Asus P4B266-E and P4S333 in the test: Intel 850 versus Intel 845 and SiS 645

Asus P4B266-E and P4S333 in the test: Intel 850 versus Intel 845 and SiS 645


Even in the feature list, it is noticeable that the P4S333 is competing for buyers in a different price range than the P4B266-E, for example. We missed a raid controller in our version, as well as onboard LAN, USB2.0 and a digital SPDIF input and output. Of course, as with the P4B266-E, it is possible to redirect the mic and line-in connection so that they function as rear and center/bass-out, and the 5.1 sound is output at least in analogue. This must be set via onboard jumpers.

P4S333 connections
P4S333 sound chip
P4S333 Board2
P4S333 power connector
P4S333 USB module

The equipment of the P4S333 is thus complete, but offers nothing special. Asus is of course still selling the P4S333 in a somewhat more extensive version with onboard LAN and digital SPDIF input and output. In no case does it match the features of the P4B266-E. At a price of around € 130 you shouldn't count on the generous scope of delivery of the P4B266-E, which costs almost € 200 and is clearly aimed at a different price range.

The biggest advantage of the P4S333 is the support of PC2700 DDR SDRAM (DDR333). DDR333 delivers a maximum data transfer of 2.7GB and closes the gap between the maximum usable bandwidth of the Pentium4 (3.2GB) and the previously available bandwidth of DDR266 (2.1GB) at least to some extent. Theoretically, the SiS645 in conjunction with DDR333 should therefore provide better performance than the i845 with DDR266. Our benchmarks will show whether this is true.


We tested the P4S333 with both the Apacer CL2 DDR266 and the brand new Crucial CL2.5 DDR333, which is one of the first met the JEDEC specifications.

The Apacer DDR266 ran in our test with the settings 'CAS Latency' 2,'RAS to CAS Delay' 2, 'RAS Precharge Delay' 2, 'Active Precharge Delay' 5, 'Command Lead-Off Time' 2. The 'Active Precharge Delay' would still have been 4 and the 'Command Lead-Off Time' “Can be lowered to 1, but the P4S333 refused to boot even if these settings were tightened. Thus the P4S333 ran in DDR266 mode with the same settings as the P4B266-E. So, to our delight, the values ​​can be compared very well.

With the Crucial DDR333, we were unfortunately not able to increase the timing quite as much. Crucial's DDR333 could not be moved to stable CL2 operation even with 2.7 volts dimming voltage, so we had to throttle it back to the sometimes much slower Cl2.5 value. Our DDR333 timings are therefore 2.5, 2, 2, 5, 2.

In terms of stability, we can use the P4S333 unfortunately not award the top grade of the P4B266-E. The board was absolutely stable, but only when it was working. For example, if we set the RAM timings so sharply that the board refused to boot completely, we heard the nice message from the Post Reporter: 'System failed memory test'. Not that tragic in itself, but the P4S333 was then very difficult to move to reboot. Several CMOS clears and the associated complete shutdown of the power often did not move the board to restart immediately. Instead, we were greeted with the nice message 'System failed CPU test' several times. However, as we checked several times, our CPU was installed correctly and in no way damaged. After a few more CMOS clears, the P4S333 suddenly started completely smoothly, although we had not changed anything on the system. Most of the time 20 minutes had passed.

We could have got over this point if it had really only been tried in connection with overclocking or if the RAM timings were too sharp,but even after we shut down the computer and only wanted to restart it, it refused to do so now and then with the now well-known message “System failed CPU test”. Asus has been offering a new beta bios for the P4S333 on their FTP for a few days, with which the problems may already have been fixed. The official version 1005c was used in our test.

But as I said, when it ran, we had no problems with the stability here either.

On the next page: Overclocking