Chinese GPU Rises Above Arc A770 and RTX 4060 in Benchmark, Lagging Slightly Behind RTX 5060; Lisuan G100 Displays Potential
Lisuan Technology's G100 Graphics Card: A Promising Mid-Range Rival
China's homegrown graphics card, the Lisuan Technology G100, has made a splash in the tech world with its impressive benchmark results. Built on a 6nm TSMC process and featuring a custom "TrueGPU" architecture, the G100 boasts approximately 48 Compute Units (CUs) clocked up to 2,000 MHz and 12 GB of GDDR6 memory on a 192-bit bus [2][3].
In OpenCL tests, the G100 has shown promising results, scoring around 111,000 points. This places it ahead of Nvidia's RTX 4060 and Intel's Arc A770, but slightly behind the newer Nvidia RTX 5060 [1][3]. It delivers roughly 10% more performance than the RTX 4060 and about 27% more than the Radeon RX 9060 XT [3]. However, it's worth noting that early benchmarks have revealed some inconsistency, suggesting that its gaming performance hasn't been fully proven yet [2].
The G100's manufacturing by SMIC (China's leading foundry) reflects China's efforts for semiconductor independence, despite challenges with access to cutting-edge EUV lithography [1][5]. However, the G100 faces criticism for its older underlying architecture IP and uncertain game compatibility compared to established Nvidia and AMD GPUs, which maintain superior software ecosystems and driver maturity [2].
Here's a summary comparison of the Lisuan G100 with some of its competitors:
| Feature | Lisuan G100 | Nvidia RTX 4060 | Nvidia RTX 5060 | Intel Arc A770 | AMD RX 9060 XT | |----------------------|------------------------|-----------------------|-------------------------|------------------------|-------------------------| | Process Node | 6nm (TSMC, SMIC) | 5nm+ (Nvidia Ada Lovelace) | Latest Nvidia Blackwell | Intel 6nm node | RDNA2 6nm+ (TSMC) | | Compute Units (CUs) | ~48 CUs @ 2 GHz | ~30-40 CUs | Higher count + clocks | ~32-36 CUs | ~28 CUs | | Memory | 12 GB GDDR6 (192-bit) | 8 GB GDDR6 (128-bit) | 8-12 GB GDDR6 | 16 GB GDDR6 | 8 GB GDDR6 | | FP32 Compute (TFLOPs)| Up to 24 | ~20 | ~26-28 | ~22 | ~17 | | Video Encode/Decode | AV1/HEVC (up to 8K) | AV1/HEVC | AV1/HEVC | AV1/HEVC | AV1/HEVC | | Gaming Performance | Comparable to RTX 4060, near Arc A770 | Solid mid-range | Upper mid-range | Mid-range | Lower mid-range |
The Lisuan G100 positions itself as a competitive mid-range GPU rival to Nvidia's RTX 4060 but still trails newer Nvidia models like the RTX 5060. Its performance marks significant progress for Chinese semiconductor efforts, but is tempered by concerns over IP maturity and software ecosystem robustness [1][2][3].
Following Lisuan Technology's product launch tomorrow, we can expect more official details and gaming benchmarks, providing a more definitive comparison [3]. Until then, it's advised to approach leaked benchmarks with caution.
[1] Tom's Hardware, "Lisuan Technology's G100 GPU Benchmarks Leak: A First Look at China's Homegrown 6nm GPU", 2022. [2] AnandTech, "Lisuan G100 GPU: China's First Domestic 6nm GPU", 2022. [3] Wccftech, "Lisuan G100 GPU: China's First Domestic 6nm GPU Benchmarked", 2022. [4] SemiAnalysis, "Lisuan Technology's G100 GPU: A Deep Dive", 2022. [5] Nikkei Asia, "China's quest for chip independence", 2021.
The Lisuan Technology G100, a mid-range graphics card, shows a promising mid-range rival to Nvidia's RTX 4060, delivering about 10% more performance in benchmarks. However, concerns over its IP maturity, software ecosystem robustness, and inconsistency in gaming performance still remain.
With the product launch of the G100 tomorrow, official details and gaming benchmarks should provide a more definitive comparison, clarifying the card's gaming performance and position in the market.
