NVIDIA Jetson Orin Nano Super vs Lisuan Tech LX PRO Comparison

Head-to-Head Benchmarks

The recorded database contains no direct head-to-head benchmark results for the NVIDIA Jetson Orin Nano Super and the Lisuan Tech LX PRO. The wins tally is zero for both entries, and the average benchmark score for each product is also zero. This absence of measured data means the comparative analysis must rely entirely on the documented specification sheets and architectural profiles.

The most striking delta between the two parts is raw compute throughput. The Lisuan Tech LX PRO delivers 24.58 TFLOPS of FP32 performance, while the NVIDIA Jetson Orin Nano Super provides 2.089 TFLOPS. That places the LX PRO at roughly 11.8 times the FP32 output of the Orin Nano Super. In FP16, the LX PRO reaches 49.15 TFLOPS, versus 4.178 TFLOPS for the Orin Nano Super, a factor of approximately 11.8 as well. These are not close figures; they represent entirely different performance tiers.

Memory bandwidth tells a similar story. The LX PRO has 432.0 GB/s of bandwidth from its 192-bit GDDR6 bus, while the Orin Nano Super has 102.4 GB/s from its 128-bit LPDDR5 bus. The LX PRO offers about 4.2 times the memory bandwidth. The pixel rate for the LX PRO is 192.0 GPixel/s, compared to 16.32 GPixel/s for the Orin Nano Super, a factor of 11.8. Texture rate shows the same ratio: 384.0 GTexel/s versus 32.64 GTexel/s, again an 11.8 times advantage for the LX PRO.

The shading resources are equally lopsided. The LX PRO has 6144 shading units, 192 texture mapping units, and 96 ROPs. The Orin Nano Super has 1024 shading units, 32 TMUs, and 16 ROPs. These are proportional scaling factors of 6.0, 6.0, and 6.0 respectively, which is consistent with the architecture design. The Tensor Core count is the one area where the Orin Nano Super has a documented feature that the LX PRO lacks entirely; the LX PRO has no tensor cores listed, while the Orin Nano Super has 32.

Architecture Differences

The two products come from different foundries and process nodes. The NVIDIA Jetson Orin Nano Super uses an 8 nm process at Samsung, while the Lisuan Tech LX PRO uses a 6 nm process at TSMC. The Orin Nano Super is based on the GA10B chip, which is part of the Ampere architecture within the Tegra generation. The LX PRO uses the 7G105 chip, built on the TrueGPU architecture within the 7G100 generation.

Die size is documented for the Orin Nano Super at 200 mm², but unknown for the LX PRO. Transistor counts are unknown for both, and transistor density is not recorded for either. The power envelope differs substantially: the Orin Nano Super has a TDP of 25 W, while the LX PRO has a TDP of 225 W, nine times higher. This explains the thermal solution differences, the Orin Nano Super is an IGP (integrated graphics processor) with board dimensions of 70 mm by 45 mm, whereas the LX PRO is a dual-slot discrete card measuring 248 mm by 118 mm by 48 mm.

Memory technology is a fundamental split. The Orin Nano Super uses 8 GB of LPDDR5 with a 128-bit bus and memory clocked at 800 MHz (6.4 Gbps effective). The LX PRO uses 24 GB of GDDR6 with a 192-bit bus and memory clocked at 2250 MHz (18 Gbps effective). The capacity difference is 3 times, and the bus width difference is 1.5 times, but the effective memory clock is 2.8 times higher on the LX PRO, which combines to produce the 4.2 times bandwidth advantage.

The PCIe interface also separates them. The Orin Nano Super uses PCIe 4.0 x4, while the LX PRO uses PCIe 4.0 x16. The LX PRO requires a 1x 16-pin power connector and a suggested 550 W power supply, while the Orin Nano Super has no power connector listed and no suggested PSU. Display outputs differ as well: the Orin Nano Super's displays are portable device dependent, while the LX PRO has 4x DisplayPort 1.4a.

Both support DirectX 12 Ultimate (12_2) and OpenGL 4.6. The Vulkan API version differs slightly: the Orin Nano Super supports Vulkan 1.4, while the LX PRO supports Vulkan 1.3. Neither product lists ray tracing cores. Release dates are far apart: the Orin Nano Super launched on 2024-12-16, while the LX PRO is dated 2026-03-16, suggesting a later arrival for the Lisuan part.

FAQ

Q: Which product has higher FP32 compute performance?

A: The Lisuan Tech LX PRO has 24.58 TFLOPS of FP32 performance, while the NVIDIA Jetson Orin Nano Super has 2.089 TFLOPS. The LX PRO delivers approximately 11.8 times the FP32 throughput.

Q: How do the memory capacities compare?

A: The LX PRO has 24 GB of GDDR6 memory, while the Orin Nano Super has 8 GB of LPDDR5. The LX PRO also has a wider 192-bit bus compared to the 128-bit bus on the Orin Nano Super, resulting in 432.0 GB/s versus 102.4 GB/s of bandwidth.

Q: Which product has tensor cores?

A: Only the NVIDIA Jetson Orin Nano Super has tensor cores, with 32 units documented. The Lisuan Tech LX PRO has no tensor core count listed in the database.

Q: What are the power consumption differences?

A: The Orin Nano Super has a TDP of 25 W and is an IGP with no power connector. The LX PRO has a TDP of 225 W, uses a dual-slot design, requires a 1x 16-pin power connector, and has a suggested 550 W power supply.

Q: Which product supports a newer Vulkan version?

A: The NVIDIA Jetson Orin Nano Super supports Vulkan 1.4, while the Lisuan Tech LX PRO supports Vulkan 1.3. Both support DirectX 12 Ultimate (12_2) and OpenGL 4.6.

Q: What are the physical size differences?

A: The Orin Nano Super measures 70 mm by 45 mm and is an integrated part, while the LX PRO is a dual-slot card measuring 248 mm by 118 mm by 48 mm.

Specification Differences

| Specification | NVIDIA Jetson Orin Nano Super | Lisuan Tech LX PRO |

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

| Chip | GA10B | 7G105 |

| Architecture | Ampere | TrueGPU |

| Generation | Tegra (Ampere) | 7G100 |

| Process Node | 8 nm | 6 nm |

| Foundry | Samsung | TSMC |

| Die Size | 200 mm² | Unknown |

| Memory Clock | 800 MHz 6.4 Gbps effective | 2250 MHz 18 Gbps effective |

| Memory Size | 8 GB | 24 GB |

| Memory Type | LPDDR5 | GDDR6 |

| Memory Bus Width | 128 bit | 192 bit |

| Memory Bandwidth | 102.4 GB/s | 432.0 GB/s |

| Shading Units | 1024 | 6144 |

| TMUs | 32 | 192 |

| ROPs | 16 | 96 |

| Tensor Cores | 32 | None listed |

| Pixel Rate | 16.32 GPixel/s | 192.0 GPixel/s |

| Texture Rate | 32.64 GTexel/s | 384.0 GTexel/s |

| FP32 | 2.089 TFLOPS | 24.58 TFLOPS |

| FP16 | 4.178 TFLOPS (2:1) | 49.15 TFLOPS (2:1) |

| TDP | 25 W | 225 W |

| Slot Width | IGP | Dual-slot |

| Power Connectors | None | 1x 16-pin |

| Suggested PSU | None | 550 W |

| Bus Interface | PCIe 4.0 x4 | PCIe 4.0 x16 |

| Display Outputs | Portable Device Dependent | 4x DisplayPort 1.4a |

| Vulkan | 1.4 | 1.3 |

| Dimensions | 70 mm x 45 mm | 248 mm x 118 mm x 48 mm |

| Release Date | 2024-12-16 | 2026-03-16 |

| Launch MSRP | 249 USD | None |

The Verdict

The data shows two products engineered for different purposes. The NVIDIA Jetson Orin Nano Super is a low-power integrated solution with a 25 W TDP, a 70 mm by 45 mm footprint, and a launch MSRP of 249 USD. Its 32 tensor cores and Vulkan 1.4 support indicate a design focus on embedded AI inference and edge deployment where power and space are constrained. The FP16 to FP32 ratio of 2:1 (4.178 TFLOPS versus 2.089 TFLOPS) suggests the architecture is tuned for mixed-precision workloads.

The Lisuan Tech LX PRO is a high-throughput discrete GPU with 24.58 TFLOPS FP32, 49.15 TFLOPS FP16, and 432.0 GB/s of memory bandwidth. Its 225 W TDP, dual-slot cooling requirement, and 550 W suggested PSU place it firmly in the desktop workstation or high-performance compute segment. The absence of tensor cores means it lacks dedicated AI acceleration hardware, but its raw shader throughput and 24 GB of GDDR6 memory make it suitable for large rendering tasks or general compute.

Benchmark results indicate no direct performance comparison exists in the database, so the verdict rests on the specification deltas. The Orin Nano Super wins on power efficiency, integration, tensor core availability, and API version currency. The LX PRO wins decisively on every raw compute metric: FP32, FP16, pixel rate, texture rate, memory bandwidth, memory capacity, and shading resources. The percentile ranking of both is 50, meaning neither is positioned above or below the other in aggregate database standing, but that figure likely reflects the absence of benchmark scores rather than actual equivalence.

Where Each One Wins

NVIDIA Jetson Orin Nano Super wins on:

• Tensor core presence: 32 tensor cores for AI workloads, which the LX PRO does not list.
• Power efficiency: 25 W TDP versus 225 W, a 9 times lower power draw.
• Physical integration: IGP form factor at 70 mm by 45 mm, suitable for embedded systems.
• API currency: Vulkan 1.4 support versus Vulkan 1.3 on the LX PRO.
• Launch MSRP: documented at 249 USD, though no pricing comparison is available for the LX PRO since it has no launch MSRP listed.
• Release timing: earlier availability (2024-12-16) versus the LX PRO's 2026-03-16 date.

Lisuan Tech LX PRO wins on:

• Raw FP32 throughput: 24.58 TFLOPS, approximately 11.8 times the Orin Nano Super's 2.089 TFLOPS.
• Raw FP16 throughput: 49.15 TFLOPS, approximately 11.8 times the Orin Nano Super's 4.178 TFLOPS.
• Memory bandwidth: 432.0 GB/s, approximately 4.2 times the Orin Nano Super's 102.4 GB/s.
• Memory capacity: 24 GB versus 8 GB, a 3 times advantage.
• Pixel fill rate: 192.0 GPixel/s versus 16.32 GPixel/s, an 11.8 times advantage.
• Texture fill rate: 384.0 GTexel/s versus 32.64 GTexel/s, an 11.8 times advantage.
• Shading resources: 6144 shading units, 192 TMUs, and 96 ROPs versus 1024, 32, and 16 respectively.
• Bus interface: PCIe 4.0 x16 versus PCIe 4.0 x4, offering 4 times the lane count.
• Display outputs: 4x DisplayPort 1.4a versus portable device dependent outputs.
• Process node: 6 nm TSMC versus 8 nm Samsung, indicating a more advanced fabrication process.

The use-case split follows directly from these numbers. The Orin Nano Super is the choice for compact, power-limited edge devices that need tensor acceleration for neural network inference. The LX PRO is the choice for compute-heavy workloads such as large-scale rendering, scientific simulation, or any task that demands maximum FP32 and FP16 throughput with generous memory capacity. The data does not support a single winner across all scenarios; the selection depends on whether the priority is tensor-accelerated efficiency or raw shader compute.