AMD Steam Deck OLED GPU vs Intel Arc B770 Comparison

FAQ

Q: What are the core architectural differences between the AMD Steam Deck OLED GPU and the Intel Arc B770?

A: The AMD Steam Deck OLED GPU uses the RDNA 2.0 architecture on a 6 nm TSMC process, while the Intel Arc B770 uses the Xe2-HPG architecture on a 5 nm TSMC process. The Steam Deck GPU is a Console GPU (Valve) generation part, whereas the Arc B770 belongs to the Battlemage (Arc 7) generation. The AMD chip is named Sephiroth, and the Intel chip is named BMG-G31.

Q: How do the memory subsystems compare?

A: Both GPUs have 16 GB of memory, but they use different types and buses. The Steam Deck OLED GPU uses LPDDR5 on a 128-bit bus with 176.0 GB/s bandwidth, while the Arc B770 uses GDDR6 on a 256-bit bus with 512.0 GB/s bandwidth. The Intel part delivers nearly three times the memory bandwidth.

Q: Which GPU has more shading units and texture mapping units?

A: The Intel Arc B770 has 4096 shading units and 256 TMUs, compared to 512 shading units and 32 TMUs on the Steam Deck OLED GPU. The Arc B770 also has 128 ROPs versus 16, and 32 ray tracing cores versus 8.

Q: What are the clock speeds of each GPU?

A: The Steam Deck OLED GPU has a base clock of 1000 MHz and a boost clock of 1600 MHz. The Arc B770 runs at a base clock of 2100 MHz and a boost clock of 2400 MHz.

Q: What is the power draw of each GPU?

A: The Steam Deck OLED GPU has a TDP of 15 W. The Arc B770 has a TDP of 225 W and requires a suggested power supply of 550 W. The Intel card uses a 1x 6-pin plus 1x 8-pin power connector configuration and is a dual-slot card.

Q: What API support does each GPU provide?

A: Both support DirectX 12 Ultimate (12_2) and OpenGL 4.6. The Steam Deck OLED GPU supports Vulkan 1.3, while the Arc B770 supports Vulkan 1.4.

Architecture Differences

The AMD Steam Deck OLED GPU and Intel Arc B770 represent two fundamentally different design targets within the database. The Steam Deck OLED GPU is built on TSMC's 6 nm process with the RDNA 2.0 architecture, a design optimized for constrained power envelopes. The chip, named Sephiroth, integrates 2,400 million transistors on a 131 mm² die, yielding a transistor density of 18.3M per mm². The Arc B770, by contrast, uses TSMC's 5 nm process with the Xe2-HPG architecture, and its BMG-G31 chip has a die size of 368 mm². Transistor count for the Intel part is not recorded in the database.

The compute resources differ by an order of magnitude. The Steam Deck OLED GPU packs 512 shading units, 32 TMUs, 16 ROPs, and 8 ray tracing cores. The Arc B770 scales that to 4096 shading units, 256 TMUs, 128 ROPs, and 32 ray tracing cores. That is an 8x increase in shading units and TMUs, and a 4x increase in ray tracing cores. The pixel rate tells the same story: 25.60 GPixel/s for the AMD part versus 307.2 GPixel/s for the Intel part. Texture rate is 51.20 GTexel/s versus 614.4 GTexel/s.

Clock behavior also differs sharply. The Steam Deck OLED GPU runs at 1000 MHz base and 1600 MHz boost, while the Arc B770 runs at 2100 MHz base and 2400 MHz boost. These clocks, combined with the wider execution resources, produce a large FP32 gap: 1.638 TFLOPS for the AMD part versus 19.66 TFLOPS for the Intel part. FP16 performance follows the same ratio, with 3.277 TFLOPS (2:1) versus 39.32 TFLOPS (2:1).

Memory architecture reinforces the performance split. The Steam Deck OLED GPU uses 16 GB of LPDDR5 on a 128-bit bus at 1375 MHz (11 Gbps effective), producing 176.0 GB/s. The Arc B770 uses 16 GB of GDDR6 on a 256-bit bus at 2000 MHz (16 Gbps effective), producing 512.0 GB/s. The AMD part is a single-chip mobile solution with one USB Type-C display output, while the Intel part is a dual-slot desktop card with PCIe 4.0 x16 and outputs of 1x HDMI 2.1a plus 3x DisplayPort 2.1.

Power delivery is the clearest architectural divergence. The Steam Deck OLED GPU is rated at 15 W TDP, befitting its integrated console role. The Arc B770 is rated at 225 W TDP with a 550 W suggested power supply. The Intel card requires 1x 6-pin plus 1x 8-pin power connectors. The AMD part has no recorded slot width or power connectors, reflecting its embedded nature.

Head-to-Head Benchmarks

The database records no direct head-to-head benchmark runs for these two GPUs. The headToHeadBenchmarks array is empty, and winsA and winsB are both 0. Consequently, the comparison rests entirely on the specification-derived metrics recorded in the database.

The largest advantage for the Intel Arc B770 appears in raw throughput metrics. FP32 compute is 19.66 TFLOPS versus 1.638 TFLOPS, a 12x difference in favor of the Intel part. FP16 compute shows the same ratio: 39.32 TFLOPS versus 3.277 TFLOPS. Texture rate is 614.4 GTexel/s versus 51.20 GTexel/s, a 12x gap. Pixel rate is 307.2 GPixel/s versus 25.60 GPixel/s, an 11x gap.

Memory bandwidth is another decisive category. The Arc B770's 512.0 GB/s is 2.9x the Steam Deck OLED GPU's 176.0 GB/s. The memory clock difference is smaller in absolute terms: 2000 MHz (16 Gbps effective) versus 1375 MHz (11 Gbps effective), but the 256-bit bus versus 128-bit bus doubles the effective transfer width.

Clock speed advantages also favor the Intel part. The Arc B770's 2100 MHz base clock is 2.1x the Steam Deck OLED GPU's 1000 MHz base clock. The boost clocks are closer in ratio: 2400 MHz versus 1600 MHz, a 1.5x difference. These clock advantages compound with the wider execution resources.

The Steam Deck OLED GPU holds advantages in power efficiency and physical footprint. Its 15 W TDP is 15x lower than the Arc B770's 225 W. The AMD die is 131 mm² versus 368 mm², and the transistor density is 18.3M per mm², though the Intel part's density is not recorded. The Steam Deck OLED GPU's dimensions are recorded as 298 mm length, 117 mm height, and 49 mm width, while the Arc B770 has no recorded dimensions.

Both GPUs share the same 16 GB memory capacity and the same DirectX 12 Ultimate (12_2) and OpenGL 4.6 support. The Arc B770 adds Vulkan 1.4 support versus Vulkan 1.3 on the AMD part. The Steam Deck OLED GPU was released on 2023-11-08, while the Arc B770's release date is recorded as 2025-12-31.

Specification Differences

| Specification | AMD Steam Deck OLED GPU | Intel Arc B770 |

|---|---|---|

| Architecture | RDNA 2.0 | Xe2-HPG |

| Process Node | 6 nm | 5 nm |

| Transistors | 2,400 million | unknown |

| Die Size | 131 mm² | 368 mm² |

| Transistor Density | 18.3M / mm² | null |

| Base Clock | 1000 MHz | 2100 MHz |

| Boost Clock | 1600 MHz | 2400 MHz |

| Memory Clock | 1375 MHz (11 Gbps effective) | 2000 MHz (16 Gbps effective) |

| Memory Type | LPDDR5 | GDDR6 |

| Memory Bus Width | 128 bit | 256 bit |

| Memory Bandwidth | 176.0 GB/s | 512.0 GB/s |

| Shading Units | 512 | 4096 |

| TMUs | 32 | 256 |

| ROPs | 16 | 128 |

| Ray Tracing Cores | 8 | 32 |

| Pixel Rate | 25.60 GPixel/s | 307.2 GPixel/s |

| Texture Rate | 51.20 GTexel/s | 614.4 GTexel/s |

| FP32 | 1.638 TFLOPS | 19.66 TFLOPS |

| FP16 | 3.277 TFLOPS (2:1) | 39.32 TFLOPS (2:1) |

| TDP | 15 W | 225 W |

| Slot Width | null | Dual-slot |

| Power Connectors | null | 1x 6-pin + 1x 8-pin |

| Suggested PSU | null | 550 W |

| Bus Interface | null | PCIe 4.0 x16 |

| Display Outputs | 1x USB Type-C | 1x HDMI 2.1a, 3x DisplayPort 2.1 |

| Vulkan | 1.3 | 1.4 |

| Release Date | 2023-11-08 | 2025-12-31 |

| Predecessor | null | Alchemist |

The Verdict

The data indicates two GPUs built for entirely different roles. The Intel Arc B770 is the overwhelmingly faster part in every measured compute and memory throughput category. Its FP32 output of 19.66 TFLOPS is 12x the Steam Deck OLED GPU's 1.638 TFLOPS. Its memory bandwidth of 512.0 GB/s is 2.9x higher. Its shading unit count of 4096 is 8x the AMD part's 512. Its pixel rate of 307.2 GPixel/s is 11x higher. The Arc B770 also carries more ray tracing cores (32 versus 8), more TMUs (256 versus 32), and more ROPs (128 versus 16).

The Steam Deck OLED GPU counters with a 15 W TDP, which is 15x lower than the Arc B770's 225 W. It achieves this with a 6 nm process and a 131 mm² die, versus 368 mm² for the Intel part. The AMD GPU's physical dimensions of 298 mm by 117 mm by 49 mm and single USB Type-C output confirm its integrated, handheld-oriented design. The Arc B770's dual-slot form factor, two power connectors, and 550 W suggested PSU confirm its desktop add-in board role.

The 16 GB memory capacity is the only major specification where the two match. Both support DirectX 12 Ultimate (12_2) and OpenGL 4.6, with the Arc B770 adding Vulkan 1.4 over the AMD part's Vulkan 1.3. The release dates also differ: the Steam Deck OLED GPU launched on 2023-11-08, while the Arc B770 is dated 2025-12-31.

For users selecting a GPU strictly from this data, the choice depends on the target system. The Arc B770 is the only one of the two that can drive high-resolution, high-refresh workloads given its 512.0 GB/s bandwidth and 19.66 TFLOPS compute. The Steam Deck OLED GPU is the only one that fits a 15 W power envelope and a compact console form factor. Neither part can substitute for the other.

Where Each One Wins

Intel Arc B770 wins on raw performance. Every throughput metric in the database favors the Intel part. FP32 compute is 19.66 TFLOPS versus 1.638 TFLOPS. Texture rate is 614.4 GTexel/s versus 51.20 GTexel/s. Pixel rate is 307.2 GPixel/s versus 25.60 GPixel/s. The 4096 shading units, 256 TMUs, and 128 ROPs give it an 8x resource advantage over the AMD part in each category. The 512.0 GB/s memory bandwidth supports these wider execution resources without bottlenecking.

Intel Arc B770 wins on memory bandwidth and capacity utilization. Both have 16 GB, but the Arc B770's GDDR6 on a 256-bit bus delivers 512.0 GB/s, which is 2.9x the 176.0 GB/s of the Steam Deck OLED GPU's LPDDR5 on a 128-bit bus. Higher bandwidth matters for 1440p and 4K texture streaming, ray tracing workloads, and compute tasks that repeatedly access large datasets.

Intel Arc B770 wins on clock speed. The 2100 MHz base and 2400 MHz boost clocks are 2.1x and 1.5x the Steam Deck OLED GPU's respective 1000 MHz and 1600 MHz clocks. Higher clocks also improve the ray tracing throughput, since the Arc B770 has 4x the ray tracing cores (32 versus 8).

AMD Steam Deck OLED GPU wins on power efficiency. The 15 W TDP versus 225 W is a 15x advantage. This is the defining metric for the AMD part, as it enables operation in a handheld console with no external power connectors and no suggested PSU requirement.

AMD Steam Deck OLED GPU wins on physical integration. The 131 mm² die on 6 nm TSMC process is less than half the 368 mm² die of the Arc B770. The recorded dimensions of 298 mm by 117 mm by 49 mm and single USB Type-C output show a self-contained unit, whereas the Arc B770 requires a dual-slot chassis, 1x 6-pin plus 1x 8-pin power connectors, and a 550 W suggested PSU.

AMD Steam Deck OLED GPU wins on transistor density. The database records 18.3M transistors per mm² for the Steam Deck OLED GPU, while the Arc B770's transistor density is not recorded. The AMD part fits 2,400 million transistors into 131 mm², which is a denser packing than the Intel part's unknown transistor count spread across 368 mm².

The data supports a simple split: the Arc B770 is for desktop performance workloads, the Steam Deck OLED GPU is for low-power integrated use. Neither part's advantages translate to the other's domain.