Biostar’s iDEQ 200P mini-barebones system

She’s a looker, she is

Note the sliding drive bay cover and the neatly recessed reset button. Very nice. Biostar is gunning for the Shuttle XPC line, and the company is clearly playing for keeps. Pull off the iDEQ’s side cover, you’ll find a design influenced by a bundle of new ideas, from a transverse-mounted, sliding 3.5″ hard drive tray to an AGP slot arrangement that will swallow a massive GeForce FX 5900 Ultra card. The iDEQ 200P also has much neater cable routing than the typical XPC and more open interior space. My initial impressions were firm: if Biostar has executed this concept well, there’s a new SFF sheriff in town.

The specs
On paper, the iDEQ 200P has the goods needed to become a state-of-the-art desktop system or a killer home-theater PC (HTPC). Have a gander at the basics.

The 200P supports all the latest gizmos and doohickeys, including Gigabit Ethernet, Serial ATA RAID, digital optical audio, and the Athlon 64 processor. Thanks to its single-chip nForce3 150 core logic and HyperTransport communications, the K8NBP motherboard ought to have a clean layout, even with its diminutive footprint and healthy complement of features. Let’s have a closer look at the box and see what we find.

Ogling the iDEQ
For starters, we put the iDEQ 200P side by side with Shuttle XPC SN45G for a comparison shot.

The rivals shoulder to shoulder

As you can see, the iDEQ and XPC share very similar dimensions in terms of height and width, but the iDEQ is just under an inch deeper than the XPC. In SFF systems, smaller is generally better, but I don’t consider the iDEQ’s extra depth a serious inconvenience. Nobody’s going to forget and call it a “cube,” though.

Up front, the 200P has a prominent power button, a well-protected reset button, and an array of useful ports. The optical audio port you see is an input, as it should be. There’s also a Firewire port and dual USB 2.0 connectors.

The iDEQ’s drive bay cover slides down to reveal… ack! ..an ugly beige CD-ROM drive! Better taste in peripherals might be recommended, but with the iDEQ, it’s not necessary. Notice how the sliding cover stops just where it should to avoid cutting off access to the port array.

Out back, the iDEQ 200P looks similar to most modern mini-PCs, with all the standard ports. There is no parallel port, but Biostar has almost been overly generous with serial ports, probably because the 200P doesn’t have any VGA outputs. (The slot cover is marked as if one or both of the serial ports were VGA ports.) You also see cutouts across the middle of the rear panel for what I presume would be parallel ports and the like, but Biostar doesn’t include any hardware with the 200P for populating these spaces.

In the belly of the Biostar

You can pop either side panel off the iDEQ 200P by removing a single thumbscrew. I’m unsure whether I prefer this arrangement to the Shuttle XPC approach, where the entire top shell of the case is made from one piece of metal secured by three thumbscrews. However, the iDEQ approach does make swapping a single component, like a DIMM or PCI card, a relative breeze.

The side panels must be removed before the top panel, because the side panels secure themselves to the top. Once all three covers are removed, the iDEQ’s tight interior is exposed. The two inches or so of space between the CD drive and the power supply in the picture above isn’t much to work with, believe me. The power and ATA connector barely fit with a regular sized CD drive, and there’s little wiggle room for sliding back the drive tray in order to remove it.

However, the arrangement for 3.5″ internal drives is just peachy. A separate drive tray slides in under the 5.25″ tray and tension-locks into place (there’s also a thumbscrew, so the computer will remain intact in the event of a killer asteroid strike.) This transverse-mounted 3.5″ drive cage will hold a pair of drives, one-upping the XPC in a critical way. I like it.

The drive cage’s proximity to the side of the case makes space very tight for Serial ATA power and I/O connectors. Fortunately, Biostar includes a neatly right-angled version of each with the iDEQ 200P. Unfortunately, one SATA drive does not a RAID array make. Biostar should have included a second set of cables.

With the drive cage removed, the iDEQ 200P seems very roomy indeed, at least for an SFF box. Oddly, though, the extra room seems little bit wasted because, as you can see, the CPU cooler still resides halfway beneath the power supply, and access to the CPU socket is rather tight.

Speaking of tight fits, I was just able to shoehorn a GeForce FX 5900 Ultra card into the iDEQ 200P, a feat simply not possible inside a Shuttle box. Once installed, the card worked as expected, although I have my doubts about running it long term off the iDEQ power supply. Once and only once, I had a 3DMark run with the 5900 Ultra abort on me with a message about insufficient power to the graphics card. However, I tightened up the aux power connector and never saw the message again.

Out of the box, nearly everything on the 200P that can be connected already is, and disconnected leads are left hanging in space at just the right spot, awaiting devices. The neatness is palpable, like in Martha Stewart’s sock drawer. The only omissions are the aforementioned Serial ATA cables; system builders will have to route these themselves, but then that’s easy.

Here now we come to a real beef. Right there next to the CMOS battery and the IDE connectors is the black CMOS clear jumper. That puppy is situated right in the middle of the motherboard, directly below the 3.5″ drive cage. If you want to access it, you’ve gotta disconnect and remove the 3.5″ internal drives. Even then, getting to it isn’t easy unless, perhaps, you have mousy little hands. My hands are fairly average sized, and I struggled to manipulate this jumper.

You will be clearing the CMOS a lot with the iDEQ 200P, for reasons I’ll explain shortly. Among them is the fact that Biostar provides no documented means of recovering from tweaking or overclocking failures other than an all-out CMOS reset.

The iDEQ’s CPU cooler looks more like a propulsion system than a thermal solution, but it has the advantage of incorporating a heat pipe design like the Shuttle ICE cooler without obviously infringing on any patents. This cooler does have an exposed fan with nothing to guard your fingers and knuckles, let alone any loose cables left hanging inside the 200P. I like the cooler retention mech, though, with a pair of simple tension clips to hold everything in place.

Here you can see the iDEQ cooler’s heatpipes as they turn and terminate into a bevy of copper fins. The fan on the front of the cooler blows toward the rear of the iDEQ, across these fins.

The iDEQ cooler hands off the air it’s moved to a rear-mounted case fan. If you look closely, you can see a clear plastic shroud intended to extend from the cooler to the rear fan. However, Biostar hasn’t secured the shroud well enough for it to stay. I tried to put it back into place, but the adhesive included on it wouldn’t do the trick. I suppose one could use superglue to make it stick.

BIOS tweaking and overclocking
The iDEQ 200P’s BIOS has the most complete set of memory timings tweaks I’ve seen this side of a manufacturer’s reference motherboard. Biostar’s BIOS developers have even put in explanations of what each memory timing setting does and, in many cases, how many nanoseconds a given delay ought to be.

Get out yer slide rules, geek-masters!

Frankly, I would have preferred fewer settings with better guidance for users. There’s no hierarchy of timings options like the helpfully cheesy “Turbo” or “Aggressive” choices in many BIOS menus, and once you choose to set memory timings manually, you’re almost completely on your own. The manual menu doesn’t even seem to base its default settings off, say, the SPD values on the current DIMMs or a reasonably conservative baseline.

For overclocking, the 200P’s BIOS allows HyperTransport speeds from 200 to 250MHz in 1MHz increments, along with a few different voltage settings each for the CPU and memory. What you won’t find, however, is any option to change AGP/PCI divisors.

The 200P’s hardware monitoring features are quite good, with a full complement of voltage and fan speed monitoring options and a CPU temperature threshold for system shutdown. I wish there were some fan failure alarms, too, but inside the 200P, the CPU temperature should probably serve that purpose. The 200P also has temperature-controlled CPU and chassis fans, which we’ll talk about next.

Smart fans, not-so-bright control
I’ll start with the good about the iDEQ’s temp-controlled fans, then move on to the bad. First, the 200P supports AMD’s Cool’n’Quiet technology, which is just another dreadful name for AMD’s PowerNow! technology, which is just another dreadful name for activity-based clock throttling. With Cool’n’Quiet enabled, the Athlon 64 adjusts its clock speed 30 times per second in response to CPU load. If you’re just checking e-mail and surfing the web, a 2GHz Athlon 64 is likely to hover around 800MHz or so, since there’s effectively almost no load on the CPU. Kick off a 3D game, though, and it cranks up to full speed almost immediately and stays there. Here’s a shot of Biostar’s System Control utility with Cool’n’Quiet in action.

This feature has the potential to keep the iDEQ 200P, err, cool and quiet.

To measure the iDEQ 200P’s noise levels, I used an Extech (model 407727) sound level meter situated approximately one inch from the front and rear of the case, outside the path of any airflow generated by fans. I took two sets of measurements, one with the iDEQ’s fans running at their minimum speeds, and another with them running full tilt.

The Biostar box shows up as surprisingly loud in these tests, a fair bit noisier than a Shuttle XPC, especially from the rear, and especially under load. Subjectively, the iDEQ 200P doesn’t seem much different, in my experience, from most systems. Still, it is a bit on the noisy side for an SFF rig.

Now for the bad news. In order for the 200P’s fan control to be effective, Biostar’s System Control utility must be running at all times. That’s bad news for Linux users and other folks looking to use the 200P in a non-Windows environment. Even in Windows, you’ve gotta keep the utility running, or you’re greeted with this dire warning:

Scarier still, a good portion of the time, the utility doesn’t seem to initialize properly on booting into windows, so the fans never drop to a lower speed setting. In order to make it kick it, you’ve got to maximize the utility window from its place in the system tray, at which point it will slow the fans down.

The Biostar utility offers three cooling modes: Normal, Quiet, and Fuzzy. The documentation for the system utility doesn’t describe the difference between these modes very well, but Normal seems to be the proper choice in all situations, because it allows the fans to auto-adjust to temperature changes.

However, the Biostar utility is a total CPU hog. It constantly uses between about five and ten percent of CPU time. Task Manager shows it like so:

That’s quite a bit of CPU utilization out of a utility that always has to be running. What’s more, the utility polls hardware at regular intervals every four or five seconds. When it does so, other tasks are interrupted. This effect seriously disrupts the computing experience, especially in something like gaming, where intermittent pauses become extremely annoying very quickly. To give you some idea about how the iDEQ utility affects the user experience, I plotted frame rates over time from our Splinter Cell benchmark, both with and without the utility. Without the utility, the frame rate graph looks about like it does for any system.

With the Biostar System Control utility running, the iDEQ 200P’s frame rates dip at regular intervals, as you can see below.

This problem essentially destroys playability on the 200P, and it would likely hamper the system’s ability to deliver a smooth computing experience in HTPC applications, as well. I tried updating to the latest revision of the System Control utility on Biostar’s website, but the new version only exacerbated the problem.

Now I understand why MSI has Core Cell and Abit has uGuru to handle onboard monitoring tasks independent of the CPU. The System Control utility is just about a deal killer for the iDEQ 200P. Essentially, to have a usable system, you have to turn off the utility and accept that the box is going to be loud.

A few more bumps on the road to happiness
Beyond the slowdowns caused by Biostar’s System Control program, I found a handful of other problems with the iDEQ 200P that made the user experience difficult. For starters, it had very serious memory-related stability problems. The iDEQ 200P would not play nice with some of the DIMMs I tried in it, including several different matched pairs of DDR400 memory from Corsair of varying sides and SPD types. When I could get the 200P to boot with a pair of XMS3200LL DIMMs installed, the system wasn’t entirely stable in Windows. I attempted to back off the timings by setting them manually, but no matter what I tried, the system wouldn’t POST with manual RAM timings.

Repeat the account above five or six times, and you have some idea what it was like working with the 200P. Magnify by the fact that every failed POST required removal of the 3.5″ hard drive cage and reset of that nearly inaccessible CMOS jumper I told you about. The iDEQ 200P wasn’t making any friends with this act, especially because Biostar offers no apparent means of recovering from a bad BIOS setting by holding down a hotkey at boot time or anything like that.

Finally, I settled on a pair of Kingmax 256MB DDR400 DIMMs with a not-very-aggressive SPD (CAS latency 2.5) for testing. They have less RAM and more relaxed timings than the DIMMs I used on our MSI K8T Neo comparison motherboard, but I had to take what the 200P was willing to give.

The iDEQ was generally stable in Windows with the Kingmax RAM, but nothing I could try—four or five different pairs of DIMMs, using a single DIMM, running the RAM at 333MHz—would allow the 200P to finish a memory benchmark. SiSoft Sandra’s memory bandwidth benchmark would produce a reboot, as did cachemem. This problem wasn’t just a quirk of my particular review unit, either. Mike Schuette of LostCircuits ran into the same problems.

I asked Biostar what to do about this issue, and I was told to try raising the memory voltage as high as possible. Unfortunately, 2.85 volts didn’t resolve the problem, and 2.9V caused the system not to POST, leaving me no choice but to clear the CMOS again.

As a result, the benchmarks you’ll see below are incomplete. They don’t include memory performance, of course, but they also don’t include some forms of testing we generally do for verification, like USB, Ethernet, audio, and disk I/O performance. We’re often just checking for basic competency in these areas, but in the case of the iDEQ 200P, we spent the majority of our time troubleshooting memory problems. Finally, our time and patience ran out. We already have enough data to know what we think of the iDEQ 200P without delving into Ethernet CPU utilization numbers.

Our testing methods
As ever, we did our best to deliver clean benchmark numbers. Tests were run at least twice, and the results were averaged.

Our test systems were configured like so:

AMD Cool’n’Quiet was disabled for all tests.

Thanks to Corsair for providing us with memory for our testing. If you’re looking to tweak out your system to the max and maybe overclock it a little, Corsair’s RAM is definitely worth considering.

The test systems’ Windows desktops were set at 1152×864 in 32-bit color at an 85Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests.

We used the following versions of our test applications:

• Cachemem 2.65MMX
• SiSoft Sandra 2004 (9.89)
• Cinebench 2003
• Sphinx 3.3
• LAME 3.95.1 (build from mitiok.cjb.net)
• Xmpeg 5.0.3 with DivX Video 5.11
• FutureMark 3DMark03 build 340
• Comanche 4 demo
• Quake III Arena v1.31
• Serious Sam SE v1.07
• Splinter Cell v1.2
• Unreal Tournament 2003 demo v.2206
• Wolfenstein: Enemy Territory v2.55

The tests and methods we employ are generally publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.

LAME MP3 encoding
We used LAME 3.92 to encode a 101MB 16-bit, 44KHz audio file into a very high-quality MP3. The exact command-line options we used were:

lame –alt-preset extreme file.wav file.mp3

DivX video encoding

The iDEQ is s few steps slower than the MSI board in our media encoding tests.

Speech recognition
Sphinx is a high-quality speech recognition routine that needs the latest computer hardware to run at speeds close to real-time processing. We use two different versions, built with two different compilers, in an attempt to ensure we’re getting the best possible performance.

There are two goals with Sphinx. The first is to run it faster than real time, so real-time speech recognition is possible. The second, more ambitious goal is to run it at about 0.8 times real time, where additional CPU overhead is available for other sorts of processing, enabling Sphinx-driven real-time applications.

Speech recognition is more of the same story. The iDEQ 200P isn’t a bad performer in the grand scheme of things, but it is pokey compared to the MSI board with the same speed CPU and the same on-chip Athlon 64 memory controller.

Cinebench 2003 lighting and rendering
Cinebench is based on Maxon’s Cinema 4D modeling, rendering, and animation app. This new revision of Cinebench measures performance in a number of ways, including 3D rendering, software shading, and OpenGL shading with and without hardware acceleration.

Cinebench isn’t much different from our other tests. The iDEQ 200P can’t quite keep up with the K8T Neo.

Quake III Arena

3DMark03

Serious Sam SE

Comanche 4

Unreal Tournament 2003

Wolfenstein: Enemy Territory

Regardless of the application, the iDEQ 200P runs demonstrably slower than the full-sized MSI K8T Neo motherboard we chose for comparison.

The iDEQ 200P was penalized in our tests by its unwillingness to run with tight memory timings and by its inability to play nicely with a couple pairs of 512MB DIMMs we tried in it. Even so, the iDEQ 200P was unexpectedly slow compared to the competition.

Conclusions
Biostar started well with the iDEQ 200P, but they couldn’t deliver the total package. The iDEQ’s intelligent enclosure design, smart cable routing, slick packaging, and endearing good looks were a heckuva nice start. Biostar has incorporated a number of clever, fresh ideas into the physical design of the iDEQ, showing us that there’s plenty of room to improve on the SFF standard set by Shuttle’s XPC lineup.

However, the fit and finish of the iDEQ 200P isn’t quite up to Shuttle’s standards. The main drive cage for external drives doesn’t fit easily into its slots, so installing it requires a little more force than I’m comfortable applying. With the covers removed on three sides, the iDEQ seemed to bend and sway too easily, making it feel flimsy, and aggravating the fit problems for the drive cage. Similarly, the cooler shroud on our 200P review sample was flimsy and not securely fastened.

Were these the only problems with the iDEQ 200P, it would still be a strong competitor for Shuttle’s XPC. Sadly, though, Biostar’s K8NBP motherboard just isn’t up to snuff. We saw too many crashes and too many memory incompatibilities with this board. One gets the impression perhaps Biostar hasn’t yet figured out how to design a motherboard for the Athlon 64 with its unique integrated memory controller. Even if the system will run without crashing, Biostar’s fan control utility spoils the iDEQ 200P as a gaming system or a home-theater PC. It’s a crying shame, because this box looks so freaking cool.

Given all the problems we experienced with memory compatibility and system stability, I can’t recommend the iDEQ 200P. I do hope Biostar finds a competent motherboard to mate with this captivating little enclosure, but based on our experience, that hasn’t happened with this model.