Lisuan Tech LX ULTRA

Lisuan Tech LX ULTRA is a 6 nm TrueGPU part from the 7G100 generation, built by TSMC and featuring 6144 shading units, 24 GB of GDDR6 memory on a 192-bit bus, and a 225 W TDP. It targets the upper mid-range of the current GPU landscape, sitting at the 50th percentile against all GPUs in the database, though its average benchmark score is listed as zero, indicating that standardized performance data is still being collected. Below is an analysis based on its architectural specifications and memory subsystem, with comparisons drawn only from its known position in the market.

How It Compares

The LX ULTRA holds a unique position in the database because its `nearestRivals` array is empty. This means there are no direct performance deltas or specific rival names provided to benchmark against. Consequently, the analysis must rely on its percentile rank and raw specifications.

Without direct rival data, the LX ULTRA's 50th percentile placement suggests it sits exactly at the median of all tracked GPUs. This implies that half of the GPUs in the database are faster and half are slower, making it a balanced middle-ground option. In practical terms, this positioning usually translates to solid 1080p and 1440p performance, but it should not be expected to dominate at 4K against higher-tier cards.

The absence of rival scores means we cannot state percentages of advantage or deficit. However, the raw compute figures provide a baseline: 24.58 TFLOPS of FP32 performance and a 432.0 GB/s memory bandwidth. Comparing those numbers to typical cards in the 50th percentile, the LX ULTRA appears to lean toward compute-heavy workloads rather than pure rasterization, though this is an inference from its specifications rather than a measured comparison.

Memory Subsystem

The LX ULTRA is equipped with 24 GB of GDDR6 memory, which is a substantial capacity for its performance class. The memory runs at 2250 MHz, translating to an effective data rate of 18 Gbps. This is paired with a 192-bit memory bus, resulting in a total bandwidth of 432.0 GB/s.

For high-resolution gaming, the 24 GB capacity is a standout feature. At 4K, modern games with high-resolution texture packs can easily exceed 12 GB of VRAM usage, and the LX ULTRA's 24 GB buffer provides significant headroom for future titles and modded experiences. The 432.0 GB/s bandwidth, however, is more modest. This figure is typical of cards with a 192-bit bus, and it may become a limiting factor in scenarios that demand extremely high texture streaming rates at 4K with maximum settings.

The 192-bit bus width means that the memory controller is narrower than what you would find on higher-end cards with 256-bit or 384-bit buses. The upside is that the 24 GB capacity compensates for the lower bandwidth in many workloads, particularly those that are capacity-bound rather than bandwidth-bound. For compute tasks like AI inference or rendering large scenes, the large VRAM pool allows for bigger datasets to reside on the card, reducing the need to spill to system memory.

At 1440p, the 432.0 GB/s bandwidth is generally sufficient for most titles, even with high refresh rate targets. At 4K, users may need to dial back texture quality or anti-aliasing settings to avoid bandwidth bottlenecks, though the large capacity means you are unlikely to run out of memory entirely. The memory subsystem is balanced: capacity is generous, bandwidth is adequate but not exceptional.

Who Should Consider It

The LX ULTRA is best suited for users who prioritize memory capacity over raw bandwidth. If you are playing at 1440p and want to max out settings with high-resolution texture packs, the 24 GB VRAM ensures you will not hit memory limits. The 24.58 TFLOPS of FP32 compute is also sufficient for smooth gameplay at this resolution in most titles, assuming the 50th percentile positioning holds true in real-world tests.

For 4K gaming, the LX ULTRA is a viable option but with caveats. The 432.0 GB/s bandwidth is not top-tier, so you may need to adjust settings like shadow quality or ambient occlusion to maintain playable frame rates. The 24 GB capacity, however, means you can enable high-resolution textures without fear of stuttering due to memory exhaustion. This card is a strong candidate for users who play heavily modded games, such as those with massive texture overhauls, where VRAM capacity is the primary constraint.

Content creators and professionals working with large 3D scenes, video editing timelines, or machine learning models will also find the 24 GB capacity appealing. The FP16 performance of 49.15 TFLOPS (2:1 ratio) indicates decent compute throughput for tasks that leverage mixed precision, making it a capable entry-level workstation card for AI inference or rendering tasks that fit within the memory footprint.

Users with lower resolution targets, such as 1080p, are likely overshooting with the LX ULTRA. The card's compute and memory are overkill for that resolution, and a less powerful GPU would suffice. This card is aimed at enthusiasts who want future-proofing in the memory department without stepping up to a flagship-tier price bracket.

FAQ

Q: What is the memory bandwidth of the LX ULTRA?

A: The memory bandwidth is 432.0 GB/s, derived from a 192-bit bus and 18 Gbps effective memory speed (2250 MHz).

Q: Does the LX ULTRA support DirectX 12 Ultimate?

A: Yes, it supports DirectX 12 Ultimate (12_2), along with OpenGL 4.6 and Vulkan 1.3.

Q: How much VRAM does the card have and what type is it?

A: It has 24 GB of GDDR6 memory, which is a large capacity for its performance class.

Q: What power supply is recommended for this GPU?

A: The suggested PSU is 550 W, and the card itself has a 225 W TDP with a single 16-pin power connector.

Q: What display outputs are available on the LX ULTRA?

A: It features 4x DisplayPort 1.4a outputs, allowing for multi-monitor setups.

Q: What is the FP32 compute performance of the LX ULTRA?

A: The FP32 performance is rated at 24.58 TFLOPS, with FP16 performance at 49.15 TFLOPS (2:1 ratio).

Benchmark Performance

The LX ULTRA's benchmark performance is currently unquantified, as its `avgBenchmarkScore` is listed as zero and the `benchmarks` array is empty. This is a notable gap in the data, meaning the card's real-world performance cannot be compared directly to rivals using exact deltas. The `nearestRivals` field is also empty, so there are no percentile deltas to report.

What we can analyze is the theoretical compute output. The card delivers 24.58 TFLOPS of FP32 performance, which is a solid figure for a card in the 50th percentile. To put this in context, a typical card at this level would likely deliver playable frame rates at 1440p in most modern titles, with 4K performance requiring some settings adjustments. The texture rate of 384.0 GTexel/s and pixel rate of 192.0 GPixel/s suggest that the card can handle high-resolution texture filtering and pixel output without obvious bottlenecks in standard rendering paths.

The lack of benchmark data means we cannot state that the LX ULTRA is "30% faster" than any specific rival. Instead, the analysis must rely on the 50th percentile ranking, which indicates a median performance level. The 24 GB memory capacity is unusual for this tier and may skew performance in memory-intensive scenarios, potentially pushing the card above its percentile in those specific workloads, even if average frame rates are median.

Until benchmark scores are populated, the LX ULTRA's performance profile is best described as "predictable but unproven." The specifications suggest it should handle 1440p gaming with ease and 4K with some compromises, but the absence of measured data means those conclusions are extrapolated from the compute and memory figures rather than verified results.

Power and Cooling

The LX ULTRA has a TDP of 225 W, which is moderate for a GPU with this compute capability. The suggested power supply is 550 W, indicating that the card is not overly power-hungry and should be compatible with a wide range of mid-range PSUs. The power connector requirement is a single 16-pin connector, which is the modern standard for GPUs in this class.

Thermally, the card is designed as a dual-slot solution, measuring 268 mm in length, 112 mm in height, and 40 mm in width. This makes it a relatively compact card for its performance level, and it should fit in most mid-tower cases. The dual-slot design suggests a capable air cooler, though the exact cooling solution is not specified. The 225 W TDP is manageable for a dual-slot cooler, and users can expect reasonable temperatures under load without excessive noise, assuming the cooler is well-engineered.

The 16-pin connector is worth noting for users with older power supplies. If your PSU lacks a native 16-pin cable, you will need to use an adapter, which is common for this connector type. The 550 W recommended PSU leaves some headroom for the rest of the system, making the LX ULTRA a practical choice for a mid-range build without requiring a high-wattage power supply.

Ray Tracing and Feature Set

The LX ULTRA supports a modern feature set, including DirectX 12 Ultimate (12_2), which is the baseline for hardware-accelerated ray tracing and mesh shaders. However, the FACT PACK does not list specific ray tracing core counts or tensor core counts, meaning the RT and AI acceleration capabilities are not quantified. The card does have 6144 shading units, which are the primary compute cores, but the absence of dedicated RT/tensor core data makes it difficult to assess ray tracing performance.

In terms of API support, the card is fully up-to-date with Vulkan 1.3 and OpenGL 4.6, ensuring compatibility with current and upcoming titles. The FP16 performance of 49.15 TFLOPS (2:1) suggests that the card can accelerate workloads that use mixed precision, which is often leveraged in machine learning and certain rendering techniques. However, without tensor core specifications, it is unclear how well the card handles DLSS-style upscaling or AI-based denoising.

The display outputs are 4x DisplayPort 1.4a, which support high refresh rates at 1440p and are capable of 4K output, though not at the highest refresh rates supported by DisplayPort 2.0. For ray tracing specifically, the lack of RT core data means we cannot predict performance. Users interested in heavy ray tracing workloads may need to rely on the raw shading unit count, which is substantial, but dedicated RT hardware typically provides significant acceleration. The 50th percentile positioning suggests that ray tracing performance is likely average for the class, but this is an assumption based on the overall ranking rather than a measured result.