New Patch Can Boost Linux ZRAM Compression Performance By Over 50% - Phoronix

A new Linux kernel patch from HONOR engineer Gao Xu could significantly boost ZRAM performance by optimizing LZ4 dictionary compression. The patch, posted to the Linux kernel mailing list, addresses a performance bottleneck in how ZRAM handles repeated dictionary loads during compression operations.

The Performance Problem

The current ZRAM implementation calls LZ4_loadDict() repeatedly during compression operations. This function performs internal dictionary pre-processing each time it's called, creating unnecessary overhead. For systems that frequently compress and decompress data in RAM—such as those using ZRAM for swap space or compressed RAM disks—this repeated preprocessing can substantially impact I/O performance.

The Solution: Template Stream Mechanism

Gao Xu's patch introduces a template stream mechanism that pre-processes the dictionary only once when it's initially set or modified. The optimized approach then efficiently copies this pre-processed state for subsequent compressions, eliminating the redundant work.

According to the engineer's testing, this optimization improves LZ4 dictionary compression performance by over 50% in ZRAM I/O tests. The patch consists of just a few dozen lines of code, making it a relatively simple yet impactful improvement to the Linux kernel's memory management capabilities.

What This Means for Linux Users

ZRAM is particularly valuable for systems with limited RAM, as it allows the creation of compressed block devices in memory. This can effectively increase available memory capacity, though at the cost of some CPU overhead for compression and decompression. The performance improvement from this patch could make ZRAM more attractive for:

• Systems with constrained memory resources
• Workloads that frequently swap data between RAM and storage
• Embedded devices and IoT applications using Linux
• Desktop users looking to optimize memory usage

Current Status

The patch is currently out for review in the Linux kernel community. If accepted, it would represent a meaningful performance enhancement for a feature that's already widely used across various Linux distributions and embedded systems.

The optimization demonstrates how targeted improvements to core kernel components can yield significant performance gains without requiring major architectural changes. For Linux users who rely on ZRAM for memory management, this patch could translate to noticeably faster system responsiveness, particularly under memory pressure.

For those interested in following the patch's progress, it can be found in the Linux kernel mailing list archives under recent ZRAM-related discussions.