So a core 2 duo has 2 processors... - HotUKDeals
We use cookie files to improve site functionality and personalisation. By continuing to use HUKD, you accept our cookie and privacy policy.
Get the HUKD app free at Google Play

Search Error

An error occurred when searching, please try again!

Login / Sign UpSubmit

So a core 2 duo has 2 processors...

Benjimoron Avatar
6y, 10m agoPosted 6 years, 10 months ago
and a quad core has 4 processors. Does that mean that a quad core is twice as fast a a core 2 duo? Do you just multiply the speed, ie a quad core at 2.5Ghz is the equivalent of a single core running at 10Ghz?

So if you see a pc listed as being dual core at 2.5Ghz does that mean that it's running stuff at 5Ghz effectively?

I've never been quite sure how these things work.

Also how much better is an i7? Are they revolutionary? Or just a tad better and possibly overpriced as they're new?

Thanks!
Benjimoron Avatar
6y, 10m agoPosted 6 years, 10 months ago
Options

All Comments

(26) Jump to unreadPost a comment
Comments/page:
1 Like #1
not necessarily. If a program or game takes advantage of 2/3/4 cores then yes it willa percentage faster, but not ie 2x for 2 cores or 4x for 4 cores. But not that many programs make use of this yet other then specialised software, many new games are now, but still many only use 2 cores.
#2
So even if it's using 2 cores it isn't twice as fast? Why is that?
banned#3
more cores mean it will be better at multi tasking.

it's clock speed that is probably the most major factor with current applications/games but there is a limit to how high you can get the clock speed so they started putting more cores on the same die.

then there is cache to think about too. a 3ghz dual core celeron will probably be slower than a 2ghz core 2 duo because of the cache is so small on celeron.


most people who use their computers for internet and emails dont need to worry about any of this though
#4
Cool, thanks for the info guys. I'm reading up on other sites and it appears that the reason is that 2 cores share the same resources so in effect they can't both be running at full speed as other parts slow the process down.
banned 1 Like #5
Benjimoron
Cool, thanks for the info guys. I'm reading up on other sites and it appears that the reason is that 2 cores share the same resources so in effect they can't both be running at full speed as other parts slow the process down.


regarding the i7 i think the 9 series uses a new architecture making it a bigger jump than the 8 series which is more common.

there is an i9 coming out soon so hopefully i7 prices will go down.

i still have a q6600 core 2 quad and it is fantastic and i probably wont upgrade for a few years still. i wouldnt get top of the range with computers. next week they wont be so you end up paying a premium for nothing :-D
1 Like #6
Benjimoron;7702301
Cool, thanks for the info guys. I'm reading up on other sites and it appears that the reason is that 2 cores share the same resources so in effect they can't both be running at full speed as other parts slow the process down.

As you're probably gathering, it's not straightforward - you can't multiply the clockspeed as some sellers do because if you're running a single program that can only run on one processor, a quad core will be no faster than a single core (in general).

Imagine the processor as an assembly line in factory, you have a conveyor belt with different robots at each section. Parts are fed onto the start of the conveyor belt, the robots transform them as needed then they exit out the end - the clock speed is how fast the conveyor belt is moving. For a quad core processor you have four independent assembly lines beside each other - if the product you're making can be split over the four assembly lines, then it can be done four times as fast. Or if you have four separate tasks to run, again it's four times as fast. But if you're only working on one task which can't be split, then you can only run one assembly line and the other three just sit idle.

Core i7's are a little more complex, they are a native quad core (or six core for the upcoming Gulftown i7) but they've a couple of tricks up their sleeves in the form of hyperthreading and turbo mode. Hyperthreading makes it look like each processor is actually two processors so a Core i7 appears as eight processors as each processor is hyperthreaded. The idea of hyperthreading is that for certain tasks the processor may not be using everything and therefore you can stuff bits of another program in to make it run more than one tasks at a time.

Going back to the assembly line analogy, imagine a part is moving up the conveyor belt through the various robots but the next bit isn't ready yet - the robots at the start of the assembly line are idle and therefore this is inefficient. So what you can do is put a different part in to go through the assembly line in between the other task so that all the robots are working as much as possilble. Hyperthreading isn't as good as having two processors but for certain tasks it can improve performance especially if you've a multithreaded program already.

Turbo mode is the ability for the processor to overclock itself if all its cores aren't in use. It has a total power/heat usage which is designed on all four cores running at full power, however if only the first two cores are running then the processor is going to be below its power/heat threshold and therefore can use this potential to go a bit faster. If only a single core is active then it can go faster again countering the normal disadvantage of having a quad core processor when the program is only running on one or two of the cores.

Unfortunately Intel have made a total and utter mess of the naming system so you need to double check what the processor actually is - some i7's are dual core rather than quad as are i5's.

John
1 Like #7
bykergrove;7702374

there is an i9 coming out soon so hopefully i7 prices will go down.

Gulftown (six cores, hyperthreading) is going to come under the i7 name rather than i9 just to ensure Intel's line up is as confusing as possible.

John
#8
Rep 4 all, especially John!
#9
I'm surprised Hyper Threading has made a comeback given it was not much use and used more power before they ditched it.

Saying that the laptop I had with a 3.06ghz p4 Hyper Threading cpu also used desktop DDR
#10
dontasciime;7703264
I'm surprised Hyper Threading has made a comeback given it was not much use and used more power before they ditched it.

Neither of those statements are particularly accurate, hyperthreading on the Pentium4 did give a noticeable performance increase (up to 20%) as well as better multitasking. The HT cost was very little as the implementation required less than 5% of the total die and only used approximately 5% more power. The main downside to the technology at the time was that SMT was virtually unheard of in the desktop market (excluding workstations) as this was before multicore processors, some software had problems with SMT. The energy penalties at the time were exaggerated by some other companies at the time, the Pentium4 was a power hungry processor without hyperthreading anyway.

In the current market SMT is far more standard so given the low die/energy cost and an architecture compatible with hyperthreading it makes sense for it to return. The HT enabled i7's outperform the non-HT i5 processrs by quite a reasonable margin depending on the software.

John
#11
Sorry: It was accurate for me at that time and the other users who also noted that there were no programs to utilise it, so no benefit was visible and other programs suffered co's of it's existence.

It's also believed to have used as much as 46% more power. That will come under your exaggerated (companies) claim though.
#12
As I have not bothered much with CPU after q6600 cannot say I have seen the benchmarks you speak of for latest cpu Intel have to offer.

Is the i7 not in nature better than the i5 anyways ? if the higher number naming means anything that is.

Why did they not compare an i7 with ht enabled with an ht disabled i7 though? Unless i5 is supposed to better than i7
#13
dontasciime;7703590
Sorry: It was accurate for me at that time and the other users who also noted that there were no programs to utilise it, so no benefit was visible and other programs suffered co's of it's existence.

All you needed to do was run more than one program at a time and you could utilise the feature, in general smoother multitasking performance was the most noticeable for home users.

It's also believed to have used as much as 46% more power. That will come under your exaggerated (companies) claim though.

Which source did you get that from exactly? It wouldn't be possible for the processor to draw that much more power as it would hit its thermal shutdown limits. In actual testing, a 6.5% increase was noted with HT enabled:

http://www.anandtech.com/showdoc.aspx?i=1746&p=8

The exaggerated claims were nothing like that, they were theoretical against dual core utilisation which smaller, simpler architectures could implementa at the time.

John
#14
dontasciime;7703713
Is the i7 not in nature better than the i5 anyways ? if the higher number naming means anything that is.

There are several different i7's across different sockets however there are socket 1156 i5's and i7's which are the same core (better turbo mode, onboard PCI-E controller etc.) aside from the i5's having HT enabled. When Intel released the first HT enabled P4's, only the 3.06Ghz had the feature enabled but the rest of the range had the hardware but it was disabled to encourage users to purchase the top of the range processor (it could be re-enabled with a socket adapter. Therefore it is possible to compare two identical processors bar the hyperthreading.

Why did they not compare an i7 with ht enabled with an ht disabled i7 though? Unless i5 is supposed to better than i7

Because there are i5's with an identical core to the i7 therefore it's the closest match.

John
#15
Johnmcl7
All you needed to do was run more than one program at a time and you could utilise the feature, in general smoother multitasking performance was the most noticeable for home users.

Which source did you get that from exactly? It wouldn't be possible for the processor to draw that much more power as it would hit its thermal shutdown limits. In actual testing, a 6.5% increase was noted with HT enabled:

http://www.anandtech.com/showdoc.aspx?i=1746&p=8

John


lol:

I did run more than 1 program at a time I had loads and just a few then 1 then 2 did a passmark and burn in no difference at all just like when you get a dual core and nothing is optimised for it it makes little diff.


eg one pipe with little yellow circles trying to go through pipe with some purple ones, had another pipe side on with 1st pipe with purple circles going through whilst yeloow ones went through first pipe smoothly. Intel ad. lol

SMT uses 46% more power according to AMT http://www.incrysis.com/forums/viewtopic.php?id=10505

http://en.wikipedia.org/wiki/Hyper-threading
#16
http://img11.imageshack.us/img11/5836/i5s.png



http://img39.imageshack.us/img39/2031/i7s.png

The above 2 images are nothing to do with anything much other than a comparison for me to see on 1 page eg the cache levels of each CPU i5/i7 etc
#17
My i7 shows 8 Processors, :thinking:


http://i50.tinypic.com/sdjjm1.png
#18
dontasciime;7703954
I did run more than 1 program at a time I had loads and just a few then 1 then 2 did a passmark and burn in no difference at all just like when you get a dual core and nothing is optimised for it it makes little diff.
eg one pipe with little yellow circles trying to go through pipe with some purple ones, had another pipe side on with 1st pipe with purple circles going through whilst yeloow ones went through first pipe smoothly. Intel ad. lol


I don't really see what relevence this has to anything - there are benchmarks which show hyperthreading did offer a performance improvement and users did notice increase multitasking performance. Just because you didn't doesn't meant it didn't exist.

SMT uses 46% more power according to AMT http://www.incrysis.com/forums/viewtopic.php?id=10505

You're misreading the information - Arm were comparing the hyperthreading implementation against a theoretical multicore x86 design (which didn't exist and wouldn't for a while yet), they were not comparing a Pentium4 with hyperthreading against one without. In reality, the power usage was approximately 5% higher as you'd expect given the slight increase in die. The claim that Intel dropped hyperthreading for the first Core processors due to the higher power usage is incorrect which is why there's no citation available.

As it currently stands, hyperthreading only adds a very slight cost to the processor and at worst on a quad core setup with SMT enabled is no slower than a non-HT processor. For those that don't want the benefits then there's the choice of the i5 although in reality if multithreaded performance isn't that important one of the dual core i3/i5's would be fine.

John
#19
bargain surfer;7704087
My i7 shows 8 Processors, :thinking:


Yes, that is normal because there are four processors each hyperthreaded which means the operating system sees four 'real' processors and four virtual ones. A hyperthreaded dual core processor will show as four cores and a single core processor with hyperthreading will show as two.
John
#20
Johnmcl7
I don't really see what relevence this has to anything - there are benchmarks which show hyperthreading did offer a performance improvement and users did notice increase multitasking performance. Just because you didn't doesn't meant it didn't exist.

You're misreading the information - Arm were comparing the hyperthreading implementation against a theoretical multicore x86 design (which didn't exist and wouldn't for a while yet), they were not comparing a Pentium4 with hyperthreading against one without. In reality, the power usage was approximately 5% higher as you'd expect given the slight increase in die.

John


In 2006 hyper-threading was criticised for being energy-inefficient. For example, specialist low-power CPU design company ARM has stated SMT can use up to 46% more power than dual core designs. Dual core processors are different than Dual CPU. Furthermore, they claim SMT increases cache thrashing by 42%, whereas dual core results in a 37% decrease.[4] These considerations are claimed to be the reason Intel dropped SMT from the Core 2

Just because you didn't doesn't meant it didn't exist


Also because I saw no benefit means I did not. I had numerous friends with same CPU they saw no benefit so we saw no benefit.

I'm not misreading anything I can see what it says in comparison to dual cores which should /were out at same time ie 2006


If you saw a benefit when you run it then great I did not (adn others I knew) when we tested it showed no benefit. When I had an Ht enabled Cpu/bios/operating system I saw no benefit.
#21
dontasciime;7704138
In 2006 hyper-threading was criticised for being energy-inefficient. For example, specialist low-power CPU design company ARM has stated SMT can use up to 46% more power than dual core designs. Dual core processors are different than Dual CPU. Furthermore, they claim SMT increases cache thrashing by 42%, whereas dual core results in a 37% decrease.[4] These considerations are claimed to be the reason Intel dropped SMT from the Core 2

Yes, I see you can quote information from Wikipedia so there's no need to do so again - as I said, the ARM claim was entirely theoretical as there were no multicore x86 processors available at the time nor would they be viable with the manufacturing technology. The 46% figure has nothing to do with the additional power consumption of the Pentium4 processor with HT enabled which was in reality a very small amount when comparing a Pentium4 with and without HT enabled. If my car could float along the ground then it would be considerably more power efficient but the technology doesn't exist so the fact it could be more power efficient doesn't matter.

Also because I saw no benefit means I did not. I had numerous friends with same CPU they saw no benefit so we saw no benefit.

I honestly what point you're trying to make, professional reviewers running a large number of benchmarks with actual evidence found the HT technology offered benefits so I'll go by them (as most people do) and my own machines and not what you and your mates thought. The fact of the matter is that hyperthreading offers a performance benefit particularly with code optimised for mulithreading with a very low increase in heat/power consumption - whether you choose to accept that or not is entirely irrelevent.

I'm not misreading anything I can see what it says in comparison to dual cores which should /were out at same time ie 2006

Ok, which x86 desktop dual core processor was available at the release of the Pentium4 3.06B then? The answer is none, the Pentium D's came out some time later and on a small manufacturing process allowing Intel to produce a dual core processor as did AMD. Despite what Arm claim, hyperthreading was added for just 5% die cost which simply wouldn't have been the case for a dual core processor (that would have needed 100% more die space) which wouldn't been economically viable.

John
#22
lol at you not wanting to believe I can tell which one of my 14 odd computers running video encoding programs and seti was better than another 1 I prefer to judge on real world perfomance and nto some paid/baised review . Quoting wiki yeah I know you asked where had those words So I posted it.

When I had my p4 there were dual core cpu's available. http://www.tomshardware.com/news/paxville-launch,1522.html I could buy them for a desktop pc.

http://www.legitreviews.com/article/184/1/

anyways I have had enough. You can believe what you want.
#23
I wasn't referring to when you bought your Pentium4, I was referring to when the technology was introduced which was when there weren't any dual core processors and is one of the main reasons for its existence - it eased the transition to dual core as it forced SMT to be brought to the desktop. The 130nm Pentium 4 3.06Ghz was an extremely expensive and power hungry processor which wouldn't have been at all viable to produce in a dual core form.

It's not about 'I can believe what I want' - I have backed up everything I've said with repeatable and quantifiable evidence whereas all your proof of hyperthreading being 'useless' is anecdotal evidence based on what you and your mates perceived. The fact remains that hyperthreading adds performance in multithreaded or multitasking use with virtually no cost in die-size, power or heat as has been verified by many professionals. It's utter nonsense to claim that these people are 'biased' especially when AT in particular were more pro-AMD at the time of the late P4's and frankly you're a bit crazy if you think anyone would believe your groundless claims against so called 'paid/biased' reviewers. The fact that hyperthreading in the i7's continue to show a performance increase over the Core i5's shows your argument is nonsense..

John
#24
Sorry, I seem to have taken a disproportionate amount of your time.

Anyway, next question, there are billions on transistors on a chip right? Who knows? Who actually counts them? How do they make them all, surely it would take millions of years unless they grew them organically?
#25
Benjimoron;7706733
Sorry, I seem to have taken a disproportionate amount of your time.


Don't worry, I was just trying to make it clear to anyone reading that there are benefits to hyperthreading - anyone considering a processor with the technology will hopefully read up first though to see if it suits their usage to be worth the money.

Anyway, next question, there are billions on transistors on a chip right? Who knows? Who actually counts them? How do they make them all, surely it would take millions of years unless they grew them organically?


On the current core i7's they're just a bit short of a billion transistors, current graphic card chips are well over a billion transistors though. The processor layout has to be designed in software first before being tapered out for manufacturing which is how they know how many transistors make up the final design. The processors are produced on large wafers (300mm for Intel at the moment), Intel recently released a PDF on the process here:

http://d.yimg.com/kq/groups/9109916/1864001592/name/Intel_Chip_Article.pdf

Or a video from AMD here: http://www.youtube.com/watch?v=-GQmtITMdas

After the processor design is etched into the silicon, the processors are tested to make sure they work correctly and then afterwards they'll be tested for their speed and power usage. Many processors are actually born equal but the results of their testing will determine which processor they will be sold as - the best processors in the batch will be sold as the highest clocked processors and then they work their way down. Even processors with defects can still be used which is where Celerons come from, the memory cache takes up a large amount of die space so if there's any faults in this area they can deactivate that portion of the cache and sell the processor as a Celeron.

The size of the manufacturing process governs how many transistors you can pack into the same area, this is how we now have quad core processors where there was previously only a single processor. Going forward, it won't be long until integrated graphics cards are actually part of the same silicon as the processor, Intel have recently moved the graphics card onto the processor but as a separate part for now.

By moving to a smaller manufacturing process (known as a die shrink) for an existing design you can pack more processors onto the same wafer which means the cost per processor goes down or you can increase the transistor count to add more in the same amount of space such as extra processors. As an example, the original PS3 had a 90nm Cell processor, the newer 40GB model changed to a smaller 65nm version - exactly the same processor but half the size making it considerably cheaper to produce and also lowering the power requirements. The PS3 slim uses an even smaller 45nm processor allowing the cost to be reduced again and the power/heat reduction reduced enough to make the PS3 itself smaller.

The ATI Rage 128 was built on the 250nm process with a transistor count of around 8 million, the current 5870 uses the 40nm process (the smallest ATI use right now) and has a staggering 2.15 billion transistors.

John
banned#26
[SIZE="7"]JUST WAIT UNTIL SKYNET!!![/SIZE]


then you ladies can fight over processors :-D

Post a Comment

You don't need an account to leave a comment. Just enter your email address. We'll keep it private.

...OR log in with your social account

...OR comment using your social account

Thanks for your comment! Keep it up!
We just need to have a quick look and it will be live soon.
The community is happy to hear your opinion! Keep contributing!