Bottlenecks?

A bottleneck is where one (or more) components slows overall performance.
The cure is to get parts that allow the machine to perform as fast as you need.
When you need more performance, identify the parts causing the bottleneck and speed them up.
If you cant speed them up/replace them with faster parts (and the configuration is fine), you are out of luck.
 
Bottlenecks are used to make it easier to control the flow of the contents inside the bottle. By limiting the amount of stuff coming out, the user has more control over the flow of the bottles contents. You can't avoid them. They're everywhere.
 
I have a few more questions about this. (sorry it's taken me so long to respond):

Is it possible to find a bottleneck just based on specs alone? If it is, what speeds/specs do you compare to find a "bottleneck"? (between ram, videocard, and processor) If it's not, is it something that has to be tested for in order for it to be found? (as in, I notice some sort of visual tearing, so my bottleneck must be etc etc)

Also, how do you identify the parts causing the bottleneck?
 
Then explain, rather than pointing out the obvious. Hey, your name is Jospeh! :p

It's my understanding that any part that results in degraded performance is considered a "bottleneck". The "Cure" to a bottleneck is to identify the parts that are causing the problem. I'm asking it again, how do you identify which parts are causing the problem?
 
You'll first need an understanding of computer hardware architecture, in order to spot a bottleneck based on specs alone. Get reading!
 
Though, some stuff may seem like a bottleneck, based on specs, but really aren't. For example, using a current gen (HD4000 and GT200 or below) PCI-E 2.0 vidcard in a PCI-E1.1 slot. Since the bandwidth of PCI-E2.0 is twice that of PCI-E1.1, someone may conclude that the slot is a bottleneck. However, reviews have shown that, despite the card's compliance with the 2.0 PCI-E spec, such bandwidth isn't needed for full speed operation. So, I don't think you should base your conclusion on specs alone. ;)

So, the real answer to your question is: its complicated and more involved then you think. :( You'll need to familiarize yourself with hardware capabilities as well as computer hardware architecture in general. Lots of learning to do... fun stuff. One thing to remember: performance on paper rarely matches real world performance.
 
Speeds and specs matter but only so far.

They have different meaning between different graphics card generations and ATI / NVidia cards.
They are useful when for example, you are looking at NVidias GT200 generation cards, like the GT250, GTX260, GTX275, GTX280, GTX285, GTX295 etc...
But you have to also look at changes in the architecture of the cards as well.

ie1: the GTX295 is 2 GTX260's on one card but also has more shader units.
ie2: the GTX275 is a faster clocked GTX260 with more shader units.
ie3: the GTX280/285 are clocked faster again with even more shaders, the 285 being the fastest.
etc etc.

ATI's card can be similarly defined by their clock speeds and features.

So for Nvidias or ATIs card ranges you can compare specs and get some meaning but only to a point.
It wont help you compare across opposing camps cards unless you have a performance baseline for each architecture and generation.
There are similarities between older generation cards that allow you to do a spec/performance analysis but its not straightforward.
Older generations can use similar ways of analysing between cards but there are other differences based on their architecture.
Most systems get the best gaming benefit by overclocking the GPU or getting a faster graphics card.
However if the system is CPU limited, a faster GPU will not help much as it needs more CPU power to use it at full speed.


CPUs have 1 to 4 cores.
Performance differences between Intel and AMD, different generations, different architectures, different number of cores with different software are not easily determined from specs alone.
To get a decent performance metric its a good idea to check out reviews of the software you use to make your own baselines.

The performance benefit from multiple cores depends on what software you are using.
Most games benefit from high clocked dual cores, quite a few will get higher performance from triple or quad core and less need highest clocked quad cores.
A lot of games games benefit from overclocking the CPU but the benefit is also affected by the power of the graphics card and its CPU requirement.



Memory speed matters more for AMD systems, current Intel systems benefit more from faster memory timings although the benefit isnt very big for either.
Large jumps in memory speed can help a little but you usually need to slow the memory timings down to achieve them so you need to clock high enough to cover the loss.
Some software is highly dependent on memory speed, some is helped by fast memory timings, most software cares little about memory speed/timings.



There is no easy comparison between systems just using hard figures.
This is why absolutely everything gets reviewed in absolutely every configuration on different sites.
 
You must log in or register to reply here.
Back
Top