- DeepSeek’s Engram separates static memory from computation, increasing efficiency in large AI models
- The method reduces high-speed memory needs by enabling DeepSeek models to use lookups
- Engram supports asynchronous prefetching across multiple GPUs with minimal performance overhead
DeepSeek, in collaboration with Peking University, introduced a new training method called Engram, designed to decouple memory storage from computational processes.
Traditional large language models require high-bandwidth memory for knowledge retrieval and basic computation, creating a bottleneck in both performance and cost.
This HBM bottleneck is widely recognized as a key reason DRAM prices rose by 5X in just 10 weeks, as hardware demand spiked to support large AI models.
You may like
Validation and technical approach
The researchers said existing models waste sequential depth on trivial operations, which could otherwise support higher-level reasoning.
Engram allows models to efficiently “look up” essential information without overloading GPU memory, freeing capacity for more complex reasoning tasks.
The system was tested on a 27-billion-parameter model and showed measurable improvements across standard industry benchmarks.
By performing knowledge retrieval through hashed N-grams, Engram provides static memory access independent of the current context.
The retrieved information is then adjusted using a context-aware gating mechanism to align with the model’s hidden state.
This design allows models to handle long context inputs more efficiently and supports system-level prefetching with minimal performance overhead.
The Engram method complements other hardware-efficient approaches, including solutions such as Phison’s AI inference accelerators.
You may like
Engram minimizes the amount of high-speed memory required by using lookups for static information, making memory usage more efficient.
Phison offers a cost-effective way to expand total memory using SSDs, supporting large AI models such as Engram or Mixture-of-Experts systems.
Combined, these approaches allow AI systems to optimize fast-memory usage while affordably increasing overall memory capacity.
It also works alongside emerging CXL (Compute Express Link) standards, which aim to overcome GPU memory bottlenecks in large-scale AI workloads.
The method separates static pattern storage from dynamic computation, enhancing the Transformer backbone without increasing FLOPs or parameter counts.
DeepSeek formalized a U-shaped expansion rule to optimize the allocation of parameters between the MoE conditional computation module and the Engram memory module.
Tests show that reallocating around 20–25% of the sparse parameter budget to Engram yields better performance than pure MoE models, maintaining stable gains across different scales.
Memory slot expansion provides predictable improvements without additional computational cost.
This confirms the scalability of conditional memory as an independent axis for sparse models.
Engram’s deterministic retrieval mechanism allows memory capacity to scale linearly across multiple GPUs while supporting asynchronous prefetching during inference.
It offloads static knowledge reconstruction from lower layers, freeing attention mechanisms to focus on global context.
Hierarchical caching of frequently used embeddings enhances efficiency, and the module works with existing GPU and system memory architectures, potentially avoiding costly HBM upgrades.
This technique may relieve pressure on expensive memory hardware, particularly in regions such as China, where HBM access lags behind competitors such as Samsung, SK Hynix, and Micron.
Early validation of Engram suggests models can expand parameter scale and reasoning capacity while managing memory demands more efficiently.
This approach may help ease memory constraints across AI infrastructure, potentially reducing sharp DDR5 DRAM price swings.
Via SCMP
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds. Make sure to click the Follow button!
And of course you can also follow TechRadar on TikTok for news, reviews, unboxings in video form, and get regular updates from us on WhatsApp too.