AMD Radeon Vega 10 Mobile vs NVIDIA Quadro K620M Comparison

The Verdict

The benchmark data presents a clear hierarchy between these two mobile graphics solutions. The AMD Radeon Vega 10 Mobile leads the NVIDIA Quadro K620M in the recorded OpenCL benchmark, scoring 6476 against 5957, a delta of 8.7%. This places the AMD part in the 37th percentile of all GPUs in the database, while the NVIDIA part sits in the 34th percentile. For users prioritizing raw compute throughput in OpenCL workloads, the AMD solution is the stronger pick.

However, the choice is not purely about performance. The AMD Radeon Vega 10 Mobile is an integrated graphics processor (IGP) with a 10 W TDP, designed to share system memory. The NVIDIA Quadro K620M is a discrete MXM module with a 30 W TDP and its own dedicated 2 GB DDR3 frame buffer with a 64-bit bus. Users who require a dedicated memory pool for professional applications, or who need a modular MXM form factor for serviceability, may prefer the NVIDIA option despite its lower raw score. The data does not indicate that the NVIDIA part wins any benchmark, but its architecture and memory configuration serve a different use case.

Where Each One Wins

The AMD Radeon Vega 10 Mobile wins the only recorded head-to-head benchmark, Geekbench OpenCL, with a score of 6476 versus 5957 for the NVIDIA Quadro K620M. That 8.7% advantage is the sole benchmark win in the database, giving AMD one win and NVIDIA zero. The AMD part also offers a higher pixel rate (10.41 GPixel/s vs 8.992 GPixel/s), a higher texture rate (52.04 GTexel/s vs 17.98 GTexel/s), and more than double the FP32 throughput (1.665 TFLOPS vs 863.2 GFLOPS). For compute-heavy tasks that scale with shading units, the AMD part has 640 shading units versus 384, and 40 texture mapping units versus 16.

The NVIDIA Quadro K620M does not win any benchmark in the recorded data, but it holds advantages in specific hardware characteristics. Its base clock of 1029 MHz and boost clock of 1124 MHz are substantially higher than the AMD base of 300 MHz and boost of 1301 MHz. The NVIDIA part also supports a higher Vulkan version (1.4 vs 1.3), though both support DirectX 12 and OpenGL 4.6. The NVIDIA part has a fixed 2 GB memory allocation, which can be an advantage in scenarios where system memory sharing is undesirable, but the database records no benchmark win for it.

Architecture Differences

The AMD Radeon Vega 10 Mobile is built on GCN 5.0 architecture, fabricated by GlobalFoundries on a 14 nm process. It integrates 4,940 million transistors on a 210 mm² die, yielding a transistor density of 23.5 million per mm². Its chip is codenamed Raven-M, and it belongs to the Vega IGP generation (Raven Ridge-M). The GPU is an integrated part with no power connectors, no slot width beyond IGP, and a bus interface of IGP. Memory is system shared, meaning size, type, bus width, and bandwidth are all system dependent. Its FP16 throughput is 3.331 TFLOPS at a 2:1 ratio, while FP32 is 1.665 TFLOPS. The base clock is 300 MHz, boosting to 1301 MHz.

The NVIDIA Quadro K620M uses Maxwell architecture, fabricated by TSMC on a 28 nm process. It packs 1,020 million transistors on a 77 mm² die, with a transistor density of 13.2 million per mm². Its chip is GM108S, and it belongs to the Quadro Kepler-M generation (Kx200M), despite the Maxwell architecture label. This is a discrete MXM module with a 30 W TDP, using the MXM-A (3.0) bus interface. It has 2 GB of DDR3 memory on a 64-bit bus, delivering 16.02 GB/s of bandwidth. The base clock is 1029 MHz, boosting to 1124 MHz, with memory running at 1001 MHz or 2 Gbps effective. The NVIDIA part has no FP16 support recorded, and its FP32 throughput is 863.2 GFLOPS.

The process node difference is stark: 14 nm for AMD versus 28 nm for NVIDIA. The AMD part integrates far more transistors (4,940 million vs 1,020 million) on a larger die (210 mm² vs 77 mm²). The AMD part also supports DirectX 12 at feature level 12_1, while the NVIDIA part supports DirectX 12 at feature level 11_0. Both support OpenGL 4.6, but the NVIDIA part supports Vulkan 1.4 versus AMD's 1.3. The AMD part is end-of-life with a successor of Navi II IGP, and the NVIDIA part is end-of-life with a successor of Quadro Maxwell-M.

FAQ

Q: Which GPU has the higher OpenCL benchmark score?

A: The AMD Radeon Vega 10 Mobile scores 6476, while the NVIDIA Quadro K620M scores 5957, giving AMD an 8.7% advantage in the recorded Geekbench OpenCL test.

Q: How do their thermal design power figures compare?

A: The AMD Radeon Vega 10 Mobile has a 10 W TDP, while the NVIDIA Quadro K620M has a 30 W TDP. The AMD part consumes one-third the power of the NVIDIA part, according to the recorded specifications.

Q: What memory configurations do these GPUs use?

A: The AMD Radeon Vega 10 Mobile uses system shared memory, with size, type, bus width, and bandwidth all dependent on the host system. The NVIDIA Quadro K620M has a dedicated 2 GB DDR3 frame buffer on a 64-bit bus with 16.02 GB/s bandwidth.

Q: Which GPU supports newer graphics APIs?

A: The AMD Radeon Vega 10 Mobile supports DirectX 12 at feature level 12_1, while the NVIDIA Quadro K620M supports DirectX 12 at feature level 11_0. Both support OpenGL 4.6. The NVIDIA part supports Vulkan 1.4, while the AMD part supports Vulkan 1.3.

Q: How do their shading resources compare?

A: The AMD Radeon Vega 10 Mobile has 640 shading units, 40 TMUs, and 8 ROPs. The NVIDIA Quadro K620M has 384 shading units, 16 TMUs, and 8 ROPs. The AMD part has 256 more shading units and 24 more TMUs.

Q: What are their production statuses and release dates?

A: Both GPUs are end-of-life. The AMD Radeon Vega 10 Mobile was released on 2019-01-07, and the NVIDIA Quadro K620M was released on 2015-02-28. The AMD part's predecessor is GCN 3.0 IGP and its successor is Navi II IGP. The NVIDIA part's predecessor is Quadro Fermi-M and its successor is Quadro Maxwell-M.

Head-to-Head Benchmarks

The only recorded head-to-head benchmark is Geekbench OpenCL, where the AMD Radeon Vega 10 Mobile scores 6476 against the NVIDIA Quadro K620M's 5957. This is an 8.7% delta in favor of AMD. Relative to the broader database, the AMD part sits at the 37th percentile, slightly below its nearest rival the NVIDIA Quadro M5000M (6481, delta of -0.1%), and just above the NVIDIA RTX PRO 5000 72 GB Blackwell (6407, delta of 1.1%). The NVIDIA Quadro K620M sits at the 34th percentile, with its nearest rival the AMD Radeon HD 8730M (5955, delta of 0%) and the Intel UHD Graphics 730 (5929, delta of 0.5%) just behind.

The AMD part's FP32 throughput of 1.665 TFLOPS is 1.93 times that of the NVIDIA part's 863.2 GFLOPS, meaning the AMD GPU delivers nearly double the single-precision compute. The texture rate difference is even larger: 52.04 GTexel/s versus 17.98 GTexel/s, a 2.89x advantage for AMD. Pixel rates are closer, with the AMD part at 10.41 GPixel/s versus 8.992 GPixel/s for NVIDIA, a 15.8% advantage. The AMD part also has a higher boost clock of 1301 MHz versus 1124 MHz, though the NVIDIA part has a much higher base clock of 1029 MHz versus 300 MHz.

The NVIDIA Quadro K620M's advantages are in memory architecture and API support. Its dedicated 2 GB DDR3 memory with 16.02 GB/s bandwidth provides a fixed resource, while the AMD part's bandwidth is system dependent. The NVIDIA part also supports Vulkan 1.4, a newer version than the AMD part's Vulkan 1.3. However, these advantages do not translate into any benchmark win in the recorded data.

Specification Differences

The table below summarizes only the fields where the two GPUs differ, based on the recorded specifications.

| Specification | AMD Radeon Vega 10 Mobile | NVIDIA Quadro K620M |

|---|---|---|

| Manufacturer | AMD | NVIDIA |

| Chip | Raven-M | GM108S |

| Architecture | GCN 5.0 | Maxwell |

| Generation | Vega IGP (Raven Ridge-M) | Quadro Kepler-M (Kx200M) |

| Process Node | 14 nm | 28 nm |

| Foundry | GlobalFoundries | TSMC |

| Transistors | 4,940 million | 1,020 million |

| Die Size | 210 mm² | 77 mm² |

| Transistor Density | 23.5M / mm² | 13.2M / mm² |

| Base Clock | 300 MHz | 1029 MHz |

| Boost Clock | 1301 MHz | 1124 MHz |

| Memory Clock | System Shared | 1001 MHz, 2 Gbps effective |

| Memory Size | System Shared | 2 GB |

| Memory Type | System Shared | DDR3 |

| Memory Bus Width | System Shared | 64 bit |

| Memory Bandwidth | System Dependent | 16.02 GB/s |

| Shading Units | 640 | 384 |

| TMUs | 40 | 16 |

| Pixel Rate | 10.41 GPixel/s | 8.992 GPixel/s |

| Texture Rate | 52.04 GTexel/s | 17.98 GTexel/s |

| FP32 | 1.665 TFLOPS | 863.2 GFLOPS |

| FP16 | 3.331 TFLOPS (2:1) | None |

| TDP | 10 W | 30 W |

| Slot Width | IGP | MXM Module |

| Bus Interface | IGP | MXM-A (3.0) |

| DirectX | 12 (12_1) | 12 (11_0) |

| Vulkan | 1.3 | 1.4 |

| Release Date | 2019-01-07 | 2015-02-28 |

| Predecessor | GCN 3.0 IGP | Quadro Fermi-M |

| Successor | Navi II IGP | Quadro Maxwell-M |

| Geekbench OpenCL Score | 6476 | 5957 |

| Percentile vs All GPUs | 37 | 34 |

Both GPUs share the same ROP count (8), the same display outputs (Portable Device Dependent), no RT cores, no tensor cores, no power connectors, no suggested PSU, no launch MSRP, no recorded dimensions, and both are end-of-life. Their average benchmark scores match their single Geekbench OpenCL scores, as that is the only recorded benchmark for each.