Intel Iris Xe Graphics G4 48EU Mobile

The Intel Iris Xe Graphics G4 48EU Mobile is an integrated graphics processor built on Intel's Generation 12.1 architecture, manufactured on a 10 nm process with a die size of 146 mm². Released in September 2020, this part has reached end-of-life status. It combines 384 shading units, 24 texture mapping units, and 12 raster output units, with base and boost clocks of 300 MHz and 1100 MHz respectively. The GPU occupies the 50th percentile of all GPUs in the benchmark database, placing it squarely in the middle of the performance distribution — a position that tells most of the story before a single benchmark is examined.

Benchmark Performance

The Iris Xe G4 48EU delivers a peak FP32 throughput of 844.8 GFLOPS, a figure that reflects its 384 shading units operating at the 1100 MHz boost clock. This places the chip in a position where it can handle light gaming and everyday graphical workloads, but the 50th percentile ranking confirms that it sits exactly at the median of all GPUs tracked in the database. In practical terms, half of all GPUs outperform it, and half perform worse. That is a remarkably central position — not a bottom-feeder, but nowhere near a performance leader.

The FP16 rate of 1.690 TFLOPS, achieved at a 2:1 ratio relative to FP32, indicates that the architecture can accelerate half-precision workloads at roughly double the rate of single-precision. This is relevant for compute tasks that leverage FP16, such as certain image processing and scientific workloads, though the lack of dedicated tensor cores means the GPU cannot accelerate AI inference in the way that dedicated hardware would. The 2:1 ratio is a sign of a modern architecture that at least acknowledges the growing importance of reduced-precision compute, even if it cannot fully exploit it.

Pixel throughput is rated at 13.20 GPixel/s, derived from the 12 ROPs at the boost clock. Texture throughput reaches 26.40 GTexel/s, driven by 24 TMUs. These figures suggest that the GPU is capable of handling 1080p gaming at low to medium settings in older titles, but modern games with heavy texture and pixel workloads will stress the chip's limited resources. The 50th percentile standing is consistent with a part that is positioned as an entry-level integrated solution rather than a discrete gaming GPU. The 844.8 GFLOPS figure, while modest by discrete standards, is enough for smooth desktop compositing, video decode, and casual gaming — but the data does not support expectations of high-refresh-rate or high-detail experiences.

Power and Cooling

The Iris Xe G4 48EU carries a TDP of just 15 W, making it one of the most power-efficient parts in the database. As an IGP, it shares the thermal and power budget of the host processor, and no dedicated power connectors are required. The slot width is listed as IGP, meaning it does not occupy an expansion slot — it is soldered onto the motherboard or integrated into the CPU package. This has profound implications for system design: there is no need for a separate cooling solution, no additional PCB area, and no power delivery circuitry beyond what the host platform already provides.

Because the GPU draws its power from the system's existing power delivery infrastructure, there is no suggested PSU rating in the data. The 15 W TDP is a fraction of what discrete GPUs demand, and the thermal output is correspondingly low. For a laptop or compact mobile system, this means cooling solutions can be minimal, and the chip can operate within the thermal envelope of a thin-and-light chassis. The 10 nm process node contributes to this efficiency, as does the modest 146 mm² die size, which keeps power density low and allows the GPU to share a heat pipe or vapor chamber with the CPU without issue. Users upgrading a system that already has this GPU should not need to change their power supply or cooling arrangement — the integrated nature of the part means it is already accounted for in the platform's design.

Ray Tracing and Feature Set

The Iris Xe G4 48EU has no dedicated ray tracing cores and no tensor cores, according to the specification data. This means hardware-accelerated ray tracing is not available on this part. However, the GPU does support DirectX 12 (12_1), which includes the feature level required for many modern rendering techniques, though not the DXR ray tracing API in a hardware-accelerated form. OpenGL 4.6 and Vulkan 1.4 are also supported, giving developers access to modern graphics APIs across multiple platforms. The API support is notably current for a 2020 part — Vulkan 1.4 in particular is a recent specification, and its presence here suggests that the driver stack has kept pace with modern standards even as the hardware has aged.

The absence of tensor cores is notable for AI-related workloads. While the FP16 throughput of 1.690 TFLOPS can be used for some compute tasks, the lack of dedicated tensor hardware means that machine learning inference and training workloads will rely on the general-purpose shader units. This is a significant limitation for users who expect to run AI-accelerated applications, such as DLSS-style upscaling or local LLM inference. The feature set is otherwise complete for traditional rasterization-based rendering, and the DirectX 12_1 support ensures compatibility with the latest Windows games that do not require ray tracing. For a 15 W integrated part, the absence of RT and tensor cores is expected — the transistor budget is simply too small to accommodate them alongside the 384 shading units.

How It Compares

The nearestRivals array in the database is empty, so there are no direct competitor scores or delta percentages to reference. The 50th percentile ranking, however, provides context: the Iris Xe G4 48EU is exactly at the median of all GPUs in the benchmark database. This means it outperforms roughly half of all GPUs tracked, which includes older integrated solutions and very low-end discrete parts, while being outperformed by the other half, which includes most discrete gaming GPUs and newer integrated solutions. The percentile is a blunt instrument, but it is a useful one — it tells a prospective buyer that this GPU will not embarrass itself in everyday tasks, but it will not impress anyone who has used a mid-range discrete card.

Without specific rival data, the comparison must be framed in terms of the chip's own specifications. The 844.8 GFLOPS FP32 throughput is the key indicator of its compute capacity. The 15 W TDP places it in the same power class as other integrated graphics solutions, but the 50th percentile shows that it is not a top performer even within that class. The 146 mm² die size and 10 nm process are typical for an integrated part of this generation, and the 2020 release date means it predates many newer integrated GPUs that have since surpassed it. The end-of-life production status further indicates that the part is no longer competitive with current offerings — it remains in existing systems, but new designs will use more recent silicon.

Memory Subsystem

The Iris Xe G4 48EU uses System Shared memory, meaning it has no dedicated VRAM. The memory type, bus width, and bandwidth are all listed as "System Shared" or "System Dependent," which indicates that the GPU accesses the same memory pool as the CPU. The bandwidth available to the GPU is therefore dependent on the system's memory configuration, including the number of channels, memory speed, and whether the system uses dual-channel or single-channel memory. This is a critical caveat for anyone evaluating this GPU: the same chip can perform noticeably differently depending on the host system's memory setup.

This shared memory architecture has significant implications for high-resolution gaming. At 1080p and above, the GPU must compete with the CPU for memory bandwidth, and the lack of dedicated VRAM means that textures and framebuffers are stored in system memory. The memory bandwidth being "System Dependent" means that a system with fast dual-channel memory will see better GPU performance than one with slower single-channel memory. At resolutions higher than 1080p, the bandwidth bottleneck becomes more pronounced, and the 13.20 GPixel/s pixel rate will likely limit frame rates in demanding scenes. For 1440p or 4K, the data suggests that the GPU would struggle not just because of compute throughput, but because the shared memory bus cannot feed the shading units fast enough to maintain acceptable frame rates.

For users considering this GPU for high-resolution workloads, the data suggests that the shared memory subsystem is a limiting factor. The GPU's 384 shading units are capable of generating frames at modest resolutions, but the memory bandwidth constraint will cap performance regardless of the compute throughput. The lack of dedicated VRAM also means that texture quality settings may need to be reduced to avoid exceeding the system memory capacity, particularly in systems with 8 GB or less of total RAM. The "System Dependent" bandwidth figure is a warning: performance is not a fixed property of the GPU, but a function of the entire platform.

Who Should Consider It

The Iris Xe G4 48EU is positioned for users who need basic graphical capabilities without the power and thermal overhead of a discrete GPU. The 50th percentile ranking, combined with 844.8 GFLOPS of FP32 performance and 13.20 GPixel/s pixel rate, indicates that this GPU is suitable for light gaming at 1080p with low settings, everyday productivity tasks, and media playback. The 15 W TDP makes it ideal for thin-and-light laptops where battery life and thermals are priorities. A user who primarily runs office applications, browses the web, and streams video will find the GPU more than adequate — these workloads rarely stress the 384 shading units beyond a fraction of their capacity.

Users who intend to play modern AAA games at high settings or resolutions above 1080p should look elsewhere, as the shared memory subsystem and modest compute throughput will not deliver acceptable frame rates. The lack of ray tracing and tensor cores further limits the GPU's appeal for enthusiasts. However, for office work, web browsing, video streaming, and esports titles at low settings, the Iris Xe G4 48EU provides a capable experience. The DirectX 12_1, OpenGL 4.6, and Vulkan 1.4 support ensures broad software compatibility, and the end-of-life status means it is still found in existing systems but is no longer being produced for new designs. In short, this is a GPU for users who know they will not push graphical boundaries — it delivers dependable, low-power performance for the tasks that make up the majority of everyday computing.