AM2 5200 to 6400 upgrade?

DonC wrote:
> "Jan Alter" <bearpuf@verizon.net> wrote in message
> news:esjnj.6278$v86.3685@trnddc08...
>>
>> "Fitz" <akfitz@gci.net> wrote in message
>> news:13pqclceu85lvf9@corp.supernews.com...
>>> I really wasn't looking at changing OS. My question was more about
>>> hardware, and whether a 20% increase in clock speed would translate
>>> roughly into a 20% increase in performance if all the other factors
>>> (harddrive, memory, graphics) were already capable of supporting it. If
>>> the single factor holding back optimal performance is the processor, is
>>> moving up to the 6400 from the 5200 a worthwhile move, or am I better off
>>> staying with what I have and moving to a quad-core or Intel based system
>>> in the future?
>>>
>>> Thanks
>>> --
>>> Your body is a temple boy,
>>> You ought to treat it well
>>> But you trash the place and rent it out
>>> Like it's some cheap motel - The Badlees
>>>
>> I would conjecture that it would be speeded up, but I don't know if it
>> would corrolate to a 20% increase. You might try Tom's Hardware for some
>> comparative benchmark tests with processors and specific processes that
>> the tests are run.
>> --
>
> FWIW, just my 2 cents: The cpu constitutes only a fraction of system
> performance. For sake of discussion lets say 25%. Three other factors are
> the memory speed (actual - not rated), hard drive speed, and video speed.
> In reality this is a crude approximation.
>
> If this is close to reality, I'd guess you'd see at best a 5% performance
> improvement by increasing cpu clock speed by 20%. (25% X 20%)
>
> That said, you would also realize a 7.8% increase in memory speed. This is
> because the 5200 has an odd (13x) multiplier which results in memory running
> at 742 Mhz rather than 800Mhz. The 6400 has a 16x multiplier results in
> 800Mhz memory.
> http://www.hardwaresecrets.com/article/328
>
> So in my primitive analysis you'd realize a 6.95% improvement (Cpu =
> 1.2 X .25 + memory = 1.078 x .25 + video = 1 x .25 + HDD = 1 x .25 -1)
>
> Before anybody jumps all over me, remember that this is for illustration
> only. Real life experience WILL be different --- but similar.
>

Use a real benchmark. I trust that more than an equation.

http://www23.tomshardware.com/cpu_2007.html

For example, if you play just the one game all the time, find
an article that evaluates the contribution of CPU and GPU.

http://www.anandtech.com/cpuchipsets...spx?i=2747&p=4

Paul