Sign in with
Sign up | Sign in
Your question

I7 processors

Last response: in Overclocking
Share
May 2, 2012 5:09:23 AM

Hi, im not very sure about this processor thing. Quad core for i7 series.

Intel® Core™ i7-980 Processor (12M Cache, 3.33 GHz, 4.8 GT/s Intel® QPI)

Intel® Core™ i7-620UE Processor (4M Cache, 1.06 GHz)

What is the difference between them? Cache, this 4.8 GT/s and etc. I dont understand the profound explanation of it. Can anyone elaborate them in layman term?

Another thing is they run with Quad core, so does that means for the i7-980, it's running at 3.33Ghz per core, and 12Ghz for all 4 cores?

More about : processors

a b K Overclocking
May 2, 2012 6:13:27 PM

Hmm. My first thought is that you are bored and trolling around the forum. Forgive me if you are not. You are comparing a chevette to a rocket. Both of these socket types have completely different architectures.

Just keeping with the basics that you are asking about,t he i7 980 has 6 cores and 12 threads. Simply put, each of the 6 cores can do the job of 2 cores at the expense of a bit more voltage and a bit more heat generated essentially giving you 12 threads. Any hardware monitoring program will show each of these threads as a separate cpu core. Each of the 12 threads runs at 3.33GHz independantly. 4.8 GT/s is the qpi(quick path interconnect). It is essentially the system bus speed(the speed at which communication between the hardware occurs). This translates into approx 2400MHz. This is quoted from wikepedia 'A CPU cache is a cache used by the central processing unit of a computer to reduce the average time to access memory. The cache is a smaller, faster memory which stores copies of the data from the most frequently used main memory locations. As long as most memory accesses are cached memory locations, the average latency of memory accesses will be closer to the cache latency than to the latency of main memory.' Essentially, more cache = faster operations. Aside from the obvious number of cores and the frequency, everything about the i7 620ue is slower than the i7 980 by a lot.
m
0
l
a b à CPUs
May 2, 2012 6:33:55 PM

The fact that they are both core-i7 means very little. i7 is just a marketing brand, not a specification. As our friend alrobichaud stated, the 980 is overall much faster and more appropriate for heavy work. The 620UE is targeted for low power devices and uses a different slot (it's cannot be used on a desktop computer and comes soldered to the motherboard)

(One correction though, i think you mixed hyper-threading with turbo-boost, HT just allows for more granular work, which is faster on threaded applications. The only increase in power draw comes from the heavier workload, not HT itself, and voltages are not touched. It also does not enable a core to do the work of two, just to do the work of 2 x 1/2, which is more efficient when the workload is too low for a singe thread.)
m
0
l
Related resources
a b K Overclocking
May 2, 2012 11:23:54 PM

Hyperthreading seems to be a source for debate as to whether or not it generates more heat and requires more cpu Vcore. My initial answer was the same as yours but I have gotten into some debates in this forum on the topic and the majority of members that I have spoken with think that hypethreading does require a slightly higher Vcore and generates more heat. I should have been more specific. I am talking about an increase of maybe 4 degrees and the difference between 1.47V vs 1.485V to get an i7 965 prime95 stable at 4.4GHz. I was not talking about turbo boost but increasing by a muliplier or two will definitley generate more heat and require more voltage than I am talking about. The way I understood how hyperthreading works is that two logical cores are created from a single core which allow a single core to perform mulitple tasks at once. I am a bit confused by your statement 'It also does not enable a core to do the work of two, just to do the work of 2 x 1/2, which is more efficient when the workload is too low for a singe thread.)' Please elaborate.

http://en.wikipedia.org/wiki/Hyper-threading

m
0
l
May 7, 2012 12:26:44 AM

So to say that i7-620UE is actually for notebooks and its not being sold individually outside? I haven seen how hyper-threading boost the system efficiency yet. Had 2 desktops but with a different specs, so cant really compare them both. Does it affect the processing speed? So what kind of application is counted as threading application?

Okay, deducing from what you said, am i correct to say that 1 core has 2 threads inside, 1 single core is able to produce the workload of 2 cores since there are 2 threads in 1 single core.

So the CPU Cache is actually to process data that are being used or loaded frequently. Does it resets itself if the user shuts the PC down, and it will have to regenerate again if the user reopens the application?
m
0
l
May 7, 2012 12:30:58 AM

Another thing is, what you mentioned about the cores running at 3.33GHz, do you mean the core or the thread itself. If it's the thread, wouldnt the core be at 6.66GHz? And am i right to say that the cores are running at 3.33GHz per core and concurrently to share the workload, instead of like accumulating them to be at 18GHz since there are a total of 6 cores? Which may sound a bit.. ridiculous though.. heheh
m
0
l
a b K Overclocking
May 7, 2012 10:34:52 AM

The i7-620ue is a mobile processor for notebooks. A single core can work on two tasks at the same time making it appear as though there are two cores working on the job. There is still only one processor core but it is doing two jobs at the exact same time making the overall process run faster than if only a single core was available. Picture it like being a clone. Using 3.33GHz as an example, a single core will run at 3.33GHz and if it were divided into 2 threads each thread will run at 3.33GHz. The combined total is not 6.66GHz. Your statement is correct

'And am i right to say that the cores are running at 3.33GHz per core and concurrently to share the workload, instead of like accumulating them to be at 18GHz since there are a total of 6 cores? Which may sound a bit.. ridiculous though.. heheh'

m
0
l
a b à CPUs
July 19, 2012 12:35:58 PM

alrobichaud said:
I am a bit confused by your statement 'It also does not enable a core to do the work of two, just to do the work of 2 x 1/2, which is more efficient when the workload is too low for a singe thread.)' Please elaborate.

http://en.wikipedia.org/wiki/Hyper-threading


It comes from how a processor works. Among the most important blocks on a processor are the execution units and scheduler. The execution unit is repsonible for executing all the logical operations. The scheduler is responsible for arranging the logical operations in an optimal order. So, for example, if the code running has memory read instrucions and some other task before it, the scheduler may request the memory read before, since it knows it will take a few cycles to get a response.

What usually happens is that execution resources end up being wasted on account of the limited maneuverability the scheduler has (not everything can be re-ordered to keep the EUs busy). To mitigate this effect, HT turns the CPU into two different logical units.

The benefit of HT is that, by interfacing with the CPU on two logical units, the OS's task scheduler can send it more tasks simultaneously, which in turn will provide the CPU with more execution tasks (and consequently, maneuverability). That way, the CPU is kept busier.

What I meant with my statement was that HT does not grant a processor any additional execution resources, so it will never be able to execute anything faster. What it does is provide a means for better use of the resources available, so if a certain task caused the CPU to idle and stall the OS for a while, you will perceive it as faster since it will be able to do something else while it waits.

Voltage and temperature rises may occur, not because of HT itself, but because more extensive use of CPU time it causes. That additional workload naturally calls for more power, and, since the voltages are usually set dinamically, that too will rise.
m
0
l
a b à CPUs
July 19, 2012 1:27:50 PM

jesperloh said:
Okay, deducing from what you said, am i correct to say that 1 core has 2 threads inside, 1 single core is able to produce the workload of 2 cores since there are 2 threads in 1 single core.

So the CPU Cache is actually to process data that are being used or loaded frequently. Does it resets itself if the user shuts the PC down, and it will have to regenerate again if the user reopens the application?


1. No. A "thread" is a software pipeline, it is the way the operational system organizes the tasks it sends to the CPU. By having two logical units on a CPU core, the OS can send tasks through two "channels", thus giving more tasks to the CPU. The CPU itself cannot execute more tasks by having HT, but it can work more and wait less by receiving more tasks.

2. The CPU Cache is an ultra-fast, volatile memory. It wipes clean on a reset, but that doesn't affect performance significantly because it is constantly erasing and writing when the CPU is working. The CPU Cache is different from the system memory in that it is not meant to store application data, only execution data. That means it keeps close to the processor the information needed to run the current execution operations on the scheduler (registers, tiny bits of data, not entire files). Reseting the computer does not affect the CPU Cache performance.

jesperloh said:
Another thing is, what you mentioned about the cores running at 3.33GHz, do you mean the core or the thread itself. If it's the thread, wouldnt the core be at 6.66GHz? And am i right to say that the cores are running at 3.33GHz per core and concurrently to share the workload, instead of like accumulating them to be at 18GHz since there are a total of 6 cores? Which may sound a bit.. ridiculous though.. heheh


alrobichaud said:
The i7-620ue is a mobile processor for notebooks. A single core can work on two tasks at the same time making it appear as though there are two cores working on the job. There is still only one processor core but it is doing two jobs at the exact same time making the overall process run faster than if only a single core was available. Picture it like being a clone. Using 3.33GHz as an example, a single core will run at 3.33GHz and if it were divided into 2 threads each thread will run at 3.33GHz. The combined total is not 6.66GHz. Your statement is correct

'And am i right to say that the cores are running at 3.33GHz per core and concurrently to share the workload, instead of like accumulating them to be at 18GHz since there are a total of 6 cores? Which may sound a bit.. ridiculous though.. heheh'


Not really. A thread has no frequency, the CPU has. Picture a thread as a pile of work the applications dump on the OS, which in turn dumps it on the CPU. By having more then one pile, the CPU can better chose the work it will do next, so that it uses most of its resources. An extra core would be closer to what you guys are describing, since that would actually give the CPU more execution hosepower.

Still, neither HT or an extra core would have an impact similar to 3.33GHz becoming 6.66GHz, because a lot of the work the CPU does is streamlined, and cannot be divided into many threads. So 2x3.33 will not equal 6.66 because the 2x3.33 will not be used 100%, while the 6.66 might (I am talking about another core here, not HT).
m
0
l
a c 185 à CPUs
a c 150 K Overclocking
July 19, 2012 3:24:31 PM

Looks at thread date.
*sighs*
m
0
l
!