Jump to content

Do all cores on a CPU run at the CPU's clock speed?

Hello, sorry about my lack of knowledge on this topic. But I was wondering how the clock speed of CPU's work with cores. Sorry if that is really badly phrased. I am just completely confused about this topic because I first was taught that if an quad core CPU runs at lets say 4GHz each core will run at 4GHz making the processor 16GHz altogether. But recently I heard that a 4GHz quad core CPU's cores would each run at 1GHz each and all of the cores would work together and be 4GHz fast. If this is true would it also not be better to get a single core CPU for applications that only take advantage of single cores? Sorry about my bad phrasing and my lack of knowledge.

Link to comment
Share on other sites

Link to post
Share on other sites

Hello, sorry about my lack of knowledge on this topic. But I was wondering how the clock speed of CPU's work with cores. Sorry if that is really badly phrased. I am just completely confused about this topic because I first was taught that if an quad core CPU runs at lets say 4GHz each core will run at 4GHz making the processor 16GHz altogether. But recently I heard that a 4GHz quad core CPU's cores would each run at 1GHz each and all of the cores would work together and be 4GHz fast. If this is true would it also not be better to get a single core CPU for applications that only take advantage of single cores? Sorry about my bad phrasing and my lack of knowledge.

both are wrong, when a chip is running at 4hgz all cores are running at 4Ghz but it doesn't stack so it's only 4Ghz. I also really don't want to meet the person(s) who taught you either of those supposed "ways" a cpu works.

-The Bellerophon- Obsidian 550D-i5-3570k@4.5Ghz -Asus Sabertooth Z77-16GB Corsair Dominator Platinum 1866Mhz-x2 EVGA GTX 760 Dual FTW 4GB-Creative Sound Blaster XF-i Titanium-OCZ Vertex Plus 120GB-Seagate Barracuda 2TB- https://linustechtips.com/main/topic/60154-the-not-really-a-build-log-build-log/ Twofold http://linustechtips.com/main/topic/121043-twofold-a-dual-itx-system/ How great is EVGA? http://linustechtips.com/main/topic/110662-evga-how-great-are-they/#entry1478299

Link to comment
Share on other sites

Link to post
Share on other sites

both are wrong, when a chip is running at 4hgz all cores are running at 4Ghz but it doesn't stack so it's only 4Ghz. I also really don't want to meet the person(s) who taught you either of those supposed "ways" a cpu works.

You claim that they do not stack, is it possible to go into more detail? If you cant its not really a problem but it would be helpful.

Link to comment
Share on other sites

Link to post
Share on other sites

both are idiotic, a 4Ghz quad core is a four core chip with each running at 4Ghz. That stacking Mhz count is pseudo marketing bullshit.

 

The only times that the clocks are different is if the chip is running in turbo mode.

Error: 410

Link to comment
Share on other sites

Link to post
Share on other sites

The only times that the clocks are different is if the chip is running in turbo mode.

to elaborate, turbo is an increase of clockspeed on 1 or 2 cores on your cpu if there is a lightly threaded load.

 

 

Hello, sorry about my lack of knowledge on this topic. But I was wondering how the clock speed of CPU's work with cores. Sorry if that is really badly phrased. I am just completely confused about this topic because I first was taught that if an quad core CPU runs at lets say 4GHz each core will run at 4GHz making the processor 16GHz altogether. But recently I heard that a 4GHz quad core CPU's cores would each run at 1GHz each and all of the cores would work together and be 4GHz fast. If this is true would it also not be better to get a single core CPU for applications that only take advantage of single cores? Sorry about my bad phrasing and my lack of knowledge.

there are 4 different cores on a cpu, each of them are running at 4 ghz. think of each core running separately on 1 thread each.

AAAAAAAAAGGGGGGGHHHHH!!!!

Link to comment
Share on other sites

Link to post
Share on other sites

I asked for assistance in this one so others should post soon.

-The Bellerophon- Obsidian 550D-i5-3570k@4.5Ghz -Asus Sabertooth Z77-16GB Corsair Dominator Platinum 1866Mhz-x2 EVGA GTX 760 Dual FTW 4GB-Creative Sound Blaster XF-i Titanium-OCZ Vertex Plus 120GB-Seagate Barracuda 2TB- https://linustechtips.com/main/topic/60154-the-not-really-a-build-log-build-log/ Twofold http://linustechtips.com/main/topic/121043-twofold-a-dual-itx-system/ How great is EVGA? http://linustechtips.com/main/topic/110662-evga-how-great-are-they/#entry1478299

Link to comment
Share on other sites

Link to post
Share on other sites

You claim that they do not stack, is it possible to go into more detail? If you cant its not really a problem but it would be helpful.

If you have a quad core 4GHz, the performance is not equal to a 16GHz one core. It is equal to 4 cores at 4GHz each doing different things. In the great words of @Whiskers

 

"A core is only one part of a CPU. Modern CPUs are usually comprised of two or more cores - however, the cores are independent and their clock speed does not 'stack'.

Basically, each core can operate (to an extent) independently. The first core can perform one task whilst the second performs another - but they cannot really work on the same specific operation simultaneously. As a result of this, the CPU itself is rated at the same clock speed as that of each of its cores. :)"

Link to comment
Share on other sites

Link to post
Share on other sites

When it says 4Ghz base clock it means that all cores will run at 4 Ghz. This doesn't mean the chip will run at 16 Ghz if it's a quad core because that's not how you calculate its speed. All the transistors in those 4 cores are switching off and on at the same rate. Turbo Core will increase a single core and when that single core is receiving most of the threaded workload it gives itself a speed boost.

Mobo: Z97 MSI Gaming 7 / CPU: i5-4690k@4.5GHz 1.23v / GPU: EVGA GTX 1070 / RAM: 8GB DDR3 1600MHz@CL9 1.5v / PSU: Corsair CX500M / Case: NZXT 410 / Monitor: 1080p IPS Acer R240HY bidx

Link to comment
Share on other sites

Link to post
Share on other sites

Now now everyone, let's look at this with a bit more care. 

 

Let's look at objective performance: throughput/how many instructions a CPU can execute in a second.

 

CPU architectures have an Instruction per Clock(or cycle) efficiency denoted as IPC between 0 and 1 (let's not argue semantics of SIMD which is a separate rat's nest of philosophical debacles).

 

CPUs have a clock rate in Gigahertz (10^9 cycles per second)

 

CPUs have a core count >1

 

IPC * Clock Rate = single core performance * core count = total performance

 

For Haswell, Intel has pushed IPC to about 0.86 whereas AMD's best FX architecture boasts only 0.68 and Kaveri only 0.70.

 

For the 4790k:

0.86 instructions/clock * 4.4*10^9 clocks/second = 3,784,000,000 instructions per second on a single core assuming no downtime and assuming an even distribution of instructions used (at this point most of the general-purpose instructions are 1 cycle, some 2, a few 3, with AMD having some of Intel's 2s in its 3s).

 

0.68 instructions/clock *5.0*10^9 clocks/second = 3,400,000,000 instructions per second on FX 9590

 

Core counts are equal on both CPUs (hyperthreading is 2x logic resources per physical core, essentially becoming an 8-core at the expense of a couple dropped clock cycles (1 in 23 becomes a no-operation instruction, but with good look-ahead circuitry it can plan them to be in places where programs are stalling to synchronize threads anyway)).

 

By the virtue of what you see above, Intel beats AMD on efficiency of instructions which is why with lower clock rates it can still win in most applications. 

 

And multiplying the clock rate by the core count to get throughput is slightly off due to synchronization tending to hinder perfect performance, but it's close enough nowadays.

Software Engineer for Suncorp (Australia), Computer Tech Enthusiast, Miami University Graduate, Nerd

Link to comment
Share on other sites

Link to post
Share on other sites

Hmm if I can remember correctly, the first amd dual cores were "sharing" the clock speed so if one core is only being used its at 2.6GHz but if two are being used its at 1.3GHz and core2duo didnt have this. But any CPU these days doesnt share the clockspeed

Link to comment
Share on other sites

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×