NVIDIA GeForce GTX 760M vs NVIDIA Quadro K620M Comparison

FAQ

Q: Which GPU has the higher OpenCL benchmark score?

A: The NVIDIA Quadro K620M scores 5,957 in Geekbench OpenCL, while the NVIDIA GeForce GTX 760M scores 5,604. The Quadro K620M wins this head-to-head by 6.3%.

Q: How does the GTX 760M compare to its nearest rivals in average benchmark score?

A: The GTX 760M has an average benchmark score of 5,236. It sits 0.5% above the NVIDIA Quadro 4000M (5,211), 0.9% below the NVIDIA GeForce 940M (5,284), and 1.3% below the NVIDIA GeForce GTX 980M (5,308). It also leads the AMD Radeon R7 M260X (5,161) by 1.5%.

Q: What is the memory bandwidth difference between the two cards?

A: The GTX 760M offers 64.13 GB/s of bandwidth, which is four times the 16.02 GB/s of the Quadro K620M. The GTX 760M also uses a 128-bit bus with GDDR5 memory, while the Quadro K620M uses a 64-bit bus with DDR3 memory.

Q: Which GPU has the higher transistor count and die size?

A: The GTX 760M has 2,540 million transistors on a 221 mm² die, while the Quadro K620M has 1,020 million transistors on a 77 mm² die. The GTX 760M's die is nearly three times larger.

Q: What is the TDP of each GPU?

A: The Quadro K620M has a TDP of 30 W, while the GTX 760M has a TDP of 55 W. The Quadro K620M is the more power-efficient option in terms of rated power draw.

Q: Which GPU supports Vulkan, and what version?

A: The Quadro K620M supports Vulkan 1.4, while the GTX 760M supports Vulkan 1.2.175. Both support DirectX 12 (11_0) and OpenGL 4.6.

The Verdict

The data points to a clear split in intended use. The NVIDIA Quadro K620M is the stronger choice for OpenCL compute workloads, as it wins the only recorded head-to-head benchmark. Its 5,957 OpenCL score beats the GTX 760M's 5,604 by 6.3%, and its 34th percentile placement among all GPUs is three points higher than the GTX 760M's 31st percentile. For users whose primary metric is OpenCL performance, the Quadro K620M is the better pick.

The GTX 760M, however, offers advantages in raw graphics throughput and memory subsystem capacity. It has double the shading units (768 vs. 384), four times the texture mapping units (64 vs. 16), and double the ROPs (16 vs. 8). Its memory bandwidth of 64.13 GB/s is four times higher than the Quadro K620M's 16.02 GB/s, and its FP32 compute rating of 1,104.4 GFLOPS exceeds the Quadro K620M's 863.2 GFLOPS by 28%. The GTX 760M also has a Vulkan benchmark score of 4,868, a test the Quadro K620M does not have recorded.

For users prioritizing OpenCL compute in a low-power mobile workstation context, the Quadro K620M is the data-backed choice. For users prioritizing memory bandwidth, shading throughput, and Vulkan support, the GTX 760M is the stronger candidate. Neither card offers a decisive all-around win; the selection depends on workload focus.

Head-to-Head Benchmarks

The sole recorded head-to-head test is Geekbench OpenCL. In this test, the Quadro K620M scores 5,957, while the GTX 760M scores 5,604. The Quadro K620M wins by 6.3%. This is the only benchmark where both cards have a directly comparable score, and it represents a clear, if modest, victory for the professional-grade mobile GPU.

The GTX 760M does have an additional recorded benchmark, Geekbench Vulkan, where it scores 4,868. The Quadro K620M has no Vulkan score in the database, so no direct comparison is possible for that API. The OpenCL result, however, is the database's primary point of comparison, and it favors the Quadro K620M.

In the broader context of rival comparisons, the Quadro K620M's OpenCL score of 5,957 places it slightly above the AMD Radeon HD 8730M (5,955) by 0%, above the Intel UHD Graphics 730 (5,929) by 0.5%, and just below the AMD Radeon HD 8750M (5,970) by 0.2% and the NVIDIA Quadro K4000 (5,982) by 0.4%. The GTX 760M's average score of 5,236, meanwhile, is 0.5% above the NVIDIA Quadro 4000M (5,211) and 1.5% above the AMD Radeon R7 M260X (5,161), but 0.9% below the NVIDIA GeForce 940M (5,284) and 1.3% below the NVIDIA GeForce GTX 980M (5,308).

The head-to-head delta of 6.3% is more significant than the percentile gap suggests. The Quadro K620M's 34th percentile versus the GTX 760M's 31st percentile shows that, in the full GPU landscape, the difference is small. But in a direct comparison, the Quadro K620M is the clear winner in OpenCL.

Specification Differences

The two NVIDIA mobile GPUs differ across nearly every core specification. The Quadro K620M uses the GM108S chip with 1,020 million transistors on a 77 mm² die, while the GTX 760M uses the GK106S chip with 2,540 million transistors on a 221 mm² die. The transistor density also differs: the Quadro K620M packs 13.2 million transistors per mm², while the GTX 760M has 11.5 million per mm².

Clock speeds vary significantly. The Quadro K620M has a base clock of 1,029 MHz and a boost clock of 1,124 MHz. The GTX 760M has a base clock of 628 MHz and a boost clock of 719 MHz. The Quadro K620M's clocks are substantially higher, but the GTX 760M compensates with more execution units.

Memory configurations are entirely different. The Quadro K620M has 2 GB of DDR3 on a 64-bit bus, yielding 16.02 GB/s of bandwidth and a memory clock of 1,001 MHz (2 Gbps effective). The GTX 760M has 2 GB of GDDR5 on a 128-bit bus, yielding 64.13 GB/s of bandwidth and a memory clock of 1,002 MHz (4 Gbps effective).

The compute resources are heavily skewed toward the GTX 760M. It has 768 shading units, 64 TMUs, and 16 ROPs. The Quadro K620M has 384 shading units, 16 TMUs, and 8 ROPs. Pixel rate is 11.50 GPixel/s for the GTX 760M versus 8.992 GPixel/s for the Quadro K620M. Texture rate is 46.02 GTexel/s versus 17.98 GTexel/s. FP32 compute is 1,104.4 GFLOPS versus 863.2 GFLOPS.

Power draw differs as well. The Quadro K620M is rated at 30 W TDP, while the GTX 760M is rated at 55 W TDP. Both use MXM Module slot width and have no power connectors.

The bus interface is different: the Quadro K620M uses MXM-A (3.0), while the GTX 760M uses PCIe 3.0 x16. Both have display outputs described as "Portable Device Dependent."

Architecture Differences

The two GPUs come from different NVIDIA architectures. The Quadro K620M is built on Maxwell architecture (chip GM108S), while the GTX 760M is built on Kepler architecture (chip GK106S). Both are manufactured on a 28 nm process at TSMC, but the transistor density differs: 13.2M per mm² for the Maxwell chip versus 11.5M per mm² for the Kepler chip.

The generation and product positioning also differ. The Quadro K620M is listed under the "Quadro Kepler-M (Kx200M)" generation, despite being a Maxwell chip, and its predecessor is the Quadro Fermi-M with the Quadro Maxwell-M as successor. The GTX 760M belongs to the GeForce 700M generation, with the GeForce 600M as predecessor and GeForce 800M as successor.

API support is nearly identical, with both supporting DirectX 12 (11_0) and OpenGL 4.6. The Vulkan version differs: the Quadro K620M supports Vulkan 1.4, while the GTX 760M supports Vulkan 1.2.175. The GTX 760M has a recorded Vulkan benchmark score of 4,868, while the Quadro K620M has no Vulkan score in the database.

The release dates show a generational gap: the GTX 760M was released on May 29, 2013, while the Quadro K620M was released on February 28, 2015. Both are marked as end-of-life production status.

The memory architecture differences are fundamental to the performance gap. The GTX 760M's GDDR5 memory on a 128-bit bus provides four times the bandwidth of the Quadro K620M's DDR3 on a 64-bit bus. This bandwidth advantage is critical for texture-heavy workloads. The Quadro K620M's higher clock speeds (1,029 MHz base vs. 628 MHz base) partially offset its lower core count in compute tasks, as the OpenCL benchmark demonstrates.

The Maxwell architecture in the Quadro K620M achieves higher transistor density (13.2M/mm² vs. 11.5M/mm²) despite having fewer total transistors. This density advantage, combined with the lower TDP (30 W vs. 55 W), indicates a more power-efficient design per transistor. The Kepler architecture in the GTX 760M uses more silicon area and more power to deliver higher raw throughput.

In terms of shader configuration, the GTX 760M's 768 shading units are double the Quadro K620M's 384. The TMU count is quadrupled (64 vs. 16), and the ROP count is doubled (16 vs. 8). These differences manifest in the texture rate (46.02 GTexel/s vs. 17.98 GTexel/s) and pixel rate (11.50 GPixel/s vs. 8.992 GPixel/s) metrics.

The FP32 output tells a similar story: the GTX 760M delivers 1,104.4 GFLOPS versus 863.2 GFLOPS for the Quadro K620M. But the OpenCL benchmark result flips the expected outcome, with the Quadro K620M scoring higher. This suggests that the Maxwell architecture's efficiency and higher clock speeds can outweigh the Kepler chip's raw resource advantage in certain compute workloads.