Compairing 2 CPUs Help Please

Hi, I am buyign my girl friend a laptop for christmas and the only difference between 2 laptops is the CPU. Here are the 2 CPUs, is there a big difference in speed?

AMD Athlon II Dual-Core P340 2.2G
vs
AMD Phenom II Quad-Core P920 1.6GHz

 

I say get the dual core.

Based on what I've seen so far and without taking the time to look at actual numbers other than cores and clocks, here's what I think:

The dual core will be faster in single-threaded stuff. This means stuff like Flash content (read: Farmville and YouTube) and anything else that mainly feeds on one thread. She'll probably get a little more battery life over a quad core.

The quad core won't slow down quite as much when doing multitasking. So if she's browsing the web, playing music, and editing photos, things will go faster. Modern multithreaded programs and Windows overall will be faster. However she'll get 10-20 minutes less battery life perhaps.

If she does a lot of photo/video editing, or is a frequent multitasker, get the quad core. For more general usage like surfing the web, stick with the higher-clocked dual core.

EDIT: Here's two links that might help with some technical comparisons:
http://www.notebookcheck.net/AMD-Athlon-II-P340-Notebook-Processor.37883.0.html
http://www.notebookcheck.net/AMD-Phenom-II-X4-P920-Notebook-Processor.30326.0.html

 

Look up the CPUs here:

PassMark Intel vs AMD CPU Benchmarks - High End

 

are they of the same architecture ? Otherwise, it is difficult to compare just the clock speed. A 3.2GHz P4 is no where near what a C2D of a much slower clock.

I would go for the Quad based on those benchmark links.

 

go for the cheapest of the two. A laptop with a CPU with two cores is more than enough for most people.

 

True statement.

 

Both processors aren't listed on there.

I would say find out what your girlfriend uses it for primarily. Although don't go cheap on your girlfriend.

 

Wrong.
Go cheap on your girlfriend.
She wouldn't know which is which anyway. The laptop is the last thing she cares about.
She wouldn't care as long as it runs.

 

Who knows? She could be on this forum with 14k posts and you wouldn't even know it.

If she renders or does very CPU intensive things, the Phenom II Quad core will help out. Find out what she does most with her laptop.

 

Indeed. Use the saved money to buy her some flower, perfume etc.

 

I agree with Tsunade_Hime in this case. You really need to know what your girlfriend likes to do with her computer. Although, given the 2 listed, I'd be a little more likely to go for the dual-core, as the quad-core is rather slow for normal use (since it lacks Intel's Turbo-Boost to speed it up when all cores are not being used). From my own experience, I upgraded my girl's (desktop) from an older Pentium dual-core to an i5-750 (which was overkill, I know). Since she's a gamer (plays Dragon Age), she was very, very, appreciative...

 

It's a good point, but I've played around with a couple notebooks with the Phenom II Quad core and they are just as "fast" as the i7 notebooks in terms of responsiveness for normal use except phenomenally cheaper. I think we had a DV7 with that quad core, 6 GB GDDR3, 640 GB HDD, and a 1GB 5650 for 650 after 100 MIR. In the age of more multi-threaded applications I would go for a quad core if the price difference is minimal.

Heh I built a new computer a few months back for my mom who was running on a AMD 3000+ HD2400 PRO Dell C521 with 1 GB RAM to a Q8200 Core 2 Quad, 6 GB DDR2 and a 9800GT graphics card. She is very happy with the upgrade as my brother who games on it also.

 

Without knowing her usage and assuming it will be 'average', I think that on the verge of 2011 a dual core is a waste of time.

As long as the notebooks are as identical as you say (except for cpu's), worth keeping for at least 2 or 3 years and her computing needs are modest, then the quadcore will be the better buy. For normal usage, it may even run cooler than the higher clocked dual core.

Don't forget that concurrent processing of threads will be increasingly important in future O/S, apps and drivers and the AMD chips you are comparing do not offer anything like Intel's hyperthreading.

Very quickly, today's dual cores will seem as antiquated and slow as trying to run a modern O/S + apps on a single core system (and will be just as frustrating too - even with an SSD inside. ).

Good luck.

 

Tell me how that quad performs on singlethreaded benchmarks and browsers compared to that dual core. How many multithreaded things will she be doing? Assuming "average" usage two cores are plenty. Clockspeedwise, 2x2.2GHz means an application will have to be using at least three full threads before it sees a boost from a 4x1.6GHz CPU. Even two-threaded applications will be faster on the dual core.

 

Your logic is quite off.

If the O/S is allowed at least one core for its housekeeping (and it needs it), and the A/V is allowed one for itself too (and it needs it) and she is listening to music while browsing the net (and they'll need it), then it isn't hard to see that the quad-cores are much better suited to a typical modern usage not only today but also more so in the future. Especially as mentioned with no HT capabilities in the AMD offerings to help a little with this typical usage.

Doesn't matter if the cores are going to 100% or not - the 'feel' of the quad core will easily eclipse the 'feel' of the dual core as more and more threads/processes become the norm for not only the O/S + Apps + A/V we need running, but also as we acquire/need new and/or updated apps/utilities in the future.

 

I was trying to think of the most intensive thing she'd likely be doing--probably farmville or watching a fullscreen youtube video while maybe virus scanning. That's one thread, each. I just think it'll probably go faster on a dual core.

 

i'd get whichever is cheaper.. i would think she would be more interested in how the laptop looks than the CPU...

 

Sure, as long as you keep yourself (and the O/S and apps) confined to two threads or less the 37.5% faster dual core will be faster.

As soon as a third thread is introduced (and my example above shows that easily, I think?) the dual core, non HT enabled, processor will start choking and the quad will still feel 'light' and 'effortless' comparatively and significantly faster with 3+ threads than the dual core can ever hope to be - even with an almost 40% clock advantage.

Even in your example above, I can see myself (if I even knew what farmwill is) using two cores as suggested and leaving my email client running, with a website or two open also (okay, more like 40+ windows/tabs running in IE9).

I know - I'm not typical of current 'normal' users - but I am pointing to what a typical 'normal' user will be in the near future.

 

Running different threads on different cores that aren't even taxing one core's full potential isn't going to do anything.

 

Yes it will. The responsiveness of each app you have running will be worlds ahead of running both threads on a single core (with HT, for example) - even if the total cpu usage is still (much) less than 100%.

 

yes. don't under estimate the effect of more core and overall system responsiveness. Higher clock duo core may get your more throughput for particular application but the overall system will 'feel' faster with more core.

 

Can you explain how?

Also can you explain how a 2.2GHz core will be less responsive under singlethreaded tasks? Or perhaps try playing a fullscreen 1080p YouTube video at 2.2GHz and 1.6GHz and compare framerates?

The way I see it, until you hit about 3 full threads of CPU power, the quadcore will be slower (3*1.6 = 4.8 vs 2*2.2=4.4), especially with a lack of technologies like Turboboost where some of the cores automatically OC.

 

I also admit I was a bit influenced by NBR's review of the Probook 4425, where they were rather unimpressed with the P920. They do admit that after bloatware removal, it was decent for "normal use", but again, we don't know what the system is going to be used for.

 

Check this thread; notably, posts #21 and #25:
http://forum.notebookreview.com/hp-...d-core-cpus-8740w-general-purpose-work-3.html
This poster did a side-by-side comparison of a quad-core and a dual-core processor and found the higher-clocked dual to be more responsive in everyday use.
Of course, that was with two very different CPUs (an i7-720QM quad and an "i7"-620M dual-core).

-----

@OP: something else to consider is the heat and battery life implications of the quad-core. Our latest review of a notebook with an AMD quad-core was not promising:
http://www.notebookreview.com/default.asp?newsID=5929&review=hp+probook+4425s
The dual-core Athlon fares much better in power consumption.

 

dtd00d,

My line of reasoning is like so: this system is for normal (avg) use, playing music, watching video's online, surfing (at least two tabs/windows), possibly an email client, office (productivity) software and running Win7x64 (Win7x64 assumption, b/c of quadcore). In addition to this, there is the manufacturer's bells and whistles (bloatware) that will be running in the background (some even useful) and to top it all off, A/V software.

In this situation, do you really think a single thread, or even two will be the only threads running on the system? With the above (admittedly) simple tasks I've listed, it is easy to see more than two or even three being used concurrently - even for a 'light' user, and again, given that A/V is monitoring the system full time.

While to me and you the difference between 2.2GHz and 1.6GHz is worth time and money, to an avg user, the continuing responsiveness of the system is worth more than finishing a single task 20-30% faster (in a 'vacuum' that does not mirror any real world use the system will generally be used in).

We are not comparing AMD vs. Intel here. We are not about ultimate performance (a few GHz short of that...), nor can we ignore the fact that with each passing day, we want/need/use more and more programs and consider that a normal or 'avg' usage scenario. Coupled with the fact that AMD is not HT enabled, this makes my viewpoint even more valid.

As an example (a bad one, I know...): simply right clicking on the task bar to bring up Task Manager is using 3 cores on my i3 (with HT) and hitting around 30-40% cpu utilization (showing kernel times). No, the cores are not maxed, but that is why the computer feels so responsive too: there is headroom to spare. So too will the quadcore feel just as responsive for the g/f of the OP here.



Judicator,
Yeah, we still need a 'reason' of use for this system by the g/f - but keep in mind that a review will keep ultimate performance as the highest priority (always). Also, they were unimpressed compared to Intel's offerings - something we're not doing here.


In the end, both processors will offer enough performance to satisfy an 'avg' user and an 'avg' user will not see the differences between the two anyway.

The quad core will be able to offer this 'good enough' level of performance for far longer. In the very near future, 2.2GHz vs. 1.6GHz is not what will determine whether it will need to be replaced - 2 cores vs. 4 cores is what will be (mostly) the deciding factor.

As the review link posted by Judicator says: remove the bloatware - and enjoy.

See:
http://forum.notebookreview.com/6723837-post163.html


Charles P. Jefferies,

Thanks for the additional link - but 16GB RAM, PS, Illustrator and 3DsMax is not really 'average use'.

As to the heat issue:

From your link above.

 

Kindly note that the heat/power comment was not referring to my first link. See the discussion for the HP review -- someone with a dual-core version of that laptop got far better battery life.

 

That is 'throughput', not system responsiveness.

 

No, that comment is from page 4 of the second link (should have clarified, sorry).

See:
http://www.notebookreview.com/default.asp?newsID=5929&p=4


In the 'Heat and Noise' section.


As far as battery life, the dual core should have longer run times - but as this is an AMD platform, I don't expect it to be significant in terms of all the other factors I've pointed out.

 

I think we're on a different page when it comes to determining system responsiveness; what's your definition?
In post #25 of the thread I linked, the poster provided timing of web page renderings, opening programs, and other smaller tasks; those little things are my definition.

At any rate, before we get off-track here, let's get some clarification from the OP on what the laptop will be used for and the importance of battery life.

 

Thanks for the reply. I see a few problems with one or two things and I'll try to put them as concisely as possible.

A Win7 x64 installation would be because of >4GB of total memory typically, but yeah I suppose. I've seen people running quads on WinXP for benchmarks. And not the 64bit edition.

Ok sure, figure a small % of constant CPU usage just for argument's sake. These are background processes and require quick bursts of processing before sitting idle for a few moments more, then another quick burst of processing.

No, but they don't even take up a full core of usage on the system, so the difference between the two CPUs is moot. The CPU is often not even clocked up to the highest multiplier and in fact sits in SLFM for the majority of idling even with these background processes. So looking at these processes are moot, since it's still within the clock scaling of the processor. If one of these processes suddenly required a full multiplier, the higher-clocked CPU would feel more responsive, as it'd be trying to complete that thread as fast as possible before dropping back to idle.

I disagree. I think from my time in Staples people would rather be able to watch a fullscreen youtube video than to have twenty things up at once. Also, most of the things that people do on computers have the CPU do bursts of processing at a time. Web browsing, for example, I think is the strongest point in my argument.
Web browsing requires a relatively large amount of memory, a fast network connection, and a CPU that will render the page as fast as possible for the best experience. I'm not aware of any major web browsers that are multithreaded to the point where they benefit from having 4 cores over 1 or 2 (I've never seen Firefox/IE use more than about 50% CPU). Usually plugins and things go in a separate thread, but even then, that's just one more thread. So a thread for the browser, a thread for the plugin... and that's about it. It can't touch more than 2 threads in that configuration.

I concur, we're looking at 2x2.2GHz vs 4x4.4GHz, accounting for technologies like Cool 'n Quiet/SpeedStep but neglecting SIMD/HT.

Not quite sure how this proves that this is better than a dual core, since there'd still be headroom.
But actually if you wanted to simulate responsiveness better, you could turn off HT, overclock by 30% (since that's about the generally accepted figure of how much HT improves performance), and see how your system runs then. I'm not sure how OC-able your particular CPU is, though.
Stupid Intel making OCing harder and harder for us... Sandy bridge is supposed to have the SATA bus tied to the memory controller or something like that, so OCing is limited to like 5% before risking HDD data loss. I think I'll be living in the Core 2 days for awhile, where we can just up the FSB and call it a day...

 

The GHz war is over! The world is going the way of cores (whether "virtual" or physical).

The way to improve performance on any modern computer is through data transfer speeds. I can take any 2GHz machine (AMD or otherwise) and stomp a computer with a slower drive or RAM.

I have an AMD Quad in my laptop (N950 2.1GHz) backed with a 500Gb 7200 RPM drive. This thing is fast! I am not sold on hyper threading one bit. Some will say "it rocks" but I don't like it and I am WAY more impressed with the quad core.

My CPU idles between 38-42 (depends on house temp), when I game it jumps up to 65-69. This machine is fast! I also get around 6 hours of battery life.

Go with the quad and make yourself future proof. Many people say that the CPU in a laptop is only good about a year. I think this is not the case anymore as before the laptop my desktop was 3+ years old with a dual core and flew just as well as any new machine. What are we talking about here? trimming 3 seconds off boot times with newer tech? I was booting windows 7 on that desktop from power press to desktop in 19 seconds. Do I really care about 3 seconds? Nope!

And another thing, benchmarks mean very little in real world use. Don't buy anything based on a benchmark alone. I assure you that this laptop here runs better the some of the fastest I7 laptops out there because I keep my machine lean and fast!

It's in the way that you use it!

 

Jesus. Why is it that so many people falsely believe that performance is commensurate with how many cores show up in the task manager? A dual core i3 with HT disabled clocked at 4Ghz will smoke the living daylights out of a six core i7 with HT enabled clocked at 1Ghz in any and all circumstances.

 

Please read what I wrote again.

Just the same a 2GHz with a good SSD drive will stomp the 4GHz whatever dual core!

 

corbintechboy,

Stomp it in what? Idling? lol...

 

Power consumption, heat.

 

And... what does that have to do with comparing 2 AMD cpu's?

 

It looks like he was merely stating his opinion on the matter -- that a quad-core is better than a dual-core. Let's try to be civilized in replying, thanks.

 

Cost.

 

Ok... I will break it down...

Any CPU can only process information as fast as the information is received (data). So speed of CPU is only a small portion of the over all performance of the whole machine.

So, with that said... This relates to this topic in this way:

A 1.6 quad core with a SSD and a lean install will indeed out do a dual core with standard laptop parts (mechanical drive and whatnot).

The argument of dual core vs quad core is a no comparison. The quad core is able to process more information then the dual even at a slower clock speed. Therefor I recommended the quad core.

People don't realize how the WHOLE picture of the machine is how real world performance is determined. Benchmarks have very little to do with anything.

My comment on hyper threading was just saying. I will elaborate on my thinking:

A CPU core is only capable of %100 utilization. This doesn't matter if the way the CPU receives throughput is from a dual data pipe or a single. So hyper threading is one core with 2 data pipes. The core is still restricted by the capabilities of the core itself. So %100 utilization is %100 utilization no matter if the data pipeline is split or not.

Run and program on a hyper threading CPU and see for yourself that if CPU is %100 utilized if it has any room for more data to pass (at a respectable speed anyhow).

A dual core can NEVER match the performance of a quad core (hyper threading or not). Benchmarks may say something else but it is impossible to get something more out of something less then you put in. Simple science there.

So on point, get the quad core and the system load will over all be much mre balanced and will always leave head room for other data to run.

Sorry if my first post seemed off topic. I was trying to insert my opinion along with my opinion.

 

You obviously don't understand hyperthreading, but that has nothing to do with this thread.

 

Do you have it out for me for some reason?

Anyway......

2 pipelines into a single core can not be more efficient then a dedicated core.

Compaq.com - Compaq ProLiant: Hyper-Threading Technology

 

That link actually doesn't say much of anything. And no durr 2 cores is better than one. When your core is at 100% in the task manager, it does not mean that all of the processor's execution units are being utilized at any given time. Many sit idle. Hyperthreading opens up these resources for another thread to utilize. It just helps to make better use of the full processing power of the core. It can have anywhere from -5% up to a 30% increase in performance.

 

No one said a thing about a single core vs. a dual core...

What I said is that a physical core yields a better performance boost then a "virtual" core.

Physics would say that %100 use of any item is at the items max use. There is no added bonus to adding some bogus virtual pipe that would allow a CPU to operate at %130 (hence you saying that there is a 5-30 percent increase in performance).

My brother-in-law has an Intel with hyper threading. His usage no matter what he does not peak above %50 usage on system load, but he is never able to do anything more at a reasonable speed that would show that there is %100 usage.

Just because I run 2 fuel lines into a single carburetor on a car does not automatically give me any increase in speed. The carb is only able to carry and burn as much fuel as it was designed to burn (explaining in simplest form).

Bulldozer is not to far off. If my theory is correct then Intel could once again placed in second by AMD.

My whole point:

If the OP is looking for the best bang for the buck then quad core is the answer. Sure if you were looking at AMD vs Intel i7 then I might say the i7 has a slight performance advantage (not because of any real speed bump but because hyper threading is indeed better at getting the data to the core for the core to process). But of course we are not saying that.

What we are looking at is a dual core with a max usability of %200 where the quad has a max usability of %400. May be flawed logic but makes sense to me.

 

You are trying to come up with an explanation for how hyperthreading works without actually knowing how it works. Hence, it is all wrong. First of all, there is no such thing as a "virtual core" and a "real core." A processor with hyperthreading duplicates its transistors that hold the architecture state, what shows up as two processors to the OS, each with their own interrupt controller. These two logical cores are identical and are equally tied into and share all other processor resources. The two logical cores share those processor resources equally. If both logical cores are requesting the same resources, they are going to share. One is not the superior that can use as many execution resources as it wants at all times and the other is not inferior that only uses execution units that are not being used by the other. They intelligently share the load so that if you run the same thing on both, they will take just as long to do the work, not one finish way faster than the other. The performance boost of hyperthreading comes from the fact that a process, even if using 100% of available clock cycles, is usually not using all of the processor's execution resources for each clock cycle. The reasons vary, but the result is the same, there is wasted potential. The duplicated architecure states help allow the unused execution resources on the processor to be used. Ananlogies like your two fuel lines into a single carburetor simply fall flat here and are completely inaccurate in describing how it works. Also, the performance difference from running hyperthreading can vary from -5% (negative 5%) to 30%.

More info, though far from all-encompassing, can be found here:
Performance Insights to Intel® Hyper-Threading Technology - Intel® Software Network
How to Determine the Effectiveness of Hyper-Threading Technology with an Application - Intel® Software Network

 

Your theory isn't correct.
With Sandy Bridge release before Bulldozer and Ivy Bridge not long after Intel will not be placed behind AMD.
AMD may have a slight edge in the netbook/low end notebook categories but that'l be about it.

 

Trottel and JKleiss summed it up pretty well about HT and the short term future of AMD vs Intel.

Also corbintechboy, you're neglecting the clockspeeds within similar architectures, which do in fact matter. 100% @ 2.2GHz = 4.4GHz vs 100% @ 1.6GHz = 6.4GHz, which is only a +45% improvement. The lower clockspeed needed to keep the TDP in check limits the quad.

 

Well for years Intel has been claiming better performance through hyper threading. This has been accomplished by sharing the resources of CPU tasks through pipelines with a shared resource. Look at the bulldozer "module" philosophy. Looks rather nice on paper anyway.

 

@Trottel

Thanks for the links. I do understand hyper threading as I have read about it many times. And yes, I read your links. Maybe "virtual" CPU was a wrong choice of words.

I stand behind my feelings as posted. If HT is so great then when the Athlon 64 came to be the P4 would of been on top, this was not the case at all.

Intel is on top right now because of the core tech of the CPU. C2D beat out AMD without HT.

 

^ great summary alot of great points made.

cpu's hard drives and memory nowdays are more than enough for most peoples useages. people are out to obtain certain benchmark status when in reality the exact useages of such a cpu that gains such noteable benchmarks, isnt ever going to be completely met by the user which imo renders having such a cpu almost useless. thats just my opinion.

hell im still running a a205 laptop with c2d and even tohugh i have updated alot of the hardware. cpu,hdd and memory this laptop will probably never exceed my needs.

 

Hyperthreading is not some magical solution. It doesn't make some crappy chip good. It simply increases performances anywhere from -5% to +30% depending upon how it is being used. Games tend to be on theHyperthreadeing didn't do much for the P4. Software was lacking to take advantage of it. They touted it as being good for multitasking for home users. Hyperthreading is but one part of a whole. The Athlon 64 was so good because it removed the one problem of the Athlon XP, memory bandwidth and latency, by integrating the memory controller on the die. And because the P4 kind of sucked.

And if you are still sticking to your explanation of Hyperthreading, you still don't understand it.