NVIDIA GeForce GTX 980M vs NVIDIA Quadro K620M Comparison

The NVIDIA Quadro K620M and NVIDIA GeForce GTX 980M represent two very different corners of the mobile GPU market from the same manufacturer. The K620M is a professional laptop part aimed at compatibility and efficiency, while the GTX 980M is a high-end consumer gaming chip. The recorded data shows a single direct head-to-head benchmark, but the full specification lists reveal a massive gulf in almost every measurable hardware category.

Head-to-Head Benchmarks

The only direct comparison available in the database is the Geekbench OpenCL test, which measures general-purpose compute performance across the GPU. The results are lopsided. The Quadro K620M scores 5,957, while the GTX 980M scores 23,832. That places the GTX 980M a full 75% ahead of the K620M in this workload. To put that in perspective, the delta is so large that the GTX 980M’s score is exactly four times higher than the K620M’s score.

This outcome is not surprising given the underlying hardware. The GTX 980M carries 1,536 shading units, while the K620M has just 384. That is a 4:1 ratio in raw shader count. The GTX 980M also has 96 texture mapping units versus 16 on the K620M, and 64 render output units versus 8. Each of those multipliers directly feeds into compute performance, which is what Geekbench OpenCL stresses.

The K620M’s nearest rivals in the database are all low-to-mid-range parts. It sits within 0.5% of the AMD Radeon HD 8730M (5,955), the AMD Radeon HD 8750M (5,970), and the Intel UHD Graphics 730 (5,929). Its largest positive delta is just 0.5% over the Intel part, and its largest negative delta is 0.4% behind the NVIDIA Quadro K4000 (5,982). This places the K620M in a very narrow performance band, essentially trading blows with integrated and entry-level discrete GPUs from 2013 to 2021.

The GTX 980M, despite its much higher raw score, has a lower overall percentile rank (31st) than the K620M (34th). That is because the GTX 980M’s average benchmark score across all recorded tests is 5,308, which is dragged down by its many PassMark tests. Its nearest rivals include the NVIDIA GeForce 930A (5,317, delta -0.2%), the NVIDIA GeForce 840M (5,322, delta -0.3%), the NVIDIA GeForce 940M (5,284, delta +0.5%), and the AMD Radeon R7 M445 (5,358, delta -0.9%). So while the GTX 980M crushes the K620M in OpenCL compute, its average score across all benchmarks places it in a similar percentile bucket, indicating that its other tests are heavily weighted toward legacy workloads where it does not excel.

The head-to-head table records exactly one benchmark, and the GTX 980M wins that one. The K620M has zero wins in direct comparisons. This is a clean sweep, though the sample size is limited to a single test.

Where Each One Wins

The GTX 980M wins the only direct benchmark comparison, and it wins by a massive margin. In OpenCL compute, it is 75% faster than the K620M. That advantage comes from having four times the shading units, six times the TMUs, and eight times the ROPs. For any workload that uses general-purpose GPU compute, whether that is rendering, physics simulation, or data processing, the GTX 980M is the clear choice.

The K620M, however, has a different set of strengths. Its average benchmark score of 5,957 is higher than the GTX 980M’s average of 5,308. That is because the GTX 980M’s average is pulled down by its PassMark results, which include scores like 35 in DirectX 10, 57 in DirectX 11, 31 in DirectX 12, and 125 in DirectX 9. These are low absolute numbers, and they drag the average down. The K620M only has one recorded benchmark, so its average is exactly its OpenCL score.

In terms of percentile rank, the K620M sits at 34, which is higher than the GTX 980M’s 31. This suggests that relative to all GPUs in the database, the K620M is closer to the middle of the pack, while the GTX 980M, despite its enormous compute power, has a wider spread of results across different test types.

For pure gaming and DirectX workloads, the GTX 980M’s PassMark DirectX 11 score of 57 is much higher than its DirectX 12 score of 31, indicating that it performs better in older APIs. The K620M has no recorded DirectX benchmark, so no direct comparison is possible there. For professional applications that rely on OpenGL or Vulkan, both cards support OpenGL 4.6 and Vulkan 1.4, but the GTX 980M’s Vulkan score of 17,703 in Geekbench shows strong compute capability in that API.

The GTX 980M also wins decisively in memory bandwidth. Its GDDR5 memory runs at 5 Gbps effective, with a 256-bit bus, yielding 160.4 GB/s. The K620M uses DDR3 at 2 Gbps effective, with a 64-bit bus, yielding just 16.02 GB/s. That is a 10x difference in bandwidth. For texture-heavy workloads, large frame buffers, or high-resolution rendering, the GTX 980M’s bandwidth advantage is critical.

The K620M’s only real advantage is efficiency. Its TDP is 30 W, while the GTX 980M has no recorded TDP in the database. The K620M also uses a smaller die (77 mm² vs 398 mm²) and far fewer transistors (1,020 million vs 5,200 million), which suggests it draws significantly less power in real-world operation. For thin-and-light workstations or embedded systems, that efficiency matters.

Architecture Differences

The two GPUs are built on the same process node, 28 nm, at the same foundry, TSMC. Both use the Maxwell architecture, but they are different revisions. The K620M uses Maxwell (specifically the GM108S chip), while the GTX 980M uses Maxwell 2.0 (the GM204 chip). This is a meaningful distinction. Maxwell 2.0 introduced several feature improvements over the original Maxwell, including better support for DirectX 12, specifically at feature level 12_1 versus 11_0 on the older Maxwell.

The transistor counts tell a story of scale. The K620M has 1,020 million transistors on a 77 mm² die, giving a transistor density of 13.2 million per square millimeter. The GTX 980M has 5,200 million transistors on a 398 mm² die, with a density of 13.1 million per square millimeter. The densities are nearly identical, which makes sense since both are on the same 28 nm process. The difference is purely die size: the GTX 980M is roughly 5.2 times larger.

The memory subsystems are entirely different. The K620M uses 2 GB of DDR3 on a 64-bit bus. The GTX 980M uses 8 GB of GDDR5 on a 256-bit bus. The effective memory clock is 2 Gbps on the K620M versus 5 Gbps on the GTX 980M. These differences compound: the GTX 980M has four times the memory capacity, four times the bus width, and 2.5 times the effective clock speed, resulting in a 10x bandwidth advantage (160.4 GB/s vs 16.02 GB/s).

The compute resources are scaled similarly. The GTX 980M has 1,536 shading units, 96 TMUs, and 64 ROPs. The K620M has 384 shading units, 16 TMUs, and 8 ROPs. The GTX 980M is exactly four times larger in shader count, six times larger in TMUs, and eight times larger in ROPs. This translates directly to pixel rate (72.13 GPixel/s vs 8.992 GPixel/s) and texture rate (108.2 GTexel/s vs 17.98 GTexel/s). The FP32 throughput is 3.462 TFLOPS on the GTX 980M versus 863.2 GFLOPS on the K620M.

The bus interfaces also differ. The K620M uses MXM-A (3.0), while the GTX 980M uses MXM-B (3.0). MXM-B is a larger module form factor, which accommodates the bigger die and more memory. Both are MXM Module slots, and both have no external power connectors, relying on the motherboard for power delivery. The GTX 980M has no recorded TDP, so the power draw difference cannot be quantified from the data, but the physical size difference suggests the GTX 980M consumes significantly more power.

The release dates are close, with the GTX 980M launching on 2014-10-06 and the K620M on 2015-02-28. The GTX 980M is the predecessor to the GeForce 10 Mobile series, while the K620M is the successor to Quadro Fermi-M and was itself succeeded by Quadro Maxwell-M. Both are end-of-life products.

FAQ

Q: Which GPU is faster in compute workloads?

A: The GeForce GTX 980M is 75% faster than the Quadro K620M in the Geekbench OpenCL test, scoring 23,832 versus 5,957.

Q: Do they use the same architecture?

A: Both use the Maxwell architecture, but the GTX 980M uses Maxwell 2.0 (GM204 chip) while the K620M uses the original Maxwell (GM108S chip). The GTX 980M supports DirectX 12_1, while the K620M only supports DirectX 12 (11_0).

Q: How much memory does each card have?

A: The K620M has 2 GB of DDR3 on a 64-bit bus, while the GTX 980M has 8 GB of GDDR5 on a 256-bit bus. The GTX 980M’s memory bandwidth is 160.4 GB/s versus 16.02 GB/s on the K620M.

Q: Which card has a higher average benchmark score?

A: The K620M has a higher average score of 5,957 compared to the GTX 980M’s 5,308. However, the GTX 980M’s average includes multiple PassMark tests with very low scores, which pulls it down.

Q: What is the transistor density of each chip?

A: The K620M has 13.2 million transistors per mm², and the GTX 980M has 13.1 million per mm². They are nearly identical because both are built on the same 28 nm process at TSMC.

Q: Are both cards still in production?

A: No, both are end-of-life products. The GTX 980M was released in October 2014, and the K620M was released in February 2015.

Specification Differences

The two cards differ in nearly every specification. The process node is the same (28 nm, TSMC), but the chip, die size, and transistor count are vastly different. The K620M uses GM108S with 1,020 million transistors on a 77 mm² die. The GTX 980M uses GM204 with 5,200 million transistors on a 398 mm² die.

Clock speeds are similar: the K620M has a base clock of 1029 MHz and a boost of 1124 MHz, while the GTX 980M has a base of 1038 MHz and a boost of 1127 MHz. The memory clocks differ significantly: the K620M runs at 1001 MHz (2 Gbps effective), and the GTX 980M runs at 1253 MHz (5 Gbps effective).

Memory capacity, type, bus width, and bandwidth all favor the GTX 980M. The K620M has 2 GB DDR3, 64-bit, 16.02 GB/s. The GTX 980M has 8 GB GDDR5, 256-bit, 160.4 GB/s.

Shading units are 384 versus 1,536, TMUs are 16 versus 96, and ROPs are 8 versus 64. Pixel rate is 8.992 GPixel/s versus 72.13 GPixel/s, and texture rate is 17.98 GTexel/s versus 108.2 GTexel/s. FP32 performance is 863.2 GFLOPS versus 3.462 TFLOPS.

The K620M has a TDP of 30 W, while the GTX 980M has no recorded TDP. Both use MXM modules, but the K620M uses MXM-A (3.0) and the GTX 980M uses MXM-B (3.0). Neither has power connectors. Both support OpenGL 4.6 and Vulkan 1.4, but the K620M supports DirectX 12 (11_0) while the GTX 980M supports DirectX 12 (12_1). The K620M has a higher percentile rank (34 vs 31), and the GTX 980M has a higher OpenCL score (23,832 vs 5,957). The GTX 980M also has additional recorded benchmarks across PassMark and Geekbench Vulkan, while the K620M only has a single OpenCL result.