Whats the FSB for all the new 64 bit processors from AMD, it says "integrated into chip" ... Does that mean it can vary or its around 800fsb or something?? Thanks in advance ! :) Asus A7N8X Deluxe 2.0 ATI 9800 Pro AMD Barton 2500+ 1024 PC3200 Ram 120gig Maxtor HD Audigy 2 420 watt psu

As I understand it, 400Mhz for an A64 and 800Mhz for an Opteron. I'd guess those numbers are arbitrary in the absence of a northbridge chip. Skip

There is no FSB on the Athlon 64 as it's built into the processor, it's called Hypertransport, if it was called FSB it would be running at 1600mhz ________________________ MSI KT8 NEO Athlon 64 3200+ @ 2.0ghz 1.0 GB DDR PC3200 2X 160 GB HDD Hightech Excalibur Radeon 9800 Pro Iceq 128mb ( Core 425 MHz and Memory 380 MHz )

Wow, thats cool to know thanks ZeroCool lol...1600 !!!!! Asus A7N8X Deluxe 2.0 ATI 9800 Pro AMD Barton 2500+ 1024 PC3200 Ram 120gig Maxtor HD Audigy 2 420 watt psu

Does that 1600mhz fsb mean 400mhz both ways, running in dual channel? Or is it really 1600mhz, so that if you got some pc4400 ram, you can hit some killer bus speeds?

This one is complicated. A p4 or amd xp uses an fsb to interface to the mobo chipset, which then interfaces to the rest of the computer. http://www.xbitlabs.com/images/chipsets/nforce2-vs-kt400/kt400-scheme.jpg That fsb is 64 bits wide and everything passes through it. It is also bi-directional, meaning upstream and downstream bits have to wait to get out of each others way. The a64 uses an integrated memory controller to interface to your memory, and hypertransport to interface to the mobo chipset. http://www.hwupgrade.it/articoli/902/athlon64_arch.gif http://www.amdboard.com/via_k8t400m_chart5.jpg Because the mobo chipset interface used to be called fsb, we tend to call the hypertransport an fsb replacement, which is misleading. The a64 memory controller replaces part of what the fsb used to do (communicate with memory), and hypertransport replaces the rest of what the fsb used to do (communicate with everything else). The memory controller is 64 bits wide just like the fsb was, but now it has separate paths for upstream and downstream so both can occur simulataneously. The speed of these paths are 200mhz ddr, meaning a rate of 400mhz. In each direction. The hypertransport speed is 800mhz in each direction which is why amd says it is 1600mhz. But it is only a 16 bit wide path. Its also not communicating with memory which is what most of us think of when we think of an fsb. To call hypertransport a 1600mhz fsb is misleading in that it is only 16 bits wide, and is not even leading to the memory. So amd is using an erroneous way to brag about their new architecture, maybe to keep it simple (?). But way more important than speeds is throughput. Below, example 1 is how computers used to look and unfortunately how we still think of them. Example 2 is how they are today. 1) Say you are in a blimp above a city looking at the cars going at 4.77 miles per hour. A new car that could go 8 miles per hour would really speed up traffic for everybody. That's why the IBM PC AT was better than the IBM PC 1 (in a nutshell). 2) Now say you are looking at that same city and the cars are travelling at 2,000 miles per hour. A car that can go 2,200 miles per hour is not going to help much because as you can imagine in this city the cars tend to sit at traffic lights all day long anyway since they go so fast between stops. What the city needs are more lanes and fewer stop lights. AMD's on-die memory controller for memory and hyptertransport for everything else are not an advancement in mhz, they are an advancement in fewer stop lights and more lanes. Other than for video processing, mhz increases are old-hat and architectural improvements that reduce waiting will take us forward.