中文
400-667-5666
Mon-Sun 9:00-21:00
MTT S4000 is a powerful GPU designed by Moore Threads for large models. It features the latest third-generation MUSA architecture, 128 Tensor Cores, 48GB of VRAM, and a blazing-fast 768GB/s memory bandwidth. Leveraging Moore Threads‘ innovative MTLink technology, it supports for multi-GPU interconnectivity to achieve clusters with thousands of cards and also accelerates computing for LLMs with a trillion parameters. MTT S4000 supports advanced graphics rendering, video encoding and decoding, and UHD 8K HDR display. With the full-featured MUSA architecture, it’s very convenient for customers to migrate their applications to MUSA platform.
LLM Training and Finetuning
Perfect platform, mature MUSA ecosystem, MTLink interconnection, high MFU, and above 91% linear speedup ratio
LLM Inference
Perfect inference platform, high throughput and low latency
KUAE Cluster
Total solution with hardware and full stack software, high reliability, ready-to-use
Advanced-Gen Tensor Cores
128 Tensor Cores to accelerate LLM training, finetuning, and Inference
Third-Gen MUSA
Full-featured MUSA architecture that is compatible with CUDA
Graphics Rendering Engine
Vulkan / DirectX / OpenGL / OpenGL ES supported
LLM Training and Finetuning
Training Platform Architecture
Moore Threads' large model training platform is fully compatible with CUDA and PyTorch. It supports distributed training frameworks such as Megatron-LM, DeepSpeed, FSDP, and Colossal-AI. Fully compatible and with high performance and flexibility, the platform is ideal for easily training mainstream large models such as GPTs, LLaMa, and GLMs with just one click. Using the KUAE thousand-GPU computing cluster for large model training, the linear speedup ratio exceeds 91%. In addition to supporting development, it also supports training supervision, automatic restart, and resumption.
Training and Finetuning Examples
MTT S4000 features 128 Tensor Cores, 48 GB of memory, and ultra-fast inter-card communication enabled by MTLink. It supports training for various LLMs, including LLaMa, ChatGPT, ChatGLM, Qwen, and Baichuan through Moore Threads' training platform. With distributed training strategies for single and multi-node systems, it accelerates the training and finetuning of LLMs with 60 to 100 billion parameters.
Cluster Scaling Efficiency
The Moore Threads KUAE Computing Cluster platform, designed for billion-parameter model pretraining, finetuning, and inference, achieves a 91% linear speedup ratio in a thousand-card cluster. The platform is optimized across applications, distributed systems, training frameworks, communication libraries, firmware, operators, and hardware. Featuring MTLink, a proprietary interconnect technology based on the architecture of MTT S4000, the platform supports MTLink Bridge connections between two, four and eight cards. MTLink achieves an inter-card bandwidth of up to 240 GB/s. This accelerates training speeds in clusters with 64 to 1024 cards and improves the linearity of multi-card interconnects.
Large Model Inference Service Platform
MTT S4000, equipped with 128 Tensor Cores and 48 GB of memory, effectively supports inference for mainstream LLMs, such as LLaMa, ChatGLM, Qwen, and Baichuan.
KUAE ModelStudio
A training, finetuning, and inference platform for developers of LLM apps. It is based on Moore Threads' GPUs and models.
MUSA Serving
A software for high-performance and distributed inference services, supporting backend models such as LLMs, image and video generation, and traditional AI.
MT Transformer
A distributed inference acceleration framework for Moore Threads' GPUs. It achieves inference acceleration for LLMs based on transformer architectures.
TensorX
An inference acceleration framework for Moore Threads’ GPUs. It is ideal for inference acceleration in image and video generation and traditional AI.
Supporting KUAE Cluster Products
MTT KUAE is Moore Threads' full-stack solution for artificial intelligence data centers. It is based on the S4000 GPU and MCCX D800 all-in-one cluster computing unit, which is equipped with eight dual-processor S4000 GPUs. This integrated solution tackles the challenges inherent in deploying large-scale GPU computing power with efficiency and effectiveness.
Advanced-Generation Tensor
Moore Threads' advanced-generation Tensor Cores assist in the training, finetuning, and inference of LLMs. MTT S4000 includes 8,192 vector cores and 128 Tensor Cores. It supports mainstream precision computing formats such as FP64, FP32, TF32, FP16, BF16, and INT8.
Third-Generation MUSA Software Stack
MUSA is Moore Threads' self-developed metacomputing unified architecture, which includes an instruction set architecture, MUSA programming model, driver, runtime library, operator library, communication library, and mathematical library. CUDA programs can be smoothly migrated to MUSA through Moore Threads' self-developed MUSIFY tool.
Fully Supports Mainstream Graphics APIs
MTT S4000 supports mainstream graphics APIs such as DirectX, Vulkan, OpenGL, and OpenGL ES. It provides versatile graphics rendering for a variety of scenarios, including digital twins, cloud gaming, cloud rendering, and content creation. It also supports large model inference capabilities, functioning as a one-stop solution for multi-modal scenarios like AIGC.
MTT S4000
Specifications
Streaming Processors
8192 MUSA Cores
FP32
25 TFLOPS
TF32 Tensor
50 TFLOPS
FP16 Tensor
100 TFLOPS
INT8 Tensor
200 TOPS
Texure Rate
768 GTexels/s
Pixel Rate
768 GPixels/s
Memory Size
48 GB
Memory Bandwidth
768 GB/s
Multimedia Encoding
H.265, H.264, AV1 48 x 1080p30
Multimedia Decoding
H.265, H.264, AV1, AVS2, VP9,… 96 x 1080p30
I/O Interface Bandwidth
240 GB/s
Bus Interface
PCIe 5.0 x16
Chip-to-Chip Interconnect
MTLink
Display Interface
4 x DisplayPort 1.4a
Security Engine
MUSA Security Engine 2.0 supports TEE and a variety of encryption and decryption algorithms
Virtualization
Hardware virtualization, GPU elastic partitioning, SR-IOV isolation
TGP
450W
Dimension
266 mm * 112 mm * 39 mm
