L ITE S WITCH is presented, a lightweight hardware-software co-design that opportunistically harvests otherwise idle cycles caused by CXL memory accesses and recovers up to 80% of the performance lost due to CXL access latency, enabling the adoption of CXL memory without prohibitive slowdowns.
The main result concerns the plain-load path: attained LDG bandwidth peaks at a small offered per-thread load (K ~ 2) and then declines, by about 35% from K=2 to K=8 at the authors' primary configuration.
Bulk memory copying (memcpy) is a dominant operation in modern data centers, driven by storage engines, in-memory databases, and caches. The emergence of heterogeneous memory (CXL, persistent memory, remote NUMA) further increases the volume of concurrent memory copy operations, making it easier to saturate the shared...
Ze-Cheng Li, Jian Zhang· Proceedings of the 18th ACM...· 0 citations
CXL-enabled memory expands server memory capacity, but introduces a page-placement problem: the operating system must decide which pages should reside in DRAM and which should reside on slower CXL memory. Existing systems make this tradeoff in one of two ways. Userspace controllers support flexible policies, but expose...
Results show that standardizing the host-interface layer reduces both the engineering cost of PIM simulator development and the energy overhead of simulation-driven workload evaluation, while reducing host CPU utilization and package energy by up to 50% on transfer-intensive workloads.
FLINT is proposed, a workload-driven HBF substrate for capacity-scalable LLM inference that integrates HBF as a memory-capacity tier alongside HBM while addressing three adoption challenges.
Geraldo F. Oliveira, Arash Tavakkol, Xiang-Yu Zhu et al.· 1 citation
PLINK (Parallel-Link), a GPU-initiated I/O platform that addresses this gap through three targeted optimizations: SQ/CQ decoupling to hide device latency, command-ID-indiscriminate polling to eliminate exhaustive CQ search overhead, and warp-level doorbell batching to minimize PCIe MMIO writes and reduce warp divergenc...
Jaewoong Jung, Hyukyul Kang, J. Choi· Proceedings of the 18th ACM...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
