This is an old revision of the document!
Table of Contents
Benchmarks
A benchmark is a test that measures the performance of a system or subsystem on a well-defined task or set of tasks. Benchmarks can be used to measure general things, like graphics, I/O, computing (integer and floating point), performance, etc. But most measure more specific tasks, like rendering polygons, reading and writing files, or performing operations on matrixes.
CPU Benchmarks
nbench
The following results were obtained using nbench 2.2.2.
All nbench results are in comparison to an AMD K6 233Mhz reference system.
| CPU | DSM-G600 revB1 (PPC) | LS1 (PPC) | LS2 (MIPSel) | HG (PPC) | LSPro (ARM9) | DNS-323 (ARM9) | AMD (K6) |
|---|---|---|---|---|---|---|---|
| Clock Speed | 170MHz | 200MHz | 400MHz | 266MHz | 400MHz | 500MHz | 233Mhz |
| Firmware | Gentoo 2005.0 | FreeLink 1.1 | OpenLink 0.52b | FreeLink 1.11 | FreeLink 0.1 dev | D-Link firmware | Linux |
| Numeric Sort | 0.70 | 0.75 | 0.79 | 1.01 | 0.84 | 1.07 | 1.00 |
| String Sort | 0.40 | 0.56 | 0.53 | 0.75 | 0.48 | 0.65 | 1.00 |
| Bitfield | 1.14 | 1.02 | 1.30 | 1.37 | 1.46 | 1.04 | 1.00 |
| FP Emulation | 1.05 | 1.11 | 1.08 | 1.49 | 2.61 | 0.99 | 1.00 |
| Fourier | 0.65 | 0.65 | 0.00 | 0.88 | 0.02 | 0.02 | 1.00 |
| Assignment | 0.69 | 0.83 | 1.17 | 1.13 | 1.97 | 1.29 | 1.00 |
| Idea | 1.12 | 1.34 | 2.50 | 1.80 | 1.10 | 1.66 | 1.00 |
| Huffman | 1.40 | 1.21 | 0.18 | 1.63 | 0.62 | 0.68 | 1.00 |
| Neural Net | 1.03 | 1.24 | 0.00 | 1.68 | 0.03 | 0.03 | 1.00 |
| LU Decomposition | 0.99 | 1.27 | 0.01 | 1.86 | 0.05 | 0.05 | 1.00 |
| Memory Index | 0.68 | 0.78 | 0.93 | 1.05 | 1.00 | 1.00 | 1.00 |
| Integer Index | 1.04 | 1.08 | 0.79 | 1.45 | 1.00 | 1.00 | 1.00 |
| Floating-Point Index | 0.87 | 1.01 | 0.00 | 1.40 | 0.03 | 0.03 | 1.00 |
Bogo Mips
| Machine | CPU | Clock | BogoMIPS |
|---|---|---|---|
| DSM-G600 revB | PPC | 170MHz | 133.12 |
| HD-HLAN (LS1) | PPC | 200MHz | 130.66 |
| HD-HGLAN (HG) | PPC | 266MHz | 173.26 |
| LS-GL | ARM | 400MHz | 266.24 |
| DNS-323 | ARM | 500MHz | 331.77 |
Throughput Benchmarks
| Protocol | Transfer Rate | Source | Sink |
|---|---|---|---|
| tcp (netcat) | 20MB/s | /dev/zero@t42laptop | /dev/null@dns323 |
| tcp (netcat) | 18MB/s | disk@t42laptop | /dev/null@dns323 |
| tcp (netcat) | 16MB/s | /dev/zero@dns323 | /dev/null@t42laptop |
| tcp (netcat) | 12MB/s | /dev/zero@t42laptop | raid1@dns323 |
| tcp (netcat) | 12MB/s | disk@t42laptop | raid1@dns323 |
| tcp (netcat) | 12MB/s | raid1@dns323 | disk@t42laptop |
| ssh | 1.6MB/s | /dev/zero@dns323 | /dev/null@t42laptop |
| ssh | 1.3MB/s | /dev/zero@t42laptop | /dev/null@dns323 |
| ftp | 12MB/s | disk@t42laptop | raid1@dns323 |
| rsync-over-tcp | 5.5MB/s | disk@t42laptop | raid1@dns323 |
| rsync-over-ssh | 1MB/s | disk@t42laptop | raid1@dns323 |
Power Consumption Benchmarks
| State | Usage (watts) |
|---|---|
| Unit powered off, power brick plugged in | 0 |
| HDs powered down (sleep), otherwise running | 7 |
| HDs spinning, little CPU activity (uPnP AV and iTunes server turned off, no network activity) | 14 |
| HDs spinning, little CPU activity, 10-100Mbps network transfer via CIFS/SMB | 16 |
| HDs up, high CPU activity (uPnP AV and iTunes turned on (v1.06)), no network activity | 22-23 |
| HDs up, high CPU, Gigabit network transfer via CIFS/SMB | 23-24 |
| HDs spinning/powering up | up to 72 for 3 seconds |
Based on P3 Kill-a-Watt meter, D-Link firmware v1.06, ffp 0.5
