Experiment: Does Intel’s Turbo Boost Trump Overclocking?

[sithlord666]

Every body hear is missing a point evin the article is missing it why did they need turbo boost in the first place it was nobody was coding fore quad cores and quads in most things ran horrable actually worse than duel cores i just tried to upgrade to a 775 quad core and i thought soumthing was broken when i played a game it was totaly unstable thats why they made turbo boost cuz quads where broken but nobody ever talks about that their still hyping the broken wones when they should be dooing a recall i sent mine back its good their fixing it now with turbo but what about all the people they ripped off telling them how good quads where fore games !!!!!!!!!!!!!

 

[xsamitt]

[citation][nom]sithlord666[/nom]Every body hear is missing a point even the article is missing it why did they needed turbo boost in the first place, it was that nobody was coding for quad cores and quads in most things ran horrible actually worse than duel cores I just tried to upgrade to a 775 quad core and i thought something was broken when I played a game it was totally unstable that's why they made turbo boost cuz quads where broken but nobody ever talks about that their still hyping the broken ones when they should be doing a recall i sent mine back its good their fixing it now with turbo but what about all the people they ripped off telling them how good quads where fore games !!!!!!!!!!!!![/citation]

Is there any weight to whats in this post?If he's right then what a rip off indeed.

 

[Gandalf]

Everybody has a right to their own opinion. However, you might take note of his username. 666 is the number of the beast
[ http://en.wikipedia.org/wiki/Number_of_the_Beast_(numerology) ] and sithlord should be self explanatory.

 

[banthracis]

Um, I'm pretty sure what sithlord wrote was in some foreign language....from what I was able to decipher, he seems to be complaining that quad core processors run programs worse than dual cores, and that turbo boost is a scam to compensate for that worse performance.

What he seems to not understand is that a program that's written to only utilize one thread will run the same on a dual core or quad core, and that if the dual core has a much higher clock speed than the quad, the program will actually run faster. The advantage of a quad core is that:

1- Many program now utilize 2 or 3 cores, gaming performance in particular increases dramatically from 1 to 2, and 2 to 3 cores for many modern games.
2- For a 1 thread program, having multiple threads allows background program to run on a separate core from the main program, allowing the program to utilize that thread in full.

The reason companies made multi core processors was because there's a limit to how fast you can clock a processor. If Intel kept going with their Pentium 4 idea, we'd have 10ghz processors that required nuclear reactors to run by now. To get around the issue of clock speed limit, they industry decided to instead of having 1 "man" do a job, to divide the job up between 2, or now 4 "men." The problem of course, is that some one, ie the programmer, has to divide this job up into 4 parts for these individual men to do.

In programs that are written in this manner, quad cores won't provide 4x faster speed, (after all, the 4 men need to spend time to piece together their individual work into a finished product), but often is close to 3x faster than a single core.

Some old games, using an OS that doesn't know how to divide up task btwn multiple cores (ie XP or earlier), will have issues. It would be the same as giving a foreman 4 men to do a job that he's always assigned to 1 man. The foreman has to spend a lot of time figuring out himself how to divide up the tasks, and may make mistakes and errors in the process.

So in summary. Multiple cores is by no means a scam. It is the only way to increase processing power without resorting to nuclear powered CPU's than run as hot as the sun. The problem is that programs must be written, and the OS must be capable of dividing up the task between multiple cores.

 

[banthracis]

Forgot to explain turbo boost in my last post lol.

The reason for turbo boost is that if you have a task that only requires 2 men, but you have 4 men standing around, why not send the two extra guys home for the day at 1/4 pay and use the money you save as a bonus to motivate the 2 men still working to work faster.

In a processor it would be disabling or lowering power to unused cores, and giving that saved power to the working cores so they work faster.


The difference between turbo boost and overclocking is that with turbo boost, the total power being used by the CPU isn't changing, only where the power is going, ie giving 3/4 of the pay of the 2 useless guys to the working guys.

When you overclock, you're increasing the total power to the CPU, ie, getting a bigger budget for the project to increase everyone's pay to motivate them to work faster.

I agree, that in the future when all programs are written in a manner that allows them to utilize multiple cores, turbo boost will be kinda useless, though combined with speed step if would be a way to save power. However, while we're still transitioning from single thread to multi threading, the feature is very useful.

 

[WR2]

Great article. Thanks for taking the time to pull it together.

 

[Ineffigy]

When testing CPU performance, the game Supreme Commander should be used. I don't know if any game has come out yet that requires as powerful of a processor before slowing down to a crawl. Can anyone think of a game with higher CPU requirements?

 

[Ineffigy]

[citation][nom]PhantomTrooper[/nom]Maybe u shouldn't interject into this conversation seeing as you don't know the origin. I was referring to the processors need for gaming. I didn't feel the need to specify that detail since the intended recipient would already know where the conversation had been.[/citation]

Even the i7 isn't fast enough for Supreme Commander in late game. It still slows down to a crawl with all the AI LUA scripts needed to run the game.

 

[steiner666]

[citation][nom]Curnel_D[/nom]Great article. On another note, the useless Mass Effect 2 ad blaring it's stupid music in my ears at every new page is really starting to piss me off.[/citation]

FF+Ad Block Plus and/or Flashblock

 

[JOSHSKORN]

[citation][nom]PhantomTrooper[/nom]No one needs a Core i7 for gaming. I'm still using a e8400 with a gtx275 and I run everything fine at 1080p, even Crysis.[/citation]
I'm running a E6750 w/ a 8800 320 MB GTS and it works alright w/ Crysis. Oh wait that's right...no one plays it. So WHO CARES. lol

 

[jowunger]

Its quiet obvious: Turboboost is overated and not worth the money...

to bad

 

[JohnnyLucky]

Tnaks for providing an explanation of turbo boost.

 

[Mike75024]

Fortunately, the motherboard we’re using here correctly scales down power and clock rate during idle periods.

How can one tell if one's mobo does that? I was considering the Gigabyte UD P4 and an earlier post alluded to that board not turboboosting correctly.

 

[banthracis]

[citation][nom]Mike75024[/nom]How can one tell if one's mobo does that? I was considering the Gigabyte UD P4 and an earlier post alluded to that board not turboboosting correctly.[/citation]

Not quite sure what the poster was alluding to. According to Everest my P55-UD4P is working correctly. I did flash to the F5 bios though, (to solve HD video and giant mouse pointer issues). I believe it was the F4 update that also fixed speedstep issues.

 

[Mike75024]

Sorry I wasn't clear. The passge I quoted above comes from the conclusion section in the article and implies that some mobos do not correctly scale down power and clock. My machine stays on a lot as it serves up music around my network, and I am looking to have one that scales down power consumption when not under heavy loads.

 

[Guest]

good information for my next computer purchase

 

[Zerk]

Great Article, but? Did they mess up on the PCMark Vantage Overall Suite Score for the i5-750 ?

Can anyone confirm that the 750 is really slower at 9842 with turbo then without at 10219?

Was it a typo or is that true?

Thanks!

 

[killdannow]

Chris Angelini,
Good job exploring the performance of the various scenarios. I have to admit though I am damn curious what was at the bottom of your troubles with first the asus maximum mobo and then the gigabyte board. The article seems to hint at either the specific cpu or memory, but not definitively specify. These are the same mobos that are overclocking and running without a hitch in some of Tom's other articles so I am left scratching my head. Thanks in advance for any follow up you can give on this.

 

[kumaiti]

A question: what is the size and running time of the "first VOB file of The Last Samurai"?

And I second 2 previous suggestions by other posters:
1. it would be interesting to see the performance of a little OC + turbo mode
2. comparing a 4GHZ processor to a "sometimes running at 3.6GHZ" processor isn't really "apples to apples" comparison. Why not compare the stock processor with Turbo Mode on and the OCed processor with Turbo off at the same frequencies?

 

[rooket]

[citation][nom]apache_lives[/nom]i play GTAIV online alot - your e8400 gets left in the dust there sorry PhantomTrooper, and theres no adverse effects having spare cores for future use with newer games etc[/citation]

I've heard GTAIV is a bug ridden game, not to mention that it isn't my preference of game. Overhyped, underdelivered. I had a ton more fun playing Vice City and I believe that one was made to run on Pentium 4s anyway. But I'd have to second Phantomtrooper. I run E8400 and everything works great on it. I have an i7 920 at work and am not overly impressed by it, it feels like much any other PC to use. Not that it is bad but it isn't mind blowing. I saved a bunch of money when I built my E8400 and don't have any regrets at all. Plus I got lifetime warranty on the ram, main board and graphics card. A lot of you knuckleheads probably still use shoddy Asus stuff

 

[rooket]

[citation][nom]cangelini[/nom]Yeah, you guys are going to get a kick out of the upcoming Clarksdale story. It's amazing how badly a Core 2 Duo E8500 gets killed by a Phenom II X4 or Core 2 Quad in some of these more optimized titles. Curnel--sorry about the ad. I also find it pretty annoying to play automatically every time I open a page for proofing. I'll ask about it.[/citation]

... and how many years did it take coders to optimize for more than 2 threads? and just for one game... lol. come on. but seriously, by the time the industry catches up I agree with the previous poster better hardware will be out and likely cheaper hardware the way things are going. I'm amazed an i7 920 still costs over $300.. that cpu has been out for a looooooong time and all these new ones are rather silly to introduce to the market. just shows intel's strong grip on the industry. if AMD is allowed to catch up, we'll see much better things coming out than what is out right now.

 

[pcchip]

Two things:

1.) Your note that the Gigabyte P55A-UD6 performs poorly with turbo boost disabled is rather alarming (!) as I am about to upgrade to a p55 system and normally only use Gigabyte. Could you please look into the issue and see if all Gigabyte boards are affected and explain the problem exactly?

2.) What if you were to find the maximum stable frequency of your i5 at your chosen voltage (lets say for example that at 1.45v , 4000MHz is the max stable frequency) and then LEAVE TURBO BOOST ENABLED, but overclock the bclk so that at MAXIMUM BIN, the chip hits 4000MHz. Therefore when the power is needed to encode / render it would automatically spin up to your max found overclock, and when idling it would spin down and save power. When I purchase my system next month I plan on testing this, is this a viable method to proceed?

 

[pcchip]

Two things:

1.) Your note that the Gigabyte P55A-UD6 performs poorly with turbo boost disabled is rather alarming (!) as I am about to upgrade to a p55 system and normally only use Gigabyte. Could you please look into the issue and see if all Gigabyte boards are affected and explain the problem exactly?

2.) What if you were to find the maximum stable frequency of your i5 at your chosen voltage (lets say for example that at 1.45v , 4000MHz is the max stable frequency) and then LEAVE TURBO BOOST ENABLED, but overclock the bclk so that at MAXIMUM BIN, the chip hits 4000MHz. Therefore when the power is needed to encode / render it would automatically spin up to your max found overclock, and when idling it would spin down and save power. When I purchase my system next month I plan on testing this, is this a viable method to proceed?

 

[benjamins_2002]

Well when I've got a P4 3.0 Ghz an i7 would be a pretty huge jump for gaming.

 

[Guest]

got my Thermalright MUX-120 here
http://www.ultimatepccooling.com/thermalright-mux-120-socket-1156-cpu-cooler.html Im happy with my results

-Steve