AMD Steam Deck OLED GPU vs Intel Arc Pro B370 Comparison

Head-to-Head Benchmarks

The recorded data for both GPUs presents a scenario where direct head-to-head benchmark scores are not available. The database shows zero benchmark entries for both the AMD Steam Deck OLED GPU and the Intel Arc Pro B370, and the win counts for each part are zero. This means a direct percentage-based comparison using measured performance results is not possible from the available information.

What the data does provide is a clear picture of theoretical peak throughput. The Intel Arc Pro B370 delivers a FP32 compute figure of 6.144 TFLOPS, while the AMD Steam Deck OLED GPU delivers 1.638 TFLOPS. That places the Intel part at roughly 3.75 times the raw FP32 output of the AMD part. In FP16, the Intel part reaches 12.29 TFLOPS versus 3.277 TFLOPS for the AMD part, again a similar multiplier. These are peak rates, not application results, but they indicate a substantial raw compute advantage for the Intel integrated GPU.

Pixel throughput tells a similar story. The Intel Arc Pro B370 achieves 48.00 GPixel/s, while the AMD Steam Deck OLED GPU achieves 25.60 GPixel/s. Texture rate follows the same pattern: 96.00 GTexel/s for Intel versus 51.20 GTexel/s for AMD. In both cases, the Intel part is roughly 1.9 times the rate of the AMD part. This suggests the Intel GPU has a meaningful edge in fill-rate-bound workloads, assuming both operate at their listed peak rates.

Clock behavior is worth addressing. The AMD part has a base clock of 1000 MHz and a boost clock of 1600 MHz. The Intel part has a base clock of 300 MHz and a boost clock of 2400 MHz. The wide gap between base and boost on the Intel side implies that sustained performance depends heavily on power and thermal headroom. The AMD part, with a much narrower clock range, likely maintains closer to its peak rate for longer periods in a constrained thermal envelope.

Memory bandwidth is a major differentiator, but the comparison is complicated by the Intel part using system shared memory. The AMD Steam Deck OLED GPU has dedicated 16 GB of LPDDR5 memory on a 128-bit bus, yielding 176.0 GB/s of bandwidth. The Intel Arc Pro B370 has no dedicated memory; its bandwidth is listed as system dependent. In an integrated configuration, the Intel part will share system memory bandwidth with the CPU, which can reduce effective GPU bandwidth versus a dedicated pool. This is a structural disadvantage for the Intel part in bandwidth-sensitive scenarios, even if the exact system dependent figure is not recorded.

Architecture Differences

The two GPUs come from different architectural generations and foundries. The AMD Steam Deck OLED GPU uses the Sephiroth chip built on RDNA 2.0, fabricated by TSMC on a 6 nm process. The Intel Arc Pro B370 uses the Panther Lake chip built on Xe3-LPG, fabricated by Intel on a 3 nm process. The Intel process node is smaller, which typically allows for higher transistor density and better power efficiency per unit area. However, Intel does not disclose transistor count or die size for this part, so a direct density comparison cannot be made. The AMD chip carries 2,400 million transistors on a 131 mm² die, giving a transistor density of 18.3M per mm².

The generation labels differ as well. AMD is listed as a console GPU for Valve, while Intel is part of the Arc Graphics-WM (Panther Lake) generation. This reflects different design targets: the AMD part is a fixed-function console GPU with a known thermal budget, while the Intel part is an integrated graphics processor intended for portable devices.

Shading resources differ sharply. The AMD part has 512 shading units, 32 texture mapping units, 16 render output units, and 8 ray tracing cores. The Intel part has 1280 shading units, 40 TMUs, 20 ROPs, and 10 ray tracing cores. The Intel part has 2.5 times the shading units and 1.25 times the ray tracing cores. This suggests the Intel architecture is designed for higher throughput per clock, assuming both can sustain their boost clocks.

API support shows a notable difference. Both support DirectX 12 Ultimate (12_2) and OpenGL 4.6. The AMD part supports Vulkan 1.3, while the Intel part supports Vulkan 1.4. The newer Vulkan revision on the Intel side may offer additional features for modern applications that adopt them, though the practical impact depends on software support.

Display output also differs. The AMD Steam Deck OLED GPU has a single USB Type-C output. The Intel Arc Pro B370 lists its display outputs as portable device dependent, meaning the actual connections are determined by the host device rather than the GPU itself. This is consistent with an integrated part that does not own its display path.

The Intel part has its predecessor listed as HD Graphics-WM, while the AMD part has no predecessor. This indicates the Intel integrated GPU line is an evolution of earlier Intel graphics, while the AMD console GPU is a standalone design within this database.

Where Each One Wins

The AMD Steam Deck OLED GPU has clear advantages in memory architecture. Its dedicated 16 GB LPDDR5 pool with 176.0 GB/s bandwidth is a fixed, known quantity. For a handheld console, this removes the variability of system shared memory and guarantees the GPU always has access to its full bandwidth. The Intel part, with system shared memory and system dependent bandwidth, cannot make that guarantee. In workloads that are heavily bandwidth-bound, the AMD part will likely maintain more consistent performance.

The AMD part also has a narrower clock spread. A base clock of 1000 MHz and boost of 1600 MHz means less performance variance between sustained and burst workloads. The Intel part, with a 300 MHz base and 2400 MHz boost, may spend more time near its lower clock in thermally limited portable devices. This makes the AMD part the steadier performer in sustained scenarios, even if its peak rates are lower.

The Intel Arc Pro B370 wins on raw throughput. Its FP32 rate of 6.144 TFLOPS, pixel rate of 48.00 GPixel/s, and texture rate of 96.00 GTexel/s are all substantially higher than the AMD part. For any workload that can scale with shading units and clock speed, the Intel part has the theoretical edge. The larger count of shading units, TMUs, ROPs, and ray tracing cores supports this.

The Intel part also has a lower TDP in terms of the listed figure: 25 W versus 15 W. Note that this is a higher power envelope, not lower. The Intel part is allowed to draw more power, which helps it reach and sustain its higher clocks. The AMD part, constrained to 15 W, must operate within a tighter power budget.

For ray tracing workloads, the Intel part has 10 ray tracing cores versus 8 on the AMD part. This is a modest advantage, but combined with the higher shading resources, it suggests the Intel part may handle ray-traced effects with less performance loss.

The AMD part wins on API maturity in one respect: both support DirectX 12 Ultimate and OpenGL 4.6, so there is no difference there. The Intel part has Vulkan 1.4 versus Vulkan 1.3, which is a feature advantage for Intel, though the real-world impact depends on driver and application support.

FAQ

Q: Which GPU has higher raw compute performance?

A: The Intel Arc Pro B370. Its FP32 rate is 6.144 TFLOPS versus 1.638 TFLOPS for the AMD Steam Deck OLED GPU. Its FP16 rate is 12.29 TFLOPS versus 3.277 TFLOPS.

Q: Do both GPUs support the same DirectX version?

A: Yes. Both support DirectX 12 Ultimate (12_2) and OpenGL 4.6. They differ in Vulkan support: the AMD part supports Vulkan 1.3, while the Intel part supports Vulkan 1.4.

Q: How much memory does each GPU have?

A: The AMD Steam Deck OLED GPU has 16 GB of dedicated LPDDR5 memory on a 128-bit bus. The Intel Arc Pro B370 uses system shared memory, with its size, type, bus width, and bandwidth all listed as system dependent.

Q: What are the clock speeds of each GPU?

A: The AMD part has a base clock of 1000 MHz and a boost clock of 1600 MHz. The Intel part has a base clock of 300 MHz and a boost clock of 2400 MHz.

Q: Which GPU has more shading units?

A: The Intel Arc Pro B370 has 1280 shading units, while the AMD Steam Deck OLED GPU has 512 shading units. The Intel part also has more TMUs (40 versus 32), more ROPs (20 versus 16), and more ray tracing cores (10 versus 8).

Q: What is the power consumption of each GPU?

A: The AMD Steam Deck OLED GPU has a TDP of 15 W. The Intel Arc Pro B370 has a TDP of 25 W.

Q: Which GPU uses a smaller manufacturing process?

A: The Intel Arc Pro B370 uses a 3 nm process from Intel. The AMD Steam Deck OLED GPU uses a 6 nm process from TSMC.

Specification Differences

The two GPUs differ across nearly every recorded specification.

The AMD Steam Deck OLED GPU uses the Sephiroth chip on RDNA 2.0 architecture, fabricated by TSMC on a 6 nm process. It contains 2,400 million transistors on a 131 mm² die, with a transistor density of 18.3M per mm². The Intel Arc Pro B370 uses the Panther Lake chip on Xe3-LPG architecture, fabricated by Intel on a 3 nm process. Its transistor count and die size are unknown.

Clock speeds differ substantially. The AMD part runs at 1000 MHz base and 1600 MHz boost. The Intel part runs at 300 MHz base and 2400 MHz boost. The Intel part has a much wider clock range.

Memory configurations are entirely different. The AMD part has 16 GB of LPDDR5 on a 128-bit bus with 176.0 GB/s bandwidth. The Intel part uses system shared memory, with system dependent bandwidth. Memory clock for the AMD part is 1375 MHz, translating to 11 Gbps effective. The Intel part has no dedicated memory clock listed.

Compute resources favor Intel. The AMD part has 512 shading units, 32 TMUs, 16 ROPs, and 8 ray tracing cores. The Intel part has 1280 shading units, 40 TMUs, 20 ROPs, and 10 ray tracing cores. The Intel part also has higher pixel rate (48.00 GPixel/s versus 25.60 GPixel/s), texture rate (96.00 GTexel/s versus 51.20 GTexel/s), FP32 (6.144 TFLOPS versus 1.638 TFLOPS), and FP16 (12.29 TFLOPS versus 3.277 TFLOPS).

Power and physical attributes differ. The AMD part has a TDP of 15 W and dimensions of 298 mm length, 117 mm height, and 49 mm width. The Intel part has a TDP of 25 W, is listed as an IGP with no power connectors, and has no recorded dimensions. The AMD part has a single USB Type-C display output. The Intel part has display outputs listed as portable device dependent.

Release timing and lineage differ. The AMD part was released on 2023-11-08 and has no predecessor. The Intel part was released on 2026-01-26 and lists HD Graphics-WM as its predecessor. Both are marked as active in production. Both have a percentile ranking of 50 against all GPUs, and neither has a launch MSRP recorded.

The API profiles are close but not identical. Both support DirectX 12 Ultimate (12_2) and OpenGL 4.6. The AMD part supports Vulkan 1.3, while the Intel part supports Vulkan 1.4.

The Verdict

The data presents two GPUs with very different design philosophies. The AMD Steam Deck OLED GPU is a fixed-function console part with a dedicated memory pool, a modest 15 W power envelope, and steady clock behavior. Its strengths are predictability and guaranteed bandwidth. The Intel Arc Pro B370 is a newer integrated part with a much higher peak throughput, a larger shading unit count, and a newer Vulkan revision, but it depends on system shared memory and has a wide clock range that may be limited by thermal constraints in portable devices.

For workloads that demand raw shading throughput, the Intel Arc Pro B370 is the clear choice based on the recorded figures. Its FP32 rate of 6.144 TFLOPS, pixel rate of 48.00 GPixel/s, and texture rate of 96.00 GTexel/s are all multiples of the AMD part's figures. Its 1280 shading units and 10 ray tracing cores give it a structural advantage in compute-heavy and ray-traced workloads.

For workloads that depend on consistent memory bandwidth and sustained performance within a low power budget, the AMD Steam Deck OLED GPU has the advantage. Its 176.0 GB/s dedicated bandwidth is a fixed resource, whereas the Intel part's bandwidth is system dependent and shared. The AMD part's narrower clock range suggests less performance drop-off under sustained load.

Neither part has recorded benchmark scores, so the choice rests on architectural specifications rather than measured application results. The Intel part targets higher peak performance at the cost of power and memory variability. The AMD part targets efficiency and stability in a constrained console environment. The data favors Intel for raw throughput and AMD for predictable, bandwidth-stable operation. The user should weigh whether peak rates or sustained consistency matter more for the intended workload.