facebook rss twitter

Review: Shuttle's XPC SB81P

by Ryszard Sommefeldt on 20 July 2004, 00:00

Tags: Shuttle, Intel (NASDAQ:INTC)

Quick Link: HEXUS.net/qazf

Add to My Vault: x

System Setup and Notes

Hardware

Motherboards

Shuttle XPC SB81P, Intel i915G, LGA775, PEG16X, DDR-I
ABIT AG8, Intel i915P, LGA775, PEG16X, DDR-I
ASUS SK8V, VIA K8T800, Socket 940, AGP8X, DDR-I

Processors

Intel Pentium 4 Extreme Edition 3.4GHz, LGA775
AMD Athlon FX-51, Socket 940

Graphics Cards

NVIDIA GeForce 6800 GT, AGP8X

Memory

1GB (2 x 512MB) Corsair XMS3500RE - 2-3-2-6 - Socket 940 system
512MB (2 x 256MB) Corsair XMS3200LLPT - 2-2-2-6 - LGA775 systems

Hard Disks

1 x 36.6GB Western Digital Raptor SATA

Software

Windows XP Professional w/ SP1
Intel Graphics Driver 14.5.0.3865 (SB81P) Intel Application Accelerator 4.0.0.6211 (LGA775 boards)
Intel 6.0.0.1014 chipset drivers (LGA775 boards)
NVIDIA Detonator 61.34 (Socket 940 board)
VIA Hyperion 4.51v (Socket 940 board)
DirectX 9.0b End User Runtime

HEXUS Pifast
Sciencemark 2.0
KribiBench v1.1
LAME 3.92MMX encoding U2's Pop album at 192CBR
Realstorm Raytracing Benchmark
3DMark 2001SE
3DMark03
Painkiller
HDTach 3.0.0

Notes

I might as well get it out in the open, but my sample SB81P died before I could get the PCI Express benchmarks done, along with some extra testing with the 6800 Ultra which fits the XPC, the BIOS screenshots and the noise-level testing. After the GMA900 on-board stuff was run, the system had a 6800 GT dropped in, which I use for PCI Express benchmarks. About ten seconds into bootup after that there was a bang, a bit of a flash, some smoke and both ring-main circuits tripped, leaving me needing a change of underpants, the master fuses needing reset and the SB81P needing a new PSU.

When said new PSU/SB81P arrives, I'll finish off the benchmarking and extra testing. So on the graphs today, you'll only see numbers run with the onboard GMA900 video.

It's (mostly) a DX9.0 part, executing parts of the API on your host CPU and the rest on its own core. Intel's reckoning is that the CPU is more than able to process vertex shaders quick enough for decent performance. When they say decent, I reckon they mean passable. We'll see.

The GMA900 on-board core was set to use 8MB of fixed video memory, allocated from the physical memory pool and used for 2D display, along with 128MB of fixed allocated memory from the physical pool when the 3D core is active, along with a further 128MB of memory the core/driver can dynamically allocate in the event of the fixed allocation not being enough. So there was a full 256MB of available memory for the 3D core, during the 3D tests.

Testing the SB81P, it became apparant that motherboard makers still haven't decided to fit high-quality DACs to on-board VGA cores. In 2004, using the most modern of on-board cores on a high-end XPC, I wasn't expecting the 2D Windows picture on my Sony G400 CRT at 1280x1024 to be a blurry, muddy mess. It was, to my utter frustration. Having to pop in a discrete graphics card to get usable 2D, a PCI Express graphics card at that, something that's very hard to get hold of at the moment, doesn't really sit well with this reviewer. That's especially true given the cost and target market of this enthusiast-level XPC.

So the SB81P, equipped with 3.4 Extreme Edition and 512MB of Corsair's finest, gets to do battle with the ABIT AG8, a similarly specced mainboard, sporting LGA775 socket, i915P core logic and DDR memory. My usual FX-51 platform makes up the numbers. Can the SB81P keep up?

CPU-Z System Information

CPU Information

Device Manager

Device Manager