CH-J Toolsserver management over SSH
Menu
All guides and articles →

Resource use: Electron 0.1.0, Rust + Tauri 0.1.1 and CH-J Tools 0.1.2 compared

Measured package size, disk space, RAM and CPU on macOS, Windows and Ubuntu for Beta 0.1.0, Alpha 0.1.1 and Beta 0.1.2, including measurement limitations.

The short version

In these measurements, Beta 0.1.2 (CH-J Tools) uses the least RAM with three modules open and during SSH and SFTP work on all three platforms, and CPU use during this work is low. The download package and installed application are slightly larger than Alpha 0.1.1, but still substantially smaller than Beta 0.1.0. We publish these numbers so you can make decisions based on facts, including the limitations listed below.

Measured values

Each cell shows Beta 0.1.0 → Alpha 0.1.1 → Beta 0.1.2. Lower is better. Values are averages. The last row was not measured on 0.1.0 because the same test requires a native file picker dialog; on 0.1.2 on Windows and Ubuntu, repeated downloading did not start during the measured phase, so the value is omitted (–).

Beta 0.1.0 ⟶ Alpha 0.1.1 ⟶ Beta 0.1.2macOSWindowsUbuntu
InstallerSize of the downloaded package.160 ⟶ 21 ⟶ 31 MB114 ⟶ 12 ⟶ 17 MB114 ⟶ 20 ⟶ 26 MB
Installed applicationDisk space after installation.352 ⟶ 61 ⟶ 73 MiB422 ⟶ 53 ⟶ 73 MiB392 ⟶ 62 ⟶ 73 MiB
RAM after unlockingThe vault is unlocked and no module is open.482 ⟶ 315 ⟶ 317 MiB353 ⟶ 374 ⟶ 405 MiB727 ⟶ 680 ⟶ 460 MiB
RAM with 3 modulesSystem Monitor, File Manager and Log Viewer are open.861 ⟶ 496 ⟶ 484 MiB664 ⟶ 1238 ⟶ 460 MiB1102 ⟶ 1935 ⟶ 1010 MiB
RAM during SSH and SFTPAdditionally a live terminal and reading over SFTP.878 ⟶ 523 ⟶ 458 MiB683 ⟶ 1292 ⟶ 354 MiB1146 ⟶ 2057 ⟶ 1048 MiB
CPU during SSH and SFTPPercentage of one core.0.3 ⟶ 2.9 ⟶ 2.42.0 ⟶ 12.4 ⟶ 2.58.7 ⟶ 23.7 ⟶ 3.0
CPU while downloading 512 MiB repeatedlyPercentage of one core; shows 0.1.1 ⟶ 0.1.2, not measured on 0.1.0.17.9 ⟶ 20.53.6 ⟶ –23.9 ⟶ –

How it was measured

Beta 0.1.0 and Alpha 0.1.1 were measured on 30 and 31 October 2026, Beta 0.1.2 on 7 November 2026, each in a single run, 40 seconds per phase after 12 seconds of settling, using the same scenario against the same type of test server. RAM is the sum of resident memory or working set across all application processes, so shared memory is counted more than once. CPU is a percentage of one core. Ubuntu ran under Xvfb without a GPU. Windows on 0.1.0 and 0.1.1 ran in a session with no logged-in user; on 0.1.2, WebView does not start in that session, so measurements were taken in a logged-in session and the Windows values are not directly comparable with the earlier ones. Windows trims the working set during measurement, so values in later phases may fall. 0.1.1 and 0.1.2 were release builds with test automation enabled, and 0.1.0 was the official package; Standby Helper was disabled on 0.1.2 for comparability. This is one measurement, not a general benchmark, and your values will differ. The very low 0.1.0 CPU value on macOS looks suspicious and has not been confirmed.

Why the numbers may look like this

In 0.1.1, each module opened its own window with its own WebView processes, whose resource use added up as more modules were opened. In 0.1.2, all windows share one family of WebView processes, which is consistent with lower RAM with three modules; a causal relationship has not been established separately. The 0.1.2 package is probably larger than 0.1.1 because of newly bundled components (Standby Helper, a macOS widget, expanded cryptography); a size breakdown was not measured. The measurement method counts shared memory repeatedly, so it may overstate values for an application with many processes.

What this means for you

In these measurements, Beta 0.1.2 is the most resource-efficient version when working with modules and over SSH and SFTP on macOS, Windows and Ubuntu, while remaining substantially smaller than Beta 0.1.0. Still, back up your data before updating.

Detailed documentation

Biometric unlock and security →