T9300 vs X7800

I have a T9300 in my FX with a GPU overclock it scores 9810, another FX owner has a X7800 running at 3.0ghz and his HD's Raid 0 except for that our laptops are identical but the best score he gets is 9727 3DMark06 with the same overclock. http://service.futuremark.com/resultComparison.action?compareResultId=5302646&compareResultType=14 His score. http://www.overclock.net/attachment...ptop-series-thread-mods-2008-02-01_151803.png My score. Why would my T9300 post better scores than a X7800 at 3.0ghz, his should easily beat mine, any thoughts on this. Edit: Noticed a big difference in the Firefly Forest scores.

 

Your CPU clock speeds are within 500MHz of each other; the T9300 should be about 10 to 20 percent more efficient per clock, and should thus only be trailing the X7800 by a small amount in sheer CPU power.

Then you said it yourself: GPU overclock. 3DMark06 is a synthetic GRAPHICS benchmark (though at times it starts to indeed be CPU dependent). That little OC you do is by all means going to make up for the slight loss in CPU strength.

And its not like your scores are HUGELY better, only by a slight bit. This isn't unusual at all.

 

Well the thing that is puzzling is I'am at 2.5ghz his is 3.0ghz both GPU's running the same overclock, his is overclocked also plus he is using RAID 0, I'am not. Is the Penryn holding somekind of advantage? I talked to him and he tried the same clocks but can't beat the Penryn.

 

raid has nothing to do with the 3dmark score.

 

So a T9300 is equal to a X7800 that is OC'ed to 3.0ghz, hard to believe.

 

Like I said before... it's not that surprising. It's only 500MHz. T9300's about ~17% slower and 10%~20% more efficient.

And of course the Penryn has an advantage. There's a reason they make these new cores, and it's not to decrease performance XD

 

What about video card driver version? RAM speed and CAS latency? That can also be a factor

 

We have identical machines, except CPU and RAID 0, he even installed the same GPU driver as mine and scored even lower. Pretty interesting to see such a comparison between two identical laptops.

 

Commander nailed it and then einhander sealed it not really testing HDD. Less than 1% difference in scores. CPU not the main weighting factor in 3Dmark, try PCmark if you want to get a better Idea. Programs and services running in the background, different anti virus. I have a 1.6Ghz CPU it really runs 1598Mhz there are variations even with the same product 1% sounds within tolerance.

Oh and sweet rig and nice upgrade!! Great score! On that other site (we don't trust) beats all but 8800m GTX SLI! Other than overclocked GTX you are King!

 

Margin of error? ..

Well the only difference is the proccesor, yours is more newer right?. Maybe better technology for gpu support. Thats the only plausible explanations i can think of

 

Every physical object has a margin of error, especially something as complex as a processor. It's usually not referred to as a margin of error, just a set minimum that the processor needs to run at.

 

It is not a margin of error if it is not an error. I am not an engineer but as Lithus said there is variation even between identically manufactured products, gets back to acceptable variations, "within tolerance" not error so not margin of of what it is not! 1% performance variation on systems composed of so many components, these two notebooks might not be exactly the same inside with all the parts some of those little insignificant parts could be from different suppliers. Calling them identical is more of a "margin of error" than the difference in scores.

 

The 3DMark score shows his CPU score higher, but his SM 2.0 is lower and the Firefly forest score is much lower. Wonder what the translation of that is? I;ve found a OC'ed GTS equal to a GTX.

 

"Margin of error" is a term used for uncertainty of a measured quantity or an acceptable range of values within the measured units. It actually has nothing to do with being wrong or the direct definition of error.

 

I found those words on googles definition lmao. You tell google that they are wrong and you are right. As a matter of fact they should replace it with what ever psycho-babble you posted below and call them idiots.

Anyway this is non-sense, back to the topic.

 

^^^I RePosted it I pulled but what is above is too much so here it is, what caused him to say above!

Wow, word games tonight? Error has nothing to do with "wrong"? What in this special circumstance? What are you saying "margin of correctness"? You completely forget about the way it is used. It is used for when people extrapolate a result from a survey (sample). Not from the variation of end results.

The margin of error is a statistic expressing the amount of random sampling error in a surveys results. The margin of error is usually defined as the radius of a confidence interval for a particular statistic from a survey.


It is a statistical term, that is it. If you want to continue to misuse it and misunderstand it, so be it. And if you want to correct people who have a better understanding of what you think you are correcting them on then by all means keep going I am not going to stop you. There are terms for what you would like to lump into "margin of error" that Lithus mentioned was in fact not correct or usual and I just tried to expound on. I introduced the correct nomenclature, I apologize my mistake.

Engineering tolerance is the permissible limit of variation in

a physical dimension,
a measured value or physical property of a material, manufactured object, system, or service,
other measured values (such as temperature, humidity, etc).

That is a better way to express what we are talking about.

Error=The difference between a measured or calculated value and a true one.
No nothing to do with that. I guess I don't understand wrong? Or the direct definition of error? Which this is.

We are not dealing with a sample here, we have two machines and the results are the results. Not a sample! So not a margin of anything.

 

You should both run PCMark05, it will give a more accurate overall representation of both systems testing all of the components, instead of mainly the CPU and GFX card.

Oh and please do I really wanna see the results!