NVIDIA GeForce GTX 765M vs NVIDIA Quadro K620M Comparison

Head-to-Head Benchmarks

The single recorded head-to-head benchmark in the database is Geekbench OpenCL, and it favors the NVIDIA GeForce GTX 765M decisively. The GTX 765M scores 7,176, while the Quadro K620M scores 5,957. That is a delta of -17%, meaning the GTX 765M outperforms the K620M by 17 percentage points in this compute workload. This is not a marginal gap; it places the two mobile GPUs in clearly different performance tiers for OpenCL compute tasks.

To contextualize the K620M's score, its 5,957 OpenCL result places it at the 34th percentile among all GPUs in the database. Its nearest rivals are tightly clustered: the AMD Radeon HD 8730M scores 5,955 (a delta of 0%), the AMD Radeon HD 8750M scores 5,970 (delta -0.2%), and the NVIDIA Quadro K4000 scores 5,982 (delta -0.4%). The Intel UHD Graphics 730 trails slightly at 5,929 (delta 0.5%). This indicates the K620M sits within a very narrow performance band, effectively tied with those four competitors, with no single rival more than half a percentage point away in either direction.

The GTX 765M, by contrast, is placed at the 32nd percentile overall, slightly lower than the K620M's percentile, despite having a higher OpenCL score. This apparent contradiction is explained by its average benchmark score across multiple tests. The GTX 765M has an average benchmark score of 5,501, which pulls its percentile down because it also records scores in Metal (2,612) and Vulkan (6,714). The OpenCL result is the strongest of its three recorded benchmarks. Its nearest rivals for the average score include the NVIDIA GeForce MX130 at 5,508 (delta -0.1%), the AMD Radeon R7 M440 at 5,483 (delta 0.3%), the NVIDIA Quadro M4000 at 5,467 (delta 0.6%), and the AMD FirePro M4000 at 5,537 (delta -0.7%). Again, the clustering is tight, but the GTX 765M's average is dragged down by the Metal result, which is substantially lower than its OpenCL and Vulkan scores.

When comparing the two directly, the OpenCL result is the only test where both have recorded scores, and it shows a 17% advantage for the GTX 765M. No other head-to-head data exists in the database, so any analysis of gaming or graphics-specific workloads must be inferred from the architecture and specification differences rather than direct measurements.

Architecture Differences

The two GPUs come from different NVIDIA architectures and different generations. The Quadro K620M is built on the Maxwell architecture, using the GM108S chip. Its generation is listed as "Quadro Kepler-M (Kx200M)", which is a curious designation given the Maxwell architecture, but the database records it that way. The GeForce GTX 765M uses the Kepler architecture with the GK106 chip and belongs to the GeForce 700M generation.

Both are manufactured by TSMC on a 28 nm process node. However, the transistor counts and die sizes differ substantially. The K620M has 1,020 million transistors on a 77 mm² die, yielding a transistor density of 13.2 million transistors per square millimeter. The GTX 765M has 2,540 million transistors on a 221 mm² die, with a density of 11.5 million per square millimeter. The GTX 765M's die is nearly three times larger, and it packs roughly 2.5 times more transistors, though at a slightly lower density.

Clock speeds tell a different story. The K620M runs at a base clock of 1,029 MHz and boosts to 1,124 MHz. The GTX 765M has a lower base clock of 797 MHz and a boost of 863 MHz. Despite the lower clocks, the GTX 765M delivers far more compute throughput because of its larger execution resources. The K620M has 384 shading units, 16 texture mapping units, and 8 raster output units. The GTX 765M has 768 shading units, 64 TMUs, and 16 ROPs, exactly double the shaders, quadruple the TMUs, and double the ROPs.

These resource differences translate into raw throughput numbers. The K620M achieves a pixel rate of 8.992 GPixel/s and a texture rate of 17.98 GTexel/s. The GTX 765M achieves 13.81 GPixel/s and 55.23 GTexel/s. In FP32 compute, the GTX 765M reaches 1,325.6 GFLOPS versus the K620M's 863.2 GFLOPS. Neither GPU has FP16 support recorded, and neither has ray tracing or tensor cores.

Memory configurations also diverge significantly. The K620M uses 2 GB of DDR3 on a 64-bit bus, producing 16.02 GB/s of bandwidth. The GTX 765M uses 2 GB of GDDR5 on a 128-bit bus, producing 64.13 GB/s of bandwidth. That is a fourfold difference in memory bandwidth, which is critical for texture-heavy workloads and compute tasks that are memory-bound. The effective memory clock is 2 Gbps for the K620M and 4 Gbps for the GTX 765M.

Power draw differs as well. The K620M has a TDP of 30 W, while the GTX 765M has a TDP of 75 W. Both are MXM modules, but the K620M uses an MXM-A (3.0) interface while the GTX 765M uses MXM-B (3.0). Neither requires external power connectors, and both have display outputs that are listed as "Portable Device Dependent."

API support is nearly identical in one respect: both support DirectX 12 (11_0) and OpenGL 4.6. However, they differ on Vulkan. The K620M supports Vulkan 1.4, while the GTX 765M supports Vulkan 1.2.175. This gives the newer Quadro a more recent Vulkan version, though the older GeForce has a recorded Vulkan benchmark score of 6,714, indicating it can execute Vulkan workloads.

FAQ

Q: Which GPU has the higher OpenCL benchmark score?

A: The NVIDIA GeForce GTX 765M scores 7,176 in Geekbench OpenCL, which is 17% higher than the Quadro K620M's score of 5,957.

Q: How does the Quadro K620M compare to its closest rivals?

A: The K620M's OpenCL score of 5,957 is effectively tied with the AMD Radeon HD 8730M (5,955), the AMD Radeon HD 8750M (5,970), and the NVIDIA Quadro K4000 (5,982). It is slightly ahead of the Intel UHD Graphics 730 (5,929) by 0.5%.

Q: What memory bandwidth does each GPU provide?

A: The Quadro K620M provides 16.02 GB/s using 2 GB of DDR3 on a 64-bit bus. The GeForce GTX 765M provides 64.13 GB/s using 2 GB of GDDR5 on a 128-bit bus.

Q: Which GPU has a higher transistor count?

A: The GeForce GTX 765M has 2,540 million transistors on a 221 mm² die. The Quadro K620M has 1,020 million transistors on a 77 mm² die.

Q: Do both GPUs support the same DirectX version?

A: Yes, both support DirectX 12 (11_0) and OpenGL 4.6. They differ in Vulkan support: the K620M supports Vulkan 1.4, while the GTX 765M supports Vulkan 1.2.175.

Q: What is the power draw difference?

A: The Quadro K620M has a TDP of 30 W. The GeForce GTX 765M has a TDP of 75 W, which is 45 W higher.

The Verdict

The data points to a clear performance hierarchy. In the only direct benchmark comparison available, the GeForce GTX 765M is 17% ahead of the Quadro K620M in OpenCL compute. The GTX 765M also has superior raw specifications: double the shading units, double the ROPs, quadruple the TMUs, and four times the memory bandwidth. Its FP32 throughput of 1,325.6 GFLOPS exceeds the K620M's 863.2 GFLOPS by a wide margin. For any workload that stresses compute, texturing, or memory throughput, the GTX 765M is the stronger part.

The Quadro K620M does have advantages that matter in specific contexts. Its TDP of 30 W is less than half of the GTX 765M's 75 W, making it far more suitable for power-constrained mobile platforms. It also has a newer Vulkan version (1.4 versus 1.2.175) and a higher base clock (1,029 MHz versus 797 MHz), though the higher clock does not compensate for the smaller execution footprint. Its transistor density is higher (13.2M per mm² versus 11.5M per mm²), reflecting the newer Maxwell design, but this does not translate into higher absolute performance in the recorded benchmarks.

For users prioritizing compute performance and memory bandwidth, the GeForce GTX 765M is the clear choice based on the database measurements. For users with strict power budgets or those needing the latest Vulkan feature set, the Quadro K620M has qualitative advantages, but it sacrifices 17% OpenCL performance and over 75% of the memory bandwidth to get them.

The percentile rankings complicate the picture slightly. The K620M sits at the 34th percentile, while the GTX 765M sits at the 32nd. This is because the GTX 765M's average benchmark score of 5,501 includes a weak Metal score of 2,612, which pulls its average below its OpenCL result. On the specific OpenCL test, the GTX 765M is unequivocally faster. The database records one win for the GTX 765M and zero for the K620M in head-to-head tests.

Specification Differences

The following fields differ between the two GPUs:

• Chip: GM108S (K620M) versus GK106 (GTX 765M)
• Architecture: Maxwell (K620M) versus Kepler (GTX 765M)
• Generation: Quadro Kepler-M (Kx200M) (K620M) versus GeForce 700M (GTX 765M)
• Transistors: 1,020 million (K620M) versus 2,540 million (GTX 765M)
• Die Size: 77 mm² (K620M) versus 221 mm² (GTX 765M)
• Transistor Density: 13.2M / mm² (K620M) versus 11.5M / mm² (GTX 765M)
• Base Clock: 1,029 MHz (K620M) versus 797 MHz (GTX 765M)
• Boost Clock: 1,124 MHz (K620M) versus 863 MHz (GTX 765M)
• Memory Clock: 2 Gbps effective (K620M) versus 4 Gbps effective (GTX 765M)
• Memory Type: DDR3 (K620M) versus GDDR5 (GTX 765M)
• Bus Width: 64 bit (K620M) versus 128 bit (GTX 765M)
• Memory Bandwidth: 16.02 GB/s (K620M) versus 64.13 GB/s (GTX 765M)
• Shading Units: 384 (K620M) versus 768 (GTX 765M)
• TMUs: 16 (K620M) versus 64 (GTX 765M)
• ROPs: 8 (K620M) versus 16 (GTX 765M)
• Pixel Rate: 8.992 GPixel/s (K620M) versus 13.81 GPixel/s (GTX 765M)
• Texture Rate: 17.98 GTexel/s (K620M) versus 55.23 GTexel/s (GTX 765M)
• FP32: 863.2 GFLOPS (K620M) versus 1,325.6 GFLOPS (GTX 765M)
• TDP: 30 W (K620M) versus 75 W (GTX 765M)
• Bus Interface: MXM-A (3.0) (K620M) versus MXM-B (3.0) (GTX 765M)
• Vulkan: 1.4 (K620M) versus 1.2.175 (GTX 765M)
• Release Date: 2015-02-28 (K620M) versus 2013-05-29 (GTX 765M)
• Predecessor: Quadro Fermi-M (K620M) versus GeForce 600M (GTX 765M)
• Successor: Quadro Maxwell-M (K620M) versus GeForce 800M (GTX 765M)
• Geekbench OpenCL Score: 5,957 (K620M) versus 7,176 (GTX 765M)
• Average Benchmark Score: 5,957 (K620M) versus 5,501 (GTX 765M)
• Percentile vs All GPUs: 34 (K620M) versus 32 (GTX 765M)