Intel Arc B570 vs NVIDIA RTX A5000 Mobile Comparison

Head-to-Head Benchmarks

The recorded data shows a clear split in benchmark victories: the NVIDIA RTX A5000 Mobile wins six of the nine head-to-head tests, while the Intel Arc B570 takes three. The largest margin belongs to the NVIDIA part in the Passmark DirectX 10 test, where it scores 115 against Intel's 65, a delta of 76.9%. This is a substantial gap that points to a significant advantage in legacy DirectX 10 workloads.

The second-largest win for the NVIDIA RTX A5000 Mobile comes in the Geekbench OpenCL test, with a score of 110,877 versus 83,514 for the Intel Arc B570, translating to a 32.8% lead. This is a strong result for compute-heavy tasks that rely on OpenCL. In the Passmark DirectX 11 test, the NVIDIA part again leads, scoring 133 against 118, a 12.7% advantage. The Passmark G3D test, which often reflects overall gaming performance across various DirectX titles, also favors the NVIDIA GPU, with a score of 15,779 versus 14,195, a delta of 11.2%. The margin in DirectX 9 is narrower, 169 versus 164, a 3% lead for NVIDIA, and the DirectX 12 test is a dead heat with both parts scoring exactly 72.

The Intel Arc B570's victories are concentrated in different areas. Its most significant win is in the Geekbench Vulkan test, where it scores 96,844 against the NVIDIA's 88,144, a delta of -9% (meaning Intel is 9% faster). This is a notable reversal, showing Intel's Vulkan driver and hardware implementation to be more efficient for that API. The Intel part also wins the Passmark G2D test, scoring 661 versus 629, a 4.8% lead, which reflects better 2D desktop performance. Finally, the Passmark GPU Compute test goes to Intel, with a score of 7,281 versus 6,945, a 4.6% advantage. This shows that in certain compute workloads, particularly those measured by Passmark, the Intel Arc B570 can outperform the NVIDIA part despite having a lower raw FP32 figure.

Looking at the overall average benchmark scores, the NVIDIA RTX A5000 Mobile sits at 24,763, placing it in the 70th percentile of all GPUs. Its nearest rival, the AMD Radeon RX 590, has an average score of 24,744, a delta of only 0.1%, making them essentially peers. The Intel Arc B570, on the other hand, has an average score of 20,556, putting it in the 65th percentile. Its closest rival is the NVIDIA GeForce RTX 3070 Mobile with a score of 20,534, a 0.1% difference, and the Intel Arc A750 is a near tie as well, with a delta of -0.1%.

Architecture Differences

The two GPUs are built on fundamentally different architectures and process technologies. The NVIDIA RTX A5000 Mobile is based on the GA104 chip using the Ampere architecture, fabricated on an 8 nm process at Samsung. The Intel Arc B570 uses the BMG-G21 chip with the Xe2-HPG architecture, built on a 5 nm process at TSMC. This process difference is significant, as the Intel part achieves a much higher transistor density of 72.1M per mm² compared to NVIDIA's 44.4M per mm², despite having more transistors overall (19,600 million versus 17,400 million) on a smaller die (272 mm² versus 392 mm²).

The execution resources differ dramatically. The NVIDIA part has 6,144 shading units, 192 TMUs, and 96 ROPs, alongside 48 RT cores and 192 tensor cores. In contrast, the Intel Arc B570 has 2,304 shading units, 144 TMUs, and 80 ROPs, with 18 ray tracing cores and no dedicated tensor cores listed. Despite the lower shader count, the Intel part runs at a much higher clock speed, with a base and boost of 2500 MHz, compared to the NVIDIA's 900 MHz base and 1575 MHz boost. This helps explain how the Intel part achieves a higher pixel rate of 200.0 GPixel/s and texture rate of 360.0 GTexel/s, against the NVIDIA's 151.2 GPixel/s and 302.4 GTexel/s.

Raw FP32 throughput favors the NVIDIA part significantly: 19.35 TFLOPS versus 11.52 TFLOPS. However, the Intel part shows a different strategy with FP16, offering 23.04 TFLOPS at a 2:1 ratio, while the NVIDIA part delivers 19.35 TFLOPS at a 1:1 ratio. Memory configurations also differ. The NVIDIA RTX A5000 Mobile has 16 GB of GDDR6 on a 256-bit bus, delivering 448.0 GB/s of bandwidth. The Intel Arc B570 has 10 GB of GDDR6 on a 160-bit bus, with 380.0 GB/s of bandwidth. Both support PCIe 4.0, but the NVIDIA uses a x16 interface while Intel uses a x8 interface.

The API support is identical, with both parts offering DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The power draw is also the same, with both having a TDP of 150 W. The NVIDIA part is now end-of-life, released in April 2021, while the Intel part is active, released in January 2025. The Intel part is a dual-slot card that requires a single 8-pin power connector and suggests a 450 W power supply, whereas the NVIDIA mobile part has no power connectors and its display outputs are described as portable device dependent.

Where Each One Wins

The NVIDIA RTX A5000 Mobile is the clear winner for compute-heavy tasks that rely on OpenCL, as evidenced by its 32.8% lead in Geekbench OpenCL. It also holds strong advantages in older DirectX APIs, with huge leads in DirectX 10 (76.9%) and solid leads in DirectX 11 (12.7%) and DirectX 9 (3%). Its overall 3D performance, as measured by Passmark G3D, is 11.2% better, which suggests it is the stronger choice for a broad range of DirectX-based games. The data indicates that the NVIDIA part is better suited for legacy software compatibility and workloads that leverage raw FP32 compute power.

The Intel Arc B570, on the other hand, demonstrates superior performance in Vulkan-based applications, with a 9% lead in Geekbench Vulkan. This is a significant advantage for modern games and applications that use Vulkan as their primary graphics API. It also wins in 2D performance (Passmark G2D) with a 4.8% lead and in the Passmark GPU Compute test with a 4.6% advantage. The Intel part's higher clock speeds and efficient architecture make it a strong option for scenarios where Vulkan is the target API or for specific compute workloads that are not dependent on massive FP32 throughput.

The tie in Passmark DirectX 12 (both scoring 72) indicates that for the latest DirectX 12 titles, there is no clear performance leader between these two, at least in that specific test. The Intel part also has the advantage of being a newer, active product, while the NVIDIA part is end-of-life. The Intel Arc B570 is a desktop card with fixed display outputs (1x HDMI 2.1a and 3x DisplayPort 2.1), while the NVIDIA part is a mobile component whose outputs depend on the laptop design.

The Verdict

The benchmark data presents a nuanced picture. The NVIDIA RTX A5000 Mobile, with its 70th percentile ranking and average score of 24,763, is the higher-performing product overall. It dominates in OpenCL compute, older DirectX APIs, and general 3D performance. For users who rely on OpenCL for professional workloads or have a software stack that depends on DirectX 10 or 11, the NVIDIA part is the clear choice based on the numbers. Its 76.9% lead in DirectX 10 and 32.8% lead in OpenCL are decisive.

However, the Intel Arc B570 is not a slouch. Its 65th percentile and average score of 20,556 still place it in respectable company, within 0.1% of the NVIDIA GeForce RTX 3070 Mobile. Its 9% victory in Geekbench Vulkan is a strong indicator for modern gaming and applications that are Vulkan-centric. The Intel part also offers better 2D performance and wins the Passmark GPU Compute test, suggesting it may be a better fit for certain compute tasks.

The choice depends entirely on the target workload. If the priority is maximum performance in OpenCL, older DirectX titles, and overall 3D throughput, the NVIDIA RTX A5000 Mobile is the better option. If the priority is Vulkan performance, modern API efficiency, and a newer product with active production status, the Intel Arc B570 has a compelling case. The DirectX 12 tie suggests that for the most current games, neither part has a definitive edge. The Intel part also comes with a launch MSRP of 219 USD. The NVIDIA part's mobile nature and end-of-life status are also factors to consider, as is the Intel part's requirement for an 8-pin power connector and a 450 W power supply, compared to the NVIDIA part's lack of power connectors.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA RTX A5000 Mobile has a higher average benchmark score of 24,763, compared to the Intel Arc B570's 20,556.

Q: How much faster is the NVIDIA RTX A5000 Mobile in OpenCL?

A: The NVIDIA RTX A5000 Mobile is 32.8% faster in the Geekbench OpenCL test, scoring 110,877 versus 83,514.

Q: In which test does the Intel Arc B570 have its largest win?

A: The Intel Arc B570's largest win is in the Geekbench Vulkan test, where it is 9% faster, scoring 96,844 against 88,144.

Q: Do these GPUs have the same TDP?

A: Yes, both the NVIDIA RTX A5000 Mobile and the Intel Arc B570 have a TDP of 150 W.

Q: What is the difference in memory bandwidth?

A: The NVIDIA RTX A5000 Mobile offers 448.0 GB/s of bandwidth, while the Intel Arc B570 offers 380.0 GB/s.

Q: Which GPU scored higher in the Passmark DirectX 12 test?

A: Neither GPU won; they tied exactly, with both scoring 72 in the Passmark DirectX 12 test.

Specification Differences

| Specification | NVIDIA RTX A5000 Mobile | Intel Arc B570 |

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

| Chip | GA104 | BMG-G21 |

| Architecture | Ampere | Xe2-HPG |

| Generation | Ampere-MW (Ax000) | Battlemage (Arc 5) |

| Process Node | 8 nm | 5 nm |

| Foundry | Samsung | TSMC |

| Transistors | 17,400 million | 19,600 million |

| Die Size | 392 mm² | 272 mm² |

| Transistor Density | 44.4M / mm² | 72.1M / mm² |

| Base Clock | 900 MHz | 2500 MHz |

| Boost Clock | 1575 MHz | 2500 MHz |

| Memory Clock | 1750 MHz, 14 Gbps effective | 2375 MHz, 19 Gbps effective |

| Memory Size | 16 GB | 10 GB |

| Memory Bus Width | 256 bit | 160 bit |

| Memory Bandwidth | 448.0 GB/s | 380.0 GB/s |

| Shading Units | 6144 | 2304 |

| TMUs | 192 | 144 |

| ROPs | 96 | 80 |

| RT Cores | 48 | 18 |

| Tensor Cores | 192 | null |

| Pixel Rate | 151.2 GPixel/s | 200.0 GPixel/s |

| Texture Rate | 302.4 GTexel/s | 360.0 GTexel/s |

| FP32 | 19.35 TFLOPS | 11.52 TFLOPS |

| FP16 | 19.35 TFLOPS (1:1) | 23.04 TFLOPS (2:1) |

| Power Connectors | None | 1x 8-pin |

| Suggested PSU | null | 450 W |

| Bus Interface | PCIe 4.0 x16 | PCIe 4.0 x8 |

| Display Outputs | Portable Device Dependent | 1x HDMI 2.1a, 3x DisplayPort 2.1 |

| Slot Width | null | Dual-slot |

| Dimensions | null | 272 mm (10.7 inches) length, 115 mm (4.5 inches) height |

| Production Status | End-of-life | Active |

| Release Date | 2021-04-11 | 2025-01-15 |

| Predecessor | Quadro Turing-M | Alchemist |

| Successor | Ada-MW | null |

| Launch MSRP | null | 219 USD |

| Percentile | 70 | 65 |

| Avg Benchmark Score | 24763 | 20556 |