| || |
| Quality and Setup of the VIA P4PB Ultra |
| You could say this board has connections... |
The VIA P4PB is both fully featured and a good looker. Set off against a blue PCB, the red jumpers and secondary IDE ports and yellow USB 2.0 headers really stand out. This makes it easier to find that jumper that you need to disable, but usually can't find quickly enough. Even though there are many features on board, everything appears to be laid out cleanly without too much crowding of any one area. The ATX connections are placed to either side of the CPU, far enough away to prevent hindering the airflow in and around the processor. Also, the majority of the capacitors reside either alongside the CPU bracket or between the socket and I/O connections, allowing for easier access to the jumpers and headers.
There are three 184-pin DDR memory slots, which in theory can support up to 3GB of DDR400, 1GB more than its Intel counterparts. We did not have the opportunity to test the board with all three banks filled, and have a strong feeling that stability would be hard to attain in such a setup with DDR400 memory. A standard gripe about the placement of the PCI slots was the close proximity to the AGP card, but this is usually the case with a board that has this many slots. Speaking of which, there are 8 slots in total: a 1.5V AGP 8x slot, 6 32-bit PCI slots, as well as one CNR slot. The AGP slot uses a standard plastic retention clip to keep the card locked in place, a feature found on most modern boards.
Other device connections are provided by two ATA133 ports placed along the edge near the middle of the board as well as the floppy drive connector. The aforementioned red secondary and RAID capable IDE ports, are placed at the far end of the board, perpendicular to airflow from a front-mounted case fan. These ports are controlled by the nearby Promise 20276 ATA133 RAID controller, a common chipset for ATA133 RAID support. As this is considered more of a "lite" controller, it only supports RAID 0 or 1, but not both.
Normally, the VT8235 SouthBridge would provide support for up to six USB 2.0 ports, but the addition of VIA's VT6202 Host Controller added an extra four USB 2.0 ports bringing the total to a whopping 10 ports! Included in the bundle was a bracket that had four ports, while the other four would be well suited to be used with cases with front mounted USB ports. Also integrated into the SouthBridge was the Ethernet MAC, which along with the VIA VT6103 Physical Layer Device provided 10/100Mbps LAN support. Rounding out the features was the on-board audio support which came from C-Media's CMI8738 6-channel audio controller. There are CD-in, AUX-in, Modem-in, and S/PDIF headers on the board, as well as another bracket that had optical and coaxial audio jacks for a complete audio solution.
The external connections included the standard PS/2, parallel, and serial ports. There also were three standard audio jacks, a game port, 2 USB 2.0 ports and the RJ-45 LAN connection. Add in the other two brackets, and you can see that a whole range of devices can easily be connected to the system. Going one step further, VIA has also included a card reader breakout box, a boon to those of us with digital cameras. It fits into a standard 3 1/2" bay and can read memory sticks, smart cards, and secure digital cards. It would be nice to see companies move away from these three types and perhaps provide readers for some other types such as Compact Flash, which are a bit more common in cameras. Another handy little addition, seen in the last photo, was a power adapter which converts the line from an older power supply unit, into the two connections needed for P4 boards. This actually came into use during this review when the PSU we were using failed, and we didn't have a P4-capable unit ready to go. We quickly dismantled an older P3 rig we had nearby, put the power adapter on, and were back up and running in no time.
THE BIOS AND SETUP:
The AWARD BIOS used for VIA P4PB Ultra was a little bland, in that it covered all of the bases without anything truly new thrown into the mix. We found the usual options for the CAS latencies and Command Rate as well as setting the AGP Mode and Aperature Size. Memory timings can be switched between 133, 166, and 200MHz and the CPU can be set anywhere between 133 and 200MHz. For use in overclocking the AGP volatges can be raised from 1.5V to 1.8V in 0.1 increments and the DRAM between 2.5V and 2.8V. Oddly, the CPU voltage choices were much more limited, only allowing us to add up to an additional 0.1V, and possibly limiting us from overclocking further. The good news is, this oversight can easily be corrected with a BIOS revision, so keep an eye out for that.
We set up the system using all of the optimal defaults but did run into a few issues. The system would not restart after making any changes in the BIOS. We had to disconnect power to the board before we could get it to boot again correctly. Although there were presets for the RAM (Fast, Turbo, etc.), we chose to manually edit the CAS ratings, putting in 2-2-5-2 with a Command Rate of 1T, and did not suffer from any stability issues using DDR333 timings. Raising the memory to DDR400 speeds would invariably cause Windows to crash, until we set the CAS Latency back to 2.5, possibly hindering performance by doing so. Since we were using good quality Corsair PC3200 RAM, we are fairly certain this was not the issue.
| || |
| Overclocking the VIA P4PB Ultra |
| A little tweak here, a little voltage there, and Voila! |
When we got around to overclocking the VIA P4PB Ultra, we decided that it would be best to use DDR333 settings for the RAM. Even though the ratings on the Corsair XMS stick of DDR we used should handle speeds over 400MHz, as you will see, the overall performance was better when set to 166/333. After having some good success overclocking Intel's i845PE boards in the past, we were somewhat at a loss as to how VIA's board would fare, especially with the limited options in raising the CPU Voltages. As we raised the speeds in the BIOS we found that we could get into Windows easily enough until we hit 148MHz, at which point the system itself would not get past the POST until we lowered the Command Rate to 2T. We continued on and finally reached as high as 157MHz for the FSB with the CPU VCore raised the full 0.1V that the BIOS allowed us. Unfortunately, the system was never fully stable and crashed from time to time while running the benchmarks. We started lowering the FSB in 1MHz increments all the way back to 150MHz before we had a testable system. While this did raise the effective CPU speed to 2.55GHz, an increase of about 13%, it wasn't quite as high as we could attain with this CPU and other i845PE boards.
System Comparisons and Sandra Scores