RAMBUS 4X

Sounds interesting to me, but it does seem a little bit overkill, no? I mean, sure it's good technology but when it comes out it will probably be extremely expensibe, like how RDRAM was when it first appeared. I mean, 12.8 GB/s bandwith? That's insane. How much bandwith can a P4 Northwood even use? Obviously the P4 right now doesn't benefit much from RDRAMs superiror bandwith over SDRAM and DDR SDRAM. Unless the reason why P4 doesn't perform well is because memory bandwith of RDRAM isn't good enough to keep up with P4. Which one is it?

"We put the <i>fun</i> back into fundamentalist dogma!"

Well then, seems like 4X Rambus is pretty pointless for single processor systems. Well, maybe not pointless but definately overkill. Here's something though that makes AMD special over Intel. On Intel multiprocessing systems, both processors share the same path to the memory. But on AMD multiprocessor systems, each processor has it's own dedicated pathway to the memory system, therefore keeping the bandwith the same to both chips as if there was only 1 chip there. Pretty nifty.

To me, either Rambus or P4 sucks. Because, obviously Rambus has huge bandwith compared to SDRAM, but a P4 doesn't seem to benefit from it too much, as performance for a P4 Rambus sytem isn't too much better than that of an Athlon DDR system in most applications. That will probably change with the Northwood, but how much is anyones guess. I'd just like to see the P4 DDR chipsets to come out to see what type of performance advantage/disadvantage P4 really has with RDRAM.

"We put the <i>fun</i> back into fundamentalist dogma!"

a few interesting facts..
AMD has no commercial SMP shipping system,
they have demo's and very expensive one off's

AMD's commercial system will NOT have dual memory paths too expensive.. and not necessary since you can have double data signal ram and 2 channels like rambus anyway..
alternates them so makes dual memory paths less advantageous..

the P4 benefits from rambus and bandwidth tremdously
its just that people do not test it right or throow enough heavy IO for it to show.

as I said the P4 outperforms the P3 and Athlon 3 fold on memory and bus throughput..

on FP with proper P4 compiles and dd3d 8a it is 50% faster

you cannot test the P4's 3.2 GPS on notepad or office or
simple looped benchmarks,
you have to do compiling, editing, 3d graphics, memory intensive apps, multitasking, Multimedia ,etc
to show the differences..

I can tell you it is very much faster than P3...
and if you run windows 2000 or ME you gain another 20 %
speed gain..
windows XP in 3 months will give another 25%
the capability is there, software, has to catch up,
that is how advanced and fast it really is

not too many things can bog it down
I have build and tested at least 50 of each for clients
and there is a difference,
don't believe what rumors you hear..
most are from people who reviewed 1 early P4 with the wrong software for like a single day.

best
CAMERON

CYBERIMAGE
<A HREF="http://www.4CyberImage.com " target="_new">http://www.4CyberImage.com </A>
Ultra High Performance Computers-

I think you have it backwards,

RAMBUS developed these technologies and patents,
for 15 years, I remember seeing a demo of it at comdex in 90

they have over 100 patents,

they are defending their patent, the other foriegn
companies are STEALING it..

it is Ironic that Micron and Hyndai and on the same side, when Micron sued them a few years back along with TI for stealing memory patents.

the Korean's are nortorious for patent infringement,
SAMSUNG expecially until they lost a few

rambus will prevail because they are in the right
the small guy against the giants
CAMERON

CYBERIMAGE
<A HREF="http://www.4CyberImage.com " target="_new">http://www.4CyberImage.com </A>
Ultra High Performance Computers-