The Sorry State of Hardware Reviews

@ 2003/12/26
Couldn't have commented on this article any better then this guy:

Quote:
Being a reviewer myself, I agree in principle with just about everything you said. The only 'beef' I have is you comment about using MBM for testing heatsinks. Yes, it is inaccurate as hell. I will not argue that point. However, for those of us who do not ahve that major cash to shell out for an accurate die simulator (which I have heard can cost up to $3000) I feel that you can use MBM and the BIOS temps for some good.

The key is to do compariative reviews. Run all yout tests and whatnot with Heatsink A and record the MBM temps several times and get the average. Keep everything else constant. Do the same for Heatsink B. Then compare the results.

Are the temps going to be accurate, hell no. Will anyone ever be able to recreate the exact same temps? Probably not. But if Heatsink A gives a temp of 50C and Heatsink B gives 45C it is a fairly safe bet that Heatsink B is the better of the two.
Comment from Tarantula @ 2003/12/27
In the bios I changed the RAM clock speed from CL2 (or auto) to CL 2.5. This is a simple way to over clock RAM without changing your FSB speed. The Ultra DDR Memory does not advertise on the package that it supports CL2.5. However, I'm proving that it does right now as I use it to write this review. The second benchmark illustrates Ultra's comparison against the same 4 configurations this time with the 2.5 configuration. As you can see, the bandwidth has greatly increased. The competition stands no chance. This might seem a bit unfair because there isn't a CL2.5 setting for those other setups, but that's what you get with THIS memory.




omg funny ****
Comment from Xploited Titan @ 2003/12/26
Indeed, if you do some comparisson, no problem if it's inaccurate...

But like everywhere: we never can reproduce the same thing twice...