[Joeteck]

Very nice. I also have a few questions...

How does it do with 1156 Intel CPUs? Some benchmarks have proven dual channel to be better...

I would love to see these same tests on a core i7 860 @ 4GHz.... Then those compared to the 965 as well...

[mdcomp]

Great article dude!


Matt

[sacha35]

Outstanding work Gautam, very intensive.

[Deanzo]

Great article and work G

[David]

Originally Posted by johan851

He did this, didn't he?

Originally Posted by EarthDog

I thought the same thing...

However I believe he wanted a COMPLETELY stripped down OS as opposed to the version he has. I will hope he clarifies soon.

Personally, I would have rather seen this run at all stock since thats how 99% of people here run their OS's.

Yeah, I mean stripped down all the way. The way Gautam did it he stripped them down to about the same point, was just wondering if some OSes lend themselves more to being stripped right the way down.

Admittedly, it's a lot of work for relatively little info.

[Deanzo]

Hey G,

Just looking at the XP64 wPrime score, did you try with and with out graphics drivers, to see if it changed ?
It looks like the kind of drop in score by not running drivers, maybe an issue with that OS and it's drivers ?

[Omega Destroyer]

This is a great atricle.

I just had one comment. Instead of putting the results of all three runs and an average, I think it might have been more useful to only show the average with some error bars.

[I.M.O.G.]

Good input. I considered that when I first looked at the graphs also. Given more thought, I like the transparency of the way G presented it - with the shades of grey and the average in Red, I think it made it very clear.

We both took the same considerations in mind however, which is interesting.

[shadowdr]

Great article, I have been thinking of installing Vista as my primary OS for a while but honestly mine still has bugs. I don't strip them down at all, and XP has a few issues of it's own mostly the .net framework but I have never been able to install SP2 for Vista. Then again there is so much software to install as well, I think I will wait untill XP is unrecoverable.

[Gautam]

Originally Posted by Deanzo

Hey G,

Just looking at the XP64 wPrime score, did you try with and with out graphics drivers, to see if it changed ?
It looks like the kind of drop in score by not running drivers, maybe an issue with that OS and it's drivers ?

Hmm...it's a nice sounding theory, but that same XP64 is the one that I used for all the other benches, every single one, and with most of them being 3D benches. And it's only in wPrime that it goes weird.

I'll take another look. All the OSes are still in their same state.

[Omsion]

Originally Posted by Omega Destroyer

This is a great atricle.

I just had one comment. Instead of putting the results of all three runs and an average, I think it might have been more useful to only show the average with some error bars.

Agree with this - average + std deviation bars presents the idea in less space (even with three samples). More samples would always be nice, though.

Anyway, the theoretical implications behind this are that the optimizations in each OS are weighted differently, leading to these results, correct? And thus Win7 + Vista are statistically indistinguishable performance-wise given these benchmarks.

Finally, I don't really have a solution for this given your dataset, but I dislike graphs not starting from zero, since that skews the scale of the real differences. Dunno what really to do here, though.

[icebob]

Great work G! I know you like Vista but the result just talk for themselves... guees I have to hit the egg and get a couple more hdd!!! One question tho do you think they (OS) will behave the same when they get the nose dripping?

[Gautam]

Originally Posted by Omsion

Agree with this - average + std deviation bars presents the idea in less space (even with three samples). More samples would always be nice, though.

Anyway, the theoretical implications behind this are that the optimizations in each OS are weighted differently, leading to these results, correct? And thus Win7 + Vista are statistically indistinguishable performance-wise given these benchmarks.

Finally, I don't really have a solution for this given your dataset, but I dislike graphs not starting from zero, since that skews the scale of the real differences. Dunno what really to do here, though.

Actually, I too was thinking about std deviation bars, but I didn't feel like trying to figure out how at first, put them like that to begin with and then just liked the way it looked. I also specifically recorded the order in which I did each run, mostly for myself. Since I did, I figured I might as well present it to the reader too. As for more data points, once again, to add even one more data point per bench means around 5-6 hours, not to be whiny about it, but that's the truth. And if you look at the results it's pretty clear that there's not even much variation between runs in each given OS. Whether I can show it statistically or not, 7 always scores lower than Vista in 05 for example. Even one run can tell you that. I could do more runs to make it more statistically sound, but pragmatically it wouldn't help a thing.

Probably plenty don't like the graphs not starting from zero, but it's conducive to what I'm trying to present. This is not supposed to be an academic paper, it's supposed to tell people clearly what OS scores the best. The real difference is tiny, yes. For example even between the averages of 7 64 and Vista 32, it's just 2.35%, which in real terms is tiny. However, on the 3DMark05 hall of fame, it's actually larger than the difference between 1st place and 5th place, which is a very big deal for anyone competitive.

[Gautam]

Originally Posted by icebob

Great work G! I know you like Vista but the result just talk for themselves... guees I have to hit the egg and get a couple more hdd!!! One question tho do you think they (OS) will behave the same when they get the nose dripping?

wha??

[icebob]

You know what I mean will they behave the same @ lets say 5.5ghz versus 4.2 ghz Let me rephrase, do you think the os will affect your max oc?

[Deanzo]

Originally Posted by Gautam

Hmm...it's a nice sounding theory, but that same XP64 is the one that I used for all the other benches, every single one, and with most of them being 3D benches. And it's only in wPrime that it goes weird.

I'll take another look. All the OSes are still in their same state.

I'm not even sure why having 3D drivers loaded matters so much on a 2D (cpu only) bench
But it could be a one off issue with that bench/OS, or not

[Gautam]

Very possible. Almost no one ever uses XP64 for benching anything, so I have nothing to go off of.

[freeagent]

Nice article hombre

A little O.T...

I like the new look of the main page, looks good

[I.M.O.G.]

Originally Posted by Gautam

Actually, I too was thinking about std deviation bars, but I didn't feel like trying to figure out how at first, put them like that to begin with and then just liked the way it looked. I also specifically recorded the order in which I did each run, mostly for myself. Since I did, I figured I might as well present it to the reader too. As for more data points, once again, to add even one more data point per bench means around 5-6 hours, not to be whiny about it, but that's the truth. And if you look at the results it's pretty clear that there's not even much variation between runs in each given OS. Whether I can show it statistically or not, 7 always scores lower than Vista in 05 for example. Even one run can tell you that. I could do more runs to make it more statistically sound, but pragmatically it wouldn't help a thing.

Probably plenty don't like the graphs not starting from zero, but it's conducive to what I'm trying to present. This is not supposed to be an academic paper, it's supposed to tell people clearly what OS scores the best. The real difference is tiny, yes. For example even between the averages of 7 64 and Vista 32, it's just 2.35%, which in real terms is tiny. However, on the 3DMark05 hall of fame, it's actually larger than the difference between 1st place and 5th place, which is a very big deal for anyone competitive.

Well stated. I also didn't care for the graphs not starting from zero, but after I digested the article, it seemed clear to me you made the right choice in presenting the data.

Originally Posted by freeagent

I like the new look of the main page, looks good

Thank you. iNet and our mods played a large part in making it what it is. Dogsoldier also deserves an immense amount of credit - he took part and won the logo design contest we held on the forums, and the entire design was centered on his logo. I owe him an article still, highlighting his artistic portfolio.

[SeanBest]

Nice read ... could've used more pictures!