How to Configure ZRAM, Swappiness, and EarlyOOM on Linux to Run Heavy IDEs Smoothly on 4GB/8GB RAM PCs

Running modern development environments—such as VS Code, Flutter/Android toolchains, or Chromium-based browsers—on compact desktop PCs equipped with 4 GB or 8 GB of RAM quickly exhausts physical memory. When the Linux kernel begins swapping memory pages out to a traditional SSD or HDD swap partition, disk I/O wait times skyrocket and the desktop UI completely freezes (a condition known as swap thrashing). Configuring ZRAM creates a compressed block device directly inside physical RAM, delivering sub-millisecond swap latency while effectively doubling usable memory capacity for text and code buffers.

1. Why ZRAM Outperforms Traditional Disk Swap

Instead of writing inactive memory pages across a SATA or NVMe bus, ZRAM compresses memory pages in-place inside RAM using high-speed algorithms like zstd or lz4. Because source code, DOM trees, and Electron heaps compress at a 2.8:1 to 3.5:1 ratio, 2 GB of data compressed inside ZRAM consumes only ~600 MB of physical RAM.

2. Step-by-Step Setup: Installing and Tuning zramswap (Ubuntu/Debian/Mint)

On Debian- and Ubuntu-based distributions, install the official zram-tools package and configure it to allocate 100% of your physical RAM capacity as the uncompressed ZRAM limit using the zstd compressor:

# 1. Install zram-tools and earlyoom
sudo apt update && sudo apt install -y zram-tools earlyoom

# 2. Configure /etc/default/zramswap
sudo tee /etc/default/zramswap <<'EOF'
ALGO=zstd
PERCENT=100
PRIORITY=100
EOF

# 3. Restart the zramswap service and verify status
sudo systemctl restart zramswap
zramctl

Running zramctl will confirm that /dev/zram0 is active with the zstd algorithm and a swap priority of 100, ensuring the kernel always prefers ZRAM over any fallback disk swap partition.

3. Kernel sysctl Tuning for Compressed Memory

Default Linux kernel VM parameters are designed for slow spinning disks, where swapping is avoided until the last moment (vm.swappiness=60) and pages are read in 8-page clusters (vm.page-cluster=3). With ZRAM, you want the kernel to proactively compress inactive background pages and fetch single pages with zero read-ahead overhead. Add these rules to /etc/sysctl.d/99-zram-tuning.conf:

# Encourage proactive compression of cold anonymous memory pages into ZRAM
vm.swappiness = 150

# Disable swap read-ahead clustering since ZRAM has zero seek latency
vm.page-cluster = 0

# Retain directory and inode caches longer to keep IDE file searches snappy
vm.vfs_cache_pressure = 50

# Prevent large dirty write buffers from stalling system memory
vm.dirty_background_ratio = 5
vm.dirty_ratio = 10

Apply the changes immediately by running sudo sysctl --system.

4. Preventing Hard System Lockups with EarlyOOM

The native Linux kernel Out-Of-Memory (OOM) killer often triggers too late, allowing a runaway compiler or browser tab to freeze the X11/Wayland compositor for several minutes. Enabling the earlyoom daemon monitors free RAM and free swap every second, gracefully terminating the single heaviest memory-consuming process before your mouse cursor ever stutters:

sudo systemctl enable --now earlyoom
sudo systemctl status earlyoom