Intel Arc Pro B60 Dual vs NVIDIA L20 Comparison

FAQ

Q: What are the core architectural differences between the Intel Arc Pro B60 Dual and the NVIDIA L20?

A: The Intel Arc Pro B60 Dual uses the Xe2-HPG architecture (Battlemage Pro Series) with a BMG-G21 chip, while the NVIDIA L20 uses Ada Lovelace with an AD102 chip. The Intel card has 2,560 shading units, 160 TMUs, 80 ROPs, and 20 RT cores. The NVIDIA card has 11,776 shading units, 368 TMUs, 128 ROPs, 92 RT cores, and 368 tensor cores.

Q: How do the memory subsystems compare between these two cards?

A: The Intel Arc Pro B60 Dual has 24 GB of GDDR6 memory on a 192-bit bus with 456.0 GB/s bandwidth. The NVIDIA L20 has 48 GB of GDDR6 memory on a 384-bit bus with 864.0 GB/s bandwidth. The NVIDIA card offers double the capacity and nearly double the bandwidth.

Q: Which card has higher compute throughput in FP32 operations?

A: The NVIDIA L20 delivers 59.35 TFLOPS of FP32 performance, which is substantially higher than the Intel Arc Pro B60 Dual's 12.29 TFLOPS. The NVIDIA card also matches its FP32 output in FP16 at 59.35 TFLOPS (1:1), whereas the Intel card reaches 24.58 TFLOPS in FP16 (2:1).

Q: What are the power requirements for each card?

A: The Intel Arc Pro B60 Dual has a TDP of 400 W and requires an 800 W suggested PSU, using a single 16-pin power connector. The NVIDIA L20 has a TDP of 275 W and requires a 600 W suggested PSU, also using a single 16-pin power connector.

Q: When were these products released and what is their production status?

A: The Intel Arc Pro B60 Dual was released on September 4, 2025, and the NVIDIA L20 was released on November 15, 2023. Both are listed as Active in production status.

Q: What benchmark data exists for the NVIDIA L20, and how does it rank?

A: The NVIDIA L20 has recorded Geekbench OpenCL score of 274,276 and a Geekbench Vulkan score of 228,018, with an average benchmark score of 251,147. It sits at the 99th percentile among all GPUs. No benchmark scores are recorded for the Intel Arc Pro B60 Dual in the database.

Architecture Differences

The Intel Arc Pro B60 Dual and NVIDIA L20 represent two fundamentally different design philosophies. The Intel card is built on the Xe2-HPG architecture from the Battlemage Pro Series, using the BMG-G21 chip. The NVIDIA L20 uses the Ada Lovelace architecture from the Server Ada generation, built around the AD102 chip. Both are manufactured on a 5 nm process at TSMC, but the similarity ends there.

The NVIDIA AD102 die is substantially larger at 609 mm² compared to the Intel BMG-G21 at 272 mm². Transistor counts reflect this disparity: the NVIDIA chip packs 76,300 million transistors versus 19,600 million for Intel. The transistor density also differs, with NVIDIA achieving 125.3M per mm² compared to Intel's 72.1M per mm².

Compute resources diverge sharply. The NVIDIA L20 contains 11,776 shading units, 368 TMUs, 128 ROPs, 92 RT cores, and 368 tensor cores. The Intel Arc Pro B60 Dual features 2,560 shading units, 160 TMUs, 80 ROPs, and 20 RT cores, with no tensor cores listed. This gives NVIDIA a 4.6x advantage in shading units and a 4.6x advantage in RT cores.

The memory architectures reflect different priorities. The Intel card uses a 192-bit memory bus with 24 GB of GDDR6, while the NVIDIA card uses a 384-bit bus with 48 GB of GDDR6. Memory bandwidth follows: 456.0 GB/s for Intel versus 864.0 GB/s for NVIDIA. Clock speeds show Intel running its base at 2000 MHz and boost at 2400 MHz, while NVIDIA runs a lower 1440 MHz base but a higher 2520 MHz boost. The Intel memory runs at 2375 MHz (19 Gbps effective) versus 2250 MHz (18 Gbps effective) for NVIDIA.

The PCIe interface differs as well. The Intel card uses PCIe 5.0 x8, while the NVIDIA card uses PCIe 4.0 x16. Display outputs are also different: Intel provides 4x mini-DisplayPort 2.1, while NVIDIA provides 4x DisplayPort 1.4a. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4.

Physical dimensions show the Intel card is longer at 300 mm (11.8 inches) versus 267 mm (10.5 inches) for NVIDIA. Heights are similar at 110 mm (4.3 inches) for Intel and 111 mm (4.4 inches) for NVIDIA. The Intel card has a width of 40 mm (1.6 inches), while the NVIDIA width is not recorded. Both are dual-slot cards.

Head-to-Head Benchmarks

The database contains no direct head-to-head benchmark results between the Intel Arc Pro B60 Dual and the NVIDIA L20. The Intel card has no recorded benchmark scores at all, while the NVIDIA L20 has two recorded results.

The NVIDIA L20's Geekbench OpenCL score of 274,276 and Geekbench Vulkan score of 228,018 combine for an average benchmark score of 251,147. This places the card at the 99th percentile among all GPUs. The Intel Arc Pro B60 Dual holds a 50th percentile ranking, though without benchmark scores to substantiate performance.

The NVIDIA L20's nearest rivals in the database provide context. It leads the NVIDIA PG506-232 by 11.6% (that card scores 225,124) and the AMD Radeon PRO W7900D by 14.2% (that card scores 219,827). The NVIDIA L40 scores 284,111, placing it 11.6% ahead of the L20, and the NVIDIA RTX 6000 Ada Generation scores 287,237, putting it 12.6% ahead.

Given the architectural differences, the recorded data strongly suggests the NVIDIA L20 would dominate compute-heavy workloads. The L20's FP32 throughput of 59.35 TFLOPS is roughly 4.8x the Intel card's 12.29 TFLOPS. The L20's texture rate of 927.4 GTexel/s is about 2.4x the Intel card's 384.0 GTexel/s. The L20's pixel rate of 322.6 GPixel/s exceeds the Intel card's 192.0 GPixel/s by a factor of 1.7.

The Intel card does hold advantages in certain areas. Its base clock of 2000 MHz runs higher than NVIDIA's 1440 MHz base. Its memory clock of 2375 MHz (19 Gbps effective) exceeds NVIDIA's 2250 MHz (18 Gbps effective). The Intel card also uses the newer PCIe 5.0 interface, though only at x8 width. These factors do not offset the massive compute and memory bandwidth advantages held by the NVIDIA card.

The Verdict

The data indicates a clear performance hierarchy. The NVIDIA L20 sits at the 99th percentile among all GPUs with an average benchmark score of 251,147. The Intel Arc Pro B60 Dual sits at the 50th percentile with no recorded benchmark scores. For compute workloads, the NVIDIA L20 delivers 59.35 TFLOPS of FP32 performance, while the Intel card delivers 12.29 TFLOPS.

Memory capacity and bandwidth favor NVIDIA decisively. The L20 offers 48 GB of GDDR6 with 864.0 GB/s bandwidth, double the Intel card's 24 GB and 456.0 GB/s. The NVIDIA card also has 368 tensor cores, a feature entirely absent from the Intel specification.

The Intel Arc Pro B60 Dual has a launch MSRP of 1,199 USD. It draws more power at 400 W TDP versus 275 W for NVIDIA, and requires a larger suggested PSU at 800 W versus 600 W. The Intel card is also longer at 300 mm versus 267 mm.

Power efficiency favors NVIDIA. The L20 delivers 59.35 TFLOPS at 275 W, while the Intel card delivers 12.29 TFLOPS at 400 W. The NVIDIA card provides roughly 0.216 TFLOPS per watt, while the Intel card provides roughly 0.031 TFLOPS per watt.

The NVIDIA L20's predecessor is listed as Server Ampere and its successor as Server Hopper, indicating a clear product lineage. The Intel card has no predecessor or successor listed. The NVIDIA card also has the advantage of established benchmark data, while the Intel card's performance profile remains unmeasured in the database.

Specification Differences

| Specification | Intel Arc Pro B60 Dual | NVIDIA L20 |

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

| Architecture | Xe2-HPG | Ada Lovelace |

| Chip | BMG-G21 | AD102 |

| Process Node | 5 nm | 5 nm |

| Transistors | 19,600 million | 76,300 million |

| Die Size | 272 mm² | 609 mm² |

| Transistor Density | 72.1M / mm² | 125.3M / mm² |

| Base Clock | 2000 MHz | 1440 MHz |

| Boost Clock | 2400 MHz | 2520 MHz |

| Memory Clock | 2375 MHz (19 Gbps effective) | 2250 MHz (18 Gbps effective) |

| Memory Size | 24 GB | 48 GB |

| Memory Bus Width | 192 bit | 384 bit |

| Memory Bandwidth | 456.0 GB/s | 864.0 GB/s |

| Shading Units | 2560 | 11776 |

| TMUs | 160 | 368 |

| ROPs | 80 | 128 |

| RT Cores | 20 | 92 |

| Tensor Cores | None listed | 368 |

| Pixel Rate | 192.0 GPixel/s | 322.6 GPixel/s |

| Texture Rate | 384.0 GTexel/s | 927.4 GTexel/s |

| FP32 | 12.29 TFLOPS | 59.35 TFLOPS |

| FP16 | 24.58 TFLOPS (2:1) | 59.35 TFLOPS (1:1) |

| TDP | 400 W | 275 W |

| Suggested PSU | 800 W | 600 W |

| Bus Interface | PCIe 5.0 x8 | PCIe 4.0 x16 |

| Display Outputs | 4x mini-DisplayPort 2.1 | 4x DisplayPort 1.4a |

| Length | 300 mm (11.8 inches) | 267 mm (10.5 inches) |

| Height | 110 mm (4.3 inches) | 111 mm (4.4 inches) |

| Width | 40 mm (1.6 inches) | Not recorded |

| Release Date | 2025-09-04 | 2023-11-15 |

| Predecessor | None listed | Server Ampere |

| Successor | None listed | Server Hopper |

| Launch MSRP | 1,199 USD | None listed |

Where Each One Wins

Compute-heavy workloads: The NVIDIA L20 dominates. Its FP32 output of 59.35 TFLOPS is nearly 5x the Intel card's 12.29 TFLOPS. The L20 also provides 368 tensor cores for accelerated AI and machine learning tasks, a capability the Intel card lacks entirely. The L20's FP16 performance of 59.35 TFLOPS (1:1) matches its FP32 output, while the Intel card's FP16 of 24.58 TFLOPS (2:1) is achieved through ratio-based conversion.

Memory-intensive applications: The NVIDIA L20 wins decisively. Its 48 GB of GDDR6 memory with 864.0 GB/s bandwidth allows for larger datasets and faster data movement. The Intel card's 24 GB and 456.0 GB/s bandwidth represents half the capacity and just over half the bandwidth.

Ray tracing: The NVIDIA L20 has 92 RT cores versus 20 for the Intel card, a 4.6x advantage. This indicates substantially stronger ray tracing performance for rendering workloads that use this feature.

Texture and pixel throughput: The NVIDIA L20 delivers 927.4 GTexel/s versus 384.0 GTexel/s for Intel, and 322.6 GPixel/s versus 192.0 GPixel/s. These figures point to faster fill-rate-bound operations on the NVIDIA card.

Power efficiency: The NVIDIA L20 requires 275 W versus 400 W for the Intel card, while delivering far higher performance. The L20 also needs only a 600 W suggested PSU versus 800 W for Intel.

Physical footprint: The NVIDIA L20 is shorter at 267 mm versus 300 mm for the Intel card, which may matter in constrained chassis. Heights are nearly identical.

Benchmark validation: The NVIDIA L20 has recorded scores of 274,276 in Geekbench OpenCL and 228,018 in Geekbench Vulkan, placing it at the 99th percentile. The Intel Arc Pro B60 Dual has no recorded benchmark scores and sits at the 50th percentile.

Interface generation: The Intel card uses PCIe 5.0 x8, a newer standard than the NVIDIA card's PCIe 4.0 x16. The NVIDIA x16 width provides more physical lanes, though the Intel card's newer generation may offer higher per-lane bandwidth.

Display connectivity: The Intel Arc Pro B60 Dual provides 4x mini-DisplayPort 2.1 outputs, which support newer display standards than the NVIDIA L20's 4x DisplayPort 1.4a outputs.

Release timing: The Intel card is newer, released September 4, 2025, versus November 15, 2023 for NVIDIA. The NVIDIA card has an established lineage with a predecessor (Server Ampere) and successor (Server Hopper), while the Intel card has neither listed.