NVIDIA GeForce GTX 670MX vs NVIDIA Quadro K620 Comparison

Head-to-Head Benchmarks

The recorded data for these two GPUs comes from two Geekbench tests: OpenCL and Vulkan. The results show a consistent winner, but the margins are worth examining closely.

In the Geekbench OpenCL test, the NVIDIA Quadro K620 scores 6693, while the NVIDIA GeForce GTX 670MX scores 6125. That gives the Quadro a 9.3% lead. This is not a small gap; it separates the two cleanly in compute workloads. The Vulkan test shows a similar story: the Quadro K620 posts 5870 against the GTX 670MX's 5316, a 10.4% advantage.

Both tests favor the Quadro K620. The database records 2 wins for the Quadro and 0 for the GTX 670MX. No benchmark in the pack shows the GeForce ahead. The closest margin is in OpenCL, where the 9.3% gap is still decisive. The Vulkan margin is slightly wider, suggesting the Quadro's advantage holds across different API paths.

Looking at the wider database context, the Quadro K620's average benchmark score is 6282, which places it at the 36th percentile of all GPUs. Its nearest rivals include the NVIDIA GeForce RTX 5070 Ti SUPER and RTX 4070 Ti SUPER AD102, both scoring 6270 with a delta of just 0.2%, and the AMD Radeon R7 M350 at 6327 with a -0.7% delta. The Quadro sits in a tight cluster, essentially trading blows with those cards.

The GTX 670MX, by contrast, has an average score of 5721, placing it at the 33rd percentile. Its nearest rivals are the Intel Iris Pro Graphics P6300 at 5712 (0.2% delta), the NVIDIA GeForce GTX 550 Ti at 5731 (-0.2% delta), the AMD Radeon HD 8790M at 5691 (0.5% delta), and the NVIDIA Quadro M500M at 5604 (2.1% delta). The GTX 670MX is firmly in the lower-middle range of the GPU field.

The delta between the two cards in head-to-head testing is clear, but the absolute numbers reveal why: the Quadro K620's average score is 6282, while the GTX 670MX's is 5721. That is a 561-point difference, or roughly 9.8% when measured against the lower score. The head-to-head deltas align with this overall trend.

Where Each One Wins

The data points to a one-sided matchup, but the reasons behind it are not uniform. The Quadro K620 wins in both API tests, so it holds the advantage in generalized compute (OpenCL) and in modern graphics API workloads (Vulkan). The GTX 670MX does not win a single recorded test.

For use-case separation, consider the workloads each card likely targets. The Quadro K620, as a professional workstation product, shows strength in compute-oriented tasks. Its OpenCL score of 6693 is its higher of the two results, indicating that raw compute throughput is its strong suit. The Vulkan score of 5870 is lower but still ahead of the GTX 670MX.

The GTX 670MX, being a mobile GeForce part, might be expected to perform better in gaming-style workloads. However, the Vulkan test, which often reflects graphics API efficiency, shows it trailing by 10.4%. The GeForce's OpenCL score of 6125 is its better result, but it still falls short of the Quadro's OpenCL number by 568 points.

In practical terms, the Quadro K620 is the card to choose for any task that relies on OpenCL compute acceleration or Vulkan rendering. The GTX 670MX, despite having more shading units and a wider memory bus, cannot overcome the Quadro's architectural advantages in these benchmarks. The data does not support a scenario where the GTX 670MX wins; every recorded measurement points the other way.

The percentile data reinforces this: the Quadro sits at the 36th percentile, three points above the GTX 670MX's 33rd. That is a modest overall ranking gap, but the head-to-head deltas show the Quadro is consistently faster in these specific tests.

Architecture Differences

The two cards come from different NVIDIA architectures and different market segments. The Quadro K620 uses the GM107 chip on the Maxwell architecture, fabricated on a 28 nm process at TSMC. It packs 1,870 million transistors on a 148 mm² die, yielding a transistor density of 12.6 million per mm². The GTX 670MX uses the GK104 chip on the Kepler architecture, also 28 nm TSMC, but with 3,540 million transistors on a much larger 294 mm² die, giving a density of 12.0 million per mm².

The transistor count difference is stark: the GTX 670MX has nearly double the transistors of the Quadro K620. Yet the Quadro wins the benchmarks. This suggests that Maxwell's architecture is more efficient per transistor, or that the GTX 670MX's lower clocks hold it back.

Clock speeds tell part of the story. The Quadro K620 runs at a base clock of 1058 MHz with a boost of 1124 MHz. The GTX 670MX runs at 601 MHz base and 601 MHz boost, with no boost headroom at all. That is a 457 MHz base clock difference in the Quadro's favor. The GTX 670MX compensates with more hardware: 960 shading units against 384, 80 texture mapping units against 24, and 24 ROPs against 16. But the Quadro's higher clocks and newer architecture win out.

Memory configurations differ significantly. The Quadro K620 has 2 GB of DDR3 on a 128-bit bus, delivering 28.80 GB/s of bandwidth. The GTX 670MX has 3 GB of GDDR5 on a 192-bit bus, delivering 67.20 GB/s. The GTX 670MX has more than double the memory bandwidth and 50% more memory capacity. Yet the benchmarks do not reflect that advantage, likely because the workloads in Geekbench are not bandwidth-limited.

The compute rates tell a nuanced story. The Quadro K620 achieves 863.2 GFLOPS FP32, 17.98 GPixel/s pixel rate, and 26.98 GTexel/s texture rate. The GTX 670MX achieves 1,153.9 GFLOPS FP32, 12.02 GPixel/s pixel rate, and 48.08 GTexel/s texture rate. The GTX 670MX has higher raw FP32 and texture throughput, but the Quadro has higher pixel throughput. The benchmark results show the Quadro winning despite lower FP32, which suggests the OpenCL and Vulkan tests are sensitive to architecture efficiency rather than raw throughput alone.

Other differences include the bus interface: the Quadro K620 uses PCIe 2.0 x16, while the GTX 670MX uses PCIe 3.0 x16. The Quadro has a 45 W TDP, the GTX 670MX has 75 W. The Quadro supports Vulkan 1.4, while the GTX 670MX supports Vulkan 1.2.175. Both support DirectX 12 (11_0) and OpenGL 4.6. The Quadro is a single-slot card with display outputs of 1x DVI and 1x DisplayPort 1.2; the GTX 670MX's display outputs are listed as "Portable Device Dependent", reflecting its mobile origins.

The Verdict

The data is unambiguous. The NVIDIA Quadro K620 wins both recorded head-to-head benchmarks, with a 9.3% lead in OpenCL and a 10.4% lead in Vulkan over the NVIDIA GeForce GTX 670MX. Its average benchmark score of 6282 is higher than the GTX 670MX's 5721, and its percentile ranking is three points higher.

Who should pick the Quadro K620? Anyone prioritizing compute performance in OpenCL or Vulkan workloads. The card's Maxwell architecture, higher clock speeds, and better per-transistor efficiency deliver wins despite having fewer shading units and less memory bandwidth. The Quadro's professional orientation likely explains its strength in these tests.

Who should pick the GTX 670MX? The data offers no benchmark-based reason. It loses both tests. However, the GTX 670MX does have 3 GB of memory versus 2 GB, and 67.20 GB/s bandwidth versus 28.80 GB/s. For tasks that are heavily memory-bound and not captured by these two benchmarks, the GTX 670MX might have an edge, but no recorded measurement supports that claim.

The verdict from the database: the Quadro K620 is the faster card in every measured category. The GTX 670MX has hardware advantages on paper, but they do not translate to wins in the recorded tests.

FAQ

Q: Which GPU has the higher OpenCL score?

A: The NVIDIA Quadro K620 scores 6693, which is 9.3% higher than the NVIDIA GeForce GTX 670MX's 6125.

Q: What is the Vulkan performance difference?

A: The Quadro K620 scores 5870, while the GTX 670MX scores 5316. The Quadro leads by 10.4%.

Q: Which card has more shading units?

A: The GTX 670MX has 960 shading units, while the Quadro K620 has 384.

Q: How do the memory bandwidths compare?

A: The GTX 670MX has 67.20 GB/s bandwidth with 3 GB of GDDR5 on a 192-bit bus. The Quadro K620 has 28.80 GB/s with 2 GB of DDR3 on a 128-bit bus.

Q: Which card has a higher clock speed?

A: The Quadro K620 has a base clock of 1058 MHz and boost of 1124 MHz. The GTX 670MX has a base and boost clock of 601 MHz.

Q: What is the average benchmark score for each?

A: The Quadro K620 averages 6282, while the GTX 670MX averages 5721.

Q: Which GPU supports a newer Vulkan version?

A: The Quadro K620 supports Vulkan 1.4, while the GTX 670MX supports Vulkan 1.2.175.

Specification Differences

| Field | NVIDIA Quadro K620 | NVIDIA GeForce GTX 670MX |

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

| Chip | GM107 | GK104 |

| Architecture | Maxwell | Kepler |

| Generation | Quadro Kepler (Kx200) | GeForce 600M |

| Transistors | 1,870 million | 3,540 million |

| Die Size | 148 mm² | 294 mm² |

| Transistor Density | 12.6M / mm² | 12.0M / mm² |

| Base Clock | 1058 MHz | 601 MHz |

| Boost Clock | 1124 MHz | 601 MHz |

| Memory Clock | 900 MHz, 1800 Mbps effective | 700 MHz, 2.8 Gbps effective |

| Memory Size | 2 GB | 3 GB |

| Memory Type | DDR3 | GDDR5 |

| Memory Bus Width | 128 bit | 192 bit |

| Memory Bandwidth | 28.80 GB/s | 67.20 GB/s |

| Shading Units | 384 | 960 |

| TMUs | 24 | 80 |

| ROPs | 16 | 24 |

| Pixel Rate | 17.98 GPixel/s | 12.02 GPixel/s |

| Texture Rate | 26.98 GTexel/s | 48.08 GTexel/s |

| FP32 | 863.2 GFLOPS | 1,153.9 GFLOPS |

| TDP | 45 W | 75 W |

| Slot Width | Single-slot | Not specified |

| Bus Interface | PCIe 2.0 x16 | PCIe 3.0 x16 |

| Display Outputs | 1x DVI, 1x DisplayPort 1.2 | Portable Device Dependent |

| Vulkan | 1.4 | 1.2.175 |

| Release Date | 2014-07-21 | 2012-09-30 |

| Predecessor | Quadro Fermi | GeForce 500M |

| Successor | Quadro Maxwell | GeForce 700M |

| Average Benchmark Score | 6282 | 5721 |

| Percentile vs All GPUs | 36 | 33 |