FMS 2024: Wolley Introduces CXL Solutions for Datacenter, AI PC, and Automotive

Wolley Inc. will demonstrate and present its CXL technologies for datacenter and AI/ML computational applications at the FMS2024 (Future of Memory and Storage) on Aug 6-8, 2024, Santa Clara, CA.

The company has demonstrated previously its PCI-SIG 5.0 certified CXL controller IP core at FMS 2023. Since then, the firm continues to create and bring the CXL into applications beyond the datacenter, such as AI PC and automotive.

The company will provide a live demo of the NVMe Over CXL and FleX solutions at the FMS2024. NVMe over CXL offers a high-performance virtualization architecture to meet the demand of huge memory capacity for AI. FleX technology brings the CXL memory to the motherboard. 

Wolly will also give 4 talks at FMS2024, ‘NVMe over CXL: Virtualizing the Integration of Memory and Storage’, ‘FleX: Bringing CXL memory to the Motherboard’, ‘CXL Native Memory: Do We Really Need DDR?’, and ‘CXL for Automotive Applications’.

NVMe over CXL: Virtualizing integration of memory and storage
Recognizing the PCIe does not support a lightweight memory protocol, Wolley has developed a solution using CXL and allowing efficient memory accesses with the CXL.mem protocol over that same PCIe physical interface while the CXL.io protocol supports all the legacy features of the NVMe without rewriting the applications. Race conditions in resource allocation are resolved by having storage and memory on the same device. Wolley will present the advantages of SSDs using NVMe Over CXL and detail the comparison to the memory semantic SSDs.

FleX: Bringing CXL memory to motherboard
CXL integration into big iron is well underway with focus on large and high-power form factors such as E3.S. However, less is said about bringing CXL to embedded environments like motherboards. The company will present a smaller implementation of CXL, called FleX, in an M.2-style module with PCIe x8 interface. FleX is a socketed module enables new classes of innovation for next-gen systems by exploiting CXL’s generalizing of interfaces to memory, storage, and accelerators

CXL native memory: Do we really need DDR?
CXL memory modules enable memory expansion, enabling larger capacities to support emerging applications such as large language models (LLMs) which demands 140GB or more of local capacity. HBM cannot enable these large memory capacities, while CXL is a logical method to expand memory but at a significant cost in terms of power consumed and bandwidth wasted. Wolley will present the CXL Native Memory to replace the inefficient DDR interface and reduce the memory latency overhead while saving power. CXL Native Memory consists of the CXL direct physical interface that drives memory cores from the CXL FLIT without protocol retranslation.

CXL for automotive applications
Cars long ago became data centers on wheels. The same architectural challenges face car designers: multiple processors with multiple cores, artificial intelligence, high data rates, private and shared resources… the list goes on. The company will present the advantages of CXL for next-gen car designs, and issues challenging the industry to meet the unique automotive requirements.