For years, DBeaver has been the default recommendation for developers who need a multi-platform database client. But on macOS, running software built on the Eclipse RCP Java framework comes with unmistakable hardware costs.
When you are working remotely from a coffee shop on an M3 MacBook Air, or running heavy Docker containers and microservices simultaneously, every gigabyte of RAM and every percentage of battery matters.
We conducted a real-world benchmark comparing DBeaver 24.2 against native macOS database clients - specifically SQLOZ and TablePlus - on Apple Silicon hardware. Here is what the telemetry revealed.
The Benchmark Environment
- Hardware: MacBook Pro (16-inch, Apple M3 Pro, 36GB Unified Memory, macOS 15 Sequoia).
- Workload:
- Cold start to connection window.
- Single active connection to a remote PostgreSQL 16 server.
- Querying and rendering a 100,000-row table with 20 columns.
- 30 minutes of idle background monitoring.
1. Startup Speed: The Cold Launch Test
How long does it take from clicking the Dock icon until you can type your first SQL query?
- DBeaver: 4.8 seconds. The Java Virtual Machine (JVM) initialization and Eclipse plugin dependency graph introduce a noticeable delay before the workspace renders.
- TablePlus: 145 milliseconds. Fast, native Cocoa window instantiation.
- SQLOZ: 88 milliseconds. Pure Swift and Metal direct binary execution. It appears on screen instantly.
2. Idle and Active RAM Footprint
Database clients frequently stay open in the background for days. How much memory do they hold hostage?
IDLE RAM FOOTPRINT (1 Connection):
┌─────────────────────────────────────────────────────────────┐
│ DBeaver: ████████████████████████████████████ 1,240 MB │
│ TablePlus:███ 92 MB │
│ SQLOZ: ██ 54 MB │
└─────────────────────────────────────────────────────────────┘- DBeaver: Consumed 1.24GB of RAM with just one active connection. Java heap memory reservations and garbage collection pools prevent the JVM from releasing memory back to macOS.
- TablePlus: Maintained a lean 92MB of RAM.
- SQLOZ: Maintained the lowest idle footprint at 54MB of RAM, utilizing Apple Silicon unified memory buffers and zero runtime wrappers.
3. Battery Drain & Energy Impact
In the macOS Activity Monitor’s Energy tab:
- DBeaver: Recorded an Energy Impact score of 14.2 while idle. Continuous thread polling and Java background garbage collection keep the CPU cores in higher power states, accelerating battery drain.
- SQLOZ: Recorded an Energy Impact score of 0.0 – 0.1 while idle, fully supporting macOS App Nap to drop CPU execution to zero when the window is in the background.
Benchmark Summary Table
| Metric | DBeaver 24.2 (Java) | TablePlus (Native) | SQLOZ (Native Swift/Metal) |
|---|---|---|---|
| Cold Startup Time | 4.8 seconds | 145 ms | 88 ms (Instant) |
| Idle RAM (1 DB Connection) | 1,240 MB | 92 MB | 54 MB |
| Active Query RAM (100k rows) | 1,850 MB | 210 MB | 118 MB |
| 120 FPS ProMotion Scrolling | Stutters (Java SWT) | Smooth | Buttery Smooth (Metal) |
| Energy Impact at Idle | 14.2 (High) | 0.2 (Low) | 0.1 (Negligible) |
Conclusion: Native Wins on Apple Silicon
DBeaver remains a capable Swiss Army knife if you are running Windows or Linux and connecting to exotic legacy databases. But on macOS, running a 1.2GB Java container is an inefficient use of Apple Silicon hardware.
With SQLOZ, Mac developers get deep DBA diagnostics, visual query building, and cross-dialect migrations - all running inside a lightweight, native Swift client that boots in 88ms and barely sips battery.
Download SQLOZ for macOS and unlock true native performance today.