These days the line between industry and academia are heavily blurred for computing. As CXL becomes more mature, the need for an ‘academic corner’ will likely diminish on this blog, but for now we’re going to spend some time here and there talking about cool papers and articles that are CXL related and CXL adjacent. Follow the anthropomorphic rabbit at the unsettling typewritter and let’s dive in:
Pointers in Far Memory A rethink of how data and computations should be organized
Summary #
This article is basically a recap/summary of different ways large scale applications use far/disaggregated memory and DRAM adjacent technologies. It’s a good refresher or jumping off point for those who haven’t been in the memory and coherency space for a while or ever. “Pointers in Far Memory” reviews the challenges of far and distributed memory both from a latency/performance standpoint and a software/utility standpoint, as well as different types of applications and algorithms which both
- need far/distributed memory due to their memory requirements and
- end up being fairly fantastic pathological use-cases for said memory given their latency challenges as compared to local (near) memory.
“Pointers” also goes into other classes of memory such as persistent memory and computational memory, as well as some of the unique qualities that come with such technologies. The article then concludes with some proposals for how to initially adopt CXL into existing application frameworks without too much upheaval in the application stack while still benefitting from the technology.
Take-aways for me #
MemOS and Twizzler are neat, interested readers should go check them out. Figure 3 has data which hints at the potential benefits of a globally addressable shared memory.
How do I do this? #
Hi there, we’re Jackrabbit Labs (JRL for short). We build open source CXL things then help you use them. If you’re finding yourself scratching your head on day 2 of your CXL deployment, please reach out. We’d love to get to know you.