Is i5/i7 just a marketing gimmick?

I have a Core2 duo 2.4 GHz. in the past which according to my understanding, is 2 cores, so virtually giving you the speed of a 4.8 GHz. processor (please correct me if I'm wrong)

Now I have an i5 2.6/2.93 GHz. Laptop, so I thought I would get blistering speeds since I have 4 cores/processor but this doesn't seem to be the case. They are not that different, even though the newer laptop has an i5 and a much better graphics chipset (ATI Radeon HD5650 11 GB)

I don't know how is the i7 also, suppose you buy a 2 GHz. laptop with an i7 processor, will it give you the speed of an 8 GHz. laptop since it has 8 cores? or is my understanding totally wrong about this because this cores BS is not getting into my head.

And why is it that on laptops we are stuck as the 2.8 GHz.. level for 2 years now..I thought that by this time we should have 4 GHz. laptop

please enlighten me and correct my understanding regarding this cores BS

 

Two things: 1. Clock Speed Frequency is not long the single-most important benchmark for CPU performance and 2. Cores are counted independently.

So first off clock speed. The reason we are stuck in 2.8 GHZ is that we need don't need massive loads of more GHZ to gain more performance. In fact the i5/i7 architecture has hardware improvements over the C2D that when both are clocked 2.8 GHZ clock for clock, the i5/i7 will still be way faster than a C2D. Clock speeds can no longer be compared apples-to-apples because of different hardware architecture. The clock speeds are just there to still distinguish the range of products being produced for this architecture and to increase performance within for a particular architecture. We no longer need to use clock speed to gain performance, it is all determinant on the new generation CPU architecture.

Now for cores. If you have four cores @ 2 GHZ, you don't get 8 GHZ. Instead you get 4x2 GHZ. Each core runs on 2 GHZ independently of each other, but they never are counted together as one. Therefore it is still considered 2 GHZ because we are still assessing clock speed per core, not total combined.

Therefore Core i5/i7 is not a gimmick. It is an improved architecture that runs with more energy efficiency, cooler temperatures and even more power from the C2D. Therefore even at 2.8 GHZ, it will still trump easily C2D's at 2.8 GHZ because of its architecture.

CPU's are no longer compared as apples-to-apples. The "gigahertz war" is over and it was over since 2004.

 

long story shoort--multiple cores are not going to make something multiple times faster. All cores operate at the same listed speed--a multicore processor will allow it to perform multiple tasks if the software is written to take advantage of it.

Think of it has a four lane highway with a speed limit of 50 miles per hour--that does not allow you to go 200 miles per hour. What it allows is for 4 cars to move at 50 miles per hour through the same stretch of land.

Now, the problem is that most programs do not take advantage of multiple cores so instead of getting four cars each moving through the same stretch of road at 50 miles per hour, such software still only allows one car through the stretch at a time (in the same lane) because the drivers "software" are not smart enough to switch lanes.

Then you also have the problem where even if you have four lanes, and four different cars' drivers are smart enough to use different lanes, they still have to come into this stretch from one lane and then merge at the other end back into one lane.

Multicore processors, therefore, tend to result in incremental boosts in performance instead of huge increases.

More Programs take advantage of multiple cores today than even a year ago (quite a few games), but again, the only thing that is sped up is the time through the processor--you still have bottlenecks before and after the processor, in the drive, on the system bus...etc

 

apparently it took me two minutes to type my explanation

 

The adding of core speed makes me laugh when I read a few ebay auctions... "Quad core which give you 42745.43GhhhZZ of RAW AWESOME speed"

The latest i7 cores are NOT quad core which is quite a nice marketing thing by Intel but the qm cores ARE quad so you have to be careful.

It matters not what cpu you have if the machine is bogged down with lots of running processes! A brand new HP machine can have nearly 110 running processes while a nicely tweaked laptop can be as low as 35/40. Also an SSD makes the world of difference as the slowest part of a laptop is usually the hard drive.

They reached a thermal limit a few years back so moved sideways to dual cores and then quad but are slowly working back towards higher and higher clock speeds as the manufacturing process improves.

P.s. There isn't an 8 core laptop... it's 4 cores and
What Is Hyper-Threading? [Technology Explained]

P.P.s To make it simple: 4 cores can undertake 4 tasks at the same time and Hyper-threading tries to handle another 4, ish, tasks.

 

Same here

 

Thank you very much for taking the time to type this.

One last thing, you said :

So what does this essentially mean to me as the end user? how does it benefit me? and where does this really help (in which situations)

I know that all the cores aren't really working during the boot up process even though I feel that my CPU usage is 100% because if 4 cores would be working during startup, I think that my Windows would have started up much earlier to a point where I can actually load a program such as a browser

I did disable unwanted services and programs from start up, but still, if we are saying 4 cores are working @ 2.93 GHz. (such as in my case) wouldn't it give a much higher performance than I would be getting now?

 

Ahh ok, that puts it in easy to understand terms...

So tell me...

Don't you think that an i5 @ 2.67/2.93 GHz. would perform better than an i7 @ 1.8/2.0 GHz?

I know the i7 has more cores, but since you said that not all apps take advantage of this feature, wouldn't it be better for me to stick to an i5 st a higher clock per core with fewer cores than an i7 with more cores but lower clock speed?

 

That is dependent of the program you are using. You see, the cores are only used when they are called up by an application or a program. Most programs these days will utilize two cores. Therefore running an internet browser, media program and productivity software will queue the OS to run the applications on one thread for one core and all OS stuff on another. Therefore, the performance levels are determinant on how the cores are being utilized by the programs. In most cases, dual cores are sufficient for most tasks these days.

However the i5/i7 architecture has better optimization of how these cores are used from C2D, hence why i5/i7 cores are more powerful than C2D's. In the end though, it is up to the applications you use and how they use the cores.

In a sense yes the i5 is better than the i7 (I presume you are talking about the i5-560M and the i7-820QM) if you are only going to utilize two cores (programs that only utilize two data threads for two cores). Once you have stuff running that needs all four cores, the i7 will beat the i5 because it is now processing four data threads at 1.8 GHZ opposed to two cores rendering two data threads at 2.67 GHZ at the same time.

 

1) it's a shame SSDs are still only available in laptops that cost an arm and a leg and they aren't exactly cheap here in Dubai where I live to buy separately. When I visit computer stores, every single laptop I check, has a 5.9 disk transfer performace in the Windows Performance thing, I know that scoring is BS but that comes to show us that the HDD is always our bottlneck

2) I thought that i3 = POS dual core , i5 = 4 cores (because I see 4 cores mentioned in the device manager), and i7= 8 cores. Please correct meh

 

It is because of HyperThreading Technology where one core can process one data thread and another in an emulated (logical) fashion. Therefore you get four cores but only two are dedicated (physical) cores and two that are virtual (emulated/logical). Only programs that are coded to work with HyperThreading can take advantage of it and can make a dual core run like a "quad" core. However in most applications, they won't take advantage of the logical or emulated cores, but only the dedicated (actual dual cores) cores.

 

Think of 2 batteries hooked together in parallel. Each battery is 12VDC. The program you run is a lightbulb. The two batteries will only feed the lightbulb with 12VDC. Hook up 3 more lightbulbs and put them in the electrical circuit. The lightbulbs will light up, but very dim and won`t last long, aka slow performance with your program.
Put in another 2 batteries in parallel with the other 2. Suddenly the 4 lightbulbs will light bright and last long. But they don`t light any more than the first experiment with 2 batteries and 1 lightbulb, since the 4 batteries still only output 12VDC to the 4 lightbulbs. They don`t feed each lightbulbs with 4x12VDC

That is the difference between a dual core (2 batteries) and a quad core (4 batteries)

 

Ahh so my i5 only has 2 cores and the i7 has 4 cores yeah?

Would you buy an i5 @ 2.67/2.93 GHz (higher clock speed per core) or would you choose an i7 @ 2 GHz. (more cores but lower clock speed?)

Seems to me the i5 would be the better options because:

1) it is cheaper

2) it has a higher clock per core so most apps that wont make use of the many cores, would at least benefit from a higher clock speed for the core they can use...am I right?

 

Yup the i5 is dual and the i7 is quad (at least the models mentioned in this thread).

The choice really depends on what you're doing. Some apps CAN make use of multiple cores but not typically the "basic" ones most computer users will use so no, most people do not need 4 cores.

I'll usually put it like this: most people who NEED more power KNOW they do and in quite a few cases, know exactly what they need (because of the fact that they need it).

 

Actually, i would say that for most normal users, the difference between an i5 and i7 won't be felt at all. For heavy multitaskers in general, the i7 is a better idea if you ask me. Even if you run multiple single threaded programs, there is nothing preventing windows from distributing the load between the 4 cores which it usually does when needed.

Then there are people like me who run specialized software that was made to use as many cores as possible who definitely benefit from quads. However, as Melody said, we know we need quad for lower crunch time.

 

Alright that clears it up! thanks a lot all

 

True to the above with the exception of "trumps" which is hardly an apt. description.

The most important variable there is. This cannot be overstated.

 

This is just an example of multi-threaded apps that will make use of all the cores of a CPU... you have dual's and quads here...

Spoiler

Even though those aren't the latest CPU's, and aren't i5/i7's, the same principles apply. If you have a 2GHz i7 and a 2.5GHz i5 with an app that will make full use of all individual cores, the i7 will whip the i5 even with a lower clock speed, because it's able to do more work in parallel.

 

most questions have already been answered, but I could probably add something as well:

- the reason we see our CPUs to top at about 3GHz or so is that going higher gives physical difficulties regarding the materials/elements that are in use today for making chips. If at some point technology moves to say using light instead of electricity in the chips, then things will be different.

due to those limitations people turned into putting more elements (transistors mainly) in the chip for increasing it's performance, and not higher clock. If they could, they'd just increase the clock as it would have been much easier, though this works somewhat fine up to about 3GHz.

people that like to overclock have probably noticed the step increase in voltages become bigger to the next multiplier once the CPU gets up in GHz rather than in the lower region (say at about 2Ghz or so), i.e. if you want to overclock from 2.0 GHz to 2.2 GHz you may need only 0.01 Volts increase, but if you go from say 3.0 to 3.2 GHz then you might as well need like 0.2V more.

 

Not only that, but the heat generated increases linearly with the clock speed, but it's more like to the power of two when voltage increases if i recall correctly. Intel tried going for very high clock speeds with netburst and look how that ended...

Anyways aren't we getting rather close to the lithography size limit for CMOS?

 

I believe it's still got another two good nodes in it. Immersion lithography will take Intel down to 11NM.

 

I know we aren't at the limit yet. I can't remember if it's around 6nm or 0.6nm though.

 

Yea, that's about it. Immersion lithography will end at 11NM, then Extreme Ultraviolet Lithography will take it down to about the sub-6NM level. Then someone better have a trick up their sleeve, or be working with carbon nanotubes/atom transistors.

 

Or carbon based polymer transistors which is being researched right now. Not talking about purely carbon structures which are nanotubes and graphene, but polymers with a carbon based backbone and functional groups to give it the desired properties. Of course, that kind of research is still at least a couple of years away from producing something viable for semiconductor manufacturing. It already gave some interesting products though like OLEDs. I was attending a polymer conference about a year and a half ago and that was really a hot topic at the time, making polymers with semi conductor properties and things like OLEDs.

If there is one thing i'm certain of is that R&D depts and universities have been working on this problem for a while, i mean they were seeing it coming after all.