AMD Instinct MI350X vs Lisuan Tech LX ULTRA Comparison

Head-to-Head Benchmarks

The recorded data shows no direct benchmark comparisons between the AMD Instinct MI350X and the Lisuan Tech LX ULTRA. Neither product has benchmark scores, average scores, or percentile rankings that differentiate them in the database; both sit at the 50th percentile versus all GPUs with an average benchmark score of 0. This absence of measured performance data means the comparison must rely entirely on architectural specifications and theoretical throughput figures.

The most substantial numerical advantage belongs to the AMD Instinct MI350X in raw compute throughput. The MI350X delivers 72.09 TFLOPS of FP32 performance, which is 2.93 times the LX ULTRA's 24.58 TFLOPS. In FP16 compute, the MI350X again posts 72.09 TFLOPS with a 1:1 ratio, while the LX ULTRA achieves 49.15 TFLOPS with a 2:1 ratio, giving the AMD part a 46.6% lead in half-precision work. The texture rate further separates the two: the MI350X reaches 2,252.8 GTexel/s versus 384.0 GTexel/s for the LX ULTRA, a 5.87x margin.

Memory capacity and bandwidth heavily favor the MI350X as well. The AMD accelerator carries 288 GB of HBM3e across an 8192-bit bus, producing 8.19 TB/s of bandwidth. The LX ULTRA has 24 GB of GDDR6 on a 192-bit bus, yielding 432.0 GB/s. The bandwidth difference is 18.96x in favor of the MI350X, while the capacity difference is 12x. The LX ULTRA does counter in pixel rate, posting 192.0 GPixel/s against the MI350X's 0 MPixel/s, a consequence of the AMD part having no ROPs (0) and being designed without a rasterization pipeline.

Clock speeds tell a mixed story. The MI350X has a base clock of 1000 MHz and a boost clock of 2200 MHz, while the LX ULTRA has no listed base or boost clocks in the database. Memory clocks differ significantly: the MI350X runs 2000 MHz with 8 Gbps effective, while the LX ULTRA runs 2250 MHz with 18 Gbps effective. The LX ULTRA's faster memory clock does not compensate for its narrower bus and smaller capacity.

Architecture Differences

The MI350X uses the CDNA 4.0 architecture, built for Instinct (MIx) generation accelerators, fabricated on a 3 nm TSMC process. The chip is designated MI350 256CU and contains 185,000 million transistors on a 2380 mm² die, resulting in a transistor density of 77.7M per mm². The LX ULTRA uses the TrueGPU architecture from the 7G100 generation, fabricated on a 6 nm TSMC process with the 7G105 chip. Transistor count and die size for the LX ULTRA are listed as unknown.

Shading unit counts differ: the MI350X has 16,384 shading units and 1,024 TMUs, while the LX ULTRA has 6,144 shading units, 192 TMUs, and 96 ROPs. The MI350X has zero ROPs, consistent with a compute-focused accelerator that has no display outputs. The LX ULTRA includes 4x DisplayPort 1.4a outputs and supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.3. The MI350X lists N/A for DirectX, OpenGL, and Vulkan, indicating no graphics API support.

Memory architecture fundamentally differs. The MI350X uses HBM3e with 288 GB capacity, 8192-bit bus width, and 8.19 TB/s bandwidth. The LX ULTRA uses GDDR6 with 24 GB capacity, 192-bit bus width, and 432.0 GB/s bandwidth. The MI350X memory clock is 2000 MHz (8 Gbps effective), while the LX ULTRA memory clock is 2250 MHz (18 Gbps effective). The MI350X's bandwidth advantage comes from the extremely wide HBM interface rather than higher clock speeds.

Process node and physical design also diverge. The MI350X is a 3 nm part from TSMC, while the LX ULTRA is a 6 nm part from the same foundry. The MI350X measures 102 mm in length and 165 mm in width, mounted as an OAM Module. The LX ULTRA measures 268 mm in length, 112 mm in height, and 40 mm in width, occupying a dual-slot form factor. The MI350X has no power connectors listed, while the LX ULTRA uses a single 16-pin connector. Suggested PSU ratings are 1400 W for the MI350X and 550 W for the LX ULTRA. The MI350X TDP is 1000 W, while the LX ULTRA TDP is 225 W.

The bus interfaces differ as well: PCIe 5.0 x16 for the MI350X and PCIe 4.0 x16 for the LX ULTRA. The MI350X has no display outputs, while the LX ULTRA has four DisplayPort 1.4a outputs. The LX ULTRA production status is Active, while the MI350X production status is not listed.

Where Each One Wins

The MI350X wins decisively in compute throughput, memory capacity, memory bandwidth, texture processing, and transistor count. Its 72.09 TFLOPS FP32 and 72.09 TFLOPS FP16 (1:1) figures position it for heavy numerical workloads like large-scale matrix operations, scientific simulation, and AI training, where the 288 GB HBM3e pool and 8.19 TB/s bandwidth allow massive datasets to remain resident on the accelerator. The 5.87x texture rate advantage and 16,384 shading units indicate a part engineered for sustained compute density rather than interactive rendering.

The LX ULTRA wins in pixel throughput with 192.0 GPixel/s versus 0 MPixel/s, in graphics API support with DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.3, and in display connectivity with four DisplayPort 1.4a outputs. The 24 GB GDDR6 memory and 432.0 GB/s bandwidth are substantially smaller but sufficient for graphics-oriented tasks, and the 49.15 TFLOPS FP16 (2:1) figure shows it can accelerate half-precision workloads, though at a lower absolute level than the MI350X.

The LX ULTRA also wins on power efficiency from the recorded specifications. Its 225 W TDP against the MI350X's 1000 W TDP means the LX ULTRA draws 22.5% of the power while delivering 34.1% of the FP32 throughput (24.58 TFLOPS versus 72.09 TFLOPS). In FP16, the LX ULTRA delivers 68.2% of the MI350X's throughput (49.15 versus 72.09 TFLOPS) at 22.5% of the power. The 550 W suggested PSU for the LX ULTRA versus 1400 W for the MI350X further reflects the lower system integration cost, though pricing is not discussed here.

The LX ULTRA has a more conventional physical footprint at 268 mm length, 112 mm height, and 40 mm width in a dual-slot design, with a single 16-pin power connector. The MI350X uses an OAM Module form factor with no power connectors listed, indicating a different integration path, likely requiring a carrier board or system-level design rather than standalone installation.

The Verdict

The data indicates two products with opposing design goals. The AMD Instinct MI350X is a high-throughput accelerator with 72.09 TFLOPS FP32, 288 GB HBM3e, 8.19 TB/s bandwidth, and 185,000 million transistors on a 3 nm process. The Lisuan Tech LX ULTRA is a graphics-capable card with 24.58 TFLOPS FP32, 24 GB GDDR6, 432.0 GB/s bandwidth, and full graphics API support.

For compute-bound workloads requiring maximum memory capacity and bandwidth, the MI350X is the clear choice based on the recorded specifications. Its 8.19 TB/s bandwidth is 18.96x the LX ULTRA's 432.0 GB/s, and its 288 GB capacity is 12x larger. The 72.09 TFLOPS FP32 figure is 2.93x the LX ULTRA's 24.58 TFLOPS. The 1000 W TDP and 1400 W suggested PSU indicate a system designed for dense compute environments, not desktop use.

For graphics rendering, display output, or workloads requiring DirectX 12 Ultimate, OpenGL 4.6, or Vulkan 1.3, the LX ULTRA is the only viable option between the two, as the MI350X lists N/A for all graphics APIs and has no display outputs. The LX ULTRA's 192.0 GPixel/s pixel rate, 96 ROPs, and four DisplayPort 1.4a outputs make it functionally a graphics card, while the MI350X cannot output video.

The LX ULTRA also serves scenarios where power draw must stay moderate. At 225 W TDP with a 550 W suggested PSU, it integrates into conventional systems, whereas the MI350X at 1000 W TDP with a 1400 W suggested PSU requires specialized power delivery. The LX ULTRA's PCIe 4.0 x16 interface is more common in existing platforms, while the MI350X uses PCIe 5.0 x16, which requires newer host support.

Release timing differs: the MI350X has a release date of 2025-06-11, while the LX ULTRA has a release date of 2026-03-16. The MI350X predecessor is listed as Radeon Instinct, and the LX ULTRA has no listed predecessor. Neither product has a launch MSRP in the database.

FAQ

Q: Which product has higher FP32 compute throughput?

A: The AMD Instinct MI350X posts 72.09 TFLOPS FP32, which is 2.93x the Lisuan Tech LX ULTRA's 24.58 TFLOPS.

Q: What is the memory bandwidth difference between the two?

A: The MI350X delivers 8.19 TB/s from 288 GB of HBM3e on an 8192-bit bus. The LX ULTRA delivers 432.0 GB/s from 24 GB of GDDR6 on a 192-bit bus. The MI350X bandwidth is 18.96x higher.

Q: Does either product support graphics APIs?

A: Only the LX ULTRA supports graphics APIs, listing DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.3. The MI350X lists N/A for all three.

Q: What are the power requirements?

A: The MI350X has a TDP of 1000 W with a suggested PSU of 1400 W. The LX ULTRA has a TDP of 225 W with a suggested PSU of 550 W.

Q: Can either product output to displays?

A: The LX ULTRA has 4x DisplayPort 1.4a outputs. The MI350X has no display outputs.

Q: What process nodes are used?

A: The MI350X uses a 3 nm TSMC process with 185,000 million transistors on a 2380 mm² die. The LX ULTRA uses a 6 nm TSMC process with unknown transistor count and die size.

Specification Differences

| Field | AMD Instinct MI350X | Lisuan Tech LX ULTRA |

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

| Chip | MI350 256CU | 7G105 |

| Architecture | CDNA 4.0 | TrueGPU |

| Generation | Instinct (MIx) | 7G100 |

| Process Node | 3 nm | 6 nm |

| Foundry | TSMC | TSMC |

| Transistors | 185,000 million | unknown |

| Die Size | 2380 mm² | unknown |

| Transistor Density | 77.7M / mm² | null |

| Base Clock | 1000 MHz | null |

| Boost Clock | 2200 MHz | null |

| Memory Clock | 2000 MHz 8 Gbps effective | 2250 MHz 18 Gbps effective |

| Memory Size | 288 GB | 24 GB |

| Memory Type | HBM3e | GDDR6 |

| Memory Bus Width | 8192 bit | 192 bit |

| Memory Bandwidth | 8.19 TB/s | 432.0 GB/s |

| Shading Units | 16384 | 6144 |

| TMUs | 1024 | 192 |

| ROPs | 0 | 96 |

| Pixel Rate | 0 MPixel/s | 192.0 GPixel/s |

| Texture Rate | 2,252.8 GTexel/s | 384.0 GTexel/s |

| FP32 | 72.09 TFLOPS | 24.58 TFLOPS |

| FP16 | 72.09 TFLOPS (1:1) | 49.15 TFLOPS (2:1) |

| TDP | 1000 W | 225 W |

| Slot Width | OAM Module | Dual-slot |

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

| Suggested PSU | 1400 W | 550 W |

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

| Display Outputs | No outputs | 4x DisplayPort 1.4a |

| DirectX | N/A | 12 Ultimate (12_2) |

| OpenGL | N/A | 4.6 |

| Vulkan | N/A | 1.3 |

| Length | 102 mm 4 inches | 268 mm 10.6 inches |

| Height | null | 112 mm 4.4 inches |

| Width | 165 mm 6.5 inches | 40 mm 1.6 inches |

| Production Status | null | Active |

| Release Date | 2025-06-11T17:00:00.000Z | 2026-03-16T17:00:00.000Z |

| Predecessor | Radeon Instinct | null |