AMD Phenom II X4 965 vs. Intel core i5-750

[dantoshi]

Which of these processors is the best one?

 

[sighQ2]

Thank you for that accurate information.
Thank you for that accurate information.
Thank you for that accurate information.

 

[keithlm]

i think it just needs to be understood this isn't benefiting programs optimized for multi-core CPU's.

In other words... any important application that somebody might be running will have minimal gain from Turbo.

But somehow it still not a gimmick and it is really important and all the cool kids want it.

EDIT: SOmeone mentioned laptops? YUCH... the last thing I would want in a laptop would be feature that automatically increases power use and heat. Just say NO.

 

[roofus]

lets be honest guys.. how many programs benefit from a quad core right now?

 

[notty22]

sighQ2- my definition of stock would be Turbo on. i think it just needs to be understood this isn't benefiting programs optimized for multi-core CPU's.

Thats partly correct.
IE the i5 750 has 21-24 multi
for dual core optimized programs its a notch down from 24 to 23, and sometimes you see 22.
3.20, 3059, 2.92,2.80

 

[roofus]

i can live with partly correct lol

 

[sighQ2]

sighQ2- my definition of stock would be Turbo on. i think it just needs to be understood this isn't benefiting programs optimized for multi-core CPU's.

Yes. but omg these useless test reports and tables generated.

Yeh. if stock is on, how can it be better in "on" result - that's pretty far out man.



"this isn't benefiting programs optimized for multi-core CPU's"
- well perhaps that is what Keithlm meant with his comment about built in obsolescence. I also wondered if spintel calculated reduced-MTBF re temp flux.

 

[warmon6]

understand what you are saying.

what is "STOCK"

Is that turbo on or off?

And how can they show on and off if we don't know if stock is on or off?

that is not very clear to this AMD kinda guy. Is that why you need to end this?

Please describe what "STOCK" means - is that some ES engineering sample? - never mind that - just tell me what stock means.

Stock = stock speed. like phenom II x4 955 or core i7 965 stock speed is 3.2 GHz.

When i was saying stock, that is turbo off. When i was saying turbo, that was meaning turbo on.

that is not very clear to this AMD kinda guy. Is that why you need to end this?

I wasn't referring to you in that comment. I was referring keithlm that seams to think that turbo boost is only a gimmick.

As i see in your post just after mine, (to me) it seams you understand that is not going useful all the time but it is useful under some situations not all.

 

[sighQ2]

Thats partly correct.
IE the i5 750 has 21-24 multi
for dual core optimized programs its a notch down from 24 to 23, and sometimes you see 22.
3.20, 3059, 2.92,2.80

I see what u r saying notty.

but inmultitask with misc proggies, some single, multi - then what.

it's all unknown it seems.


- so for benchmarking - o wow
- the bench selected yields specific win results, or specific lose results?
Is that a fair statement?

It seems that several benches of varying threadage, would be required to accurately report. and even then - still going to vary every run. or every testor. or every ambient temp. it's unpredictable.

 

[randomizer]

Regarding JF-AMD's comments about Turbo, he actually said the same thing on AMDZone in mid December (Bulldozer blog thread).

As far as LGA1366 goes, Turbo can be pretty much ignored. The second speed bin that is supposed to exist for single-threaded applications can only really be reached if you disabled 3 cores or if you assign affinity for every task, including background tasks, to a single core, effectively killing any performance benefit from that extra 133MHz. The first speed bin increase, as I'm sure you all know, happens whenever load is high enough regardless of the number of loaded cores as long as thermals allow for it. For LGA1156 it's a different story as Turbo is much more aggressive.

When benchmarking, if you want to compare clock-for-clock performance then you need to clock all chips the same. If you want to compare out-of-the-box performance (ie. what most people care about) then you need to leave everything as it is, including Turbo. It does not matter to most people how Intel's chip takes the lead, it only matters that it does. Due to how Turbo works, one would need to find benchmarks of the same or similar software to what they use to see if it will really benefit them though. Because an i7 860 can rip up the competition in SuperPi doesn't mean it's going to do the same in POV-Ray

 

[Chad Boga]

Regarding JF-AMD's comments about Turbo, he actually said the same thing on AMDZone in mid December (Bulldozer blog thread).

.
.
Because an i7 860 can rip up the competition in SuperPi doesn't mean it's going to do the same in POV-Ray

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3641&p=5

 

[randomizer]

http://www.anandtech.com/cpuchipsets/showdoc.aspx?i=3641&p=5

http://img.photobucket.com/albums/v46/JOSEPHSARANDOS/20076.png

Thanks for proving my point.

 

[Chad Boga]

Thanks for proving my point.

How so?

You were saying that just because the i7-860 rips up the competition in SuperPi, that doesn't mean it is going to do so in Pov-Ray.

I then produce results showing it does rip up the competition in Pov-Ray.

 

[randomizer]

I see it almost exactly the same as an i7 920 which just so happens to have a very poor Turbo Boost. In fact, even the 870 which is at a somewhat higher clockspeed than the 920 is not doing much better. I'm not sure what is causing the 975 to do so well as it's not at a significantly faster clock speed than the 870.

I'd wager that putting these chips into a SuperPi contest would put the 860 and 870 at the top, or very close to the 975.

 

[Chad Boga]

I see it almost exactly the same as an i7 920 which just so happens to have a very poor Turbo Boost. In fact, even the 870 which is at a somewhat higher clockspeed than the 920 is not doing much better. I'm not sure what is causing the 975 to do so well as it's not at a significantly faster clock speed than the 870.

I'd wager that putting these chips into a SuperPi contest would put the 860 and 870 at the top, or very close to the 975.

So in a thread where we have AMD fanboys trying to fud Turbo boost, you make comments about how an i7-860 can benefit over "the competition" in Super Pi, but may not do so in Pov Ray, and one is supposed to think that by "the competition", you mean against other members of the i7 family, rather than AMD processors?

 

[randomizer]

I mean both AMD's CPUs and Intel's CPUs. Usually when you benchmark a CPU it competes against every other CPU in the line-up, not just another manufacturer's product. Otherwise there would be no reason to include half of the CPUs in the article.

Had you read my entire post, you may have realised that I focussed largely on LGA1366 and LGA1156, not AM3. I would have thought that context had a greater importance than a preconceived idea of what "the competition" is.

 

[Chad Boga]

I mean both AMD's CPUs and Intel's CPUs. Usually when you benchmark a CPU it competes against every other CPU in the line-up, not just another manufacturer's product. Otherwise there would be no reason to include half of the CPUs in the article.

So how doesn't the i7-860 rip up the AMD offerings in POV-Ray, which was part of your initial contention.

 

[Chad Boga]

Had you read my entire post, you may have realised that I focussed largely on LGA1366 and LGA1156, not AM3. I would have thought that context had a greater importance than a preconceived idea of what "the competition" is.

The lead an i7-860 has over an i7-920 in Super Pi is not so great that the term "rip up" would be applicable.

However with the amount by which an i7-860 beats anything from AMD, then the term "rip up" would be applicable.

 

[randomizer]

However with the amount by which an i7-860 beats anything from AMD, then the term "rip up" would be applicable.

Yes, well that happens in the case of Core 2 as well. The program doesn't really like AMDs architecture much. It really needs to die... or at least be given an overhaul.

 

[Guest]

Yes, well that happens in the case of Core 2 as well. The program doesn't really like AMDs architecture much. It really needs to die... or at least be given an overhaul.

Super Pi...compiled with the intel compiler I believe?

 

[BadTrip]

Super Pi...compiled with the intel compiler I believe?

nope, sorry.

 

[Chad Boga]

It is too bad that AMD might actually start using the same kind of failed marketing mechanism. Luckily there is always "disable". I don't need the marketing gimmicks. I prefer predictable and repeatable performance results. Too bad you can't seem to comprehend the concept because you and many other Intel fanboys are too busy trying to insist that everybody accept your fanboy opinions or automatically be wrong.

Hey Keith, I dunno why, considering the tremendous technical points you raised against Turbo boost, but not only are the engineers from Intel and AMD in favour of Turbo boost, so too are the engineers at IBM.

Check out Page 42 of this IBM document for Power 7.

http://www.ibm.com/developerworks/wikis/download/attachments/104533501/POWER7%2B-%2BThe%2BBeat%2BGoes%2BOn.pdf


So on one hand we have the Great Mind of Keith(saying Turbo Boost is a gimmick) & the non-existent mind of ZootyGay vs the designers of Intel, AMD and IBM.

Who is likely to know what they are doing and who is just talking out their arse? :lol:

 

[keithlm]

Hey Keith, I dunno why, considering the tremendous technical points you raised against Turbo boost, but not only are the engineers from Intel and AMD in favour of Turbo boost, so too are the engineers at IBM.

Check out Page 42 of this IBM document for Power 7.

http://www.ibm.com/developerworks/wikis/download/attachments/104533501/POWER7%2B-%2BThe%2BBeat%2BGoes%2BOn.pdf


So on one hand we have the Great Mind of Keith(saying Turbo Boost is a gimmick) & the non-existent mind of ZootyGay vs the designers of Intel, AMD and IBM.

Who is likely to know what they are doing and who is just talking out their arse? :lol:

As has been pointed out this "feature" is really only useful for old single threaded or badly programmed applications. For any serious application it is worthless. Good job trumpeting the benefits of something that is basically non-relevant and thus useless as I've been saying.

But you just keep telling yourself... "It's not a gimmick... it's not a gimmick... it's not a gimmick." Then close your eyes and click your ruby slippers together.


EDIT:::::
And btw:

However with the amount by which an i7-860 beats anything from AMD, then the term "rip up" would be applicable.

It is not uncommon for Intel fanboys to use the term "rip up" to describe a difference of 1% or even less. We're basically used to that type of tactic. So basically the descriptive term of "rip up" means nothing.

 

[someguy7]

You can't stand the fact that somebody will not agree with you on the value of this marketing gimmick. It is making you apoplectic. So you pretend that this is all because I have an irrational hate for a brand in a veiled attempt to belittle my opinion. I've already told you the exact reason I find this marketing gimmick to be less than acceptable. I will be saying the same thing if or when AMD adds this garbage to their CPU. That is something that is really making your head crazy. Since I'm not basing my opinion on a brand your brain can't accept the fact that I could have another non-brand specific reason.

Some people might be willing to accept gimmicks that compromise the system. I do not find it acceptable and I will not lower my standards low enough to find it acceptable in any brand.



How many times did you say it was not a gimmick in your post? Perhaps you need to say it half a dozen more times like a mantra. Then you can pretend that your opinion is reality and there is no way anybody can have an opposing opinion.

It is my opinion Turbo-Boost is a benchmarketing gimmick. I am not the only person that has this opinion on this issue. For some reason you guys just can not accept the fact that people do not share your collective opinion and that this has nothing to do with brand and everything to do with functionality.

There is no doubt we will continue to get a bunch of rabid Intel fanboys joining together to defend their opinion at all costs. It must really bother you guys that you can't disprove the fact that it is a benchmarking gimmick. Instead of just accepting the fact that this "feature" is not acceptable to some people Intel fans often resort to personal attacks in failed attempts to defend their opinion against the godless people that do not agree with them.

Actually I find the fact that Intel fans can't accept that some people consider it a gimmick to be utterly pathetic.

So you Intel fans should just keep making incorrect assumptions and deny others the freedom to have their own opinion. The more defensive you get the more pathetic the arguments become.

EDIT: And BTW... it is a gimmick.

It does not matter how many times I said it. In my post I explain how it is not a gimmick and will be a vital part of cpus in the future.

You keeping saying you and others do not like the feature and will disable it. That is your choice and I'm fine with you making that choice. But because you will disable the feature does not mean that it is a gimmick. If it is not acceptable to you that does not make it a gimmick. Cool and quite is not a gimmick. Speed step is not a gimmick. HT is not a gimmick.

You keep saying that the people that have the opinion that this is a not a gimmick are rabid Intel fanboys. Which is false. You, me and everybody else knows that all the people that think this feature is a gimmick or a tool to cheat on benchmarks are part of the AMD cult.

You admit you're AMD biased but yet when somebody points out your biased crap in your post they are labeled Intel fanboys. You're a joke

 

[Chad Boga]

As has been pointed out this "feature" is really only useful for old single threaded or badly programmed applications. For any serious application it is worthless. Good job trumpeting the benefits of something that is basically non-relevant and thus useless as I've been saying.

Accept that you are wrong.

There is heaps of code that isn't optimised for more than Dual Core, thus a 2.8Ghz i7 can run at 3.46Ghz when only one core is needed and up to 3.33Ghz when 2 cores are all that is needed in those situations.

But you just keep telling yourself... "It's not a gimmick... it's not a gimmick... it's not a gimmick." Then close your eyes and click your ruby slippers together.

Are the designers from AMD and IBM also doing this Keith?

The way you respond, it is though it just hasn't registered to you that AMD and IBM are taking up Intel's innovation here.

Can you explain why IBM is putting this technology into POWER 7 if it is only a gimmick?

Like I and others have told you, this will be an ever increasing aspect of CPU design technology, but because right now Intel have it and AMD don't, you just cry "gimmick".

Do you realise how stupid you make yourself look?

 

[keithlm]

Do you realise how stupid you make yourself look?

Not as stupid as somebody attempting to pretend that everybody will adopt and use this "feature".

It really is too bad when marketing causes compromising features to be added to hardware.

(It's even more pathetic when people that actually do have a bit of intelligence buy into the tripe.)