System setup and testing notes
Motherboard | Foxconn G33M-S | Foxconn G965MA-8EKRS2H | Sapphire PI-AM2RS690MHD |
---|---|---|---|
CPU | Intel Core 2 Duo E4300 (1.8GHz, 2MiB L2 cache, 800MHz FSB, LGA775) | AMD Athlon 64 X2 4200+ (2.2GHz, 2 x 512KiB L2 cache, 200MHz HTT, Socket AM2) | |
BIOS revision | 722N1S01 (17/08/2007) | 631F1P36 (12/08/2006) | 2K070211A (11/02/2007) |
Memory | 1GiB (2 x 512MiB) Corsair XMS2-5400UL | ||
Memory timings and speed | 4-4-4-9 @ DDR2-667.2 | 4-4-4-9 @ DDR2-670 | 4-4-4-8 1T @ DDR2-628.6 |
Integrated graphics | Intel GMA 3100 8MiB fixed 256MiB DVMT | Intel GMA X3000 8MiB Fixed 256MB DVMT | Radeon Xpress 1250 256MB HyperMemory |
Disk drive(s) | Seagate 160GB SATAII (ST3160812AS) | ||
Mainboard software | Intel INF 8.300.1013 | Intel INF 8.1.1.1010 | CATALYST 7.1 |
Graphics driver | PV 14.29 | PV 14.27 | CATALYST 7.2 BETA |
Operating system | Windows XP Professional 32-bit with SP2 |
Tests
Benchmarks |
ScienceMark 2.0 memory bandwidth
HD Tach 3.0.1.0 - SATA average read speed
Far Cry v1.33 - 640x480 and 1024x768 |
---|
We put the Foxconn G33M-S up against the Sapphire PI-AM2RS690MHD, which we looked at in our recent AMD 690G comparisons. The Sapphire board was chosen because 690G represents the strongest IGP currently available. We've also added a Foxconn G965-based motherboard into the mix, too.
Testing notes
The Foxconn board tested was not off the production line but, instead, a pre-production sample.
Testing was carried out with the S01 BETA BIOS, as previous BIOS versions did not allow memory-timing adjustments. However, even with the S01 BIOS, we could not set the TRAS timing below 9, as opposed to the 8 of the AMD 690G system.
The S01 BIOS also had issues with the OCZ memory we first tried, necessitating a switch to Corsair's XMS2-5400UL modules.
Also, the Foxconn G33M-S failed to recognise our Mitsumi FA-401 USB FDD + 7-in-1 card reader. This, combined with the lack of an on-board FDD connector, meant we were unable to install AHCI drivers for Windows XP. Consequently, testing was carried out with the controller set to IDE mode.
However, we fully appreciate the pre-production nature of the board and expect these problems to be ironed out in due course.