More Intel PAT:
Today we have more news about PAT. We have more [H]'er mail coming in today about Intel's Performance Acceleration Technology and more disinformation being spread.
First, let’s address the issues of the -100MB/s scores when using two DIMMs per channel on an 875P chipset. It seems as though new information has come to light. Our hypothesis about latency issues giving us the -100MB/s was proved wrong last night by Michael over a LostCircuits (he is a great memory guru). The 875P apparently has a multifunction that has yet to be discovered. PAT does in fact have at least two different modes of operation. When a user inserts a second DIMM on one channel, PAT changes its memory timing optimizations in order to give the chipset solid stable operation that so many of us are used to seeing in Intel products. This was explained to me as PAT "balancing maximum performance with robust operating stability". So yes in fact using more than one DIMM per channel does change the way PAT works, but it is far from turning it off or disabling it as our graphs here proved yesterday.
Secondly, we are getting reports of Asus is stating that they have enabled PAT on Springdale/865PE chipsets.
Taipei, Taiwan; May 28, 2003 – ASUSTeK Computer Inc. (ASUS), the worldwide leader of motherboards, today announced its P4P800 series motherboards, based on Intel’s latest 865PE chipset, are able to active Intel’s Performance Acceleration Technology (PAT) for robust computing. Currently, ASUS’ 865PE motherboards are the only solutions in the market that provide this advanced feature.
This press release makes statements that are simply not true. Intel's Performance Acceleration Technology is very much a function of hardware. The resources to enabled PAT on a 865PE chipset would simply be too great if it is possible at all. So get it out of your head that Asus is enabling PAT on their Springdale chipset, as that is not a reality. I will have to categorize this as terribly bad and misleading marketing rather than an outright lie. What is true is that Asus has enabled some incredibly aggressive memory optimizations on their P4P800 865PE mainboard.
We finished testing their 865Pe board earlier this week and quite simply when we tested the memory bandwidth; we thought that there was an issue with it running faster than it was reporting. The scores are quite incredible and we should have the entire review on line for you next week. To give you a bit of insight though, and to give you a word of warning before you go out and buy the Asus 865PE board, we have found that when you overclock the memory bus and are running asynchronously, like you would be doing if you were running a Pentium 4 2.4C at 3GHz, that there is a very real negative impact on the real world scores. Yes, so far we are seeing 3GHz 1GHz FSB scores being outstripped by 3GHz 800MHz FSB scores. This is something every enthusiast will need to take into consideration if think you might be running the board at something other than stock settings.
