The Ultimate Fight AMD vs Intel.

I was comparing the the prescott and northwood at for example 3.2GHz... The Prescott only starts to pull ahead when there is an increase in clock speed. But if you pit them together at 3.2GHz, Northwood is faster. Intel needed to crank up the speed in order to make up for the deeper pipelines...

 

that's true. Unfortunately, Prescott is too hot to be able to crank up the speed.

 

yeah, it's still hot even after they revamped it. it used to be a lot worse, i heard

 

Yeah. Its just insanely hot. The mobo gets even hotter from what Phok told me....

 

I think I read somewhere mention that it actually melt certain mobo.

 

yes. but then u have to remember that they dun design the processor in a day time. it takes months to years. back then, 90nm process technology was still a new thing to them. and considering their excellent record in processor technology transition, i think they never expected heat issue would hold them back so deadly.

yes I know u are comparing them at 3.2ghz. again as I've said, it's app dependable. Doom3 runs faster on prescott. it's really up to what kinda app do u use u see..can't just compare them apple to apple..

 

I guess you have a point on the apples to apples bit...

 

it's funny, either TH try to screw intel or they really got a faulty setup. More problem is recorded for intel's system.

 

it's amazing how just a touch to the hsf and 500 rpm can alter temperatures.

wow, a double whammie, both the hsf and mobo turned the fan down

 

Well.. never know what happens. There are a LOT of ways to influence the outcome.. remember there's no one to audit the results at all.

 

Well, that's true if you use this program for work. But no one works with a benchmark. Even display rendering, there're so many ways to influence to outcome.. different drivers, custom drivers, custom BIOS, different versions of DirectX, different versions of OpenGL, whether or not SSE/MMX/3Dnow is used, what the system load is like, whether the OS is updated or not, whether the card is modded or not.. so many ways. Anyway one of each is not a significant sample

 

not only that, but remember that something could just blow out on you...unlikely, but still the possibility.

i'm sure that THG tried to get the best drivers for everything

 

sure the benchmark is not designed to measure day to day usage (as most modern system don't distinguish that much among daily softwares). Still you need a way to measure the performance and benchmark is a good way. Besides reading the benchmark number, I think us as a reader also have to be careful reading them, don't just trust all that's being said and be cautious about the data and read more, if 4-5 or more sites report similar finding then, it means that the data are more trustworthy.

 

they should also test more than once, if they dont do that already. sometimes my score on 3dmark01 changes by around 500 points, with no changes to computer

 

that's true, you should always have duplicate or even triplicate data. Although that would take quite a long time, I think this benchmark is just a overview of all cpu, we shouldn't take too much on the numbers but look at the relative number and see the trends on the cpu.

 

yeah, for instance, a XP-M beating a fx55? not likely, just that the fx55 probably had to read from the hdd for a few more seconds than the xp-m

 

You mean run the same test a few times and get the average? Yeah, that would ensure that its consistent. But it takes a long time...

 

don't repeat what I have already said . I don't think TH's intention is really show the performance rather the improvement and development. The benchark is a way to indicate how far cpu has come to.

 

Averaging and duplicating is different...

Isn't this just a reliability test...