AMD Radeon 8065S vs Intel Arc Pro B390 Comparison

AMD Radeon 8065S and Intel Arc Pro B390 are both integrated graphics solutions designed for mobile platforms, yet they represent fundamentally different approaches to processor design and graphics execution. The Radeon 8065S, built on AMD’s Gorgon Halo chip, uses the RDNA 3.5 architecture on a 4 nm TSMC process. The Arc Pro B390, built on Intel’s Panther Lake chip, uses the Xe3-LPG architecture on a 3 nm Intel process. Both are active production parts, with the Radeon 8065S releasing on December 31, 2025, and the Arc Pro B390 releasing on January 26, 2026. The database records no benchmark scores for either part, so the analysis below is drawn entirely from the architectural specifications, clock behavior, and compute resources listed in the recorded data.

Head-to-Head Benchmarks

The recorded data contains no direct head-to-head benchmark results, so a comparison must be constructed from the measured specifications. The most significant single number separating these two parts is FP32 compute throughput. The AMD Radeon 8065S delivers 15.36 TFLOPS of FP32 performance, while the Intel Arc Pro B390 delivers 7.680 TFLOPS. That is exactly a 2.00x advantage for the AMD part in raw single-precision floating-point work. In practice, this means any workload that scales with shader math, such as general compute tasks, physics simulations, or graphics shading, will see the AMD part complete roughly twice as much work per cycle.

The gap in pixel throughput is even larger. The Radeon 8065S records a pixel rate of 192.0 GPixel/s, while the Arc Pro B390 records 60.00 GPixel/s. That gives AMD a 3.20x advantage in fill-rate-bound scenarios, such as high-resolution rasterization with heavy overdraw. Texture rate tells a similar story: the Radeon 8065S reaches 480.0 GTexel/s, while the Arc Pro B390 reaches 120.0 GTexel/s, a 4.00x difference. These are not subtle margins; they indicate that the AMD part is in a different performance class for traditional graphics workloads.

The FP16 comparison is where the picture changes. Both parts record exactly 15.36 TFLOPS of FP16 throughput, but they achieve it differently. The AMD Radeon 8065S does FP16 at a 1:1 ratio with FP32, meaning it processes half-precision data at the same rate as single-precision. The Intel Arc Pro B390 does FP16 at a 2:1 ratio, meaning it processes half-precision data at twice the rate of its FP32 throughput. The measured result is identical at 15.36 TFLOPS, so in FP16-heavy workloads, such as certain machine learning inference tasks or media processing pipelines, the two parts land at the same theoretical ceiling.

Clock speeds also differ substantially. The AMD part runs at a base clock of 1295 MHz and boosts to 3000 MHz. The Intel part runs at a base clock of 300 MHz and boosts to 2500 MHz. The AMD base clock is 4.32x higher than Intel’s base clock, and the AMD boost clock is 1.20x higher. These numbers matter for sustained workloads, where a higher base clock suggests the AMD part can maintain more throughput without relying on boost headroom. The Intel part’s very low 300 MHz base clock indicates a design that idles or drops to minimal activity quickly, relying on its 2500 MHz boost for bursts of performance.

The Radeon 8065S also carries more execution resources. It has 2560 shading units, 160 TMUs, and 64 ROPs. The Arc Pro B390 has 1536 shading units, 48 TMUs, and 24 ROPs. That means AMD has 1.67x more shading units, 3.33x more TMUs, and 2.67x more ROPs. The ray tracing hardware follows the same pattern: AMD has 40 RT cores, while Intel has 12 RT cores, a 3.33x difference. For ray-traced scenes, the AMD part has substantially more dedicated hardware to draw from.

Power draw flips the comparison. The Radeon 8065S is rated at 55 W TDP, while the Arc Pro B390 is rated at 80 W TDP. This is notable because the AMD part delivers higher raw throughput across nearly every measured metric while consuming 25 W less power. The efficiency advantage is baked into the numbers: the AMD part uses 68.75% of the Intel part’s power budget while delivering 2.00x the FP32 throughput.

The Verdict

The data points to a clear split between absolute performance and power efficiency. The AMD Radeon 8065S wins on every compute and graphics throughput metric recorded in the database, with the exception of FP16, where the two parts are tied at 15.36 TFLOPS. The AMD part also does this at a lower TDP of 55 W versus 80 W. For any user prioritizing raw graphics performance in a mobile integrated part, the Radeon 8065S is the stronger choice based on the recorded specifications.

The Intel Arc Pro B390 does not win any recorded throughput comparison. Its only equal standing is in FP16, where both parts record 15.36 TFLOPS. Its lower clock speeds, fewer shaders, fewer TMUs, fewer ROPs, and fewer RT cores place it behind in every other measured category. However, the Intel part uses a 3 nm process from Intel’s own foundry, while the AMD part uses a 4 nm process from TSMC. The Intel part also has a much lower base clock of 300 MHz, which may indicate a design that prioritizes power gating and idle efficiency over sustained throughput. That is a qualitative observation from the clock data, not a measured efficiency score, since the database does not record idle power or efficiency benchmarks.

The percentile data places both parts at the 50th percentile against all GPUs, with an average benchmark score of 0 for both. That means neither part has recorded benchmark submissions in the database, so the percentile is a neutral midpoint rather than a performance ranking. The absence of benchmark data means the verdict rests on architectural and specification comparison.

For a user who needs maximum graphics throughput from an integrated solution, the Radeon 8065S is the recorded choice. For a user who prioritizes the lower base clock behavior or the Intel 3 nm process, the Arc Pro B390 is the only option, but the data does not show any performance benefit from that choice.

Architecture Differences

The two parts use different architectures, different processes, and different foundries. The AMD Radeon 8065S uses RDNA 3.5, built on a 4 nm process at TSMC, with a die size of 308 mm². The Intel Arc Pro B390 uses Xe3-LPG, built on a 3 nm process at Intel, with an unknown die size. The AMD part is part of the Navi Mobile (RX 8000M) generation, while the Intel part is part of the Arc Graphics-WM (Panther Lake) generation. Their predecessors also differ: the Radeon 8065S follows Polaris Mobile, and the Arc Pro B390 follows HD Graphics-WM.

Memory architecture is shared in concept but not in detail. Both parts use System Shared memory, meaning they rely on the host system’s RAM rather than dedicated VRAM. The memory type, bus width, and bandwidth are listed as System Shared or System Dependent for both. This means the actual memory throughput is determined by the host platform, and the database does not record a fixed bandwidth for either part. Neither part has a dedicated memory interface.

Compute resource allocation differs significantly. The Radeon 8065S has 2560 shading units, 160 TMUs, 64 ROPs, and 40 RT cores. The Arc Pro B390 has 1536 shading units, 48 TMUs, 24 ROPs, and 12 RT cores. The AMD part does not record tensor cores, and the Intel part also does not record tensor cores, so neither part has a dedicated AI acceleration block listed in the database.

Clock behavior is a key architectural difference. The Radeon 8065S records a base clock of 1295 MHz and a boost clock of 3000 MHz. The Arc Pro B390 records a base clock of 300 MHz and a boost clock of 2500 MHz. The gap between base and boost on the Intel part is 8.33x, while the gap on the AMD part is 2.32x. This suggests the Intel part is designed to spend more time at low clock states, while the AMD part maintains a higher floor.

Power delivery also differs. The Radeon 8065S is rated at 55 W TDP and uses no power connectors, consistent with an integrated graphics processor. The Arc Pro B390 is rated at 80 W TDP and also uses no power connectors. Both are listed as IGP slot width, meaning they are integrated into the processor package rather than installed as discrete cards. The bus interface differs: the Radeon 8065S uses PCIe 5.0 x16, while the Arc Pro B390 uses IGP. Display outputs for both are listed as Portable Device Dependent.

API support is identical across both parts. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. There is no difference in the recorded API feature set, so software compatibility at the API level should be equivalent.

The process node difference is notable. Intel’s 3 nm process is one step ahead of TSMC’s 4 nm process in the recorded lithography. The Intel part achieves its 80 W TDP on that process, while the AMD part achieves a lower 55 W TDP on the 4 nm process. Transistor counts are unknown for both parts, and the die size is unknown for the Intel part, so density cannot be compared directly.

FAQ

Q: Which GPU has higher FP32 performance?

A: The AMD Radeon 8065S records 15.36 TFLOPS of FP32, which is exactly 2.00x the 7.680 TFLOPS recorded for the Intel Arc Pro B390.

Q: Are these two GPUs equal in any performance metric?

A: Yes. Both record exactly 15.36 TFLOPS of FP16 performance. The AMD part achieves this at a 1:1 ratio with FP32, while the Intel part achieves it at a 2:1 ratio.

Q: What is the power draw of each GPU?

A: The AMD Radeon 8065S is rated at 55 W TDP, and the Intel Arc Pro B390 is rated at 80 W TDP. The AMD part consumes less power while delivering higher throughput in most recorded metrics.

Q: Do these GPUs have dedicated video memory?

A: No. Both use System Shared memory, with the memory type, bus width, and bandwidth listed as System Shared or System Dependent. Neither has a dedicated VRAM interface.

Q: What are the clock speeds of these GPUs?

A: The AMD Radeon 8065S has a base clock of 1295 MHz and a boost clock of 3000 MHz. The Intel Arc Pro B390 has a base clock of 300 MHz and a boost clock of 2500 MHz.

Q: Which GPU has more ray tracing hardware?

A: The AMD Radeon 8065S has 40 RT cores, while the Intel Arc Pro B390 has 12 RT cores. That gives the AMD part a 3.33x advantage in dedicated ray tracing hardware.

Where Each One Wins

The AMD Radeon 8065S wins in every raw throughput category recorded in the database. Its FP32 rate of 15.36 TFLOPS is double the Intel part’s 7.680 TFLOPS. Its pixel rate of 192.0 GPixel/s is 3.20x the Intel part’s 60.00 GPixel/s. Its texture rate of 480.0 GTexel/s is 4.00x the Intel part’s 120.0 GTexel/s. It has 2560 shading units versus 1536, 160 TMUs versus 48, 64 ROPs versus 24, and 40 RT cores versus 12. Its boost clock of 3000 MHz is 1.20x higher than Intel’s 2500 MHz, and its base clock of 1295 MHz is 4.32x higher than Intel’s 300 MHz. It also uses less power at 55 W versus 80 W.

The Intel Arc Pro B390 wins in exactly one recorded category: process node. It is built on Intel’s 3 nm process, while the AMD part uses TSMC’s 4 nm process. That is a lithographic advantage, but it does not translate into a performance win in any recorded metric. The Intel part also ties the AMD part in FP16 at 15.36 TFLOPS and matches the API feature set with DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. In terms of use cases, the AMD part is the choice for graphics-heavy workloads, compute-heavy workloads, and any scenario where pixel fill rate or texture throughput matters. The Intel part is the choice only if the 3 nm process or the lower 300 MHz base clock behavior is the deciding factor, since no performance metric in the database favors it.

Both parts sit at the 50th percentile against all GPUs with an average benchmark score of 0, reflecting the absence of recorded benchmark submissions. The wins and ties above are entirely derived from the architectural specification data. The AMD Radeon 8065S is the faster part on paper, and the Intel Arc Pro B390 is the more conservatively clocked part on a newer process node.