AMD Radeon Vega 6 Mobile

The AMD Radeon Vega 6 Mobile is an integrated graphics processor from the Raven Ridge-M generation, built on the GCN 5.0 architecture and fabricated by GlobalFoundries on a 14 nm process. The die packs 4,940 million transistors across 210 mm², for a transistor density of 23.5M per square millimeter. The GPU operates with a base clock of 300 MHz and a boost clock of 1101 MHz. It was released on January 7, 2018, and is now marked as end-of-life. As an IGP, it relies on system-shared memory, and its display outputs are portable-device dependent.

Benchmark Performance

The Vega 6 Mobile's compute configuration consists of 384 shading units, 24 texture mapping units, and 8 raster operation units. At the 1101 MHz boost clock, FP32 throughput reaches 845.6 GFLOPS, while FP16 performance scales to 1.691 TFLOPS via a 2:1 ratio. The pixel fill rate is 8.808 GPixel/s, and the texture fill rate is 26.42 GTexel/s. These raw figures place the part at the 50th percentile among all GPUs in the database, a dead-center position that is unusual for an integrated part.

The benchmark array for this GPU is empty, and the average benchmark score is 0, meaning no measured performance data has been recorded. The analysis therefore rests on the specification-derived metrics. The 845.6 GFLOPS of FP32 compute is the defining performance figure. With 384 shaders operating at up to 1101 MHz, the GPU can handle shader-bound workloads that would overwhelm lower-end IGPs, but it is not in the territory of discrete graphics. The 8 ROPs cap the pixel throughput at 8.808 GPixel/s, which becomes the limiting factor in fill-rate-heavy scenes such as high-resolution post-processing or heavy overdraw.

The texture pipeline, with 24 TMUs delivering 26.42 GTexel/s, outpaces the pixel rate by a wide margin, indicating that texturing is not the bottleneck in most scenarios. The 2:1 FP16 ratio is a notable architectural feature: the 1.691 TFLOPS half-precision figure allows compute shaders that use FP16 to run at double the FP32 rate, though the practical benefit depends on application support. API support spans DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.3, covering the modern graphics API landscape.

The 50th percentile ranking is the single most informative data point. In a database that spans entry-level IGPs to flagship discrete cards, the Vega 6 Mobile sits exactly at the median. That means half of all GPUs are faster and half are slower. For an integrated part, this is a strong showing — many IGPs sit lower in the distribution. However, the percentile is derived from the specification-based ranking, not from measured benchmarks, so real-world performance will vary with the host platform.

Power and Cooling

The Vega 6 Mobile carries a TDP of 15 W, a modest figure that reflects its integrated design. The slot width is listed as IGP, meaning it does not occupy an expansion slot, and it requires no power connectors — power is delivered through the host socket. There is no suggested PSU listed, consistent with a part that never needs a standalone power supply. The 14 nm GlobalFoundries process is responsible for keeping power draw at 15 W while housing 4,940 million transistors on a 210 mm² die.

The thermal design is straightforward: the 15 W TDP can be managed by the cooling solution already present in the host laptop or portable device. No aftermarket cooler is needed, and the database lists none. The transistor density of 23.5M per square millimeter indicates a moderately packed die, but at 15 W, the heat density remains within what a typical thin-and-light chassis can dissipate. The base clock of 300 MHz is very low, suggesting deep idle states that conserve power, while the 1101 MHz boost clock provides the performance headroom when workloads demand it. The wide gap between base and boost is characteristic of a power-constrained integrated part that ramps up only as thermal and power budgets allow.

The absence of a suggested PSU is notable. For an IGP, the power delivery is entirely handled by the system's existing voltage regulator circuitry, and the 15 W figure is the graphics portion's total board power. This makes the Vega 6 Mobile suitable for systems where a discrete GPU would be impractical due to space or thermal constraints.

Who Should Consider It

The Vega 6 Mobile is an integrated part for portable devices, as its portable-device-dependent display outputs indicate. The 50th percentile ranking means it offers mid-pack performance relative to the entire GPU database, which is a meaningful reference for users who want to know what their laptop can do. For everyday tasks — desktop compositing, video playback, and light 2D work — the 845.6 GFLOPS of FP32 compute is ample. For 3D workloads, the 8.808 GPixel/s pixel rate and 26.42 GTexel/s texture rate define the practical ceiling.

At high display resolutions, the system-shared memory becomes the primary constraint. The GPU has no dedicated VRAM, so the frame buffer competes with the operating system for the same memory pool. The 384 shading units are capable of producing the 845.6 GFLOPS of compute, but the memory subsystem can bottleneck before the compute units are fully utilized. The 2:1 FP16 ratio of 1.691 TFLOPS offers headroom for compute tasks that support half precision, but games typically rely on FP32.

The end-of-life status means this is not a part for new system builds. It is relevant for owners of existing Raven Ridge-M platforms who want to understand the capabilities of their hardware. The 15 W TDP makes it well-suited to thin-and-light laptops where battery life and thermals are priorities. Users who plan to run modern 3D games at high detail settings will find the pixel rate and shading throughput limiting, but those who stick to lighter titles or lower detail levels will find it adequate. The DirectX 12 (12_1) and Vulkan 1.3 support ensure compatibility with modern graphics APIs.

How It Compares

The nearestRivals list for the Vega 6 Mobile is empty, so there are no direct rival comparisons with scores or delta percentages available in the database. The comparison must instead be drawn from its generational position. The predecessor is the GCN 3.0 IGP, and the successor is the Navi II IGP. The Vega 6 Mobile's GCN 5.0 architecture represents a generational step forward from the older GCN 3.0 IGP, with the architectural improvements that GCN 5.0 introduced — though the database does not quantify the difference.

The 50th percentile standing is the only quantitative comparison available. In a database of all GPUs, half are faster and half are slower. For an integrated part, that is a respectable position; many IGPs sit lower in the distribution. The fact that the successor is the Navi II IGP indicates that AMD has moved to a newer architecture family, and the end-of-life status confirms that the Vega 6 Mobile is no longer in production. The 14 nm process node is a generation behind the more advanced nodes used by newer parts, but the database does not list those nodes for comparison.

The empty rivals list also means there are no deltaPct values to cite. The comparison to the predecessor and successor is qualitative: the GCN 3.0 IGP would be expected to deliver lower performance given the older architecture, while the Navi II IGP would be expected to be faster — but no numbers back those expectations in the database. The 50th percentile is the single data point that anchors this part's position in the overall GPU landscape.

Memory Subsystem

The Vega 6 Mobile uses system-shared memory for its frame buffer. The memory size, type, and bus width are all listed as "System Shared," and the bandwidth is "System Dependent." This is the defining characteristic of an integrated GPU: there is no dedicated VRAM, no dedicated memory bus, and no fixed bandwidth figure. The GPU borrows from the host system's main memory, and performance is entirely dependent on the platform's memory configuration.

The absence of a dedicated bus width means the memory interface is whatever the host CPU provides. In practice, the effective bandwidth is limited by the system's memory channels and speed, but the database does not specify those parameters. The "System Dependent" bandwidth designation is a caveat for anyone evaluating this GPU: the same Vega 6 Mobile can perform at very different levels depending on whether it is paired with fast dual-channel memory or slower single-channel memory. The 8.808 GPixel/s pixel rate and 26.42 GTexel/s texture rate are compute-side figures that do not account for memory bandwidth constraints.

For high-resolution workloads, the system-shared memory is a significant limitation. At higher resolutions, the frame buffer footprint grows, and the shared memory must accommodate both the operating system and the graphics data. The 845.6 GFLOPS of FP32 compute is sufficient for many tasks, but the memory subsystem can become the bottleneck before the compute units do. The 2:1 FP16 ratio of 1.691 TFLOPS does not help with memory bandwidth; it only accelerates the math portion. The practical upshot is that the Vega 6 Mobile's performance is heavily influenced by the host platform, and any benchmark result must be understood in the context of the system it ran on.