The Gigahertz Dupe: Dell SmartStep 250N – THG.RU

A Dell Novelty: A Desktop CPU in the Notebook

Gericom made it big in Europe by doing it. Toshiba held back for a long time before finally bowing to Gericom’s “powers of persuasion.” We’re talking about using a desktop CPU in a notebook.

Dell took a longer time considering its options. But then in July came the news that, for the first time, the direct shipper would offer customers a notebook with a desktop CPU and – also a novelty – an (almost) fixed configuration. The result, according to the heady texts of the marketing strategists, is “… the perfect companion for almost all conceivable applications…” They add: “… regardless of whether it’s multimedia or DVD films, MP3 editing or CD burning, or whether you need intricate presentations, calculations, the Internet or e-mail: Even with modern 3D games, the speeds reached by the 250N are impressive.”

A glance at the list of features presented with the SmartStep 250N we tested really does suggest that there’s a powerful beast inside this “people’s notebook.”

The maximum brightness of the SmartStep’s 15.1″ display was found to be 120 candelas per square meter, with a contrast ratio of 128:1 (black-and-white contrast). The test device screen’s brightness distribution is good. All in all, THG gives the SmartStep 250N’s display quality a good rating.

Dimensions and Weight: Smart But Not Slim

“Fat Boy:” the SmartStep tips the scales at a hefty 3.7 kilograms.

The silver finish of the plastic case gives the SmartStep an elegant appearance. But the case’s haptics and scratch-resistance are not convincing. With dimensions of 32.5 x 4.8 x 27.5 cm3, the device is not exactly slim, either. Including the power supply, the SmartStep weighs a hefty 4.26 kilograms (device 3.7kg, power supply 0.56kg). In practice, both factors tend to limit mobility considerably.

But in terms of connections, there’s no room for complaint with the SmartStep 250N. The back features parallel and serial ports, a VGA connector, a PS2 jack and two USB connectors and S-Video output.

Connector-Friendly: the reverse of the notebook.

Located on the left-hand side of the case are the audio connectors and the volume control; next to them is the FireWire connector.

Multimedia center: audio and FireWire connectors.

The two slots for PC cards are on the right-hand side of the case.

The Eternal Debate: Desktop or Notebook CPU?

The advantages of a mobile Intel CPU are as clear as day. Less power consumption and special power-saving features, such as Enhanced Speedstep, permit longer battery life and thus greater mobility for the user. In addition, the lower power consumption results in lower heat release from the processor. Accordingly, the processor’s cooling system has slimmer dimensions and is less power-hungry (and quieter). It seems that the use of mobile CPUs means that notebooks can come in smaller packages. But why use a desktop CPU in this notebook? The answer to that is simple: the desktop CPU is considerably cheaper.

Or, put another way: if a manufacturer wants to aim for a certain price, it has two possibilities. Either it can use a mobile CPU and cut corners with equipment, or it can lavish the mobile PC with wide-ranging accessories and a desktop CPU. Nevertheless, the customer still has to accept some restrictions as far as mobility goes. But the loss of mobility is not as severe as often claimed. In normal operation, a notebook with a mobile CPU (using the processor’s power-saving modes) runs on average 30 minutes longer than a device with a desktop CPU. A good design can mean it can run for up to 50 minutes longer.

The relatively slight difference in operation time is easily explained. On the one hand, on the face of it, the CPU running in normal operation (running typical office applications) has little to do, so power consumption (known as “average power”) is far under the Thermal Design Power for mobile and desktop CPUs. On the other hand, the processor is not the only energy consumer in a notebook. Using up 30 percent, the display accounts for a huge chunk of the energy needs of the entire system. The graphics processor and memory modules also prove themselves voraciously hungry for power.

Operating Time in Battery Mode and System Performance

Normally, we rely on Sysmark 2002 to give us a reading of overall system performance. But, due to a problem with the graphics subsystem driver on the test system, meaning that we couldn’t work out a valid drive, we used Content Creation Winstone 2002. The battery operation time was determined with Business Winstone 2002, while Business Disk Winmark 99 delivered the key figures for hard-drive performance.

No complaints about battery operation time. Comparably equipped notebooks with mobile CPUs fail to keep up with this benchmark either. A little over 25 points with Content Creation Winstone is certainly a respectable score. But the score could have just as easily been attained with a slower CPU and a quicker hard drive.

CPU Performance: PC Mark 2002

PC Mark 2002 revealed some interesting behavior by the system:

In battery mode, it showed that the CPU is only half as fast as it is when running on the mains.

CPU Performance: Sisoft Sandra 2002 SP1

The Arithmetic Benchmark from Sisoft Sandra 2002 underlines the results of PC Mark 2002:

Again, the CPU only delivers 50 percent of its performance capacity in battery mode. Obviously, the CPU only operates internally at half-speed when running on the battery.

SmartStep’s Low-Energy Setting: Clock Throttling

In principle, there are two methods of reducing a CPU’s energy consumption. Either you can reduce the operating voltage or you can lower clock speed. Reducing the operating voltage is a whole lot more effective, because a CPU’s power consumption is proportional to the square of its voltage. The clock speed is merely a linear value in its power consumption. Both methods are combined to save energy, for example, in the Enhanced Speedstep mechanism of the mobile Pentium 4 M processor.

The desktop variant of the P4 doesn’t “recognize” Speedstep. Thankfully, the Advanced Power and Configuration Interface (ACPI) has another legal method of saving electricity up its sleeve. Within the system state “System working” (G0), there are altogether four processor modes (C0-C4), according to ACPI specs. In C0 status (normal operation), the processor executes commands. But within this status the processor can switch between two states, “Full Speed” and “Throttling.” In “Full Speed” mode, the processor runs at full speed. In “Throttle” mode, the CPU maintains clock speed for a limited time only. The rest of the time, clock supply is prevented via the “Stopclock” signal. To lower clock speed by half using this method, for example, the chipset would send a “Stopclock” signal via the corresponding pin on the processor with a mark space ratio of 50 percent. That means that the processor only receives “its” speed half of the time. This behavior is controlled via the chipset. The disadvantage of this power-saving mechanism is that the user cannot influence it. In other words: in SmartStep, the processor runs only at half-speed in battery operation. Although that needn’t be an issue when running simple office applications, anyone who uses the CPU primarily to run intensive applications – and wants to do this on the road – will be bitterly disappointed by the device’s performance.

Application Benchmarks: MP3 Encoding and Archiving

Based on two typical tasks that modern users frequently perform, it is plain that the substantially reduced CPU performance in battery mode conceals a considerable speed deficiency in practice. What’s particularly annoying for the user is that the speed cannot be raised when running on battery. Even more annoying is the fact that Dell documents this energy-saving feature neither on its German-language homepage nor anywhere in the handbook. The English homepage, in the meantime, features a footnote: “Power management features limit processor speed when running on battery.” This is pretty well hidden, however.

MP3 Audio Encoding: Lame MP3 Encoder 3.92

Lame MP3 Encoder was used under Windows XP to convert a 178-MB sound file from a WAV format to a “MPEG-1 Layer 3” format.

Here, too, the SmartStep failed to perform well in battery mode.

Archiving: WinACE 2.2

Archiving is a very practical application. WinACE 2.11 was used under Windows XP to archive a 178-MB WAV file while the clock was running.

Gaming Performance: Quake 3 Arena

As the demo runs with Quake Arena III showed, graphics tend to fall by the wayside relatively quickly at high resolutions. Although the system is good enough for the occasional quick game now and then, real gamers will be expecting more power, like the kind they get from a Mobility Radeon 7500/9000 von ATI, for example, or NVIDIA’s Geforce4 Go440/Go460 chips.

Summary: Half-Speed in Battery Mode

Owing to the really lavish configuration, Dell’s SmartStep 250N could be considered a real bargain. A closer look, however, reveals the system’s weaknesses: with a traveling weight of altogether 4.26 kilograms (device 3.7kg, power supply 0.56kg) and dimensions of 32.5 x 4.8 x 27.5 cm3, it is at best suited for short trips. The graphics resolution with the Mobility Radeon M6 does not represent the latest in technology and, far more seriously, in battery mode, the CPU runs at half-speed only. The user does not have the option of running the SmartStep on battery at the advertised speed of 2.2GHz at the expense of battery-running time. The advantage of the quick processor is thus lost in mobile operation. We find the fact that Dell fails to clearly point out this special “feature” pretty shameless, to put it mildly.

What did leave a good impression were the two hours and 47 minutes of battery-running time and the complete set of equipment.