NVIDIA Quadro P620

The NVIDIA Quadro P620 is a professional-grade graphics card built on the Pascal architecture, utilizing the GP107 chip manufactured on Samsung's 14 nm process. Released in early 2018, this end-of-life product occupies a specific niche in the workstation market, and the benchmark data reveals a complex positioning story that demands careful analysis.

Memory Subsystem

The Quadro P620 ships with 2 GB of GDDR5 memory across a 128-bit bus interface. This configuration produces a memory bandwidth of 80.13 GB/s, a figure that directly constrains the card's capabilities at higher resolutions. The memory operates at 1252 MHz, translating to 5 Gbps effective speed, which is typical for this class of professional card.

The 2 GB capacity is the most significant limitation in this subsystem. For modern professional workloads, particularly those involving large textures or complex 3D scenes, this capacity will be exhausted quickly. The 128-bit bus width, while adequate for the card's compute capabilities, means that memory-intensive operations will experience bottlenecks. The 80.13 GB/s bandwidth is sufficient for 1080p workloads but becomes a critical constraint when pushing toward 1440p or 4K resolutions, where texture data demands far exceed what this subsystem can deliver.

The FP16 performance of 21.66 GFLOPS, operating at a 1:64 ratio to FP32, indicates this card was never designed for compute-heavy tasks that leverage reduced precision. The pixel rate of 21.66 GPixel/s and texture rate of 43.33 GTexel/s are modest figures that align with the card's entry-level positioning. For high-resolution rendering, the data suggests the memory subsystem will be the primary limiting factor, not the compute units.

Who Should Consider It

The Quadro P620's benchmark scores paint a clear picture of its intended use case. With a Passmark G3D score of 3674 and a percentile rank of 19 among all GPUs, this card sits firmly in entry-level territory. The data indicates it is suitable for 1080p professional applications that do not require extensive texture memory or complex shading.

Users working with 2D CAD applications, basic 3D modeling, or office productivity suites that leverage GPU acceleration will find the P620 adequate. The 4x mini-DisplayPort 1.4a outputs suggest multi-monitor setups are a primary use case, and the card can drive multiple 1080p displays without issue. However, for 1440p gaming or professional rendering at higher resolutions, the memory subsystem and compute performance will prove inadequate.

The Geekbench OpenCL score of 12064 and Vulkan score of 11018 indicate reasonable compute capability for basic tasks, but the Passmark GPU Compute score of 1561 reveals limitations in parallel processing workloads. The data suggests this card is best suited for legacy professional applications, basic visualization, and multi-display productivity rather than demanding creative workloads.

Benchmark Performance

The benchmark results show a card that performs consistently across different test suites but never excels. The Passmark DirectX 11 score of 28 is notably higher than the DirectX 10 score of 14 and DirectX 12 score of 15, suggesting the card's architecture is better optimized for older API versions. The DirectX 9 score of 69 is by far the strongest result, indicating this card's Pascal architecture handles legacy workloads with relative ease.

The average benchmark score of 3221 places the P620 within a tight competitive cluster. The data shows a 0.2% advantage over the NVIDIA GeForce GT 750M and a 0.4% lead over the Intel HD Graphics 4600. These deltas are negligible in real-world terms, meaning the P620 offers essentially identical performance to these rivals. Against the NVIDIA GeForce 920M, the P620 trails by 1.4%, and it lags the NVIDIA GeForce GT 740 by 2.5%.

The performance parity with integrated graphics from the HD 4600 generation is particularly telling. This means the P620's dedicated memory and professional drivers do not translate into meaningful performance advantages over well-implemented integrated solutions in the same era. The 2D performance score of 545 suggests the card handles basic desktop acceleration competently, but this is expected of any modern GPU.

How It Compares

NVIDIA GeForce GT 750M: The P620 leads this mobile GPU by a razor-thin 0.2% margin. Both cards deliver virtually identical average scores of approximately 3215-3221. This parity suggests the P620's professional drivers do not unlock additional performance headroom over a mobile consumer chip of the same generation.

Intel HD Graphics 4600: The P620's 0.4% advantage over this integrated solution is statistically insignificant. The HD 4600's average score of 3208 demonstrates that even mature integrated graphics can match the P620's performance in general benchmarks. This comparison raises questions about the value proposition of dedicated entry-level professional cards.

NVIDIA GeForce 920M: The P620 trails this mobile consumer card by 1.4%, with the 920M scoring 3266. This deficit, while small, indicates the P620 is not even the fastest entry-level NVIDIA solution available. The performance gap, though minor, is consistent across most benchmark scenarios.

NVIDIA GeForce GT 740: The largest delta appears against this desktop card, with the GT 740 scoring 3304, a 2.5% advantage. This margin, while still modest, shows the P620 at the bottom of its competitive cluster. The GT 740's higher score suggests desktop-oriented designs have a slight edge over this professional variant.

Power and Cooling

The Quadro P620 has a TDP of just 40 W, making it an exceptionally power-efficient card. NVIDIA recommends a 200 W power supply, which is remarkably low and allows integration into nearly any system. The card requires no power connectors, drawing all its power from the PCIe 3.0 x16 slot. This makes it ideal for upgrading pre-built systems with limited power delivery capabilities.

The single-slot design, measuring 145 mm in length and 69 mm in height, ensures compatibility with compact chassis. The lack of power connectors simplifies installation, and the low TDP means cooling requirements are minimal. The 14 nm process node contributes to this efficiency, with the 3,300 million transistors packed into a 132 mm² die. This power profile suggests the card can operate quietly in workstation environments where acoustic performance matters.

The 40 W TDP is a significant advantage over many consumer cards, enabling silent operation and reduced system heat output. The data indicates this card is an excellent choice for dense workstation configurations or systems where power consumption is a critical constraint.

FAQ

Q: Does the Quadro P620 support DirectX 12?

A: Yes, the card supports DirectX 12 with feature level 12_1, though benchmark scores suggest its DirectX 12 performance is modest compared to legacy APIs.

Q: What display outputs are available on this card?

A: The P620 provides 4x mini-DisplayPort 1.4a outputs, supporting multi-monitor professional setups.

Q: How does the P620 compare to the GeForce GT 750M?

A: The P620 holds a 0.2% performance advantage, with average scores of 3221 versus 3215, making them effectively equivalent in performance.

Q: What is the memory bandwidth of the Quadro P620?

A: The card features 80.13 GB/s of memory bandwidth from 2 GB of GDDR5 memory on a 128-bit bus.

Q: Is the Quadro P620 suitable for high-resolution gaming?

A: The data indicates the 2 GB memory capacity and modest compute performance limit it to 1080p workloads; higher resolutions will strain the memory subsystem.

Q: What power supply is recommended for this card?

A: NVIDIA recommends a 200 W PSU, and the card requires no external power connectors, drawing power solely from the PCIe slot.

Ray Tracing and Feature Set

The Quadro P620 does not include dedicated ray tracing cores or tensor cores, as evidenced by the null values in the specifications. This is consistent with its Pascal architecture, which predates NVIDIA's RTX technology. The card supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, providing broad API compatibility for professional applications.

The lack of hardware ray tracing means the P620 cannot accelerate modern ray-traced workloads. Any ray tracing would need to be handled through compute shaders, which the FP32 performance of 1,386.5 GFLOPS suggests would be impractically slow. The 512 shading units and 16 ROPs provide basic rasterization capabilities, but the card is clearly designed for traditional rendering pipelines.

The Vulkan 1.4 support is notable, as it allows access to modern graphics features when used with compatible applications. However, the absence of tensor cores eliminates any AI-accelerated features, such as DLSS. The 14 nm process and 3,300 million transistors indicate this is a mature design optimized for efficiency rather than current-generation features. The card's feature set is adequate for basic professional visualization but lacks the specialized hardware found in newer workstation GPUs.