Core i3 350M vs Core i3 370M

I want to compare these two processors: Intel Core i3 350M and Intel Core i3 370m. They are almost the same according to this link:

Compare Intel® Products

The 370M has a higher Clock Speed (2.4 GHz vs 2.26 GH), and a faster System Bus (4.8 GT/s vs 2.5 GT/s)

My question is that is the 370M significantly better in performance ? I wonder what System Bus is and does it really matter ?

 

You'll barely notice a difference, you could probably only tell through benchmarks. Just go with the cheaper one as the prices rise rapidly for a small increase in performance.

 

Whats the price difference?

 

usually the 350m is offered instead of 330m as stock.. 370m should be $20-30 more.. but really there isn't much difference..

 

With a small speed bump of just over 6% but a major (92%) bump in how fast the CPU communicates internally; what I would highly favor the 370M in is a system with only integrated graphics (as opposed to only a dedicated graphics chip and/or both integrated and dedicated (Optimus) technology).

In such a system (only integrated graphics), I would suspect that the 370M would prove a superior chip compared to the 350M in terms of multitasking and/or 'snap'.

At least, that is what the numbers would lead us to believe, right?

 

Most important question is: what do you do with your notebook?

If you are an average user that surfs the web, uses office programs, plays some music and video there is zero benefit in getting the faster CPU because a faster CPU is only faster during CPU intensive tasks.

If you are a user that runs video editing, math calculations or 3d design software the faster CPU may be worthwhile.

What's the price difference and what are the rest of the specs?

 

I have to disagree that a faster cpu is only faster during cpu intensive tasks and that there is zero benefit in normal or average use.

As my post just above hints at: a faster cpu can have internal optimizations that will make it faster at anything; not just when it is going 100%.

It doesn't matter what you do with your notebook (even if it is only office apps, websurfing and playing music) if you value the 'snap' your system has, a faster processor ensures that it has enough headroom to always respond to your input(s) in real time.

 

a Core i3 370m will not feel snappier than a Core i3 350m when doing Microsoft Word or watching a movie.

The CPU is simply not a bottleneck in those operations.

 

I'm not saying it will be.

But, if we were using MS Word on a notebook with only integrated grahics and a complete Win 7 install including MS Security Essentials (or your AV software of choice) and a few other utilities/apps that get loaded by each notebook manufacturer (think 'bloatware')... I'm willing to bet that the higher internal communications of the cpu/gpu et al will give the 370M an edge in 'snap'. Even if you choose to believe otherwise.

The CPU is always the bottleneck when it comes down to it - this is what makes the computer 'go' or not.

 

The CPU is a bottleneck when it's maxed out. Otherwise it's other components that are the bottleneck.

And like it or not, the difference between 350M and 370M is not noticeable during normal usage.

Higher internal communication with an integrated GPU can help for gaming and 3d rendering.

 

A CPU does not have to be maxed out to be a bottleneck in a system's performance.

Try MS Word on an Atom and on an i7 extreme cpu.

The CPU is what makes the computer 'go' - every other component gets faster with a faster CPU. Some components more than others and some get very little benefit with a faster CPU, but the computing 'experience' is always enhanced with a faster processor because it allows everything else to operate at the maximum.

Note: I'm not saying a 350m and a 370m will be noticeably different - I'm saying that they very well might be.

Gaming and 3D rendering are not all an integrated gpu is used for - simply drawing to the screen at almost twice the speed (theoretically...) will give the impression of more 'snap'.

This is why a long time ago I preferred Matrox video cards because their 2D capabilities were above even the best 'gaming' cards at that time.

I have the 350M (with an Optimus graphic subsystem) and although the clock speed is the same as my 'ancient' P8400, the 350M is easily much more 'snappy' - even with the same 7K500 HD installed.

How would you explain that difference - both CPU's are not pegged, yet there is an obvious difference between them (at the same clock speed).

 

You are comparing completely different systems (different memory speeds and size, different FSB speed, different GPUs) and then you attribute the difference to the CPUs.

In your second example you are also comparing different systems.

But I think you've made it clear enough now what you believe. Let's have some other people give their opinions now.

 

While I can somewhat understand your logic, your argument is completely flawed. There will be no discernible difference in the applications you suggested. Even a benchmark for these scenarios will yield insignificant deviation, which itself could be attributed to a margin of error.

While there is certainly some additional overhead created by using a dedicated GPU, this will only become apparent during applications where the GPU is the bottleneck. This assumes, however, that we are considering two equal graphics chips, which is an infeasible scenario, given various constraints for a single chip. You cannot simply integrate the current generation's most powerful graphics solution into the CPU. Throwing cost aside, greater power requires a more effective cooling solution, which becomes very problematic when you integrate two processing units that already need considerable heat dissipation. In GPU design, you must make a compromise between performance, power consumption, and thermal requirements, which makes any "apples to apples" comparison a fallacious endeavor.

As Phil earlier indicated, the component that acts as the bottleneck limits performance and the perception of "snappiness". For the applications you referenced, the CPU, quite simply, is not the bottleneck, and will not demonstrate any apparent effect on responsiveness. You suggest that the CPU "is always the bottleneck", but this is a false statement for many common computer tasks. For example, loading time for boot up and launching applications is dictated primarily by your storage medium. In this case, the drive containing the operating system is the bottleneck. A faster CPU is unable to deliver greater read performance for an HDD/SSD, and thus loading time is driven by storage technology, not the central processing unit. If you have ever used an SSD, you will quickly discover that the perception of "snappiness" is very much associated with non-CPU-intensive tasks, since it applies to general OS performance.

Sure, a faster CPU will offer greater performance, but only when it represents the limiting factor. Examples of scenarios where the CPU is typically the bottleneck includes rendering, encoding, and mathematical applications (i.e. calculating pi, determining prime numbers, or optimization techniques).

Your posts carry a sense of dogmatic faith that a CPU alone determines the responsiveness and capability of the system. This is true, but only when it represents the limiting factor. The integrated GPU, in opposition to a dedicated GPU, and the perception that the former necessarily will yield greater performance, is quite simply an irrelevant claim for reasons I have already discussed. Most of this post is merely an elaboration on what Phil has already stated. Try reading his posts with an open mind. You just might gain some insight in this debate

 

I think what he fails to realise is that the CPU has to "wait" for other components or processes a lot of the time before it can pass information to the RAM or other things like that. In this time frame, it doesn't make a tiny bit of difference how long the CPU take to do whatever calculation it's been tasked with provided it can get it done before the other thread it's waiting for allows it to do anything with it.

For example say one thread tells another to do a seperate calculation while at the same time it is busy with something which might take say 50ms, in that 50ms the original thread is completely busy and can't really do much so the second thread can kick its heals provided it meats the deadline (in a matter of speaking, in these such situations the second thread would usually be carried out as quickly as possible but the effect is the same).

 

Looking at the difference between the two, the only time you'd notice the difference between the 2 processors would be during something like 'Handbrake' or using an application to manipulate HD video. However, like others have said, you won't "notice" the better performance in day-to-day tasks, save if you're using something like Google Chrome with multiple tabs open at once (an application that creates multiple forks from a proc-tree).