A fast-growing trend in computing today is wireless networking, and the SB65G2 embraces that trend by including 802.11b connectivity out of the box. In typical XPC fashion, that connectivity is accomplished elegantly using a solution specifically designed for XPCs. But is this added functionality enough to differentiate the SB65G2 in a growing sea of small form factor systems? Let's find out.
Before we get any further along, here are the technical specifications for the SB65G2.
|CPU support||Socket 478-based Intel processors with 400/533/800MHz front-side bus|
|Form factor||Flex ATX (Shuttle form factor)|
|North bridge||Intel 82865PE MCH|
|South bridge||Intel 82801EB ICH5|
|Interconnect||Accelerated Hub (266MB/s)|
|PCI slots||1 32-bit/33MHz|
|AGP slots||4X/8X AGP|
|Memory||2 184-pin DIMM sockets|
Maximum of 2GB of DDR200/266/333/400 SDRAM
|Storage I/O||Floppy disk|
2 channels ATA/133
2 channels Serial ATA 150
|Audio||6-channel audio via 865PE ICH5 and RealTek ALC650 codec|
|Networking||10/100 Ethernet via RealTek RTL8100B|
802.11b via Shuttle PN11 wireless module
|Ports||1 PS/2 keyboard, 1 PS/2 mouse,|
1 serial, 6 USB 2.0 (2 front, 4 rear), 2 IEEE 1394 (1 mini front, 1 rear), 1 RJ45 Ethernet
2 line out/front out (1 front, 1 rear), 1 rear out (rear),
|Bus speeds||100-355MHz in 1MHz increments|
(400-1420MHz quad pumped)
|Monitoring||Voltage, fan status, and temperature monitoring|
The SB65G2 packs all the typical Springdale goodies, including support for dual-channel DDR400 memory, an 800MHz bus, and Hyper-Threading processors. Other highlights include a total of six USB 2.0 ports and two channels of Serial ATA. Speaking of mass-storage controllers, the SB65G2 keeps the SB61G2's ICH5 south bridge, so there's no RAID joy. You might discount the idea of RAID on a cube, but there are two 3.5" bays here, and besides, Shuttle popped for the ICH5R on the SB75G2 and the SB62G2.
Another important inclusion here is the 802.11b wireless networking capability, which comes courtesy of Shuttle's PN11 module. The PN11 is standard with the SB65G2, but it's also available as an add-on for a number of XPCs, and starts at around $55 on Pricewatch. We'll take a closer look at the PN11 module during our photo tour of the SB65G2.
In October of last year, Shuttle announced an intriguing new feature for their 865-based XPCs in the form of P-BIOS. P-BIOS is intriguing largely because, to the best of my knowledge, Shuttle has never explained exactly what it is. The press release announcing it speaks only to its effects ("P-BIOS gives user’s as much as 30 percent performance increases"), saying nothing about its method of operation. While preparing this review, I wrote Shuttle twice asking for information on exactly what P-BIOS is, but received no reply.
Of course, anyone who's familiar with these things hears about a BIOS tweak for 865 chipsets with the letter P in the name, and immediately thinks PAT. If you're not familiar with PAT, you can read about it here. Basically, it's a feature of the 875P Canterwood chipset which reduces memory latency within the memory controller when the system is using dual-channel DDR400 memory and an 800MHz front-side bus. Shortly after the release of the 875P and 865PE chipsets, several motherboard makers announced BIOS upgrades which would enable PAT on their 865PE Springdale boards (though all were careful not to use the term PAT when describing this functionality). So, the question is, is P-BIOS an implementation of PAT?
Well, to quote from the Magic 8 Ball, "Outlook not so good." There are a number of factors that make it very unlikely that P-BIOS is enabling PAT. First, the P-BIOS feature is available on all 865-based XPCs, including the SB61G2 and SB62G2, which are based on the 865G chipset. As far as I know, all the hacks to enable PAT have been restricted to the 865PE chipset, so if Shuttle is enabling PAT on the 865G, they're really breaking new ground. Much stronger evidence is found in the help screen for the feature, which is called "Turbo" in the BIOS screens. Take a look:
There are two possible settings for the Turbo function: Auto or Enable. As you can see, Auto mode enables the mysterious Turbo function when the system is using a processor with a 533MHz front-side bus, while Enable extends the feature to 800MHz front-side bus processors, as well. This is a huge strike against the PAT theory, because PAT only operates in conjunction with an 800MHz front-side bus and dual-channel DDR400 memory.
Of course, the best way to judge the question might be to look at the effect of the Turbo feature on performance, which we'll see in the benchmarking section later. It will be interesting to see how the benchmarks compare to Shuttle's claims of an up to thirty percent performance increase.