Intel Data Center GPU Max Subsystem vs Lisuan Tech LX ULTRA Comparison

Head-to-Head Benchmarks

The database contains no recorded benchmark scores for either the Intel Data Center GPU Max Subsystem or the Lisuan Tech LX ULTRA. Both entries show zero wins in head-to-head testing, and their average benchmark scores are zero. Their percentile rankings against all GPUs are identical at 50, placing both squarely in the middle of the database distribution despite their vastly different designs.

What the recorded data does show is a clear split in raw computational capability. The Intel Data Center GPU Max Subsystem delivers 52.43 TFLOPS of FP32 performance and 52.43 TFLOPS of FP16 performance with a 1:1 ratio. The Lisuan Tech LX ULTRA delivers 24.58 TFLOPS of FP32, which is roughly half of Intel's figure, but reaches 49.15 TFLOPS in FP16 through a 2:1 ratio. That means Intel's FP32 throughput is approximately 2.1 times higher than Lisuan's, while Intel's FP16 output is only about 1.07 times higher. The FP16 gap is narrow because Lisuan's architecture doubles its FP32 rate when operating in FP16 mode, while Intel's architecture processes both formats at the same rate.

Memory bandwidth tells a similar story of Intel dominance. The Intel subsystem carries 128 GB of HBM2e memory across an 8192-bit bus, producing 3.21 TB/s of bandwidth. The Lisuan card uses 24 GB of GDDR6 on a 192-bit bus, yielding 432.0 GB/s. Intel's bandwidth advantage is roughly 7.4 times that of Lisuan. Texture rate follows the same pattern: Intel reaches 1,638.4 GTexel/s against Lisuan's 384.0 GTexel/s, a 4.3 times difference. Pixel rate, however, is a category where Lisuan wins outright: Intel records 0 MPixel/s because it has no ROPs, while Lisuan delivers 192.0 GPixel/s from its 96 ROPs.

Where Each One Wins

The Intel Data Center GPU Max Subsystem wins in every compute-heavy category recorded in the database. Its FP32 throughput of 52.43 TFLOPS nearly doubles Lisuan's 24.58 TFLOPS, making it the stronger choice for workloads that rely on single-precision math, such as scientific simulation, physics processing, or general data-center compute tasks. Its FP16 performance of 52.43 TFLOPS edges out Lisuan's 49.15 TFLOPS, so even in mixed-precision AI workloads where Lisuan's 2:1 FP16 advantage helps, Intel still holds a slight lead. Memory capacity is another decisive win: 128 GB versus 24 GB means Intel can hold far larger datasets in local memory, while Lisuan's smaller pool requires more frequent host-side transfers. Bandwidth follows capacity, with Intel's 3.21 TB/s dwarfing Lisuan's 432.0 GB/s, which matters for memory-bound kernels like large matrix operations or graph analytics.

The Lisuan Tech LX ULTRA wins in categories tied to rendering and display output. It has 96 ROPs producing 192.0 GPixel/s, while Intel has zero ROPs and outputs no pixels. The Lisuan card also offers 4x DisplayPort 1.4a outputs, making it usable for direct display connection, while Intel provides no display outputs at all. In API support, Lisuan lists DirectX 12 Ultimate (12_2) and Vulkan 1.3, while Intel lists DirectX 12 (12_1) and no Vulkan entry. That gives Lisuan a feature-level advantage for graphics workloads, particularly ray tracing or modern game engines that require DX12 Ultimate features. Lisuan's 192 texture units deliver 384.0 GTexel/s, which is lower than Intel's 1,024 texture units at 1,638.4 GTexel/s, so Intel retains the texture-processing crown, but Lisuan's architecture is clearly oriented toward rasterization and frame output rather than pure compute density.

FAQ

Q: Which GPU has higher FP32 performance?

A: The Intel Data Center GPU Max Subsystem delivers 52.43 TFLOPS of FP32, which is about 2.1 times the Lisuan Tech LX ULTRA's 24.58 TFLOPS.

Q: Can either card output video to a display?

A: Only the Lisuan Tech LX ULTRA can. It provides 4x DisplayPort 1.4a outputs and has a pixel rate of 192.0 GPixel/s. The Intel subsystem has no display outputs and a pixel rate of 0 MPixel/s.

Q: What is the memory configuration difference?

A: Intel uses 128 GB of HBM2e on an 8192-bit bus with 3.21 TB/s bandwidth. Lisuan uses 24 GB of GDDR6 on a 192-bit bus with 432.0 GB/s bandwidth.

Q: How do the FP16 rates compare?

A: Intel achieves 52.43 TFLOPS FP16 (1:1 ratio), while Lisuan achieves 49.15 TFLOPS FP16 (2:1 ratio). Intel is about 1.07 times faster in FP16.

Q: What are the power requirements?

A: Intel lists a TDP of 2400 W with a suggested PSU of 2800 W. Lisuan lists a TDP of 225 W with a suggested PSU of 550 W.

Q: Which card supports newer graphics APIs?

A: Lisuan supports DirectX 12 Ultimate (12_2) and Vulkan 1.3. Intel supports DirectX 12 (12_1) and does not list Vulkan support in the database.

Specification Differences

The two cards differ across nearly every specification field in the database. The Intel Data Center GPU Max Subsystem uses a 10 nm process from Intel's own foundry, while the Lisuan Tech LX ULTRA uses a 6 nm process from TSMC. Intel's die size is 1280 mm² with 100,000 million transistors, giving a transistor density of 78.1M per mm². Lisuan's die size and transistor count are listed as unknown, with no density figure recorded.

Clock behavior splits sharply. Intel lists a base clock of 900 MHz and a boost clock of 1600 MHz. Lisuan lists no base or boost clock at all, only a memory clock of 2250 MHz with 18 Gbps effective data rate. Intel's memory clock is 1565 MHz with 3.1 Gbps effective. Memory type also differs: Intel uses HBM2e, Lisuan uses GDDR6. Bus width is 8192 bits for Intel versus 192 bits for Lisuan.

Shading units favor Intel at 16,384 against Lisuan's 6,144. Texture mapping units are 1,024 versus 192. Raster output units are 0 for Intel and 96 for Lisuan. Intel lists 128 ray tracing cores; Lisuan lists none. Pixel rate is 0 MPixel/s for Intel and 192.0 GPixel/s for Lisuan. Texture rate is 1,638.4 GTexel/s versus 384.0 GTexel/s. FP32 is 52.43 TFLOPS versus 24.58 TFLOPS, and FP16 is 52.43 TFLOPS (1:1) versus 49.15 TFLOPS (2:1).

Power and interface also differ. Intel has a TDP of 2400 W and a suggested PSU of 2800 W. Lisuan has a TDP of 225 W and a suggested PSU of 550 W. Both use a single 16-pin power connector and a dual-slot cooler. Intel runs on PCIe 5.0 x16; Lisuan runs on PCIe 4.0 x16. Intel has no display outputs; Lisuan has 4x DisplayPort 1.4a. API support: Intel lists DirectX 12 (12_1) and OpenGL 4.6, with no Vulkan entry. Lisuan lists DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.3.

Physical dimensions are close in length: Intel is 267 mm (10.5 inches), Lisuan is 268 mm (10.6 inches). Lisuan also records a height of 112 mm (4.4 inches) and width of 40 mm (1.6 inches); Intel does not list height or width. Production status is Active for both. Intel's release date is January 2023, while Lisuan's is March 2026. Intel lists a successor, H3C Graphics, in the database; Lisuan has no successor listed.

Architecture Differences

The Intel Data Center GPU Max Subsystem is built on the Ponte Vecchio chip, which uses Intel's Generation 12.5 architecture. The database classifies its generation as "Data Center GPU (Ponte Vecchio)." The Lisuan Tech LX ULTRA uses the 7G105 chip under an architecture called "TrueGPU," from the 7G100 generation. These are fundamentally different design philosophies: Intel's is a massive data-center accelerator with 16,384 shading units and 128 ray tracing cores, while Lisuan's is a more conventional GPU with 6,144 shading units and no dedicated ray tracing cores.

Process technology diverges as well. Intel uses a 10 nm node at its own foundry, while Lisuan uses a 6 nm node at TSMC. Intel's transistor count is listed at 100,000 million across a 1280 mm² die, which the database calculates as 78.1M transistors per mm². Lisuan's transistor count and die size are unknown, so no density comparison is possible. The smaller process node for Lisuan suggests a more modern manufacturing approach, but the database does not provide enough data to confirm a density advantage.

Memory architecture reflects their different roles. Intel uses HBM2e across an 8192-bit interface, which is a wide, high-bandwidth design typical of server accelerators. Lisuan uses GDDR6 on a 192-bit bus, a narrower configuration common in consumer graphics cards. The 128 GB capacity on Intel versus 24 GB on Lisuan further separates their target workloads: Intel's memory pool is built for large model residency, while Lisuan's is sized for typical graphics frame buffers and moderate compute tasks.

Feature sets differ in key ways. Intel lists 128 ray tracing cores, while Lisuan lists none, suggesting Intel has dedicated hardware for ray traversal that Lisuan lacks. Lisuan, however, supports DirectX 12 Ultimate and Vulkan 1.3, while Intel only lists DirectX 12 (12_1) and no Vulkan. The presence of display outputs on Lisuan and their absence on Intel reinforces that Lisuan is designed for direct rendering to screens, while Intel is a compute-only accelerator that requires external handling for any visual output.

The Verdict

The Intel Data Center GPU Max Subsystem is the clear choice for raw compute density and large-scale data processing. Its FP32 output of 52.43 TFLOPS is more than double Lisuan's 24.58 TFLOPS, and its 128 GB HBM2e pool with 3.21 TB/s bandwidth provides a memory capacity and throughput that Lisuan's 24 GB GDDR6 at 432.0 GB/s cannot approach. For workloads like scientific computing, large model inference, or any task where memory footprint and bandwidth dominate, the data shows Intel's accelerator is in a different class. Its 2400 W TDP and 2800 W suggested PSU also indicate it is designed for rack-mounted data center environments, not desktop use.

The Lisuan Tech LX ULTRA is the only one of the two that can function as a display-capable graphics card. It provides 4x DisplayPort 1.4a outputs, 96 ROPs, and a pixel rate of 192.0 GPixel/s, while Intel has none of those. Its DirectX 12 Ultimate and Vulkan 1.3 support gives it modern graphics API coverage that Intel lacks. The 225 W TDP and 550 W suggested PSU make it practical for standard PC builds, and its 268 mm length fits typical tower cases. For any application that requires rendering to a screen, gaming, or graphics workstation use, Lisuan is the only viable option from the recorded data.

The choice between them is therefore straightforward based on the database. If the workload is compute-only, runs in a server context, and needs massive memory and bandwidth, the Intel Data Center GPU Max Subsystem dominates every relevant metric. If the workload involves frame output, display connectivity, or modern graphics APIs, the Lisuan Tech LX ULTRA is the sole candidate because Intel offers no display path at all. The two cards do not compete in the same market segment; they serve different physical and logical roles, and the benchmark data, while sparse, confirms that split.