32-bit Linux can use PAE to allow >3GB of memory. We have 6GB to 8GB on a few servers here that are still on 32-bit Linux.

Windows can't do this as lots of drivers are too crappily written to handle PAE.

Ian

I have a Core i7 with 8GB of Dual Channel memory and a Radeon 4850 graphics card.

With Windows Vista Ultimate 64bit I'm having numerous problems. I've had the machine 6 months and I'm still experiencing graphics driver issues (card has been swapped out) to the point where I will need to return the unit.

Hence, the recommendation to use a 64 bit operating system, while providing access to more memory, may cause problems in itself and needs careful investigation.

I think your review shows up a limitation in PCMark rather than triple channel memory not providing much benefit over dual; I agree that the average user wouldn't see the benefit but then again the average user isn't going to be building a high spec Nehalem based system and worrying about overclocking are they?

A review where the conclusion is to use as much RAM as possible, the latest processor and most up to date software.

A day well spent.

Cool, I skipped from page 1 to page 7, scrolled down, and saw what I had already in place with my 965 extreme, 4 months ago. :)

Any review based on PCMark as your main "real world" test sucks. Also, DDR3 1600 at loose timmings is cheaper than 1333 at tight timmings, and about the same price as good 1066 memory. Plus it's slightly 'real world' faster than both as well. Also (I'm guessing here) it probably has more range when overclocking (speed, voltage and latency.)

I would actually tend to say that your problem is the graphics card itself. You stated the drivers are whats causing trouble not the CPU or the RAM. ATi is notorious for shitty drivers. Before you return it try an Nvidia card and see if the problems persist then check your results.

You are wrong, Windows Server has had PAE since the Server 2000 days, and I can be enabled on the desktop variety of Windows as well. Bing it next time before you open your ignorant yap.

... which says that more bog-standard memory trumps less but faster memory every time.

Slightly surprised that 3-channel offers no noticeable advantage. Perhaps it's time will come with future iterations and speed steps of Intel's new architecture. Anyway, there's a financial advantage: 12Gb without needing to buy expensive 4Gb DIMMS.

PAE and multicore CPUs means that 8Gb or even 12Gb may be sensible with 32-bit Linux: 2Gb or 3Gb per process, each running flat out in its own core. But if you aren't constrained by some sort of historical relic, 64-bit Linux should be today's default. I doubt I'll be doing many new 32-bit installs in the future.

I think you are absolutely correct. Alas, the product was sold with the card by the manufacturer and they won't swap it for an NVIDIA. They keep on getting me to reinstall the OS and try the latest drivers from the ATI site.

I intended (obviously poorly) to make the point that 64bit drivers may be an issue and hence to move to a 64 bit O/S after doing some research.

Ian didn't say Windows "couldn't do PAE - period". He said it couldn't do PAE because too many drivers couldn't handle it, having not been written with PAE in mind.

Here is what Microsoft say on the issue -

http://www.microsoft.com/whdc/system/platform/server/pae/paedrv.mspx

And the wikipedia page on Physical Address Extension says -

"However, desktop versions of Windows (Windows XP, Windows Vista) limit physical address space to 4 GB for driver compatibility reasons."

Microsoft themselves confirm that >4GB is a no-go with 32-bit XP and Vista.

http://tinyurl.com/n279v6

So, very limited PAE with Windows on the desktop and deeply scary compatibility issues with PAE on both servers and desktop. We've tried PAE on desktop and server Windows and it quickly became clear that the pain of moving to 64-bit was less than the pain of trying to get PAE stable and effective.

With Linux, we installed a PAE kernels, rebooted, and the servers all worked exactly as before but with much more memory. We are now moving to 64-bit (with virtualisation where required) but it has bought us a few years.

Ian

This is very worthwhile reporting!

It is little known that the only fully performant memory configuration for dual processor AMD Opterons has been exactly 8 DIMMS of identical density, 4 on each socket, at least according to my tests. Other configurations give poorer measured performance, which may or may not be reported by the BIOS.

I have only tested with an in house tool, Opteron versions up to Barcelona. Anybody concerned about memory performance should repeat the tests on more modern hardware.

I find it amazing that such basic information is not clearly documented and is also rarely tested and reported-...

Yes, I find it truly surprising that an 8-DIMM dual-opteron setup was not tested in this article for Core i7 memory configs!!