RSS Feed

HCW Tech Blog

For the latest info on computer hardware, tech, news, video games, software tips, and Linux, check out our new improved front page: HCW Tech Blog

Reviewed by: Carl Nelson [06.02.03]
Manufactured by: Intel

Discuss this article in the forum!
Registration NOT Required!


The Tests

All tests were done on identical hardware.  The only thing that changed was the motherboards, and in the case of the 850E, RDRAM was used of course.

We ran each chipset in the following configurations:

We'll put this table on the top of every page for reference.  This order will be maintained in all tests, so try memorizing the order! Keep in mind the following things:

  • 875P results will be in red.
  • 865PE results are in blue.
  • For those 2 chipsets, the brighter colours indicate the 800FSB CPU's.
  • DDR333 is utilized when 533 FSB processors are in use.

Windows was reinstalled for each chipset.  All the testing software we used that is publicly available can be found on our Downloads Page.

Let's kick things off with what will be affected most - Memory Bandwidth!

Memory Tests

We always like to start things with SiSoft Sandra's Memory Benchmarks.  Let's keep with that tradition:

This is where the most profound results will be, so pay close attention to this chart.  As you can see, Dual Channel DDR400 and 800 FSB together make a HUGE difference in these tests.  PAT doesn't seem to be doing much more than a few hundred MB/s in bandwidth here.  Also not how far behind Single Channel DDR333 makes.  Dual Channel DDR333 is right inline with RDRAM, regardless of PAT.

Aida32 is a wonderful PC Information suite that has a built-in memory benchmark tool that measures read and write capabilities independantly.

Memory Write time benefits wonderfully from Dual Channel memory, both on DDR and RDRAM.  It looks like PAT affects Memory Write more significantly here, as the 875P's DDR333 scores are faster than 865PE's DDR400 scores.

On the Read side of things, we're seeing similar results to the SiSoft scores above.  PAT looks to be making more of a difference here, and once again you can see how much single channel memory suffers.

To round things off for memory tests, let's have a look at PCMark's Memory Test scores.  PCMark isn't exaclty the most robust benchmarking tool, and can actually be quite flaky, particularly with the HDD tests.  But it uses real world tests to give an overall theoretical score.

Once again, similar numbers to the raw memory tests.  PAT is showing up again (welcome!), and once again RDRAM is keeping up with 865 with the 3.06 GHz 533 FSB P4.  845PE doesn't look so embarassing here.

Next Page: (4)