GoKawiilMeta engineer's CRAM project speeds up Linux memory compression up to 452x
Gregory Price and a team at Meta have built CRAM, a new Linux memory-compression technique that keeps compressed data directly in RAM instead of routing it through the swap layer like existing tools zswap and ZRAM. CRAM uses a private NUMA node to let Linux manage compressed memory with its normal memory semantics, including migration and ballooning, and includes a mechanism called the 'Chicken Bit' to pause CRAM during allocation management. Early benchmarks cited show read performance gains of up to 452 times over ZRAM.
GoKawiil's interpretation of the reporting above, not reported fact.
By avoiding the swap-layer fault handling that reportedly causes most of the slowdown in existing compression schemes, CRAM suggests a path to much faster effective memory capacity without added hardware, which could matter for memory-constrained servers and devices. Because this originates from Meta's infrastructure needs, it may eventually be proposed for mainline Linux, though it is not yet clear from the reporting whether or when that will happen. The complexity around data compressibility and allocation handling implied by features like the Chicken Bit suggests real-world performance could vary by workload.
- CRAM keeps compressed memory entirely in RAM rather than treating it as swap.
- It reportedly delivers up to 452x the performance of ZRAM on reads.
- Developed by Gregory Price's team at Meta, using Linux's existing NUMA and memory-management mechanisms.
Source: tomshardware.com — Zak Killian, 2026-10-07
Published there as: “New Linux tech compresses memory in RAM, as RAM, for 452x speedup”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.