The Spin off Smoothness Thread

[Hornet331]

after owning a core 2 duo, a core 2 quad, a phenom 9850, and now a core i7 I personally saw absolutely nothing smoother about the amd setup over the intel setup. I actually never noticed the difference going from one cpu to the next.

tell that the guys "who are seeing things"
Im in the same boat as you are. Ever since the switch to dualcores i noticed no more increase in "smoothness" over the time.

riped from the review thread:

I'm sure this whole "smoothness" thing would be a lot easier to sell to an engineer at a server OEM- they work on a little bit different platfom integration level than just slapping some crap together and shooting zombies

Sure, but you dont hear such comments form that crowd. Its the gamers that claim its "more smoother."

Regardless of whether you believe it has an effect on gaming or not, intel and AMD platforms running specific applications differently is a FACT, not a MYTH.

I dont deny that, every architecture has its strengh and weaknesses, just look at archivers, ever since AMD introduced its IMC it was king in such apps that needed a lot of bandwidth. Thats a fact and can be proven by numbers and data.


Imho "smoothness" is the same as with the "noone needs more then 30fps".
Some guys see it, some guys dont, you cant quantify things that rely on human perception, regardless of how sophisticated your test suite, measurment gear, data loging etc. is, noone has succeded to factor in human perception off certain things till now.

Addition:

Thats also why the whole benches will fail, the only thing i have seen here has nothing to do with smoothness rather then how good the cpu or more like the os can handle/manage multiple threads with heavy load.

this is a example:
[QUOTE=sundancerx;3544242]

EXAMPLES:
1. Run benchmark.
2. Run benchmark with 4xPrime95.
3. Run benchmark with Virus scan or defrag.
4. Run benchmark while compressing.
5. Run benchmark while encoding.

The second test don't really tells you anything about smoothness, but rather how good the os is in allocating processor time. If the priority of the benchmark and prime is the same, you loose fps. But if the priority of the benchmark is higher then for prime it will run faster (but still slower as if would run without prime in the backround).

Same goes for test 3, but now the dominant factor is the HDD, a 6ms VR would fell much "smoother" then a standard 12ms HDD, with that test your really benching your HDD and not the CPU. Also depending on the benchmark, you wouldn't even notice any drops at all.

Same for the other 2 scenarios, your testing the cpu while one subsystem is beeing stressed and turned into a bottleneck.

Imho the biggest problem is, that there is no explicit definition for "smoothness". It not possible to test for something if the objective is not clear in the first place.