So in otherwords, you should consider clock speed and core count when estimating the performance of a processor.
caption your avatar "HA!" and it perfectly suits that post.
Rollback Post to RevisionRollBack
Proud member of the MCF AWA war of '13! if someone suggests Alienware or Cyberpower, wait for a custom-built list from someone who knows their stuff. Meh Rig
Correct, clock speed and cores do not matter in the way we are discussing in this thread.
And overall, they can largely be ignored when selecting a CPU, especially for gaming.
So yes, a (lets just throw random examples out here) 2.0GHz sandy bridge processor with only 1 core enabled will peform the same as, say, an athlon II x4 at 3.0GHz with a single threaded app. It might even perform better. However, if we're comparing, say, an IB i5 against a modern FX, then the per core and per clock (GHz) difference is so minimal that you need to go down to architecture and CPU design to differentiate the two, or even see which is more powerful on one end or the other.
However, that brings in another case, in that cores and GHz also do not matter for many multithreaded apps if the architecture is not as good, a hexa core at 5GHz with a multithreaded app and poor architecture will perform worse than a quad core at 3GHz with a multithreaded app. You eventually hit a plateau of performance with GHz and cores and need to look deeper. We saw this with the flawed BD chips (cache bug aside).
What part of this are you not understanding?
Do they 'matter'? Sort-of. In the same way that the color of your mobo 'matters' or the brand of your CPU 'matters'.
However after that point you start to get into complicated matters of the inner workings of the CPU, in which case, it is far easier and less annoying to just look at benchmarks.
Correct, clock speed and cores do not matter in the way we are discussing in this thread.
And overall, they can largely be ignored when selecting a CPU, especially for gaming.
So yes, a (lets just throw random examples out here) 2.0GHz sandy bridge processor with only 1 core enabled will peform the same as, say, an athlon II x4 at 3.0GHz with a single threaded app. It might even perform better. However, if we're comparing, say, an IB i5 against a modern FX, then the per core and per clock (GHz) difference is so minimal that you need to go down to architecture and CPU design to differentiate the two, or even see which is more powerful on one end or the other.
However, that brings in another case, in that cores and GHz also do not matter for many multithreaded apps if the architecture is not as good, a hexa core at 5GHz with a multithreaded app and poor architecture will perform worse than a quad core at 3GHz with a multithreaded app. You eventually hit a plateau of performance with GHz and cores and need to look deeper. We saw this with the flawed BD chips (cache bug aside).
What part of this are you not understanding?
Do they 'matter'? Sort-of. In the same way that the color of your mobo 'matters' or the brand of your CPU 'matters'.
However after that point you start to get into complicated matters of the inner workings of the CPU, in which case, it is far easier and less annoying to just look at benchmarks.
Yes that is what I have been maintaining. Higher cores or Ghz doesn't mean better performance nor worse performance but is is simply just a factor when looking at CPU's.
Rollback Post to RevisionRollBack
I always wanted to argue with a brick wall, I suppose the internet is the second best option.
Yes that is what I have been maintaining. Higher cores or Ghz doesn't mean better performance nor worse performance but is is simply just a factor when looking at CPU's.
But it shouldn't be a deciding factor unless the CPU is by the same company, made with the same architecture, in the same generation, etc.
e.g. you could compare an i5 3570k @ 3.4GHz and an i5 3570k @ 4.3GHz and, with complete confidence, say the latter i5 is better than the former. You cannot, however, compare an i5 to an i7, or i3 to an AMD chip using GHz. Hell, you couldn't even compare two different i5s using GHz.
But it shouldn't be a deciding factor unless the CPU is by the same company, made with the same architecture, in the same generation, etc.
e.g. you could compare an i5 3570k @ 3.4GHz and an i5 3570k @ 4.3GHz and, with complete confidence, say the latter i5 is better than the former. You cannot, however, compare an i5 to an i7, or i3 to an AMD chip using GHz. Hell, you couldn't even compare two different i5s using GHz.
Yes. Exactly. However Ghz has to be considered as does every other factor.
Rollback Post to RevisionRollBack
I always wanted to argue with a brick wall, I suppose the internet is the second best option.
Considered yes. But very, very minimally, and as a last resort. Not one of the deciding factors between processors.
Yes. I know.
I don't know where this idea that I apparently assume that Ghz is the absolutely most important thing ever came from.
If a CPU has inferior design (IPC, Instruction set, etc) and a lower clock, then it is possible to estimate that it will perform at a lesser level.
However if you look the IPC, instruction set, efficiency, and overall logic and design, it is not possible to estimate the CPU's performance in relation to others w/o considered Ghz and cores.
Yes. I know.
I don't know where this idea that I apparently assume that Ghz is the absolutely most important thing ever came from.
If a CPU has inferior design (IPC, Instruction set, etc) and a lower clock, then it is possible to estimate that it will perform at a lesser level.
However if you look the IPC, instruction set, efficiency, and overall logic and design, it is not possible to estimate the CPU's performance in relation to others w/o considered Ghz and cores.
The only real comparisons of CPUs is through multiple, professional benchmarks. You have to be exceedingly knowledgeable about CPU designing and microelectrical engineering to be able to compare CPUs without benchmarks.
The only real comparisons of CPUs is through multiple, professional benchmarks. You have to be exceedingly knowledgeable about CPU designing and microelectrical engineering to be able to compare CPUs without benchmarks.
Oh yes so lets all give up now right? Perhaps the average consumer doesn't understand or wish to look up bench marks. Perhaps they want to have some basic knowledge of comparing CPU's.
But lets all just tell them that they will never be able to ever even begin to have any hope of making a decision.
Rollback Post to RevisionRollBack
I always wanted to argue with a brick wall, I suppose the internet is the second best option.
Clock speed does not matter. Cores do not matter. Other variables you are ignoring matter.
caption your avatar "HA!" and it perfectly suits that post.
if someone suggests Alienware or Cyberpower, wait for a custom-built list from someone who knows their stuff. Meh Rig
So a 400 Mhz Single Core Processor has the chance to beat a 4Ghz Six Core?
If it is a cell based processor, or a parralex "GPU like" core of modern articture, then yes.
If the 400 MHz processor is of this generation and has superior architecture, yes it can.
What fm87 is saying is that you cannot compare two processors of different models, generations, etc. using GHz in any form.
Two processors of the EXACT same brand, model, generation, etc can be compared by GHz.
"Programmers never repeat themselves. They loop."
You mean:
And overall, they can largely be ignored when selecting a CPU, especially for gaming.
So yes, a (lets just throw random examples out here) 2.0GHz sandy bridge processor with only 1 core enabled will peform the same as, say, an athlon II x4 at 3.0GHz with a single threaded app. It might even perform better. However, if we're comparing, say, an IB i5 against a modern FX, then the per core and per clock (GHz) difference is so minimal that you need to go down to architecture and CPU design to differentiate the two, or even see which is more powerful on one end or the other.
However, that brings in another case, in that cores and GHz also do not matter for many multithreaded apps if the architecture is not as good, a hexa core at 5GHz with a multithreaded app and poor architecture will perform worse than a quad core at 3GHz with a multithreaded app. You eventually hit a plateau of performance with GHz and cores and need to look deeper. We saw this with the flawed BD chips (cache bug aside).
What part of this are you not understanding?
Do they 'matter'? Sort-of. In the same way that the color of your mobo 'matters' or the brand of your CPU 'matters'.
However after that point you start to get into complicated matters of the inner workings of the CPU, in which case, it is far easier and less annoying to just look at benchmarks.
Yes that is what I have been maintaining. Higher cores or Ghz doesn't mean better performance nor worse performance but is is simply just a factor when looking at CPU's.
But it shouldn't be a deciding factor unless the CPU is by the same company, made with the same architecture, in the same generation, etc.
e.g. you could compare an i5 3570k @ 3.4GHz and an i5 3570k @ 4.3GHz and, with complete confidence, say the latter i5 is better than the former. You cannot, however, compare an i5 to an i7, or i3 to an AMD chip using GHz. Hell, you couldn't even compare two different i5s using GHz.
Click the picture!
-Derek Shunia
Yes. Exactly. However Ghz has to be considered as does every other factor.
Considered yes. But very, very minimally, and as a last resort. Not one of the deciding factors between processors.
"Programmers never repeat themselves. They loop."
i5 6600k 4.6ghz / MSI 280X / 8Gb 2666 DDR4 / Gigabyte Z170X-UD5 / TX550M / 500Gb 850 EVO / NZXT S340 / Corsair K65 / Corsair M60
I honestly have no idea.
"Programmers never repeat themselves. They loop."
Yes. I know.
I don't know where this idea that I apparently assume that Ghz is the absolutely most important thing ever came from.
If a CPU has inferior design (IPC, Instruction set, etc) and a lower clock, then it is possible to estimate that it will perform at a lesser level.
However if you look the IPC, instruction set, efficiency, and overall logic and design, it is not possible to estimate the CPU's performance in relation to others w/o considered Ghz and cores.
The only real comparisons of CPUs is through multiple, professional benchmarks. You have to be exceedingly knowledgeable about CPU designing and microelectrical engineering to be able to compare CPUs without benchmarks.
"Programmers never repeat themselves. They loop."
Oh yes so lets all give up now right? Perhaps the average consumer doesn't understand or wish to look up bench marks. Perhaps they want to have some basic knowledge of comparing CPU's.
But lets all just tell them that they will never be able to ever even begin to have any hope of making a decision.
That is not how the average consumer is.
The average consumer couldn't give less of a damn about benchmarks. They just want a computer "that works".
If you care about which CPU is better than another, then you are knowledgeable enough to look up benchmarks.