What Makes the cores of a CPU "Weak"

Jul 30, 2018
23
0
10
Hello, i would like to know what people mean when they say , its a good processor but it has weak cores, is it just the ghz or there is more? tell me everything please
 
Solution
Another way to say it would be that clock speed just tells you how often "bundles" of instructions are sent through the CPU per second and chip architecture (which in theory should be improved with each generation) is how efficient and how many instructions can be done per cycle. Faster clock speed does't necessarily mean faster processor. This is only true if its the same processor. An 8700k at 5.0Ghz is of course faster than an 8700k at 4.3Ghz. But when you compare generations usually you get more computation power per clock with a new architecture.

Dunlop0078

Titan
Ambassador
I wouldn't just say weak, weak single thread performance would be a better way to put it. It has to do with how many instructions a CPU can perform per clock. So for example a single core out of an 8700k can do more calculations at say 1ghz in a given time than an single core out of an old intel Q6600 could.

That's the way I understand it anyway.
 
To be honest I've never seen anyone say that. Processors can be well designed or poorly designed, like anything else. In certain types of undemanding tasks it's possible that both a well designed CPU and a poorly designed one can seem to function the same. So, in theory, you could say that's a 'good' processor about both of them.

Then you try to run a game and you see that the poorly designed one can't keep up. As to the technical reasons why, no one can answer that without knowing which specific processor you're talking about. There can be any number of reasons why it's 'weak'.
 
Jul 30, 2018
23
0
10


what if all my cores say they are running at 4.2ghz , is there anything else that makes it weak? my friend who has an fx-8320 gets told by people that his single thread performance is weak, which is true, but what makes it weak? he is running at 4.2 ghz with good cooling system
 
Jul 30, 2018
23
0
10


fx-8320 (friend's cpu)

 

Dunlop0078

Titan
Ambassador


Exactly what I descried in my first post. If you took one core or even module out of an fx 8320 at 4.2ghz and one core out of an i7 2600k at 4.2ghz the i7 2600k core would perform better because it can perform more instructions per clock. If you want to know why that is you need to take a deep dive into cpu architecture and manufacturing process and likely many other factors that determines a CPU's IPC.
 
The clock is related to how quickly data moves between things in the cpu.
The CPU is made of billions of transistors and they have performance as well.
Moore's Law is based on transistor count, which is a better rough estimate of performance when comparing different lines. CPU memory is very important. Certain tasks need more of it.

AMD Bulldozer was behind on nearly everything, but it's poor architecture design really bottlenecked it. It couldn't utilize the resources it had. That was AMD's worst line ever. 4core Sandy bridge is better than the whole line.
 

mjbn1977

Distinguished
Another way to say it would be that clock speed just tells you how often "bundles" of instructions are sent through the CPU per second and chip architecture (which in theory should be improved with each generation) is how efficient and how many instructions can be done per cycle. Faster clock speed does't necessarily mean faster processor. This is only true if its the same processor. An 8700k at 5.0Ghz is of course faster than an 8700k at 4.3Ghz. But when you compare generations usually you get more computation power per clock with a new architecture.
 
Solution
Jul 30, 2018
23
0
10


Thanks, i'd give you the solution mark but the other guy got it right too, lets help him earn the cpu badge too , dont you think?
 
Jul 30, 2018
23
0
10


how do we count the efficiency and the number of instructions that can be done per cycle
 


passmark is a decent bench for general use. finding real use performance for specific tasks is the only real way, but it's hard to find the data. It's difficult to know what the bottleneck is. single core performance has been fairly stagnant. intel has made real strides with their new turbo settings. some of the new ones hit 2700+ passmark. AMD is still quite a bit behind there, but their chips are really close and offer a lot of value at certain price points now with Ryzen.
 
Jul 30, 2018
23
0
10


Yes, my question is what does the 800mhz processor have that the other doesnt have, whats the measurement
 


There are billions of parts to a cpu. If you are looking at bulldozer, they tried to implement something where a piece was processed on one core and another waited on it to finish it. They couldn't feed it through fast enough and it bottle necked them really bad. People were contesting their core count as well in court. It was closer to SMT then full cores.