AMD Playstation 4 Pro GPU

How It Compares

The AMD PlayStation 4 Pro GPU, based on the Neo chip, occupies a unique position in the benchmark database. Its percentile rank of 50 places it exactly at the midpoint of all recorded GPUs, meaning it outperforms half of the tracked graphics hardware while trailing the other half. This is a console-class part, not a discrete desktop card, so its comparative standing should be interpreted through that lens.

The FACT PACK lists no nearest rivals for this entry, which underscores its isolation in the database. Without direct comparison points, the analysis must rely on absolute specifications and architectural traits rather than head-to-head deltas. The GPU’s end-of-life production status further distinguishes it from actively competing modern parts, as it represents a fixed hardware target from a specific console generation rather than an evolving product line.

Ray Tracing and Feature Set

The architecture is GCN 2.0, a design that predates dedicated ray tracing hardware. The FACT PACK explicitly lists no RT cores and no tensor cores, confirming that this GPU has no hardware acceleration for ray tracing or AI-based tensor operations. Any ray tracing workload would fall back to compute shaders on the 2304 shading units, a method that is functional but substantially less efficient than dedicated hardware.

The API support includes OpenGL 4.6 and Vulkan 1.1, with no DirectX version listed. This is notable because the GPU is a console part, so its API exposure is mediated through the PlayStation 4 Pro system software rather than a desktop driver stack. Vulkan 1.1 support allows for low-level hardware access, which developers can leverage for optimized rendering paths. The absence of DirectX in the FACT PACK does not imply a hardware limitation but rather reflects the console’s software environment.

The feature set is therefore defined by compute and rasterization capabilities from the GCN 2.0 era. The 2304 shading units provide a solid base for traditional forward and deferred rendering, while the 144 texture mapping units and 32 render output units handle texturing and pixel output. There is no hardware video encode or decode block listed, which is typical for a console GPU where fixed-function media blocks are integrated into the system-on-chip but not exposed as discrete GPU features.

Benchmark Performance

The average benchmark score in the FACT PACK is 0, which means there are no recorded benchmark results to analyze numerically. This is an unusual situation for the database, and it requires careful interpretation. The percentile rank of 50 is derived from the full GPU population, but without actual scores, the rank cannot be broken down into specific workload deltas.

What can be analyzed are the theoretical throughput figures. The FP32 performance is 4.198 TFLOPS, which places this GPU in the mid-range of the database’s tracked hardware. The FP16 rate of 8.396 TFLOPS at a 2:1 ratio means that half-precision workloads can run at double the single-precision rate, a feature that GCN 2.0 supports for certain compute tasks. This is not a 1:1 FP16 implementation, but the 2:1 ratio still offers a meaningful speedup for compatible shaders.

The pixel rate of 29.15 GPixel/s is derived from the 32 ROPs and the memory clock. The texture rate of 131.2 GTexel/s comes from the 144 TMUs. These figures are consistent with a GPU designed for 1080p to 1440p rendering at console-quality settings. The memory subsystem provides 217.6 GB/s of bandwidth from 8 GB of GDDR5 on a 256-bit bus, which is adequate for the texture and geometry workloads typical of this performance class.

Without nearest rivals, a relative performance statement cannot be made with exact percentages. The data shows that this GPU sits at the 50th percentile, meaning it is neither a high-end part nor a low-end one. In practical terms, this suggests it can handle most games at 1080p with high settings and some titles at 1440p with adjusted settings, but it will struggle with 4K at high framerates or with heavy post-processing effects.

Power and Cooling

The thermal design power is 150 W, which is modest for a GPU with 5,700 million transistors on a 16 nm process from TSMC. The die size is 322 mm², and the transistor density is 17.7 million per square millimeter. These figures indicate a mid-size chip that operates at moderate power levels. The 150 W TDP includes the GPU core and memory, but the console’s overall power consumption is higher due to the CPU, storage, and system components.

The FACT PACK lists no suggested PSU wattage and no power connector requirements. This is because the GPU is soldered into a console motherboard with a fixed power delivery system. There is no user-accessible power connector, and the PSU is internal to the PlayStation 4 Pro chassis. The cooling solution is likewise integrated into the console, with a single fan and heatsink designed to dissipate the combined heat from the GPU, CPU, and other components.

The physical dimensions of the console are 327 mm in length, 295 mm in height, and 55 mm in width. These dimensions accommodate the internal power supply and cooling system. The display output is a single HDMI 1.4a port, which supports up to 4K resolution at 30 Hz or 1080p at 60 Hz, depending on the HDMI implementation. The lack of multiple display outputs confirms this is a console GPU with a fixed output configuration.

FAQ

Q: Does this GPU support ray tracing?

A: No. The FACT PACK lists no RT cores, and the GCN 2.0 architecture predates hardware ray tracing. Any ray tracing would need to run on the 2304 shading units via compute shaders, which is significantly slower than dedicated hardware.

Q: What is the memory configuration?

A: The GPU has 8 GB of GDDR5 memory on a 256-bit bus, providing 217.6 GB/s of bandwidth. The memory clock is 1700 MHz, which translates to 6.8 Gbps effective.

Q: What APIs are supported?

A: The FACT PACK lists OpenGL 4.6 and Vulkan 1.1. No DirectX version is provided, which is consistent with a console GPU that uses the PlayStation 4 Pro system software for graphics API access.

Q: What is the FP32 performance?

A: The FP32 throughput is 4.198 TFLOPS. The FP16 rate is 8.396 TFLOPS at a 2:1 ratio, meaning half-precision compute runs twice as fast as single-precision.

Q: What is the TDP and power connector requirement?

A: The TDP is 150 W. There is no power connector listed because this is a console GPU with an integrated power delivery system. No suggested PSU is provided for the same reason.

Q: What is the production status and launch date?

A: The production status is end-of-life. The release date is 2016-11-09, and the launch MSRP is 399 USD.

Who Should Consider It

The 50th percentile ranking indicates this GPU is suited for gamers targeting 1080p resolution with high settings in most titles. The 4.198 TFLOPS of FP32 performance and 217.6 GB/s of memory bandwidth are sufficient for 1080p rendering at 30 to 60 frames per second, depending on the game’s optimization. For 1440p, the GPU can handle it with adjusted settings, but demanding titles may require reducing shadow quality, draw distance, or anti-aliasing to maintain playable framerates.

The FP16 2:1 ratio of 8.396 TFLOPS is a feature that developers can use for certain compute-based effects, such as lighting calculations or post-processing, which can free up FP32 resources for other tasks. This makes the GPU somewhat more capable in compute-heavy workloads than raw FP32 numbers suggest, but it does not change the fundamental rasterization limits.

Gamers who prioritize 4K output should note that the HDMI 1.4a port caps at 4K 30 Hz, which is below the 60 Hz standard for modern displays. This GPU is not intended for 4K gaming at high framerates; it is a 1080p-to-1440p part. The end-of-life status means no further driver or firmware updates are planned, so performance will not improve over time.

The 150 W TDP and integrated cooling make this a quiet and efficient solution for a console form factor. Users who already own a PlayStation 4 Pro are the only audience for this GPU, as it is not a standalone product. For those considering a purchase of a used console, the 50th percentile performance is respectable for the price point, but the lack of ray tracing and the fixed HDMI 1.4a output are clear limitations compared to modern dedicated GPUs.