Shuttle's SS51G Mini Barebones System

Shuttle's SS51G Mini Barebones System - Page 2

Shuttle's SS51G XPC Mini Barebones System
The Baddest Mini of Them All!

By, Jeff Bouton
August 18, 2002


Setup & Quality
Now We're Talkin'

The SS51G is the second of the mini PC product line to don a new and improved look.  This time around, Shuttle took the same face as we first saw in our SS40G review and gave it a powder blue hue.  The face plate is cut from a tinted piece of Plexiglas material that is then anchored to the front of the case.  Built around a SiS 651 North Bridge and a 962L South Bridge, SS51G is the most versatile and powerful XPC Shuttle has offered to date.  With support for USB 2.0, IEEE 1394, ATA 133 and PC2700 DDRRAM, the SS51G is a culmination of the best features available in a single, compact system.  This system is also the first of its kind to incorporate an AGP expansion slot onto the system board.  This is a major change not found in any of the previous minis and something reviewers and users alike have been asking for since the first mini PC was introduced.  While the SS51G had a similar feature set to the SS40G, we turned the unit around and found that something was fundamentally different.

Aside from the AGP slot, the most noticeable difference was the lack of a TV-Out option, an item found on each previous model of mini PC.  With this version of the XPC, Shuttle opted to incorporate the TV-Out function into an add-on AGP card, making it an optional component.  At first we were less than thrilled with this decision, thinking that the user may have to sacrifice one feature if they were planning to add on an AGP Graphics card to the SS51G.  Then we thought about it a little more and realized just how common a TV-Out feature has become with a wide range of new video cards.  In the end, the likelihood of losing the TV-out feature with a video card upgrade is remote.  The included AGP adapter also provides a digital out connection as well as a DVI port for a flat panel display, two items frequently found on newer video cards as well.  So what seemed to be a strike against the SS51G at first, is now looking like a well thought out decision to offload some onboard features that would easily be replaced with a video upgrade.

The inside of the SS51G is a familiar scene if you've seen the last few models of mini PC we've reviewed.  One thing that Shuttle began to address with the SS40G is the common complaint that the units were a bit too noisy.  The first step they took was to develop a new heat pipe cooling system for the CPU.  Dubbed the I.C.E Technology CPU heat-pipe, this system allowed Shuttle to reduce the fan count of the SS40G and SS51G respectively, from two to one, greatly reducing system noise.  While this process made a marked improvement on the system noise, there was still room for improvement.  So Shuttle took some extra steps to continue the battle against ambient noise by upgrading the system fan and giving the unit a BIOS that had greater control over fan speed.  Now the BIOS monitors the system and CPU temperature of the SS51G and throttles the fan speed based on these conditions.  The end result is a system that is so quiet that the first time we ran the unit we thought that the fan wasn't working.

As we first addressed in the SS40G review, although we like the custom face that Shuttle has been applying to the XPC line, we found that the "wow effect" was greatly diminished once we installed our beige colored drives. We soon found that this could easily be taken care of with a little spray paint and about 30 minutes of our time.  To wrap things up, we thought we'd try to match the drive color a little closer to the color of the unit to give the system a more balanced appearance.  Like the SS40G, we were pleased by the end result.  Although not an exact match, the system has a much more appealing appearance than if we left things alone. 

Next we'll give you a run down on how we set up the SS51G before we put the system to work in the benchmark arena.

The Hot Hardware Test Systems
Little but mean!

Shuttle SS51G XPC Mini Barebones System
eVGA GeForce4 Ti4600
Intel Pentium 4 "Northwood" 2.26GHz. (533MHz. Bus)
256MB Corsair XMS PC2700 DDR RAM CAS 2.5
Western Digital 30GB UDMA/100 7200 RPM Hard Drive
Sony 52X CD-ROM Drive
Standard 3.5" Floppy
On-Board NIC
On-Board Sound

Microsoft Windows XP Professional
(With all current Critical Updates)
SiS Video Drives (v2.07)
SiS AGP Drivers  (v1.10)


Set Up:
The first thing we did was enter the system BIOS and set the board to its "Optimized Default" settings and the Memory CAS Latency to 2T.  We set the CPU:DRAM Clock Ratio to 4:5, allowing the memory to run at full 333MHz. (166MHz. DDR).  The hard drive was then formatted, and Windows XP Professional was installed. We went to Windows Update site and downloaded all of the available updates, with the exception of the ones for Windows Messenger. Then we installed all of the necessary drivers, removed Windows Messenger, disabled Auto-Updating, disabled System Restore and set a 768MB permanent swap file. Lastly, we set the Visual Effects to "best performance", installed all of the benchmarking software and defragged the hard drive.

When deciding how we were going to test the SS51G, there were two things we needed to consider.  For one, we thought that the main reason someone would buy the SS51G was because of the added AGP slot, giving the system excellent graphics potential.  Then there is the crowd that insists on seeing overclocking scores with any piece of hardware we review.  So we felt we should run the benchmarks not only with onboard graphics but also with a quality AGP video card installed to show the system's gaming abilities.  Next we considered doing some overclocking benchmarks, but as you'll see below, this wasn't meant to be.


Overclocking the SS51G XPC
Not What We Hoped For...

When it came to overclocking the SS51G, we ran into a few problems.  Normally when we are overclocking a system, a good gauge to show we've gone too far is that we encounter errors when WindowsXP begins to load.  Often times we'll see a BSOD with some cryptic information on it or a message that a file is corrupt and we need to run the Recovery Console.  These are not unusual things to see when pushing hardware beyond its intended limits.  In the end, we normally drop the settings in the BIOS back a little and the errors go away.  With the SS51G, however, things were a little more difficult.

As we increased the system's bus speed to around 142MHz., we encountered the aforementioned error regarding a corrupt file and the Recovery Console.  However, when we reduced the system's settings or returned them to their defaults, we continued to receive the same message.  Normally the Recovery Console can be used for extracting files from the installation CD to replace a corrupt file, but the error would reference the entire WindowsSystem directory.  So we went ahead and did another fresh installation of WindowsXP, set up the system with the latest drivers and BIOS, and gave the overclocking one more attempt.  This time we were able to increase the Bus to 140MHz. and complete a round of scores in SiSoft Sandra, but when we rebooted the system, the error reappeared.

At this point we're not sure why the system is so sensitive to overclocking, but when we factor in how low the bus speed was set when it did fail, it didn't make much sense to continue.  Perhaps this is an isolated issue with the SS51G, or an issue with the hard drive or SiS chipset, we're not sure.  However, it's difficult to conclude that the fault lies solely with the SS51G.  We can't help but wonder if the outcome may have been different if voltage adjustments were available in the BIOS.  So we're going to forego any more overclocking attempts and focus on the system's stock performance using the onboard AGP graphics and with a GeForce4 Ti4600 installed.

Performance With SiSoft Sandra 2002
Sinfully Synthetic

SiSoftware's SANDRA (the System ANalyzer, Diagnostic and Reporting Assistant) is a very popular benchmarking, information and diagnostic utility.  We began our testing with four of the built-in sub-system tests that are part of SiSoftware's Sandra 2002 benchmarking suite:  CPU, Multimedia, Memory and File System.  We ran each test using both onboard AGP graphics and the GeForce4 Ti4600.  There are two major effects to consider after we installed the Ti4600.  First, the SS51G reserves 32MB of onboard memory for video when using the onboard AGP, so adding a video card to the system has the same effect as adding 32MB of memory to the motherboard.  Second, with less graphics data being sent to the system memory, more bandwidth is available for for other processes, increasing overall memory performance.  Lets see what the SS51G has to offer with both configurations...

    Onboard AGP            Vs.          GeForce4 Ti4600

CPU Test


Multimedia Test

With both the CPU and Multimedia tests, we saw no noticeable difference in performance.  Where we really expected to see a difference was with the remaining two tests.

File System Test


Memory Test

With the File System benchmark, we saw a decent increase of roughly 400 points with the Ti4600.  This may be considered too close to call by some, but if it is a true gain resulting from the addition 32MB of memory, we're looking at a 2.3% increase.  In the Memory benchmark, however, the increased bandwidth made for a nice increase in the system's memory performance.  With the Ti4600 installed, we saw a 4.5% boost in the Integer test, while the Floating Point test posted a gain of 4.6%.  Clearly the effects of adding a video card to the SS51G are significant, and this is only the start.  Once we get to the gaming tests, the difference should become much more apparent.

More Performance With PC Mark 2002
CPU, MEM and HD...

Next up we have MadOnion's new PCMark2002 benchmarking suite.  This program runs a series of 62 separate tests to score a system's performance in three key areas: CPU, Memory and Hard Drive. 

The SS51G is shaping up to be the first mini barebones system that has the potential to compete with full-sized high-performance systems.  The system also posted nominal gains across the board with the Ti4600 video card installed.  Clearly the memory bandwidth walked away with the most improvement, tipping the scales with a 5% increase.


Some More Benchmarking and Final Thoughts


Related content