Intel UHD Graphics 620 Mobile

Intel UHD Graphics 620 Mobile is an integrated graphics processor built on Intel’s Generation 9.5 architecture, specifically the Whiskey Lake GT2 chip fabricated on a 14 nm process. It operates with a base clock of 300 MHz and a boost clock of 1000 MHz, featuring 192 shading units, 24 texture mapping units, and 3 raster output pipelines. The GPU relies entirely on system memory, with its size, type, bus width, and bandwidth all marked as “System Shared” or “System Dependent,” making its performance heavily contingent on the host system’s RAM configuration. As an end-of-life product released in late August 2018, it occupies the 50th percentile among all GPUs in the database, with an average benchmark score of 0 and no nearest rivals listed for direct comparison. This places it in a peculiar position: it is neither a standout performer nor a complete outlier, but rather a baseline reference point for integrated graphics from its era.

How It Compares

Without any nearest rivals provided in the data, the comparison must rely on the GPU’s own specifications and general positioning within the 50th percentile. The lack of rival entries means there are no direct score deltas to cite, so the analysis focuses on what the hardware itself indicates. Against discrete GPUs from its generation, the UHD 620’s 384.0 GFLOPS of FP32 compute and 3.000 GPixel/s pixel rate are modest figures, suggesting it trails even entry-level dedicated cards by a wide margin. The 24.00 GTexel/s texture rate further underscores its limited throughput, which would make it unsuitable for anything beyond light gaming or basic multimedia tasks.

In the context of other integrated solutions, the 620’s 192 shading units are a typical count for Intel’s HD Graphics lineup of that period, but its boost clock of 1000 MHz is conservative. The data shows no specific competitor names or percentages, so the practical takeaway is that this GPU sits at the bottom of the performance hierarchy, roughly where one would expect an IGP from 2018 to land. Its 50th percentile ranking is more a reflection of the database’s inclusion of many older and weaker parts than a sign of competence; in real-world terms, it would struggle to maintain playable frame rates in any modern 3D title, even at the lowest settings and resolutions.

Who Should Consider It

Given the benchmark data, the Intel UHD Graphics 620 is not aimed at gamers or users requiring any level of 3D acceleration. The 384.0 GFLOPS FP32 throughput and 3.000 GPixel/s pixel rate are sufficient only for 2D desktop environments, video playback, and extremely light productivity workloads like document editing or spreadsheet work. At 1080p resolution, the GPU would be hard-pressed to run even older or esports titles at playable frame rates, as the texture rate of 24.00 GTexel/s and the shared memory bandwidth create severe bottlenecks. Users should consider this GPU exclusively for basic office tasks, web browsing, and streaming video, where its capabilities are adequate.

For those who might attempt light gaming, the realistic ceiling is 720p with the lowest graphical presets and titles from before 2015 or highly optimized 2D games. The 3 ROPs are a critical limitation, capping fill-rate-dependent operations and making any anti-aliasing or high-resolution rendering impractical. The system-shared memory further compounds the issue, as the GPU must compete with the CPU for bandwidth, and performance will vary dramatically depending on whether the system is equipped with single-channel or dual-channel RAM. In short, this is a chip for users who know they will never game, not for those seeking a stopgap solution.

Benchmark Performance

The absence of benchmark scores and nearestRivals in the data prevents any quantitative comparison against specific competitors. The average benchmark score of 0 and the empty nearestRivals array mean there are no exact percentage deltas to report. What the numbers do show is the raw theoretical limits: 384.0 GFLOPS FP32, 768.0 GFLOPS FP16 (at a 2:1 ratio), 24.00 GTexel/s, and 3.000 GPixel/s. These figures, when viewed together, paint a picture of a GPU that is roughly one-tenth to one-twentieth the compute capability of even modest discrete GPUs from the same era, though no specific rival names or scores are available to cite.

The boost clock of 1000 MHz is low by modern standards, but it is the 3 ROPs that truly hamper performance. With such a low pixel throughput, any resolution above 1366x768 will likely cause frame rates to collapse in 3D applications. The 192 shading units, while not negligible, are paired with a texture rate that limits complex scene rendering. In practical terms, the benchmark data suggests this GPU is best described as a video decoder with basic 2D acceleration, rather than a 3D performer. The 50th percentile ranking is misleading; it reflects the long tail of very old GPUs in the database, not any meaningful capability relative to contemporary hardware.

FAQ

Q: What is the maximum memory bandwidth of the Intel UHD Graphics 620?

A: The memory bandwidth is listed as “System Dependent,” meaning it varies based on the host system’s RAM configuration and cannot be specified as a fixed number.

Q: Does the Intel UHD Graphics 620 support DirectX 12?

A: Yes, it supports DirectX 12 with feature level 12_1, along with OpenGL 4.6 and Vulkan 1.3.

Q: What is the thermal design power (TDP) of this GPU?

A: The TDP is 15 W, which is typical for an integrated graphics processor.

Q: What is the pixel fill rate of the Intel UHD Graphics 620?

A: The pixel rate is 3.000 GPixel/s, which is very low and limits high-resolution rendering.

Q: Is the Intel UHD Graphics 620 still in production?

A: No, it is marked as end-of-life in the database, with a release date of 2018-08-27.

Q: How much VRAM does the Intel UHD Graphics 620 have?

A: It has no dedicated VRAM; the memory size is “System Shared,” meaning it uses a portion of the system’s main RAM.

Memory Subsystem

The memory subsystem of the Intel UHD Graphics 620 is entirely dependent on the system’s main memory, with the size, type, and bus width all listed as “System Shared.” This is a critical limitation because the GPU has no dedicated VRAM to draw upon, and its bandwidth is marked as “System Dependent.” In practice, this means performance scales directly with the speed and channel configuration of the host system’s RAM; dual-channel memory can roughly double available bandwidth compared to single-channel, but even the best-case scenario leaves the GPU severely constrained.

For high resolutions like 1440p or 4K, the shared memory architecture is a fatal flaw. The system must allocate a portion of RAM for the GPU, reducing available memory for the CPU and potentially causing stuttering in multitasking scenarios. The 3.000 GPixel/s pixel rate further limits the ability to drive high-resolution displays in 3D applications, as the ROPs simply cannot fill frames fast enough. Even for 1080p video playback, the GPU relies on fixed-function decode blocks rather than its shading units, which is the only reason it remains usable for media consumption. The lack of a dedicated bus width number underscores that this is not a performance-oriented part; it is a convenience feature integrated into the CPU.

Power and Cooling

The Intel UHD Graphics 620 has a TDP of 15 W, which is the only power-related figure provided in the data. This low power draw is a double-edged sword: it means the GPU generates minimal heat and requires no dedicated cooling solution, as it is an IGP integrated into the processor package. The slot width is listed as “IGP,” confirming that it is not a discrete card and occupies no expansion slot. There is no suggested PSU rating, and power connectors are listed as null, which is expected for an integrated GPU that draws power from the motherboard’s CPU power delivery.

From a cooling perspective, the 15 W TDP is well within the capabilities of even the most basic laptop or small-form-factor PC cooling solutions. The GPU shares its thermal envelope with the CPU, so sustained loads on both components could lead to thermal throttling in thin-and-light laptops. However, for the workloads this GPU is suited for, web browsing, office applications, video streaming, the power draw will rarely approach the 15 W limit. Builders should note that there are no additional power connector requirements, and the Ring Bus interface means it communicates with the rest of the system through the CPU’s internal interconnect, further simplifying integration. In essence, this is a zero-effort component from a power and cooling standpoint, which is its primary virtue.