Samsung Electronics Co. Invests in Next-Generation Memory Technology

Samsung Electronics Co. is dedicating significant resources to establish a presence in the upcoming memory technology, CXL (Compute Express Link). During a press briefing held on Thursday, Jangseok Choi, vice president of Samsung's new business planning team, revealed the company's plans to start mass-producing 256 GB DRAM supporting CXL 2.0 by the end of this year. CXL technology is expected to revolutionize high-performance server systems by offering a unified interface for accelerators, DRAM, and storage devices used alongside CPUs and GPUs.

The company anticipates that CXL technology will boost memory capacity per server by eight to ten times, representing a major advancement in computing capabilities. Samsung's extensive investment in CXL development is nearing completion, with the company currently conducting performance tests on products in collaboration with partners. Samsung recently established the industry's first CXL infrastructure certified by Red Hat. Choi expressed optimism about the CXL market, predicting significant growth starting in the second half of this year and accelerating from 2028 onwards. He emphasized the technology's potential to expand memory capacity and bandwidth beyond current constraints.

Executive Director Choi provided insights into Samsung's new Compute Express Link (CXL) Memory Module (CMM) technology. He described CMM as a product that enables the addition of memory in the space typically reserved for SSDs, enhancing the CPU's data transmission capabilities around the main memory and complementing existing DRAM functionality.

One notable feature of Samsung's CXL 2.0 DRAM, developed in May 2023, is the memory pooling feature. This technology allows the creation of a shared memory pool by combining multiple CXL memories on a server platform. Hosts can access and utilize memory from this pool as needed, optimizing resource allocation. The result is a more efficient utilization of the entire CXL memory capacity, eliminating empty spaces, reducing data transfer bottlenecks, and enhancing overall system performance and flexibility.