The Spin off Smoothness Thread

[TL1000S]

I'm sorry... if you were actually making that point it was not something that could be ascertained by reading your posts since you did not actually communicate that concept.

And you really should stop playing the i7 -- 8 thread card. It is only a quad core processor. Those 4 added threads don't help as much as many people are advocating.

Every bench/usage that take advantage of multiple threads shows real numbers for performance.
Intel implementation of HT on Nehalem/ci7 also shows a clear improvement in scaling compared wiht HT on P4.
This is no subjective "feeling of smoothness" but objective facts.

You are one of the most well known trolls on this forum. Just for your information: I have read all the posts in this thread. But apparently you have not. Either that or you have difficulty understanding the actual discussion... and as such you should NOT be participating.

And now you are going to jump into your ridiculous taste example in an attempt to derail the thread even more. You really don't have a clue do you? Let me refresh your memory. WE DO NOT WANT TO TALK ABOUT SUBJECTIVE THINGS THAT CAN NOT BE MEASURED. Does that help you? (I'm sure it didn't help because you have shown the concept is beyond your comprehension.)

I wish we had a moderator to keep trolls like you and the Zucker out of our threads.

This is utter BS as this just echoes back to yourself(and others).

Could it rather be that you want a thread to "discuss" how smooth your favourite systems runs and ban out any poster that dare to disagree?
To disagree and post a message about that is *not* trolling.
Please go back to your starting post in this thread and correct it accordingly if you only want "yay", "+1 posts".

And lastly:
I do not intend to question anyone subjective feelings regarding "smoothness".
They are as real as objective numbers for them.
And I respect that, but please allow discussion/debate in a forum without calling "disbelievers" names.
And when I build my first PHII-rig (920) I will not use it for the smoothness, but for OC and benching. Like always.
If that is "trolling" and Intel fanboism, so be it.

[rk7p5]

It's not just smoothness, when i switch back and forth from intel and amd machines, amd rigs feel more responsive. I'm not 100% sure why exactly but it's not just me, lots of other people experience this "phenomena" as well.

[Final8ty]

But its what happens, just look at the max/min fps for Phenom/C2, they are sometimes huge on the C2 and not as high on the Phenom. Meaning that Phenom is limiting the graphics card eralier then a C2.

If smoothness in games steams from that huge discrepancies between max and min fps, then its possible to optain the same smoothness on both systems with a simple checkbox and all that discussion was pointless.

snip.

Its the difference between the max & min that matters on a system and not how high they are on a system as to the exercise that people are trying to do here & you can only sensibly measure with Vsync off as that is an artificial limitation, how high is irrelevant as that is not the exercise.
Hypothetically speaking If the AMD is pushing 150max 125min fps & the intel is pushing 250max 175min fps then the AMD is producing more constant fps & is statistically smoother the perception is another matter.
If the difference between max and min did not matter but just how high they were then there would be no need for Vsync to exist, your LCD would not try to output at a constant 60hz & fluctuate from 70-100hz at random as higher is better right? & films would not be 25fps or 30fps but fluctuate from 50-100fps at random because higher is more important than consistency?.


Application multi tasking is also important to many & the smoothness in how it handles them.

[roofsniper]

snip.

Its the difference between the max & min that matters on a system and not how high they are on a system & you can only sensibly measure with Vsync off as that is an artificial limitation, how high is irrelevant as that is not the exercise.
If the AMD is pushing 150max 125min fps & the intel is pushing 250max 175min fps then the AMD is producing more constant fps & is statistically smoother the perception is another matter at those levels.

not necessarily. if one system has just one problem then the min will be set low and won't be a good representation. its more of what is happening in split seconds over the period of the entire trial.

[demonkevy666]

phenom II 940 7,150 cpu points 5.4ghz

i7 4.1ghz 7,393

Q9650 4.5ghz 7,133

don't for get that AMD Always scores lower on cpu points but higher in graphics scores. scaling isn't always there either with more threads (just one example I could find with close scores. 3dmark06 ain't the best of scaling.)

[Calmatory]

not necessarily. if one system has just one problem then the min will be set low and won't be a good representation. its more of what is happening in split seconds over the period of the entire trial.

Thats why the benchmarks/tests should be repeated to make sure no such artifacts remain in the given scope.

don't for get that AMD Always scores lower on cpu points but higher in graphics scores. scaling isn't always there either with more threads (just one example I could find with close scores. 3dmark06 ain't the best of scaling.)

There is no way one can draw any conclusion regarding smoothness from that. Neither can one draw any conclusions from the CPU vs GFX test results since the other(may it be CPU or GFX) is written in a way in which it flavours either AMD K10 or Intel Core µarch - there is no way to make a "unbiased" test. Look at SuperPI, the way it is written(and ran inside a CPU) (unintentionally)flavours the Intel µarch. It just happens to run better on Intel parts. It must be - "smoother" there.

[roofsniper]

if someone with a core 2 could run fraps and record the frames times and run 3d mark 06 it would be nice. then you could see the differences between each frame instead of just the fps values.

anyone willing to do this?

[keithlm]

But its what happens, just look at the max/min fps for Phenom/C2, they are sometimes huge on the C2 and not as high on the Phenom. Meaning that Phenom is limiting the graphics card eralier then a C2.

If smoothness in games steams from that huge discrepancies between max and min fps, then its possible to optain the same smoothness on both systems with a simple checkbox and all that discussion was pointless.

Cause you optain exactly that, a smooth constant fps distribution, if you limit the fps to a ceratin maximum.



No, let me help you,
You try to quantify something that is in the so called "eye of the beholder" aka subjective preception.

Why do you think theres such a huge debate about what fps rate is visible to the human eye or can be seen etc. and what not...
This is exactly the same.

Again what data do you want to use, that isn't already there?

We have fps (outlines over time) max/min avg. yet they give no clue about whats smoother and whats not.

Just look at the graphs accord99 linked to, according to some theories in this thread the INTEL quadcore should be smoother then then the AMD quad.

The only thing so far you have suggested so far is to run a benchmark and then run several other activities in the background and see who does better.
That dont addresses the so called smoothness, this adresses the multitasking capabilities of a system.

Cause the thing thats always said about smoothness in conjunction with amd rigs is, that this systems are more smooth while daily work/gaming.

I dont think many gamers run havy tasks in the background when they want to game for some hours. So where do these smoother feeling steam from?

A bit OT:
Its funny that you accuse all people of trolling that dont agree with you imidiantly on your goal/opinion. They are people that have had quite some experience with different systems, yet they say things that are simmilar to what i say? All trolls and intel fanboys.... i dont think so.

Oh and please cut that name calling and shouting... you want to keep the thread clean and your the only one shouting and calling out pwople trolls...

Why yes. I was the one that kept talking about subjective things long after it was said that we need to find methods to measure this attribute objectively by using scientific methods and experimentation.

Oh wait... that wasn't me. That was you.

Yet somehow now you've flipped things around and want to now pretend you were not being a troll earlier when you kept doing this and that it is not okay for me to call you out on it. (EDIT: Although that is not the actual reason I call you a troll. I base that opinion on your baited comments as seen in your first post in this thread #42. Although I must admit everyone was ignoring the baited comments until Zucker2k posted.)

To purport that this is all subjective and that there is no way to ascertain any objective measurement and then to stand by that decision without anything other than your opinion is slightly contrary to all scientific methods. To only look at the data available and not attempt to find new methods to test something that appears to be experienced by a multitude of people would also be contrary to scientific principles. To give up before doing anything and ridicule anyone that wants to experiment is vacuous.

If the current data does not fit the model then you need to find new data using new methods. With as many people that talk about this attribute I would not be willing to lightly dismiss it as merely a subjective experience. Therefore the data does not fit the model. So the data must change or the model must change.

[ryboto]

so, before the site maintenance, there was only 2 pages, now there's quite a few more...excuse me for not reading them all, but my statement probably isn't related to anyone else's comments.

I've owned two AMD dual core systems but my most recent system is composed of an Intel Q9400 and G45 chipset. I never vigorously overclocked the AMD systems, I used my 4400 X2 at 2.5ghz.

I have to say, the AMD systems, using the same RAM and hard drives, had better desktop response, and I've noticed no subjective difference in gameplay using the Q9400. I'm comparing of course with fresh installs of Vista 64, and each system had a very similar configuration in regards to programs installed, since all of the systems were at one point my main PC. I thought it was a bit of a myth, then again, I wont call it "smoothness", but more consistent responsiveness.

[Final8ty]

True.

Before Throwed or whoever you are come and flame me off, I must tell that my older brother told me that when he switched from 3000+ Barton to 3000+ NewCastle, he told me how the NC was way faster and had no lags anywhere. He said it felt smoother. However, to burst the bubble, I must say that any fresh installation of Windows always feels smooth regardless of hardware. Even old P2 @ 266 feels "smooth" when opening start-menu and MyComputer etc.

less Lag is a good description.
Hypothetically lets way Intel system is like a Turbo which is faster than a Supercharger AMD system.
But you can get Turbo lag under curtain conditions where as the Supercharger does not.

What people are asking, whether its a fact or not between Intel system & AMD system.
I will leave the rest of you to work that out.

[Final8ty]

not necessarily. if one system has just one problem then the min will be set low and won't be a good representation. its more of what is happening in split seconds over the period of the entire trial.

Aint that the point one system has that problem more often than the other Hypothetically.

[roofsniper]

Aint that the point one system has that problem more often than the other Hypothetically.

maybe but if it is an outlier compared to the rest of the data then it shouldn't be considered. the min and max can paint a picture but it doesn't always come down to that. anyone have any ideas on how to test this idea? i had one idea but i don't have a core 2 to test it.

[G0ldBr1ck]

Look at it this way, if there was actualy a way to show this with proof AMD would have used it in there marketing. The fact that they have not used this as a marketing point tells me we will not find a way to bench this smoothness.

[Final8ty]

Look at it this way, if there was actualy a way to show this with proof AMD would have used it in there marketing. The fact that they have not used this as a marketing point tells me we will not find a way to bench this smoothness.

Its a point the majority don't care about so no point in AMD making a point about it.

[roofsniper]

Look at it this way, if there was actualy a way to show this with proof AMD would have used it in there marketing. The fact that they have not used this as a marketing point tells me we will not find a way to bench this smoothness.

if amd actually advertised maybe we would hear about it.

[Final8ty]

Results 1 - 50 of about 104,000 for amd smoother than intel. (0.23 seconds)

Well

[Glow9]

so, before the site maintenance, there was only 2 pages, now there's quite a few more...excuse me for not reading them all, but my statement probably isn't related to anyone else's comments.

I've owned two AMD dual core systems but my most recent system is composed of an Intel Q9400 and G45 chipset. I never vigorously overclocked the AMD systems, I used my 4400 X2 at 2.5ghz.

I have to say, the AMD systems, using the same RAM and hard drives, had better desktop response, and I've noticed no subjective difference in gameplay using the Q9400. I'm comparing of course with fresh installs of Vista 64, and each system had a very similar configuration in regards to programs installed, since all of the systems were at one point my main PC. I thought it was a bit of a myth, then again, I wont call it "smoothness", but more consistent responsiveness.

Prolly cause both could max out your video card. Using a 4870x2 things prolly would have been noticeable when ones bottlenecking way before the other.

[iandh]

edit: Actually, nvm. I also now agree this is pointless. I have made my decision to use the hardware that I choose based upon extensive personal experience, and that's that.

[BenchZowner]

phenom II 940 7,150 cpu points 5.4ghz

i7 4.1ghz 7,393

Q9650 4.5ghz 7,133

don't for get that AMD Always scores lower on cpu points but higher in graphics scores. scaling isn't always there either with more threads (just one example I could find with close scores. 3dmark06 ain't the best of scaling.)

You're missing that the CPU Test is a totally different usage scenario from the Graphics Tests in 3D Mark.
Having equal CPU Scores doesn't make the performance of the CPU's equal in the Graphics Tests that are using less resources of the CPU and are coded differently.

The second part and point of my post, that I'm trying to keep short since I have to go real soon is that there's no actual way to test the systems for "smoothness" as some of you call it.
But if you'd really like to "test" this "phenomenon" then the only solution that I can foresee is what people usually call a "blind test".
Build and config 5 AMD Phenom II systems, 5 Intel Core 2 Quad systems, and 5 Intel Core i7 systems with the exact same components ( except the ones that you can't really use, like the motherboard and the RAM in the Core i7 system that needs to be DDR3 ).
Configure the hardware & the software on every system and make sure the settings are the same for each one.
Mix the PCs, and have 10 people test them ( 1 person each time ) and ask them to tell you in the end which CPU is in each system.

This is the only way to "test" this theory/phenomenon out.
However it is indeed hard to obtain all that hardware ( unless you can borrow it from a retailer/shop/manufacturer ), and of course raise the "testees".

I thought of doing that in the past, but the people who claimed that their AMD systems were smoother than the Intel based ones of their friends suddenly refused to participate in the blind test the last day before I decided to go and pickup the stuff.

p.s. If you're going to compare desktop "smoothness/responsiveness" like the Start Menu for example, then you need to make sure that both systems have the same Start Menu appearance delay, and other stuff too.

And no, keithlm, Hornet331 IS NOT trolling and surely not a troll

[Hornet331]

This is the only way to "test" this theory/phenomenon out.
However it is indeed hard to obtain all that hardware ( unless you can borrow it from a retailer/shop/manufacturer ), and of course raise the "testees".

I thought of doing that in the past, but the people who claimed that their AMD systems were smoother than the Intel based ones of their friends suddenly refused to participate in the blind test the last day before I decided to go and pickup the stuff.

p.s. If you're going to compare desktop "smoothness/responsiveness" like the Start Menu for example, then you need to make sure that both systems have the same Start Menu appearance delay, and other stuff too.

Actually i dont think you would need that much hardware, 3 or 4 computers with 2 dualcores (amd/intel) and 2 quadcores (amd/intel) maybe a third intel quad (Ci7) and then let people test it one after another.

But i think the outcome would be simmilar to the Mojave Experiment from MS.

If you have made up your mind about a thing, even if its only subconsciously, there will be a "difference".

[TL1000S]

Thanks for posting screenies demonkevy666. It really give hopes for a good sale of PHII.

The huge OC-ability for PHII alone is reason enough for buying/building a PHII-rig (for me).

But, it is "wrong" to conclude what you do regarding scores when the CPU-clock is that different.
If the Ci7-rig was able to clock to PHII-level you would see a significant boost in GPU-scores also.
That is my experience with lotsa 3DM06-benches.
But I think you're right that AMD rigs/CPU's (K8/K10) scores well in the GPU-benches even though the CPU-score is much lower than a similary OC'ed C2D/C2Q.
We see the same pattern in games also.

[demonkevy666]

Thats why the benchmarks/tests should be repeated to make sure no such artifacts remain in the given scope.



There is no way one can draw any conclusion regarding smoothness from that. Neither can one draw any conclusions from the CPU vs GFX test results since the other(may it be CPU or GFX) is written in a way in which it flavours either AMD K10 or Intel Core µarch - there is no way to make a "unbiased" test. Look at SuperPI, the way it is written(and ran inside a CPU) (unintentionally)flavours the Intel µarch. It just happens to run better on Intel parts. It must be - "smoother" there.

my post was about more threads = to more performance. not really about smoothness. the "generic CPU" bench is pretty good to not bias at all, and while I know you made one/postedone I haven't tried it out on my O.C yet (I'm trying to sort out other issues I'm having first before I do.)

You're missing that the CPU Test is a totally different usage scenario from the Graphics Tests in 3D Mark.
Having equal CPU Scores doesn't make the performance of the CPU's equal in the Graphics Tests that are using less resources of the CPU and are coded differently.




This is the only way to "test" this theory/phenomenon out.
However it is indeed hard to obtain all that hardware ( unless you can borrow it from a retailer/shop/manufacturer ), and of course raise the "testees".

I thought of doing that in the past, but the people who claimed that their AMD systems were smoother than the Intel based ones of their friends suddenly refused to participate in the blind test the last day before I decided to go and pickup the stuff.

p.s. If you're going to compare desktop "smoothness/responsiveness" like the Start Menu for example, then you need to make sure that both systems have the same Start Menu appearance delay, and other stuff too.

And no, keithlm, Hornet331 IS NOT trolling and surely not a troll

I've got to the conclusion that the phenom graphic scores are all part of the clock speed(be it gpu or cpu It don't know) after reading a few threads about 3dmark06. but this don't change the fact Tony's Q9650 was getting lower scores in graphics to and equal clocked phenom 9850 and an 8800 ultra. also if I recall VIA benched there chip using both codes and the AMD code was far below then intel code for 3dmark06. It seem strange that there would be any real reason to change coding for a cpu unless you run it a 64 bit problem.

Thanks for posting screenies demonkevy666. It really give hopes for a good sale of PHII.

The huge OC-ability for PHII alone is reason enough for buying/building a PHII-rig (for me).

But, it is "wrong" to conclude what you do regarding scores when the CPU-clock is that different.
If the Ci7-rig was able to clock to PHII-level you would see a significant boost in GPU-scores also.
That is my experience with lotsa 3DM06-benches.
But I think you're right that AMD rigs/CPU's (K8/K10) scores well in the GPU-benches even though the CPU-score is much lower than a similary OC'ed C2D/C2Q.
We see the same pattern in games also.

I'd love to know why no one uses an SSD with phenom ??? no one bench with it at all ever.
sure cpu point are going to vary, but when I bench 3dmark I have to turn off all the back ground task that I don't need and bingo added 300 more points to my cpu. of coruse I'm using vitsa 64 bit so I'm already 400 behind.

[BenchZowner]

I've got to the conclusion that the phenom graphic scores are all part of the clock speed(be it gpu or cpu It don't know) after reading a few threads about 3dmark06. but this don't change the fact Tony's Q9650 was getting lower scores in graphics to and equal clocked phenom 9850 and an 8800 ultra. also if I recall VIA benched there chip using both codes and the AMD code was far below then intel code for 3dmark06. It seem strange that there would be any real reason to change coding for a cpu unless you run it a 64 bit problem.

Most part of your post doesn't seem to make any sense mate.The second part at least.
The tests that you are referring to were done on PC Mark04 or PC Mark05 ( can't recall which one for sure, I think it was 05 ) and it was a comparison between Intel's Atom and Via's Nano CPUs, and they changed the CPU string to make the program think that it's another CPU and that probably triggered different routines/optimizations and/or instruction-sets to be used.

There are several benchmarks in the soon to be released Clock per Clock review from HWbox.gr, and 3D Mark06 is included, and there are direct comparisons of both Phenom I vs Phenom II and Phenom II vs the competition ( Kentsfield, Yorkfield, Bloomfield )

[Enisey]

My first CPU was Athlon XP 1600+ (1400 mhz), the next one was Celeron D336 (2800 mhz). When I was running iTunes and then starting some other task, the music would jerk and become repetitive for a few seconds. I remember that surprised me at the time since I couldn't remember something like that on my old Athlon. On the other hand I don't remember if I used the same version of the software or not, with the time iTunes was becoming more and more of a resourse hog.

Still on the Celeron now choosing between Q9550 and P2 940.

[Ghostbuster]

Like I said, this whole thread is epic fail, in my opinion... Everytime the thread starter hits a brick wall on his arguments, he starts calling people trolls.

Talking about low FPS, you can check this very interesting article from many moons ago.

http://www.firingsquad.com/hardware/$500_gaming_pc_upgrade/page7.asp

The average FPS all looks similar until you see the minimum FPS in the table down the page! I call this real "smoothness" data...