTX 680M’s advantage here is consistent with its much larger silicon: it packs 1,344 shading units, 112 texture mapping units, and 32 ROPs, versus the WX 3100’s 512 shading units, 32 TMUs, and 16 ROPs. Even at significantly lower clocks — 719 MHz base and 758 MHz boost versus 925 MHz base and 1219 MHz boost — the NVIDIA chip’s sheer width in execution resources carries the day.

However, the story does not end with a single win. The AMD Radeon Pro WX 3100 has its own benchmark results, just not in the same test category. It scores 7827 in Geekbench Vulkan, a test the GTX 680M has no entry for in the FACT PACK. The GTX 680M counters with a Geekbench Metal score of 4815, which AMD has no corresponding result for. So while NVIDIA dominates in OpenCL, AMD holds an unopposed position in Vulkan. The average benchmark score across all tests tells a similar but slightly different tale: the WX 3100 averages 7580, while the GTX 680M averages 7023. That flips the head-to-head result, showing that when you include Vulkan and Metal, AMD’s newer architecture actually comes out ahead on the aggregate.

The percentile rankings underscore how close these two are in the broader GPU landscape. The WX 3100 sits at the 41st percentile of all GPUs, and the GTX 680M sits at the 39th. That is a two-point gap, which is negligible in real-world terms. The nearest rivals confirm this: the AMD Radeon R7 250 scores 7557 (0.3% from the WX 3100), and the NVIDIA GeForce GTX 675M scores 6946 (1.1% from the GTX 680M). Both cards are firmly in the same performance tier, even if the single OpenCL test shows a lopsided result.

Where Each One Wins

The NVIDIA GeForce GTX 680M wins decisively in OpenCL compute workloads. Its 9230 score versus 7333 is a 20.6% lead, which is substantial for any application that relies on OpenCL for general-purpose GPU computing. This is the card’s clear domain. The architecture backs this up: Kepler was designed with high shader counts and high memory bandwidth in mind. The GTX 680M has a 256-bit memory bus delivering 115.2 GB/s, compared to the WX 3100’s 128-bit bus at 96.00 GB/s. That 19.2 GB/s bandwidth advantage, combined with more than double the TMUs (112 vs 32) and double the ROPs (32 vs 16), gives NVIDIA a structural edge in fill-rate and texture-heavy tasks. Its pixel rate is 21.22 GPixel/s and texture rate is 84.90 GTexel/s, both higher than AMD’s 19.50 GPixel/s and 39.01 GTexel/s.

The AMD Radeon Pro WX 3100 wins in efficiency and modern API support. Its 14 nm process node, built by GlobalFoundries, is a full two generations ahead of the GTX 680M’s 28 nm TSMC node. This shows up in power: the WX 3100 has a 65 W TDP, while the GTX 680M draws 100 W. That is a 35 W difference, meaning AMD delivers competitive average performance at 35% lower power consumption. The WX 3100’s FP32 output is 1,248.3 GFLOPS, while the GTX 680M outputs 2.038 TFLOPS, but AMD achieves that lower figure with far less energy. The WX 3100 also supports Vulkan 1.3, whereas the GTX 680M is capped at Vulkan 1.2.175. For modern games or compute that leverage Vulkan, AMD has the only validated path in this comparison.

The Verdict

The data points to a clear split decision. For any workload that runs on OpenCL, the NVIDIA GeForce GTX 680M is the superior choice. Its 20.6% advantage in that specific test is too large to ignore, and its higher texture and pixel rates reinforce that lead in graphics-heavy scenarios. The GTX 680M also has a higher raw FP32 throughput (2.038 TFLOPS vs 1,248.3 GFLOPS), which matters for scientific or rendering tasks that are not bound by API overhead.

For everything else, the AMD Radeon Pro WX 3100 is the better pick. Its average benchmark score of 7580 beats the GTX 680M’s 7023, and it has a working Vulkan 1.3 implementation, which the NVIDIA card lacks. The WX 3100 also uses 65 W instead of 100 W, making it far easier to cool and power in a small chassis. Its 14 nm process and 2,200 million transistors on a 103 mm² die give it a transistor density of 21.4M / mm², versus the GTX 680M’s 12.0M / mm² on a 294 mm² die. That density advantage translates directly into efficiency. If the choice is for a modern workstation or a system where power draw matters, the WX 3100 is the only rational option. If the choice is for raw compute in a legacy OpenCL pipeline, the GTX 680M is the undisputed winner.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon Pro WX 3100 has an average benchmark score of 7580, while the NVIDIA GeForce GTX 680M averages 7023. The AMD part is ahead by 557 points, or roughly 7.9%.

Q: How much faster is the GTX 680M in OpenCL?

A: The GTX 680M scores 9230 in Geekbench OpenCL, versus 7333 for the WX 3100. That is a delta of -20.6% for AMD, meaning NVIDIA is about 20.6% faster in this specific test.

Q: Does the WX 3100 support Vulkan 1.3?

A: Yes. The AMD Radeon Pro WX 3100 lists Vulkan 1.3 support. The NVIDIA GeForce GTX 680M is limited to Vulkan 1.2.175.

Q: What is the memory bandwidth difference?

A: The GTX 680M has a 256-bit bus with 115.2 GB/s bandwidth. The WX 3100 has a 128-bit bus with 96.00 GB/s bandwidth. NVIDIA has a 19.2 GB/s advantage.

Q: Which card draws less power?

A: The AMD Radeon Pro WX 3100 has a 65 W TDP, while the NVIDIA GeForce GTX 680M has a 100 W TDP. AMD uses 35 W less power.

Q: Are these cards considered end-of-life?

A: Yes. Both the AMD Radeon Pro WX 3100 and the NVIDIA GeForce GTX 680M have a production status of "End-of-life" in the data.

Architecture Differences

The architectural gap between these two GPUs is vast, reflecting a five-year generational split. The AMD Radeon Pro WX 3100 is built on the GCN 4.0 architecture, using the Lexa chip, and belongs to the Radeon Pro Polaris generation. It is fabricated on a 14 nm process at GlobalFoundries, with 2,200 million transistors packed into a 103 mm² die. That yields a transistor density of 21.4M / mm². The NVIDIA GeForce GTX 680M, by contrast, uses the Kepler architecture with the GK104 chip, part of the GeForce 600M generation. It is built on a 28 nm process at TSMC, with 3,540 million transistors spread across a much larger 294 mm² die, giving a density of only 12.0M / mm².

The core configurations are equally divergent. The WX 3100 has 512 shading units, 32 TMUs, and 16 ROPs. The GTX 680M has 1,344 shading units, 112 TMUs, and 32 ROPs. NVIDIA’s chip has more than double the shading units, more than triple the TMUs, and double the ROPs. The AMD card compensates with higher clocks: 925 MHz base and 1219 MHz boost, versus 719 MHz base and 758 MHz boost for NVIDIA. Memory clocks also differ, with AMD running at 1500 MHz (6 Gbps effective) and NVIDIA at 900 MHz (3.6 Gbps effective). Both use GDDR5 memory, but NVIDIA’s wider 256-bit bus outweighs AMD’s faster clock on a 128-bit bus.

The API and output capabilities are another major split. AMD supports DirectX 12 (12_0), OpenGL 4.6, and Vulkan 1.3. NVIDIA supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The WX 3100 has dedicated display outputs (1x DisplayPort 1.4a and 2x mini-DisplayPort 1.4a), while the GTX 680M’s outputs are listed as "Portable Device Dependent," meaning it relies on the laptop’s embedded display. The AMD card has a single-slot form factor, while NVIDIA uses an MXM Module. Neither card has ray tracing or tensor cores, and both are end-of-life products.

Specification Differences

The two GPUs differ across nearly every measurable specification, so only the fields where they diverge are listed below.

| Specification | AMD Radeon Pro WX 3100 | NVIDIA GeForce GTX 680M |

|---|---|---|

| Chip | Lexa | GK104 |

| Architecture | GCN 4.0 | Kepler |

| Generation | Radeon Pro Polaris (WX x100) | GeForce 600M |

| Process Node | 14 nm | 28 nm |

| Foundry | GlobalFoundries | TSMC |

| Transistors | 2,200 million | 3,540 million |

| Die Size | 103 mm² | 294 mm² |

| Transistor Density | 21.4M / mm² | 12.0M / mm² |

| Base Clock | 925 MHz | 719 MHz |

| Boost Clock | 1219 MHz | 758 MHz |

| Memory Clock | 1500 MHz (6 Gbps effective) | 900 MHz (3.6 Gbps effective) |

| Memory Bus Width | 128 bit | 256 bit |

| Memory Bandwidth | 96.00 GB/s | 115.2 GB/s |

| Shading Units | 512 | 1344 |

| TMUs | 32 | 112 |

| ROPs | 16 | 32 |

| Pixel Rate | 19.50 GPixel/s | 21.22 GPixel/s |

| Texture Rate | 39.01 GTexel/s | 84.90 GTexel/s |

| FP32 | 1,248.3 GFLOPS | 2.038 TFLOPS |

| FP16 | 1,248.3 GFLOPS (1:1) | null |

| TDP | 65 W | 100 W |

| Slot Width | Single-slot | MXM Module |

| Bus Interface | PCIe 3.0 x8 | MXM-B (3.0) |

| Display Outputs | 1x DisplayPort 1.4a, 2x mini-DisplayPort 1.4a | Portable Device Dependent |

| DirectX | 12 (12_0) | 12 (11_0) |

| Vulkan | 1.3 | 1.2.175 |

| Dimensions | 168 mm / 6.6 inches length, 69 mm / 2.7 inches height | null |

| Release Date | 2017-06-11 | 2012-06-03 |

| Predecessor | Radeon Pro GCN | GeForce 500M |

| Successor | Radeon Pro Vega | GeForce 700M |

| Launch MSRP | 199 USD | null |

The AMD card is newer, smaller, denser, more power-efficient, and has better API support. The NVIDIA card is larger, has more execution resources, higher bandwidth, and wins the OpenCL test by a wide margin. The specification sheet tells the story of two completely different design philosophies: AMD optimizing for efficiency and modern features, NVIDIA optimizing for raw throughput on an older process.