Krova
Star Download
Menu

Performance

What Krova uses, one run per state.

3 idle runs, one for each state below, on one Windows PC, 30 October 2026 to 31 October 2026. This is v0.1: a single run each, with a known gap list below, and the numbers will change. No other app was measured, so there is no comparison here.

In plain words. With no window open, the whole app was 2.4 MB of private memory in a single process (average CPU 0.34 % of one core). With a window open it was 199.3 MB across 7 processes (average CPU 0.49 %). The extra 6 processes are the embedded web view (Microsoft WebView2 on Windows) that draws the editor and settings. In the current app, closing the window ends that web view. These are separate runs, possibly of different builds (see the date and state of each run below), so read the gap as a rough size, not an exact cost.

Memory

One bar for each measured state, in private memory: the memory that belongs only to Krova’s own process and the web view processes it started. This is the headline number. The table adds the working-set figures.

Private memory of all Krova processes while idle, in MB: Tray only, no window, 31 Oct 2026: 2.4 MB in 1 process; Editor window open, 31 Oct 2026: 199.3 MB in 7 processes; Editor window open, 30 Oct 2026, earlier build: 230.6 MB in 7 processes. 0 50 100 150 200 250 MB Tray only, no window 1 process · 31 Oct 2026 2.4 MB Editor window open 7 processes · 31 Oct 2026 199.3 MB Editor window open 7 processes · 30 Oct 2026, earlier build 230.6 MB
Private memory of all Krova processes while idle, in MB: Tray only, no window, 31 Oct 2026: 2.4 MB in 1 process; Editor window open, 31 Oct 2026: 199.3 MB in 7 processes; Editor window open, 30 Oct 2026, earlier build: 230.6 MB in 7 processes. 0 50 100 150 200 250 MB Tray only, no window 1 process · 31 Oct 2026 2.4 MB Editor window open 7 processes · 31 Oct 2026 199.3 MB Editor window open 7 processes · 30 Oct 2026, earlier build 230.6 MB
Memory while idle, in MB. Every bar is private memory, summed over all of that run’s processes.
The numbers behind the memory chart
State Processes Private, all (MB) Krova process, working set (MB) Working sets added (MB)
Tray only, no window 31 Oct 2026, idle-2026-10-31-tray.txt 1 2.4 range 2.4 to 2.4 13.0 13.0
Editor window open 31 Oct 2026, idle-2026-10-31-editor.txt 7 199.3 range 198.4 to 201.2 25.7 402.5–422.3
Editor window open 30 Oct 2026, earlier build, idle-2026-10-30.txt 7 230.6 range 230.5 to 230.6 42.7 469.2–473.3
  • Private, all: the median of the run’s samples, summed over every process of that run. It counts only memory those processes own, and it is the headline number.
  • Krova process, working set: the RAM krova-app.exe itself has in use, including pages it shares with other programs.
  • Working sets added: not used as a headline. It counts pages shared between processes more than once.

CPU

Share of one CPU core used by all of a run’s processes together, sampled at a fixed interval. Each run has its own step chart, one step per sample, nothing smoothed.

Tray only, no window31 Oct 2026

CPU use over 60 seconds, as a percentage of one core, one sample every 2 seconds. It is 0.00 in 19 of 30 samples; the others are 1.31 percent at 12 seconds and 0.69 percent at 24 seconds and 0.69 percent at 26 seconds and 0.70 percent at 36 seconds and 0.69 percent at 46 seconds and 0.69 percent at 48 seconds and 2.08 percent at 50 seconds and 0.70 percent at 52 seconds and 1.41 percent at 54 seconds and 0.70 percent at 56 seconds and 0.69 percent at 58 seconds. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 % of one core 0 20 40 60 seconds 1.31 % at 12 s 0.69 % at 24 s 0.69 % at 26 s 0.70 % at 36 s 0.69 % at 46 s 0.69 % at 48 s 2.08 % at 50 s 0.70 % at 52 s 1.41 % at 54 s 0.70 % at 56 s 0.69 % at 58 s
CPU use over 60 seconds, as a percentage of one core, one sample every 2 seconds. It is 0.00 in 19 of 30 samples; the others are 1.31 percent at 12 seconds and 0.69 percent at 24 seconds and 0.69 percent at 26 seconds and 0.70 percent at 36 seconds and 0.69 percent at 46 seconds and 0.69 percent at 48 seconds and 2.08 percent at 50 seconds and 0.70 percent at 52 seconds and 1.41 percent at 54 seconds and 0.70 percent at 56 seconds and 0.69 percent at 58 seconds. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 % of one core 0 10 20 30 40 50 60 seconds 1.31 % at 12 s 0.69 % at 24 s 0.69 % at 26 s 0.70 % at 36 s 0.69 % at 46 s 0.69 % at 48 s 2.08 % at 50 s 0.70 % at 52 s 1.41 % at 54 s 0.70 % at 56 s 0.69 % at 58 s
CPU use, % of one core, one step per 2-second sample over 60 seconds. The machine has 12 logical cores, so 2.08 % of one core is about 0.17 % of the whole CPU.

Editor window open31 Oct 2026

CPU use over 60 seconds, as a percentage of one core, one sample every 2 seconds. It is 0.00 in 21 of 30 samples; the others are 1.35 percent at 2 seconds and 0.66 percent at 4 seconds and 1.99 percent at 8 seconds and 0.66 percent at 10 seconds and 6.06 percent at 32 seconds and 2.02 percent at 34 seconds and 0.67 percent at 36 seconds and 0.66 percent at 56 seconds and 0.68 percent at 60 seconds. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 % of one core 0 20 40 60 seconds 1.35 % at 2 s 0.66 % at 4 s 1.99 % at 8 s 0.66 % at 10 s 6.06 % at 32 s 2.02 % at 34 s 0.67 % at 36 s 0.66 % at 56 s 0.68 % at 60 s
CPU use over 60 seconds, as a percentage of one core, one sample every 2 seconds. It is 0.00 in 21 of 30 samples; the others are 1.35 percent at 2 seconds and 0.66 percent at 4 seconds and 1.99 percent at 8 seconds and 0.66 percent at 10 seconds and 6.06 percent at 32 seconds and 2.02 percent at 34 seconds and 0.67 percent at 36 seconds and 0.66 percent at 56 seconds and 0.68 percent at 60 seconds. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 % of one core 0 10 20 30 40 50 60 seconds 1.35 % at 2 s 0.66 % at 4 s 1.99 % at 8 s 0.66 % at 10 s 6.06 % at 32 s 2.02 % at 34 s 0.67 % at 36 s 0.66 % at 56 s 0.68 % at 60 s
CPU use, % of one core, one step per 2-second sample over 60 seconds. The machine has 12 logical cores, so 6.06 % of one core is about 0.51 % of the whole CPU.

Editor window open30 Oct 2026, earlier build

CPU use over 60 seconds, as a percentage of one core, one sample every 2 seconds. It is 0.00 in 28 of 30 samples; the others are 1.39 percent at 8 seconds and 0.69 percent at 46 seconds. 0.0 0.5 1.0 1.5 2.0 % of one core 0 20 40 60 seconds 1.39 % at 8 s 0.69 % at 46 s
CPU use over 60 seconds, as a percentage of one core, one sample every 2 seconds. It is 0.00 in 28 of 30 samples; the others are 1.39 percent at 8 seconds and 0.69 percent at 46 seconds. 0.0 0.5 1.0 1.5 2.0 % of one core 0 10 20 30 40 50 60 seconds 1.39 % at 8 s 0.69 % at 46 s
CPU use, % of one core, one step per 2-second sample over 60 seconds. The machine has 12 logical cores, so 1.39 % of one core is about 0.12 % of the whole CPU.
The numbers behind the CPU charts
State Samples Average (% of one core) Peak Samples above zero
Tray only, no window31 Oct 2026 30 over 60 s 0.34 2.08 1.31 at 12 s, 0.69 at 24 s, 0.69 at 26 s, 0.70 at 36 s, 0.69 at 46 s, 0.69 at 48 s, 2.08 at 50 s, 0.70 at 52 s, 1.41 at 54 s, 0.70 at 56 s, 0.69 at 58 s. The other 19 of 30 were 0.00.
Editor window open31 Oct 2026 30 over 60 s 0.49 6.06 1.35 at 2 s, 0.66 at 4 s, 1.99 at 8 s, 0.66 at 10 s, 6.06 at 32 s, 2.02 at 34 s, 0.67 at 36 s, 0.66 at 56 s, 0.68 at 60 s. The other 21 of 30 were 0.00.
Editor window open30 Oct 2026, earlier build 30 over 60 s 0.07 1.39 1.39 at 8 s, 0.69 at 46 s. The other 28 of 30 were 0.00.

Method and limits

Machine
AMD Ryzen 5 5625U with Radeon Graphics, 12 logical cores, 13.8 GB RAM, Microsoft Windows 11 Home Single Language
Tray only, no window
Krova 0.1.0 (the published v0.1.0 installer, installed per-user), running in the tray only: no window opened since launch, no window being dragged. One run, 60 seconds, one sample every 2 seconds, 1 process counted (krova-app.exe alone; no web view process). Raw output: docs/measurements/idle-2026-10-31-tray.txt. Not covered by this run: a drag in progress, the editor window open in this build (see idle-2026-10-30.txt for an earlier build with it open), macOS.
Editor window open
Krova 0.1.0 (the published v0.1.0 installer, installed per-user), editor window OPEN (opened by starting Krova a second time), idle, no window being dragged. One run, 60 seconds, one sample every 2 seconds, 7 processes counted (krova-app.exe plus 6 web view processes started for Krova). Raw output: docs/measurements/idle-2026-10-31-editor.txt. Not covered by this run: a drag in progress, macOS.
Editor window open
Krova 0.1.0 release build (local NSIS install), running, Settings/Editor window OPEN, no window being dragged. One run, 60 seconds, one sample every 2 seconds, 7 processes counted (krova-app.exe plus 6 web view processes started for Krova). Raw output: docs/measurements/idle-2026-10-30.txt. Earlier build: taken before closing the window ended the web view, and before the keyboard shortcuts and window memory were added. Not covered by this run: the window closed/hidden, a drag in progress, macOS.
Limits
A single run for each state, on one machine, so there is no spread to speak of. No window drag was measured. Nothing was measured on macOS. The runs may be of different builds; each run names its build and state above, and its date. Every number on this page is read from the raw files when the site is built.

Re-run it yourself. The script is docs/measurements/measure-idle.ps1: start Krova, put it in the state you want to measure, then run powershell -NoProfile -ExecutionPolicy Bypass -File measure-idle.ps1 -Label "Tray only, no window" powershell -NoProfile -ExecutionPolicy Bypass -File measure-idle.ps1 -Label "Editor window open" Both are in the repository. The result depends on what state you leave Krova in, so give each run a label and write down the build.

What we have not measured yet

  • Drag latency: how long it takes from release to the window landing.
  • CPU and memory while a window is being dragged, or while a keyboard shortcut or a monitor change is being handled.
  • macOS, in any state.
  • More than one run per state, more than one machine, and any other app.

These are v0.1 numbers. They will change as the app does, and this page is regenerated from the measurement files whenever a new one is added.

Back to the overview