AMD Instinct MI300A vs Lisuan Tech LX ULTRA Comparison

Head-to-Head Benchmarks

The recorded database contains no executed benchmark runs for either the AMD Instinct MI300A or the Lisuan Tech LX ULTRA. Both entries show an average benchmark score of zero, and the head-to-head benchmark array is empty. Consequently, there are no frame rates, compute scores, or throughput measurements to compare directly. The absence of measured data means that any quantitative comparison must rely entirely on the architectural specifications recorded in the database.

The AMD Instinct MI300A lists a FP32 compute rate of 61.29 TFLOPS, while the Lisuan Tech LX ULTRA records 24.58 TFLOPS. This places the MI300A approximately 2.5 times higher in single-precision floating-point throughput. The MI300A also reports a texture rate of 1,915.2 GTexel/s versus 384.0 GTexel/s for the LX ULTRA, a difference of roughly 5 times. Pixel rate figures diverge sharply: the MI300A shows 0 MPixel/s, whereas the LX ULTRA delivers 192.0 GPixel/s. The MI300A’s zero pixel rate reflects its design as a compute-oriented module without rasterization outputs, while the LX ULTRA includes 96 ROPs and four display outputs.

Memory bandwidth presents another large gap. The MI300A uses 128 GB of HBM3 across an 8192-bit bus, yielding 5.32 TB/s. The LX ULTRA uses 24 GB of GDDR6 across a 192-bit bus, yielding 432.0 GB/s. The MI300A’s bandwidth is over 12 times higher. Clock speeds differ as well: the MI300A runs a base of 1000 MHz and boost of 2100 MHz on the GPU core, while the LX ULTRA has no recorded base or boost clock for the core. Memory clocks are 1300 MHz (5.2 Gbps effective) for the MI300A and 2250 MHz (18 Gbps effective) for the LX ULTRA.

Shading units count 14,592 on the MI300A versus 6,144 on the LX ULTRA. Texture mapping units number 912 versus 192. The MI300A has no ROPs, while the LX ULTRA has 96. No tensor core counts or ray tracing core counts are recorded for either product.

Where Each One Wins

Based on the recorded specifications, each product wins in distinct domains that align with its intended role.

The AMD Instinct MI300A dominates in raw compute throughput and memory capacity. Its FP32 performance of 61.29 TFLOPS is the highest recorded figure in this comparison. The 128 GB memory capacity is more than five times the LX ULTRA’s 24 GB, and the 5.32 TB/s bandwidth is more than twelve times the 432.0 GB/s. These attributes position the MI300A for large-scale scientific computing, dense linear algebra, and workloads where data residency and memory bandwidth are the primary constraints. The 153,000 million transistors on a 1017 mm² die, fabricated on a 5 nm TSMC process, indicate a design optimized for maximum compute density. The transistor density of 150.4M per mm² further underscores the aggressive integration.

The Lisuan Tech LX ULTRA wins in graphics pipeline capability and practical deployment flexibility. It has 96 ROPs and a pixel rate of 192.0 GPixel/s, while the MI300A records 0 MPixel/s and no display outputs. The LX ULTRA provides 4x DisplayPort 1.4a outputs, enabling direct video output. It supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.3, whereas the MI300A lists N/A for all three APIs. The LX ULTRA also has a dual-slot form factor and a 1x 16-pin power connector, with a suggested PSU of 550 W. The MI300A uses an OAM module form factor, has no power connectors, and suggests a 1150 W PSU. The LX ULTRA’s 225 W TDP is one third of the MI300A’s 750 W TDP, making it far more accessible in conventional chassis configurations.

The LX ULTRA also records FP16 throughput of 49.15 TFLOPS (2:1), a feature absent from the MI300A’s data. While the MI300A has higher FP32, the LX ULTRA’s FP16 mode could be relevant for mixed-precision workloads where supported. The LX ULTRA’s 6 nm TSMC process and 7G105 chip design with TrueGPU architecture indicate a different design philosophy: one that balances compute with rasterization and display functionality.

Architecture Differences

The two products diverge fundamentally at the architectural level, as recorded in the database.

The AMD Instinct MI300A is built on CDNA 3.0 architecture with the chip codename Aqua Vanjaram. It belongs to the Instinct (MIx) generation. The process node is 5 nm from TSMC. The die measures 1017 mm² and contains 153,000 million transistors, resulting in a transistor density of 150.4M per mm². The memory subsystem uses HBM3 with a 8192-bit bus and 128 GB capacity. The bus interface is PCIe 5.0 x16. The MI300A has no display outputs, no power connectors, and is designed as an OAM Module. Its TDP is 750 W, and the suggested PSU is 1150 W. The FP32 compute rate is 61.29 TFLOPS, texture rate is 1,915.2 GTexel/s, and pixel rate is 0 MPixel/s. The shading units are 14,592, TMUs are 912, and ROPs are 0. The release date is recorded as 2023-12-05. Its predecessor is listed as Radeon Instinct. No successor is recorded.

The Lisuan Tech LX ULTRA is built on TrueGPU architecture with the chip codename 7G105. It belongs to the 7G100 generation. The process node is 6 nm, also from TSMC. Transistor count and die size are listed as unknown. The memory subsystem uses GDDR6 with a 192-bit bus and 24 GB capacity. The bus interface is PCIe 4.0 x16. The LX ULTRA provides 4x DisplayPort 1.4a outputs. It has a dual-slot form factor, a 1x 16-pin power connector, and dimensions of 268 mm length, 112 mm height, and 40 mm width. Its TDP is 225 W, and the suggested PSU is 550 W. The FP32 compute rate is 24.58 TFLOPS, FP16 is 49.15 TFLOPS (2:1), texture rate is 384.0 GTexel/s, and pixel rate is 192.0 GPixel/s. The shading units are 6,144, TMUs are 192, and ROPs are 96. The release date is recorded as 2026-03-16. Production status is Active. No predecessor or successor is recorded.

The API support differs entirely. The MI300A lists DirectX, OpenGL, and Vulkan as N/A, reflecting its compute-only orientation. The LX ULTRA lists DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.3, indicating full graphics driver support. The MI300A’s memory clock is 1300 MHz (5.2 Gbps effective), while the LX ULTRA’s memory clock is 2250 MHz (18 Gbps effective). The MI300A’s base and boost clocks are 1000 MHz and 2100 MHz respectively, while the LX ULTRA has no recorded base or boost clocks.

Both entries show a percentile vs all GPUs of 50 and no nearest rivals, meaning the database has not yet positioned either product relative to other GPUs in terms of measured performance.

The Verdict

The data indicates two products with opposite design priorities. The AMD Instinct MI300A is a compute accelerator with no display path, no graphics API support, and a massive memory subsystem. Its 61.29 TFLOPS FP32, 5.32 TB/s bandwidth, and 128 GB capacity make it suitable for workloads that require sustained high-throughput arithmetic and large in-memory datasets. The 750 W TDP and OAM module form factor confirm that it is intended for server racks and specialized compute nodes, not desktop systems. The absence of pixel rate and ROPs reinforces this: the MI300A does not rasterize.

The Lisuan Tech LX ULTRA is a conventional graphics card with a display output, full graphics API support, and a standard dual-slot form factor. Its 24.58 TFLOPS FP32 and 49.15 TFLOPS FP16 provide capable compute, while 192.0 GPixel/s pixel rate and 96 ROPs enable rasterization. The 24 GB GDDR6 memory with 432.0 GB/s bandwidth is sufficient for graphics workloads and moderate compute tasks. The 225 W TDP and 550 W suggested PSU make it deployable in standard consumer or workstation builds. The PCIe 4.0 x16 interface is compatible with most modern platforms.

The database records no benchmark scores, so a direct performance verdict is impossible. The specification comparison, however, shows that the MI300A is decisively ahead in FP32 throughput, memory bandwidth, memory capacity, texture rate, and transistor count. The LX ULTRA is ahead in pixel rate, ROP count, display outputs, API support, power efficiency (225 W versus 750 W), and physical compatibility with standard chassis.

A user or system integrator selecting between these two must first determine whether the workload requires graphics output and rasterization. If yes, the LX ULTRA is the only viable option, as the MI300A has no display outputs and no graphics APIs. If the workload is purely compute and memory-bound, the MI300A’s recorded specifications indicate a far higher ceiling for arithmetic throughput and data movement, though at the cost of a higher power draw and a non-standard module form factor.

The production status of the LX ULTRA is recorded as Active, while the MI300A’s production status is null. The MI300A’s release date of 2023-12-05 precedes the LX ULTRA’s 2026-03-16 by more than two years. No pricing data is recorded for either product.

FAQ

Q: Which product has higher FP32 compute performance?

A: The AMD Instinct MI300A records 61.29 TFLOPS FP32, while the Lisuan Tech LX ULTRA records 24.58 TFLOPS FP32. The MI300A is approximately 2.5 times higher.

Q: Does the AMD Instinct MI300A support display output?

A: No. The MI300A lists display outputs as “No outputs” and has a pixel rate of 0 MPixel/s. The Lisuan Tech LX ULTRA provides 4x DisplayPort 1.4a outputs and a pixel rate of 192.0 GPixel/s.

Q: What are the memory capacities and types of each product?

A: The MI300A has 128 GB of HBM3 with an 8192-bit bus and 5.32 TB/s bandwidth. The LX ULTRA has 24 GB of GDDR6 with a 192-bit bus and 432.0 GB/s bandwidth.

Q: Which product supports DirectX and Vulkan?

A: The Lisuan Tech LX ULTRA supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.3. The AMD Instinct MI300A lists all three APIs as N/A.

Q: What is the power draw difference between the two?

A: The MI300A has a TDP of 750 W and suggests a 1150 W PSU. The LX ULTRA has a TDP of 225 W and suggests a 550 W PSU.

Q: What form factors do they use?

A: The MI300A is an OAM Module with no power connectors. The LX ULTRA is a dual-slot card with a 1x 16-pin power connector and dimensions of 268 mm length, 112 mm height, and 40 mm width.