omarman Wrote:I posted that because micro stuttering is a term used to describe what happens when you use dual GPU setups, that's what I meant by incorrect vocabulary. I think the word you are looking for is "lag", or dropped frames.
Well then you don't understand what the term means. Microstuttering isn't exclusive to dual GPU setups. Microstuttering just means stuttering with a short duration. The prefix "micro" means small (in this case referring to a small interval of time). And framerate stuttering is just an increase in frametime. In other words a microstuttering in video games is a short increase in frametime. Which is what we're measuring here.
Lag and dropped frames have nothing to do with this. Dropped frames are frames that are not decoded from a video stream. Video games don't produce dropped frames since frames don't have to be decoded. "Lag" is a catch all term used to refer to any increase in latency. Video game lag comes in many different flavors (network, framerate, or stuttering) with many different sources (memory transfers, input, GPU, CPU, etc.).
omega_rugal Wrote:if i`m upgrading from an older AM2/AM3 CPU, or whichever is cheaper or available in the country i live? or someone who doesn`t plan to run demanding games on it? those are 4 reasons, or maybe people like me, who simply don`t buy intel anymore for my very own reasons (their customer support sucks).
Well you're in a different scenario.
As I said in my last sentence:
NaturalViolence Wrote:It's about evaluating which product is right for you and why.
@Garrlker
We're only talking about emulation and PC gaming right now, not general benchmarks.
(07-02-2013, 02:21 PM)garrlker Wrote: [ -> ]http://www.anandtech.com/bench/Product/697?vs=701
I5 3570k wins the majority on these benchmarks. And processors can have micro stuttering.
The way I read it was basically when the cpu goes from not processing much like just the character walking or driving, and all of the sudden there's an explosion and tons of physics and other things to process all at once. That spike can cause it. I can't really remember exactly how it was worded but that's the gist of it. I'm not describing the cpu bottlenecking or anything either. Just a spike in a lot more stuff to process than what it was processing can cause it for a split second. I think NV will know what I'm talking about and elaborate.
At the top of the comparison it shows that the two CPUs are almost identical with the intel in the lead by a little, except it costs 30$ more, and that's why I chose the FX 8350, it's almost as good as the i5 3570k and 30$ cheaper, 30$ I didn't have.
omarman Wrote:Now I agree with both arguments, but I began a new one.
Argument 3: Although Intel CPUs are better than AMD CPUs, you cannot notice the difference in real world settings. (eg; 1 GPU, 1080p, fps cap at 30/60)
Where do you currently stand on this argument? Still agree or have I changed your mind?
(07-02-2013, 04:55 PM)NaturalViolence Wrote: [ -> ]omarman Wrote:Now I agree with both arguments, but I began a new one.
Argument 3: Although Intel CPUs are better than AMD CPUs, you cannot notice the difference in real world settings. (eg; 1 GPU, 1080p, fps cap at 30/60)
Where do you currently stand on this argument? Still agree or have I changed your mind?
I'm going to check it out, I have most of my parts already, just waiting for my Hyper 212+ , case, and solid state to arrive. I will see if I notice any stuttering at all.
I plan to OC to at least 4.5ghz
P.S. If Dolphin wasn't so poorly optimized for multi-cores, we could run fair-er tests.
Well if
you can work out how to make algorithms such as the following be parallelised, we could fix that
Example algorithm:
Quote:Take a number
Double it
Take off 1
Square it
Divide it by 3
find its cubic root
Spit this out
Basically, each instruction relies on the result of the previous instruction, and so cannot be run simultaneously on a second core. Most of Dolphin is in a similar kind of situation.
(07-02-2013, 04:59 PM)omarman Wrote: [ -> ]just waiting for my Hyper 212+
Why didn't you get the EVO?
(07-02-2013, 04:59 PM)omarman Wrote: [ -> ] (07-02-2013, 04:55 PM)NaturalViolence Wrote: [ -> ]omarman Wrote:Now I agree with both arguments, but I began a new one.
Argument 3: Although Intel CPUs are better than AMD CPUs, you cannot notice the difference in real world settings. (eg; 1 GPU, 1080p, fps cap at 30/60)
Where do you currently stand on this argument? Still agree or have I changed your mind?
I'm going to check it out, I have most of my parts already, just waiting for my Hyper 212+ , case, and solid state to arrive. I will see if I notice any stuttering at all.
I plan to OC to at least 4.5ghz
P.S. If Dolphin wasn't so poorly optimized for multi-cores, we could run fair-er tests.
Poorly optimized for multi-cores?
Woah there pal,
Optimizing an application to scale across multiple threads is hard as hell!.
Heck, even more so in emulation!
Emulation is the process of "emulating something" (Which is, in our case, the Wii/GC using Dolphin Emu.) and thus, needing to emulate the hardware of the system that we want to emulate.
Dolphin is very accurate in emulating the Wii and GC consoles.
Dolphin is a Dual Core Application!! (Well, it can use up to 3 cores though) and making it run on more than three threads will most probably break a lot of things which will make the emulation pointless.
Think twice before posting that "Dolphin is poorly optimized for multi-core" crap.
I got the Hyper 212+ for the same reason I got the FX chip, it's cheaper.
Also about the whole dolphin is poorly optimized for multi-cores, it's true. I'm not saying it's a bad program, or that I could do any better, I think it's an awesome program, and fine the way it is. But if it were able to accurately emulate while utilizing 8 cores, it would make more fair tests between AMD and Intel , that is all I was getting at.
This has been discussed countless times on the forums, if they were to code Dolphin to use more cores it would end up just slowing it down. The 2 core/3 core with LLE is the best way to do it. Any more cores and it would only slow the emulator down. Optimizing a Cpu,Gpu,And audio threaed to 8 cores instead of 3 would be alot more difficult also and would break a lot of stuff for a while.
omarman Wrote:P.S. If Dolphin wasn't so poorly optimized for multi-cores, we could run fair-er tests.
"If it doesn't take advantage of the differences in my architecture then it's not a fair test."
This logic was used by those defending the Itanium and Netburst microarchitectures back in the day. It is poor logic. It needs to die. Different workloads require different algorithms, this will always be true. Some algorithms are not going to be well suited to the design choices used in a given microarchitecture. You cannot evaluate cpu performance based off of "what if" scenarios. The only fair way to evaluate cpu performance is to test software that actually exists and is actually used by users. Some algorithms simply can't be redesigned to run well on certain microarchitectures that emphasize things that the algorithm doesn't need and/or can't make use of.
Like others have said what you ask also happens to be impossible in this particular case due to the way emulation works. If you use the forum search you'll find lots of long posts about this topic including some by the devs. I've written about this topic myself so many times that I'm not going to do it again here. But to sum it up, it's because the GC/Wii only has 3 chips that do most of the processing work that we can emulate. And they're all single threaded. If the GC/Wii had an 8 core cpu we would have no problem making dolphin use 8 cores, but it doesn't. We can't emulate something in parallel if the real hardware doesn't do it in parallel.
@tiktakt0w
Your attitude isn't helping.
omarman Wrote:I'm going to check it out, I have most of my parts already, just waiting for my Hyper 212+ , case, and solid state to arrive. I will see if I notice any stuttering at all.
You can do that. But whether you notice it doesn't really matter. Especially when you don't have another system in front of you to compare it against.