NVIDIA GeForce GTX 675M vs NVIDIA Quadro K620M Comparison

The Verdict

The benchmark database shows a clear overall winner in the GeForce GTX 675M, which beats the Quadro K620M by 16.6% in the only recorded head-to-head test, the Geekbench OpenCL benchmark. This places the GTX 675M at the 39th percentile of all GPUs, while the Quadro K620M sits at the 34th percentile. For users prioritizing raw compute throughput, the GTX 675M is the stronger choice. However, the Quadro K620M is not without merit: it is a Maxwell-generation part with a much smaller die, lower power draw, and a dedicated professional workstation pedigree. The data suggests the GTX 675M for users who need maximum floating-point performance, while the Quadro K620M suits those who value efficiency and a more modern feature set in a mobile workstation context. The GTX 675M wins the only benchmark, but the K620M’s architectural advantages may matter more in specific professional workloads.

Where Each One Wins

The GeForce GTX 675M wins the only recorded OpenCL benchmark outright, scoring 6946 against the Quadro K620M's 5957. That is a 16.6% lead, and it also outperforms its closest rivals in the database: it sits 1.7% ahead of the AMD FirePro M5100, and it trails the AMD Radeon R5 M240 by only 0.4%, the GeForce GTX 680M by 1.1%, and the NVIDIA T600 by 1.3%. In practical terms, the GTX 675M’s advantage comes from higher memory bandwidth (96.00 GB/s vs 16.02 GB/s), a wider 256-bit bus versus 64-bit, and double the texture units (64 vs 16) and ROPs (32 vs 8). These characteristics make it better suited for graphics-intensive tasks where memory throughput and pixel fill rate dominate.

The Quadro K620M wins in efficiency and architectural modernity. Its 30 W TDP is a third of the GTX 675M's 100 W. It is built on a 28 nm process versus 40 nm, and its transistor density is more than double (13.2M per mm² versus 5.9M per mm²). It also supports Vulkan 1.4, while the GTX 675M does not list a Vulkan version in the database. The K620M’s nearest rivals include the Intel UHD Graphics 730 (0.5% ahead) and the AMD Radeon HD 8730M (exactly tied), suggesting that it competes in a lower-performance tier but with a more efficient design. For users who prioritize battery life, thermal management, or compatibility with modern API standards, the K620M has a case.

Architecture Differences

The two GPUs come from different NVIDIA architectures and generations. The GeForce GTX 675M uses the GF114 chip, based on Fermi 2.0, and belongs to the GeForce 600M series. It is fabricated on a 40 nm process at TSMC, with 1,950 million transistors on a 332 mm² die. The Quadro K620M uses the GM108S chip, based on Maxwell, and is listed under the Quadro Kepler-M (Kx200M) generation. It is also TSMC-made, but on a 28 nm node, with 1,020 million transistors on a much smaller 77 mm² die. This explains the density difference: 5.9M transistors per mm² for Fermi versus 13.2M for Maxwell.

Both GPUs have 384 shading units, but the similarity ends there. The GTX 675M has 64 texture mapping units and 32 ROPs, while the K620M has 16 TMUs and 8 ROPs. The memory subsystem differs sharply: the GTX 675M uses 2 GB of GDDR5 on a 256-bit bus, achieving 96.00 GB/s, while the K620M uses 2 GB of DDR3 on a 64-bit bus, achieving only 16.02 GB/s. Clock behavior also differs: the K620M has a base clock of 1029 MHz and a boost clock of 1124 MHz, while the GTX 675M lists no base or boost clock in the database, only a memory clock of 750 MHz (3 Gbps effective). The GTX 675M’s memory clock is lower than the K620M’s 1001 MHz (2 Gbps effective), but the wider bus compensates.

The API support shows another split. Both support DirectX 12 (11_0) and OpenGL 4.6, but only the K620M lists Vulkan 1.4 support. The GTX 675M has no Vulkan version recorded. The bus interface also differs: the GTX 675M uses MXM-B (3.0), while the K620M uses MXM-A (3.0). Both are MXM modules with no power connectors and portable-device-dependent display outputs. The release dates are about three years apart: the GTX 675M launched in March 2012, the K620M in February 2015. Both are end-of-life products now.

FAQ

Q: Which GPU has the higher OpenCL benchmark score?

A: The GeForce GTX 675M scores 6946, which is 16.6% higher than the Quadro K620M's 5957 in the Geekbench OpenCL test.

Q: How does each GPU compare to its nearest rivals?

A: The GTX 675M is 1.7% ahead of the AMD FirePro M5100, while trailing the AMD Radeon R5 M240 by 0.4%, the GeForce GTX 680M by 1.1%, and the NVIDIA T600 by 1.3%. The K620M is exactly tied with the AMD Radeon HD 8730M, trails the AMD Radeon HD 8750M by 0.2% and the NVIDIA Quadro K4000 by 0.4%, and leads the Intel UHD Graphics 730 by 0.5%.

Q: What are the power consumption figures?

A: The GTX 675M has a TDP of 100 W, while the K620M has a TDP of 30 W, making the K620M significantly more power-efficient.

Q: Do both GPUs have the same number of shading units?

A: Yes, both have 384 shading units. However, the GTX 675M has 64 TMUs and 32 ROPs, whereas the K620M has 16 TMUs and 8 ROPs.

Q: Which GPU supports Vulkan?

A: Only the Quadro K620M lists Vulkan 1.4 support. The GTX 675M does not have a Vulkan version recorded in the database.

Q: What is the memory configuration difference?

A: The GTX 675M uses 2 GB of GDDR5 on a 256-bit bus with 96.00 GB/s bandwidth, while the K620M uses 2 GB of DDR3 on a 64-bit bus with 16.02 GB/s bandwidth.

Head-to-Head Benchmarks

The database contains a single head-to-head benchmark: Geekbench OpenCL. The GTX 675M scores 6946, and the K620M scores 5957. The delta is 16.6% in favor of the GTX 675M. This is a substantial margin, and it aligns with the architectural differences. The GTX 675M’s 256-bit memory bus and 96.00 GB/s bandwidth provide six times the bandwidth of the K620M’s 64-bit bus and 16.02 GB/s. The GTX 675M also has four times the texture fill rate (39.68 GTexel/s versus 17.98 GTexel/s) and a slightly higher pixel rate (9.920 GPixel/s versus 8.992 GPixel/s). Its FP32 compute is 952.3 GFLOPS versus 863.2 GFLOPS, a 10.3% advantage in raw floating-point throughput.

In the context of the nearest rivals, the GTX 675M’s score is competitive: it is within 1.3% of the NVIDIA T600 and within 1.1% of the GeForce GTX 680M, while slightly below the AMD Radeon R5 M240. The K620M’s score, by contrast, is closely matched with mid-range parts like the AMD Radeon HD 8730M and HD 8750M, and it edges out the Intel UHD Graphics 730 by only 0.5%. The GTX 675M’s win is not just a narrow margin over its rival; it is a consistent performance tier above the K620M’s peer group.

The only benchmark win goes to the GTX 675M, with 1 win and 0 for the K620M. That said, the K620M’s lower TDP and newer architecture mean the benchmark score does not tell the whole story for mobile workstation use cases. The data shows the GTX 675M dominates in raw compute and memory throughput, but the K620M offers a more efficient and modern platform.

Specification Differences

The two GPUs differ in nearly every physical and clock specification. The process node is 40 nm for the GTX 675M versus 28 nm for the K620M. Transistor count is 1,950 million versus 1,020 million, and die size is 332 mm² versus 77 mm². Transistor density is 5.9M per mm² versus 13.2M per mm². The GTX 675M has no recorded base or boost clock, while the K620M has a base of 1029 MHz and a boost of 1124 MHz. Memory clock is 750 MHz (3 Gbps effective) for the GTX 675M, and 1001 MHz (2 Gbps effective) for the K620M. Memory type is GDDR5 for the GTX 675M and DDR3 for the K620M. Bus width is 256-bit versus 64-bit, and bandwidth is 96.00 GB/s versus 16.02 GB/s.

Shading units are identical at 384, but TMUs differ (64 versus 16) and ROPs differ (32 versus 8). Pixel rate is 9.920 GPixel/s versus 8.992 GPixel/s, texture rate is 39.68 GTexel/s versus 17.98 GTexel/s, and FP32 is 952.3 GFLOPS versus 863.2 GFLOPS. TDP is 100 W versus 30 W. The bus interface is MXM-B (3.0) for the GTX 675M and MXM-A (3.0) for the K620M. The GTX 675M has no Vulkan version listed, while the K620M supports Vulkan 1.4. Both support DirectX 12 (11_0) and OpenGL 4.6. The GTX 675M is from the GeForce 600M generation, with a predecessor in GeForce 500M and successor in GeForce 700M. The K620M is from the Quadro Kepler-M (Kx200M) generation, with a predecessor in Quadro Fermi-M and successor in Quadro Maxwell-M. Both are end-of-life products, with the GTX 675M released in March 2012 and the K620M in February 2015. Neither has a recorded launch MSRP.