I apologize in advance if my question seems trivial. I have read several articles on benchmark results for dual channel vs single channel, but I am still not sure about what I want to know. Here goes. I just went from a Dell Optiplex GX110 with 256MB ram and PIII 666mhz processor to a GX270 with PIV 3Ghz, hyperthreading capable, and 512 MB of DDR PC3200 400Mhz CL3 ram. I purchased an additional 512MB as well before the computer was even delivered. What I did not know (since regular SDRAM is all I had ever dealt with) is the new dual capability that I ran across in the bios. After much reading, I am now familar with what it means and how it works. Unfortuneately, I am now stuck with 4 different sticks of ram. Luckily, 2 of the 256 sticks are compatible enough to run in dual mode. There is another 256Mb stick (all 3 256 sticks are single sided with 8 chips) and finally a 512 MB stick with 8 chips on both sides; and so as I understand it (and my computer verfies) is a deal breaker and reverts to single mode. The program I run a LOT is called Mathematica, a software for performing complex mathematical calculations (like factoring a 50 digit number for example). Based on the mathematical software I run, 1) What is best for me? Using two of my 256 Mb sticks (512MB) in dual mode, or using the 3 256 and 512 (1280MB) in single mode. I know 1.28G of memory is better than 512MB in terms of memory storage, but if the data can't get through the memory controller fast enough for the 800fsb - well, which is the better option? Way more memory single channel better, or dual channel way less memory better? I do not plan to buy any more, so I want to know which I should go with. Also question 2) Am I better of enabling or disabling hyperthreading in the bios? I read very conflicting opinions on hyperthreading. Thanks for your time with these now obsolete concepts (but new to me). Kindly, Chris.
go with the most ram that works and use hyperthreading.
Is this just an opinion? If not, what is your response based upon just out of curiosity? Incidentally, I finally broke down and just tried all 4 configurations: 512MB Dual and 1280MB Single with Hyperthreading enabled and disabled. I ran a lengthy computation within Mathematica. To my surprise there was virtually no change in output time (down to 1/10 of a second) between any of the 4 configurations. So shockingly it appears that there is no difference for my application. Another surprise is that the processor is only working at about 50%-60% regardless of the setup. I foolishly assumed that Mathematica would try to use all the processing speed (and RAM) available. This doesn't appear to be the case.