Time for a new thread!

[powerpack]

I did make mention of in a post but why not put it where it belongs. Here it is! :)

Make a GPU sticky. The information would be the actual speed of the cards. Core and memory. Why? Well Nvidia and AMD/ATI make the chips/chipsets? But MFG's set the speeds. So would be useful to know. If there is a resource I have not found yet I will withdraw but please give me the address.

Notebookcheck is the only site I have found that makes any serious attempt at gauging notebook GPU's and it is flawed. And from Germany I think? No problems with Germans or anything, except when links send me to the German language version, I don't read German. :D

Also they only list the chipset and do not differentiate by MFG and what they set speeds at. Also they use 3DMark but not sure if they control settings it is run at. Both are critical bits of information to make an intelligent informed decision.

If you guys ran GPU-Z it would tell you core and memory. You could also run 3DMark (if you wanted) only the free versions of the application. If you guys got way inspired sure you could run other settings to see comparative marks with things like AA which cards handle different?

You have for the most part two kinds of customers; those who know a good amount about notebooks and those who are learning and soon will know much.

We all know a disproportionate amount of Xotic's customers are Gamers or Power users. That bunch loves info/crunch numbers, in general they/we care about stats others don't.

I really do not see it as a lot of work. It could become an authority on the topic. Notebookreview which would be a great place to do this. Lacks the ability to have a controlled environment. I think that is notebookcheck's entire problem also. You have that luxury.

If you did share this info? The logic tree/matrix as the "Owners lounge" is (give it time, do not laugh, I feel a ton/onslaught of Asus owners ready to post once they find the enter button) by brand then model then stats. As you cycle out models move down on the list.

Other option by card of course higher performing being higher on list. Then stickys with model.

I am sorry I know you guys have jobs so building notebooks is job #1 but a little time and well could be good and build traffic?

Also as mentioned in my post (the one that inspired this) Just as you cannot just compare AMD/ATI vs Nvidia on clocks because they have very different design. Be careful comparing DDR2 to GDDR3. First if heat and power are your concern (if so Xotic is the wrong site for you, we sell the NP9262, anymore questions? Get your sissy bag as you leave and thanks for coming) yea I think it wins. And yes it does have the ability to flush/dump it's storage. But just as system RAM suffers from higher latency's so does GDDR3. I wish you guys would post/talk/explain. I am saying other than the dump thing 500Mhz DDR2 (with lower latency)would likely kick 600Mhz DDR3 to the curb?

Have a great night all! Happy holidays too all! :)

[Woody]

The biggest problem with benchmarking mobile GPUs is platform compatibility. With desktop cards you can benchmark and compare every card ever made on a single reference platform by swapping the cards. Such comparisons aren't possible with notebooks and usually aren't relevant anyway because you usually don't have more than one or two options to choose from in any given model at best.

For this reason it's usually more useful to post and compare performance results from different configurations for direct comparisons between different models and configurations rather than comparing the individual components. The real difficulty comes when choosing between two notebooks that have identical or nearly identical components apart from the graphics card in which case it's difficult to know which one is the better value. It gets even more complicated when trying to compare ATi to NVidia due to the architectural differences that cause a wide range of performance variations from one game or benchmark to the next.

And to add to all that confusion, notebooks offer a much more personal and asthetic choice for individuals due to the nature of their use...they are in plain veiw and don't hide under the desk so purely subjective factors that have nothing to do with performance come into play.

Notebookcheck does a better job than anyone at cutting through this confusion but they don't make it very clear just how much variation or error they will have on their website simply due to the complexities listed above.

What I'm saying really is....good luck with that powerpack! ;-)

[powerpack]

Woody I understand your points and agree with them. I was pushing with the 3DMark stuff as well some argue the relevance of (which I understand). As you clearly stated it can be a game too game issue. Nvidia better on that game AMD/ATI better on this one. So yes 3DMark is by no means any kind of "Holy Grail". But well it does exist and well what the heck?

Your comment on identical test beds is also spot on and can't be done. That said I would still like to know the core/memory speeds of the cards offered. What those would mean? Sure up for interpretation. But if you have the same card different speeds say Asus vs Force good to know? Xotic does say DDR2 vs GDDR3 already? So they give information that is for many reasons you have stated not comparing "apples to apples". But they give, so why not share more?

Woody I do have to agree with what you have said as "status quo" my suggestion is improving on what has been done before. I actually have spent much time at notebookcheck. I interpret their results. I think I am reasonably smart. I have no training in statistics but recognize thing such as throwing out the lowest and the highest score can at times increase the accuracy of results. I also recognize when every bench is around 7,000 say six? That if one is 4,000 it is flawed. Now you may say it is valid as that is what that notebook tests at? Fine OK but being a GPU test sight the flaws that cause it to perform so badly may not be GPU, if such it has no business being on that sight for comparative purposes. That actually gets at the heart of your test bed issue. I mean if someone figures out how to put a 8600m GT on a PII system that is beyond any reasonable variance and as such should be scrubbed. Identical testbeds yea I know. But chipset/CPU/RAM is a good start. Also 3DMark06 includes CPU in the composite so notebookcheck is flawed even though I agree at this point the best out there.

Woody jump on board! Don't say no say yes. Instead of saying "why" say "why not"!

On your comment of aesthetics? Yes, I wanted a 14" or smaller form factor with as much power as I could find. So I agree with you on that but do not find my suggestions to be in conflict. I believe more information is never a bad thing. I also do not believe in protecting people from themselves by limiting information.I very much believe in sharing information and teaching/helping when it is misunderstood.

Woody on a final note, I have a copy of GPU-Z so if you have any weight/pull at Xotic and can arrange them shipping notebooks to me for testing? I would be more than happy to find core/memory speeds.:D

[fallenxlionheart]

Woody I understand your points and agree with them. I was pushing with the 3DMark stuff as well some argue the relevance of (which I understand). As you clearly stated it can be a game too game issue. Nvidia better on that game AMD/ATI better on this one. So yes 3DMark is by no means any kind of "Holy Grail". But well it does exist and well what the heck?

Your comment on identical test beds is also spot on and can't be done. That said I would still like to know the core/memory speeds of the cards offered. What those would mean? Sure up for interpretation. But if you have the same card different speeds say Asus vs Force good to know? Xotic does say DDR2 vs GDDR3 already? So they give information that is for many reasons you have stated not comparing "apples to apples". But they give, so why not share more?

Woody I do have to agree with what you have said as "status quo" my suggestion is improving on what has been done before. I actually have spent much time at notebookcheck. I interpret their results. I think I am reasonably smart. I have no training in statistics but recognize thing such as throwing out the lowest and the highest score can at times increase the accuracy of results. I also recognize when every bench is around 7,000 say six? That if one is 4,000 it is flawed. Now you may say it is valid as that is what that notebook tests at? Fine OK but being a GPU test sight the flaws that cause it to perform so badly may not be GPU, if such it has no business being on that sight for comparative purposes. That actually gets at the heart of your test bed issue. I mean if someone figures out how to put a 8600m GT on a PII system that is beyond any reasonable variance and as such should be scrubbed. Identical testbeds yea I know. But chipset/CPU/RAM is a good start. Also 3DMark06 includes CPU in the composite so notebookcheck is flawed even though I agree at this point the best out there.

Woody jump on board! Don't say no say yes. Instead of saying "why" say "why not"!

On your comment of aesthetics? Yes, I wanted a 14" or smaller form factor with as much power as I could find. So I agree with you on that but do not find my suggestions to be in conflict. I believe more information is never a bad thing. I also do not believe in protecting people from themselves by limiting information.I very much believe in sharing information and teaching/helping when it is misunderstood.

Woody on a final note, I have a copy of GPU-Z so if you have any weight/pull at Xotic and can arrange them shipping notebooks to me for testing? I would be more than happy to find core/memory speeds.:D

hugely off topic i know but i just wanted to point out....powerpack you are everywhere!

[Stormbreaker]

"hugely off topic i know but i just wanted to point out....powerpack you are everywhere! "

You do know that the last post in this topic is six months old, right? xD

[shadow]

when benchmarking is done, how many times is each one ran?

[fallenxlionheart]

"hugely off topic i know but i just wanted to point out....powerpack you are everywhere! "

You do know that the last post in this topic is six months old, right? xD

thats exactly what i mean...hes in all the old and new threads lol...hes everywhere