Idiots at THG Test TwinMos and Buffalo Tech DDR466 on an AthlonXP????

[bilydkid1970]

http://www6.tomshardware.com/motherboard/20030508/index.html

With all Intel/AMD brand loyalties put aside, isn't it generally accepted that AthlonXPs don't see such great returns from higher speed memory????

Why wouldn't a tech site examine DDR466 chips on a P4 system that can better capitalize on additional memory bandwidth???

We all know athlons have a memory bandwidth bottleneck. Showing us this bottleneck once again with DDR466 doesn't show us what DDR466 can do in a hospitable environment.

I was all jazzed to read about DDR466 performance and all I got was another "Athlon Asynchronous Memory Can't Handle the Bandwidth" article.

What a bummer.

I found this to be a lame excuse for a memory review.

Sometimes THG has articles I find informative. This is actually the first time I thought something was just plain stupid.

~BdK

 

[hoytron]

I agree with you completely. I was excited when I first saw this article, as I want to use that particular memory for a new P4 system. I am interested in seeing how high it will clock. I got really disappointed after reading on and not seeing any P4 tests.

 

[deathstar13]

guys im gonna comment on this from a differant angle.
first an amd system is the best judge of any memory test.
for memory to really be judged it has to be run synch with the fsb.

its a known issue in amd land that memory that runs 233 asynch most times never will do it running synch with the fsb.so it is a stronger and harsher way to test.

tom did the testing like tom always does,his head in the toilet.
he did all his testing asynch which is useless.on an amd if the fsb and memory arent run in synch the scores suck.thats what his graphs show also.

why he did this i dont have a clue. he cant be that dumb,well mabey he can? lol oc.com has stated this a million times over the last year so it isnt new news by any means.

sure p4 will gain big time with any high memory setting.
also its unfair also as with a fsb and mem run in sysnch even on an amd there isnt much differance from cas2 over cas2.5 which tom stated incorrectly.well when you run it asysnch it does make a differance but the losses in performance are due to running asynch.
so really pc3700 if run sysnch at cas2.5 which is also widley known doesnt like cas2, will beat the pc3500.

this article was flawed from page one. heck i just lost 8 iq points skimming the pages for facts!

 

[larva]

AMD's need to be run in sync mode, Intel chipsets do not. It's true that it would be much more meaningful to test the ram with an 845pe or something, as it's well able to run the ram past 450MHz and test the claims made for this ram.

Nforce2's struggle past 210MHz on the fsb no matter how good the ram. 845pe's will run the fsb up to 200MHz easily, and with the very effetive 4:5 an 3:4 modes you can run the ram anywhere from 200 to 533MHz if the cpu will allow. Mobile P4's with the 12x multiplier would be preferable for such a test.

This doesn't tell you which ram best crutches the Nforce2 instabilities that set in soon after 200fsb, but it would allow you to test the limits of the ram and the claims made for it.

As an example my PC2700 runs ~470MHz with ease on my 845pe motherboard, where a Nforce system with the same stick in it would have no chance for running above 220fsb.

 

[bilydkid1970]

guess that means:

Guess that means you guys agree with me that THG was stupid to test highspeed RAM the way they did.

Like I said, THG has IMHO some good articles. This one was just STUPID! STUPID! STUPID!

-BdK