NVIDIA GeForce GTX 670MX vs NVIDIA Quadro K3100M Comparison

Head-to-Head Benchmarks

The recorded data places the NVIDIA Quadro K3100M ahead in both available benchmark comparisons, though the margins are modest. In Geekbench OpenCL, the Quadro K3100M scores 6154 against the GTX 670MX's 6125, a delta of 0.5%. This is a narrow victory, nearly a statistical tie in practical terms. The Vulkan test shows a slightly larger gap: the Quadro K3100M posts 5484 versus 5316 for the GTX 670MX, a 3.1% advantage. Neither result represents a dominant win, but the Quadro takes both rounds, giving it a 2-0 record in head-to-head matchups.

Context matters here. The GTX 670MX's average benchmark score across all recorded tests is 5721, placing it in the 33rd percentile of all GPUs in the database. Its nearest rivals include the Intel Iris Pro Graphics P6300 (5712, 0.2% behind), the GeForce GTX 550 Ti (5731, 0.2% ahead), the AMD Radeon HD 8790M (5691, 0.5% behind), and the Quadro M500M (5604, 2.1% behind). This is a tight cluster; the GTX 670MX sits within 2% of four other products, which indicates it performs in line with mid-range mobile graphics of its era.

The Quadro K3100M, by contrast, holds a 30th percentile ranking with an average score of 5154. Its nearest rivals are the AMD Radeon R7 M260X (5161, 0.1% ahead), the Quadro 4000M (5211, 1.1% ahead), the GeForce GTX 760M (5236, 1.6% ahead), and the AMD Radeon R7 240 (5063, 1.8% behind). Notably, the Quadro's average score is dragged down by a Geekbench Metal result of 3823, a test the GTX 670MX did not run. Without that Metal result, the average would be higher; the OpenCL and Vulkan scores alone (6154 and 5484) are competitive with or above the GTX 670MX's corresponding numbers.

Interpreting the head-to-head data, the Vulkan delta of 3.1% is the most meaningful difference. Vulkan measures low-level API performance, and a 3.1% advantage suggests the K3100M has slightly better compute throughput per clock, despite having fewer shading units and TMUs, which is a counterintuitive result worth noting. The OpenCL result, with a 0.5% delta, is effectively parity. The GTX 670MX does not post a Metal score, so no comparison can be drawn there, but the Quadro's Metal result of 3823 is substantially below its own other scores, indicating that Metal is not its strong suit.

For buyers or reviewers, the practical takeaway is that these two mobile GPUs are closely matched in synthetic compute tests. The Quadro K3100M edges out the GTX 670MX in both recorded tests, but the margins are well within what could be considered run-to-run variance. The GTX 670MX is not embarrassed in any test, and its 33rd percentile rank versus the Quadro's 30th shows that the average score across all benchmarks reverses the head-to-head result, but that is partly due to the inclusion of the Metal test.

FAQ

Q: Which GPU wins the Geekbench OpenCL test?

A: The Quadro K3100M wins by a narrow margin, scoring 6154 against the GTX 670MX's 6125, a delta of 0.5%.

Q: What is the larger benchmark gap between the two cards?

A: The Vulkan test shows the biggest difference: the K3100M scores 5484 versus 5316 for the GTX 670MX, a delta of 3.1%.

Q: How does the GTX 670MX compare to its nearest rivals?

A: It sits within a tight cluster. The GeForce GTX 550 Ti is 0.2% ahead, the Intel Iris Pro Graphics P6300 is 0.2% behind, the AMD Radeon HD 8790M is 0.5% behind, and the NVIDIA Quadro M500M is 2.1% behind.

Q: What is the Quadro K3100M's average benchmark score relative to its rivals?

A: Its average score is 5154. It is 0.1% behind the Radeon R7 M260X, 1.1% behind the Quadro 4000M, 1.6% behind the GeForce GTX 760M, and 1.8% ahead of the Radeon R7 240.

Q: Does the GTX 670MX have a higher or lower percentile rank?

A: The GTX 670MX holds a 33rd percentile rank, while the Quadro K3100M sits at the 30th percentile.

Q: Are there any tests where the GTX 670MX wins?

A: In the recorded head-to-head data, the GTX 670MX wins zero tests. The Quadro K3100M wins both the OpenCL and Vulkan tests.

The Verdict

Based strictly on the recorded data, the NVIDIA Quadro K3100M is the superior performer in direct comparison. It wins both head-to-head benchmarks, with the Vulkan gap (3.1%) being more than three times larger than the OpenCL gap (0.5%). For any user who prioritizes raw benchmark scores, the K3100M is the safer pick. Its 4 GB GDDR5 memory on a 256-bit bus, giving 102.4 GB/s of bandwidth, also positions it better for larger workloads, though the benchmark scores do not directly reflect memory capacity.

However, the GTX 670MX is not a poor choice. Its average benchmark score of 5721 is actually higher than the Quadro's 5154, but that is misleading because the Quadro's Metal score (3823) drags its average down. In the two tests where both cards are directly compared, the Quadro comes out ahead. The GTX 670MX's lower 33rd percentile versus the Quadro's 30th is a minor difference, and the GTX 670MX's nearest rival cluster shows it is within 2% of several other cards, indicating it is a solid mid-range performer.

The two GPUs share the same 75 W TDP and the same 28 nm TSMC process, but the K3100M has a higher base clock (706 MHz vs 601 MHz) and a wider memory bus (256-bit vs 192-bit). These specs likely explain its slight edge in compute tests. For an application that is compute-bound, such as OpenCL or Vulkan workloads, the Quadro K3100M is the better choice. For general use, the GTX 670MX is not a bad option, but the data shows the K3100M is ahead in every direct comparison.

Specification Differences

The two cards differ in several key specification fields. The Quadro K3100M has a higher base clock at 706 MHz, while the GTX 670MX runs at 601 MHz. Both cards have the same boost clock as their base clock, so the Quadro's advantage is consistent across all frequencies. Memory clock also favors the Quadro: 800 MHz (3.2 Gbps effective) versus 700 MHz (2.8 Gbps effective) on the GTX 670MX.

Memory capacity differs: the Quadro K3100M has 4 GB of GDDR5, while the GTX 670MX has 3 GB. Bus width is a major difference, with the Quadro using a 256-bit interface versus the GTX 670MX's 192-bit, resulting in a significant bandwidth gap: 102.4 GB/s versus 67.20 GB/s. The shading unit count is higher on the GTX 670MX with 960 units, but the Quadro K3100M has fewer, 768 units. Similarly, the GTX 670MX has 80 TMUs and 24 ROPs, while the Quadro has 64 TMUs and 32 ROPs.

Pixel rate and texture rate reflect these differences. The GTX 670MX posts 12.02 GPixel/s and 48.08 GTexel/s, while the Quadro K3100M has 11.30 GPixel/s and 45.18 GTexel/s. The GTX 670MX is slightly ahead in fill rates, but the Quadro has a higher FP32 compute: 1,084.4 GFLOPS versus 1,153.9 GFLOPS for the GTX 670MX, which is a 6% advantage for the GTX 670MX.

Bus interface also differs. The GTX 670MX uses PCIe 3.0 x16, while the Quadro K3100M uses MXM-B (3.0) and is a MXM module. The GTX 670MX's power connectors are none, and the Quadro also has none, but the Quadro's slot width is listed as MXM Module. Release dates differ: the GTX 670MX came out on 2012-09-30, while the Quadro K3100M was released on 2013-07-22.

The GTX 670MX has a predecessor of GeForce 500M and a successor of GeForce 700M. The Quadro K3100M's predecessor is Quadro Fermi-M and successor is Quadro Maxwell-M. Neither card has a launch MSRP listed.

Architecture Differences

Both cards are based on NVIDIA's Kepler architecture, using the GK104 chip, fabricated by TSMC on a 28 nm process. They share the same transistor count of 3,540 million and a die size of 294 mm², giving a transistor density of 12.0M per mm². The core architecture is identical, which explains the similar benchmark behavior.

The key architectural difference lies in the compute layout. The GTX 670MX has more shading units (960) and TMUs (80) but fewer ROPs (24), whereas the Quadro K3100M has fewer shading units (768) and TMUs (64) but more ROPs (32). This distribution explains the slight fill rate advantage for the GTX 670MX (12.02 GPixel/s vs 11.30 GPixel/s) and the texture rate advantage (48.08 GTexel/s vs 45.18 GTexel/s). However, the Quadro has a higher FP32 throughput of 1,084.4 GFLOPS, versus the GTX 670MX's 1,153.9 GFLOPS, which is surprising given the lower core count, but the higher clock speed compensates.

Memory architecture is a major divergence. The Quadro K3100M uses a 256-bit memory bus, double the width of the GTX 670MX's 192-bit bus, and also runs at a higher effective memory clock (3.2 Gbps vs 2.8 Gbps). This results in a 102.4 GB/s bandwidth, which is 52% higher than the GTX 670MX's 67.20 GB/s. For memory-intensive compute tasks, the Quadro has a clear architectural advantage.

Both cards support the same APIs: DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. Neither has ray tracing or tensor cores. The GTX 670MX belongs to the GeForce 600M generation, while the Quadro K3100M belongs to the Quadro Kepler-M (Kx100M) generation. The production status for both is end-of-life.

In terms of power, both have a TDP of 75 W and no power connectors. The GTX 670MX uses PCIe 3.0 x16, while the Quadro is an MXM-B (3.0) module, which is a physical form factor difference. The display outputs are portable device dependent for both, meaning they vary by laptop implementation.

The Quadro's higher clock and wider memory bus are the defining architectural differences. The GTX 670MX's higher shading unit count does not translate to a benchmark win in the recorded tests, suggesting the Quadro's memory bandwidth and clock speed are more impactful for the workloads measured. The GTX 670MX is a capable part, but the data shows the Quadro K3100M is the more effective GPU for OpenCL and Vulkan compute.