Specifications

abit NF7-M
Written by Stephen
Friday, 04 September 2009 10:42
this will truly only be useful when the onboard graphics are enabled as the system’s front side
bus can only handle a given amount of bandwidth. Anymore than that becomes, in essence,
useless. The nForce2 chipset is capable of up to 6.4 GB/sec of bandwidth with a Dual DDR400
setup. In modern graphics cards, we’ve seen the incredible amount of bandwidth they require to
keep the data flowing, so it’s safe to say that the extra bandwidth created from the 128-bit wide
memory bus will only aid the onboard graphics.
The key to my heart with DualDDR deals with the performance optimization. NVIDIA has
optimized the nForce2 to put out maximum performance when the CPU Front Side Bus and
Memory Clock run in synchronous mode. In case you’re unfamiliar with this term, this means
that the FSB and Memory Clock will run at identical frequencies, while asynchronous mode
means that each will operate at its own independent speed. Well, why is this fact the key to my
heart? Simple, it can save you the one thing you love the most, well almost, money. With
NVIDIA’s nForce2 platform, if you’re running a 266MHz FSB AthlonXP, you only need DDR266
memory for optimal performance. If you’re running a 333MHz FSB AthlonXP, you only need
DDR333 for optimal performance. Instead of having to buy the absolute fastest, and more than
likely most expensive RAM on the market, you can save a good deal of money when purchasing
“slower” RAM, or you could use the extra cash to stock up on more RAM, since as we all know,
you can never have enough RAM.
Well, why exactly does the nForce2 platform allow for maximum performance when the FSB
and Memory Clock run at the same speeds? Shouldn’t DDR333 provide a higher level of
performance, and so forth with DDR400? Well, to put it plainly, no. When the CPU FSB and
Memory Clock are running at the same speed, the CPU Read Latency will be at its lowest. The
address that the CPU sends to the memory and the data returned are sent back and forth at the
same speed within the IGP/SPP’s memory controller. This means that the CPU does not have
to wait for any conversions, commonly termed as “overhead,” to take place before it can be sent
the necessary data. When the FSB and Memory Clock are running in asynchronous mode and
are operating at different frequencies, this overhead penalty requires the data being sent back
and forth to be synchronized with the clock domain that it will be entering. If data is going from a
clock domain of 133MHz to 166MHz, the data must be resynchronized, therefore using up clock
cycles and increasing latency. This is exactly why you’ll want to run your nForce2 system FSB
and Memory Clock in sync.
..:: Dynamic Adaptive Speculative Pre-Processor (DASP) ::..
6 / 7