AMD Radeon Vega 6 Embedded

Memory Subsystem — VRAM size/type, bus width, bandwidth and what it means for high resolutions

The AMD Radeon Vega 6 Embedded is an integrated graphics processor (IGP) built on the Raven chip, and its memory subsystem is entirely dependent on the host system. The VRAM size is "System Shared," the memory type is "System Shared," and the bus width is likewise "System Shared," meaning there is no dedicated video memory on the graphics die. Instead, the Vega 6 Embedded borrows from the system’s main memory pool, with bandwidth listed as "System Dependent." This design implies that the effective memory performance will vary significantly based on the platform’s RAM speed, channel configuration, and capacity allocation.

For high-resolution workloads, the lack of dedicated VRAM is a limiting factor. At 1080p or 1440p, the system must allocate a portion of the main memory for frame buffering, texture storage, and geometry data, which can reduce the memory available to the CPU. Benchmark results indicate that the pixel rate is 10.24 GPixel/s, and the texture rate is 30.72 GTexel/s, both of which are modest figures for an IGP. At higher resolutions, the memory bandwidth bottleneck becomes more pronounced because the shared bus must handle both graphics and system traffic concurrently. The data shows that the Vega 6 Embedded sits at the 50th percentile among all GPUs, suggesting a mid-pack position, but its shared-memory architecture means that users with fast dual-channel system memory will see better performance than those with slower single-channel configurations. There is no dedicated bandwidth figure to quote, so the practical takeaway is that the memory subsystem is a hybrid solution—capable for light to moderate 1080p gaming or desktop use, but unlikely to sustain high-resolution textures or multi-monitor setups without noticeable performance degradation.

Power and Cooling

The thermal design power (TDP) for the AMD Radeon Vega 6 Embedded is 15 W. This is a low-power integrated solution, which means it does not require a dedicated cooling solution beyond what the host system provides. The slot width is listed as "IGP," and the bus interface is "IGP," confirming that this is not a discrete add-in card but rather a graphics core embedded within the processor package. There is no suggested PSU rating provided in the data, and no power connector requirements are listed, which is consistent with an IGP that draws power from the motherboard’s CPU power delivery system rather than a separate PCIe power cable. The power connectors field is null, indicating that no external power is needed. For system builders, this means the Vega 6 Embedded places no additional burden on the power supply unit; a standard PSU capable of supporting the host CPU will suffice. The 15 W TDP also implies that the thermal envelope is small, allowing for passive or low-profile cooling in embedded or compact systems. The process node is 14 nm, manufactured by GlobalFoundries, with a die size of 210 mm² and 4,940 million transistors, which is a dense but power-efficient design. The transistor density is 23.5 million per square millimeter. This combination of low TDP and integrated form factor makes the Vega 6 Embedded suitable for thin-and-light laptops, mini PCs, or embedded industrial systems where power draw and heat dissipation are critical constraints.

Benchmark Performance

The FACT PACK lists an average benchmark score of 0, with a percentile rank of 50 among all GPUs. However, the nearestRivals array is empty, so there are no direct comparative scores or deltaPct values to analyze against specific competing products. The benchmark data is sparse, which makes it challenging to provide exact percentage deltas. What can be stated is that the Vega 6 Embedded achieves a FP32 performance of 983.0 GFLOPS and a FP16 performance of 1.966 TFLOPS (at a 2:1 ratio). These raw compute figures place it in the lower-to-mid tier of integrated graphics solutions. The pixel rate of 10.24 GPixel/s and texture rate of 30.72 GTexel/s are consistent with a GPU that has 384 shading units, 24 texture mapping units, and 8 raster operation units. The boost clock is 1280 MHz, with a base clock of 300 MHz, indicating a wide dynamic range that helps conserve power when idle. Without rival data, the interpretation must rely on the percentile rank: the 50th percentile suggests that the Vega 6 Embedded performs better than half of all GPUs in the database, but this includes a vast range of integrated and discrete parts. In practical terms, the benchmark results indicate that this IGP is suited for 1080p gaming at low-to-medium settings in older titles, or for light productivity tasks such as video playback and office applications. It is not designed for high-refresh-rate gaming or content creation workloads that demand sustained FP32 throughput.

How It Compares

The nearestRivals field is empty, so there are no direct competitor scores, names, or deltaPct values to reference. Based on the available FACT PACK data, the Vega 6 Embedded’s position must be inferred from its own specifications and percentile rank. Its predecessor is listed as "GCN 3.0 IGP," and its successor is "Vega II IGP," which provides a generational context: the Vega 6 Embedded belongs to the Vega IGP (Raven Ridge) generation, built on the GCN 5.0 architecture. The 14 nm process node is typical for its era, and the 15 W TDP is low compared to discrete GPUs. The lack of RT cores and tensor cores means it does not compete with modern ray tracing-capable graphics cards. In the absence of rival data, the analysis must note that the 50th percentile ranking places it in the middle of the database, but this is a broad measure. The FP32 throughput of 983.0 GFLOPS is roughly half of what a typical mid-range discrete GPU from the same period would offer, but the IGP’s advantage is its minimal power draw and integrated form factor. The data shows no benchmark scores, so any comparison to specific rivals is impossible; the reader should interpret the Vega 6 Embedded as a baseline integrated solution rather than a performance-oriented part.

Ray Tracing and Feature Set

The Vega 6 Embedded does not include dedicated ray tracing cores or tensor cores; both fields are null in the FACT PACK. This means that any ray tracing workloads would have to be handled via compute shaders or other software methods, which would be inefficient given the modest FP32 performance of 983.0 GFLOPS. The architecture is GCN 5.0, which predates the dedicated hardware found in later AMD RDNA or CDNA products. The API support includes DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3. The DirectX 12_1 feature level indicates support for conservative rasterization, rasterizer-ordered views, and other advanced rendering features, but not the full DirectX 12 Ultimate feature set that includes ray tracing and mesh shaders. Vulkan 1.3 support provides access to modern GPU-driven rendering techniques, which can help mitigate the lack of dedicated RT hardware by allowing developers to optimize compute-based effects. The display outputs are "Motherboard Dependent," meaning that the actual video outputs (HDMI, DisplayPort, etc.) are determined by the motherboard or system manufacturer, not the GPU itself. The shading units are 384, with 24 TMUs and 8 ROPs, which is a limited configuration for advanced effects. The FP16 performance of 1.966 TFLOPS is double the FP32 rate, indicating support for half-precision compute, which can be used selectively in shaders to improve throughput. However, without tensor cores, AI-accelerated features like DLSS are not available; any upscaling would rely on the API-level features in DirectX or Vulkan.

FAQ

Q: What is the memory size of the AMD Radeon Vega 6 Embedded?

A: The memory size is "System Shared," meaning it uses the host system’s main RAM rather than dedicated VRAM.

Q: Does the Vega 6 Embedded support ray tracing?

A: No, the RT cores field is null, and the GPU does not include dedicated ray tracing hardware.

Q: What is the TDP of the Vega 6 Embedded?

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

Q: What DirectX version does the Vega 6 Embedded support?

A: It supports DirectX 12 (12_1), along with OpenGL 4.6 and Vulkan 1.3.

Q: What is the FP32 performance of this GPU?

A: The FP32 performance is 983.0 GFLOPS, with FP16 at 1.966 TFLOPS (2:1 ratio).

Q: Does the Vega 6 Embedded require an external power connector?

A: No, the power connectors field is null, and the bus interface is IGP, so it draws power from the motherboard.

Who Should Consider It

The AMD Radeon Vega 6 Embedded is best suited for users who require a low-power integrated graphics solution for everyday computing, basic multimedia, or light 1080p gaming. The 15 W TDP makes it ideal for compact embedded systems, thin laptops, or fanless designs where power efficiency is paramount. Given the 50th percentile ranking and the compute figures of 983.0 GFLOPS FP32, this GPU can handle older or less demanding games at 1080p with low-to-medium settings, but it will struggle with modern AAA titles at high resolutions or detail levels. The memory being system-shared means that performance is highly dependent on the host platform’s RAM speed and configuration; users with fast dual-channel memory will get better results than those with single-channel modules. The lack of ray tracing and tensor cores means that this is not a choice for gamers seeking hardware-accelerated ray tracing or AI upscaling features. For office productivity, video streaming, or light photo editing, the Vega 6 Embedded is more than adequate. It should not be considered for 1440p or 4K gaming, as the pixel rate of 10.24 GPixel/s and texture rate of 30.72 GTexel/s will bottleneck at those resolutions. The production status is end-of-life, so it is primarily relevant for existing systems or budget-oriented builds that prioritize energy efficiency over raw performance. The 14 nm process and 4,940 million transistors indicate a mature design, but the architecture lacks modern features. In summary, consider the Vega 6 Embedded if you need a competent, low-power IGP for basic tasks and occasional light gaming, but not for demanding workloads or future-proofing.