NVIDIA GeForce GTX 670M vs NVIDIA Quadro K620M Comparison

The NVIDIA GeForce GTX 670M and NVIDIA Quadro K620M are both end-of-life mobile graphics solutions, but they serve very different purposes. The GTX 670M is a gaming-oriented part from the GeForce 600M generation, while the K620M is a professional mobile workstation GPU. The benchmark data shows a clear, though not overwhelming, performance gap between them.

Head-to-Head Benchmarks

The database records a single direct comparison for these two GPUs: the Geekbench OpenCL test. In this test, the NVIDIA GeForce GTX 670M scores 6513, while the NVIDIA Quadro K620M scores 5957. This gives the GTX 670M a win with a delta of 9.3% over the K620M.

This 9.3% lead is the entire story in the head-to-head data. The GTX 670M is the clear winner in raw compute performance as measured by this test. For the K620M, a 9.3% deficit puts it in a position where it is trailing its rival, but it is not a catastrophic gap. The K620M’s score of 5957 places it very close to its own nearest rivals, which include the AMD Radeon HD 8730M at 5955 (a 0% delta) and the AMD Radeon HD 8750M at 5970 (a -0.2% delta). This indicates that the K620M is right in the middle of its performance tier, whereas the GTX 670M is slightly higher up.

The GTX 670M’s score of 6513 also puts it in a tight cluster with its own nearest rivals. It is only 0.3% ahead of the NVIDIA GeForce GT 555M (6493), 0.5% ahead of the NVIDIA Quadro M5000M (6481), and 0.6% ahead of the AMD Radeon Vega 10 Mobile (6476). It trails the Intel UHD Graphics P750 (6554) by 0.6%. This shows that while the GTX 670M wins this head-to-head, it is not a dominant force; it is just at the top of a very closely packed group.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 670M has a higher average benchmark score of 6513, compared to the NVIDIA Quadro K620M’s 5957.

Q: What is the performance difference in the recorded head-to-head test?

A: In the Geekbench OpenCL test, the GTX 670M wins by 9.3% over the K620M.

Q: How does each GPU rank among all GPUs in the database?

A: The GTX 670M ranks in the 38th percentile, while the K620M ranks in the 34th percentile.

Q: Are these GPUs from the same architecture generation?

A: No. The GTX 670M uses the Fermi 2.0 architecture, while the K620M uses the Maxwell architecture.

Q: Which GPU has a higher memory clock speed?

A: The GTX 670M has a memory clock of 750 MHz (3 Gbps effective), while the K620M has a higher base clock of 1029 MHz and boost clock of 1124 MHz, but its memory runs at 1001 MHz (2 Gbps effective). The GTX 670M’s memory clock is lower in MHz, but its effective data rate is higher.

Q: What are the memory capacities of these two GPUs?

A: The GTX 670M has 1536 MB of GDDR5 memory, while the K620M has 2 GB of DDR3 memory.

Architecture Differences

The architectural divide between these two is significant, representing two distinct design philosophies from NVIDIA. The GeForce GTX 670M is built on the Fermi 2.0 architecture with the GF114 chip, manufactured on a 40 nm process at TSMC. This chip is large, with a die size of 332 mm² and a transistor count of 1,950 million, resulting in a transistor density of 5.9M per mm². The Quadro K620M, on the other hand, uses the Maxwell architecture with the GM108S chip, built on a more modern 28 nm process, also at TSMC. It is a much smaller chip, with a die size of 77 mm² and 1,020 million transistors, giving it a much higher transistor density of 13.2M per mm².

This difference in node and die size explains much of the performance and efficiency gap. The newer 28 nm process allows the K620M to pack more transistors per square millimeter, but the overall chip is far smaller and simpler. The GTX 670M’s larger, older chip is built for raw throughput with more functional units. The K620M supports Vulkan 1.4 in its API list, while the GTX 670M has no Vulkan support listed. Both support DirectX 12 (11_0) and OpenGL 4.6.

The Verdict

The data points to a simple conclusion: the NVIDIA GeForce GTX 670M is the faster GPU in raw compute performance. It wins the only head-to-head benchmark recorded, and its average score is higher. For any task that relies heavily on raw shader throughput or general compute, the GTX 670M is the better choice.

The Quadro K620M, however, is not without merit. Its key advantage is efficiency. The data shows it has a TDP of 30 W, which is less than half of the GTX 670M’s 75 W. It also has a higher base clock of 1029 MHz and a boost clock of 1124 MHz, compared to the GTX 670M’s unspecified base and boost clocks. This means the K620M can achieve a respectable score while drawing significantly less power. For a mobile workstation where battery life and thermal management are critical, this efficiency is a major factor.

The choice depends on the priority. If the user needs maximum compute performance for tasks like rendering or GPU-accelerated applications, the GTX 670M is the clear winner, despite its higher power draw. If the priority is a cooler, more power-efficient system that still provides adequate professional performance, the K620M is the more practical option. The benchmark scores do not show a huge gap, but the architectural differences suggest the GTX 670M has more headroom for heavy workloads.

Specification Differences

The two GPUs differ across nearly every core specification. Here are the fields where the data shows a difference:

• Chip: GF114 (GTX 670M) vs GM108S (K620M)
• Architecture: Fermi 2.0 vs Maxwell
• Generation: GeForce 600M vs Quadro Kepler-M (Kx200M)
• Process Node: 40 nm vs 28 nm
• Transistors: 1,950 million vs 1,020 million
• Die Size: 332 mm² vs 77 mm²
• Transistor Density: 5.9M / mm² vs 13.2M / mm²
• Base Clock: Not specified vs 1029 MHz
• Boost Clock: Not specified vs 1124 MHz
• Memory Clock: 750 MHz (3 Gbps effective) vs 1001 MHz (2 Gbps effective)
• Memory Size: 1536 MB vs 2 GB
• Memory Type: GDDR5 vs DDR3
• Bus Width: 192 bit vs 64 bit
• Bandwidth: 72.00 GB/s vs 16.02 GB/s
• Shading Units: 336 vs 384
• TMUs: 56 vs 16
• ROPs: 24 vs 8
• Pixel Rate: 8.372 GPixel/s vs 8.992 GPixel/s
• Texture Rate: 33.49 GTexel/s vs 17.98 GTexel/s
• FP32 Performance: 803.7 GFLOPS vs 863.2 GFLOPS
• TDP: 75 W vs 30 W
• Bus Interface: MXM-B (3.0) vs MXM-A (3.0)
• Vulkan Support: Not listed vs 1.4
• Release Date: 2012-03-21 vs 2015-02-28
• Predecessor: GeForce 500M vs Quadro Fermi-M
• Successor: GeForce 700M vs Quadro Maxwell-M

Where Each One Wins

The GTX 670M wins the compute benchmark, but the K620M has specific advantages that make it a better fit for certain scenarios.

NVIDIA GeForce GTX 670M Wins:

• Raw Compute Performance: The 9.3% lead in Geekbench OpenCL is the primary win. This translates to better performance in any OpenCL-accelerated application.
• Memory Bandwidth: With a 192-bit bus and GDDR5, the GTX 670M offers 72.00 GB/s of bandwidth, which is 4.5 times the 16.02 GB/s of the K620M. This is critical for texture-heavy workloads and large data sets.
• Texture and Pixel Fill: The GTX 670M has a significantly higher texture rate (33.49 GTexel/s vs 17.98 GTexel/s) and a similar pixel rate (8.372 GPixel/s vs 8.992 GPixel/s). The texture rate advantage is notable for 3D rendering.

NVIDIA Quadro K620M Wins:

• Power Efficiency: The K620M’s 30 W TDP is a massive advantage over the 75 W TDP of the GTX 670M. This means less heat, longer battery life, and the ability to fit in thinner, lighter laptops.
• Shading Unit Count: With 384 shading units, the K620M has more than the GTX 670M’s 336. This gives it a theoretical edge in compute tasks that scale well with shader count, even if the overall benchmark score is lower.
• FP32 Performance: The K620M has a higher raw FP32 throughput of 863.2 GFLOPS, compared to the GTX 670M’s 803.7 GFLOPS. This is a single metric where the K620M is ahead.

In practice, the GTX 670M is the choice for a user who needs the fastest possible mobile graphics performance and is willing to accept the higher power draw. The K620M is the choice for a professional who needs a quiet, cool, and efficient workstation GPU for CAD, modeling, or light rendering, where the lower power consumption is more valuable than the 9.3% performance deficit. The data shows that while the GTX 670M is faster, the K620M is a more modern, efficient part that trades a small amount of speed for a large reduction in power.