What do you think of AMD's Thuban

[DXcellence718]

You guys probably already heard the news, but now that amd's six core processor is coming out what are your thoughts. Personally im not really impressed, from what i've heard its only equivalent to an i7 quad but cost under $300. It is still the cheapest hex core a consumer can get, so i guess there still good for price/performance ratio(for a six core processor anyways)

 

[Chad Boga]

You're absolutely confusing the term best bang for the buck. It is not subjective. If best bang for the buck was to be completely subjective then there would be no point in saying best of anything. Do you understand? You can NARROW your range and compete dual cores or hex cores or quad cores and label best bang for the buck within that spectrum but I was not doing that. I was talking about the ENTIRE cpu spectrum. You don't seem to understand the concept of best bang for the buck so I'm going to leave your failed line of logic to yourself.

I understand how many people use the term and how they arrive at it, but it also seems that some people are incapable of understanding why for many consumers, they would be unhappy with the best bang for buck processor, because it still doesn't meet their required performance needs.

 

[werxen]

I understand how many people use the term and how they arrive at it, but it also seems that some people are incapable of understanding why for many consumers, they would be unhappy with the best bang for buck processor, because it still doesn't meet their required performance needs.

Are you extra slow today, Chad?

 

[jimmysmitty]

GADS... you actually said something relevant! CONTRATULATIONS. But then it doesn't really matter. The fact that the Turbo-marketing gimmick is going to stop all of the pretenders is very useful. Besides... the only person that will be looking at results from the slower chips are people obsessed with finding something less than optimal. (Probably predominantly i5-750 owners that are used to pretending their chip faster than it really is.)

Keith... what the hell is a CONTRATULATIONS?

BTW to all i5-750 owners, Keith has proclaimed that your CPU is not really that fast. So when your i5-750 rips a BD to a .MKV really fast, it is a lie much like the cake because Keith has proclaimed it and nothing you see or think can be true!!!!

 

[Chad Boga]

Are you extra slow today, Chad?

No, I have a Quad core compared to your crappy Dual core, I am a blazingly fast. :lol:

 

[werxen]

No, I have a Quad core compared to your crappy Dual core, I am a blazingly fast. :lol:

Too bad your blazingly fast quad core can't help your reading comprehension. Maybe you should switch to a 980x.

 

[Chad Boga]

Too bad your blazingly fast quad core can't help your reading comprehension. Maybe you should switch to a 980x.

It fails the bang for buck test.

 

[werxen]

It fails the bang for buck test.

Good boy. You're making progress. I see a bright future ahead of you where you could add positive insight to these forums. Keep trucking.

 

[Chad Boga]

Good boy. You're making progress. I see a bright future ahead of you where you could add positive insight to these forums. Keep trucking.

You will be able to as well, once you dump your silly dual core.

 

[werxen]

You will be able to as well, once you dump your silly dual core.

I don't need a quad. The difference is I am smart enough to recognize it.

 

[Chad Boga]

I don't need a quad. The difference is I am smart enough to recognize it.

You are now just trying to justify the mistake you made.

 

[yannifb]

Well keith i have to admit Hyperthreading and turbo actually are kinda useful and are therefore not just gimmicks.

Also, chad, what did you do to gain such a superior knowledge of tech compared to that Kieth fellow who (if he was being serious) has been 'programing in assembly since the 70's? Because honestly your arguments consist of the following:

:lol:

LOL
your just a crazy fanboy
lol sure
and many others.

Also don't call me a fanboy because i tend to point out the flaws of the more unpleasant forum members, or because i own an AMD chip. I got it because my machine is for pure entertainment and didn't need anything better. Which is why i'm not buying Thuban.

 

[Chad Boga]

Well keith i have to admit Hyperthreading and turbo actually are kinda useful and are therefore not just gimmicks.

Also, chad, what did you do to gain such a superior knowledge of tech compared to that Kieth fellow

You answered your own question in the first sentence of your post.

Keith is so AMD obsessed, that if Intel has something that AMD doesn't, then that thing is a gimmick only.

Additionally, turbo technology is about the most efficient use of the power envelope your processor has, and as CPU's increase their core count, then turbo implementations will become an increasingly important aspect of modern CPU design.

Keith refuses to accept this, so if that isn't a failure to understand the technology, what would you call it.

 

[yannifb]

You answered your own question in the first sentence of your post.

Keith is so AMD obsessed, that if Intel has something that AMD doesn't, then that thing is a gimmick only.

Additionally, turbo technology is about the most efficient use of the power envelope your processor has, and as CPU's increase their core count, then turbo implementations will become an increasingly important aspect of modern CPU design.

Keith refuses to accept this, so if that isn't a failure to understand the technology, what would you call it.

I agree that turbo is amazing for the user who doesn't really know much about computers, and for people like us who know about over clocking and how to do it it's still pretty damn convenient.

Also, don't take some of my posts personally. Im just kind of on edge today.

 

[werxen]

Turbo = win
Hyperthread = meh

 

[fazers_on_stun]

Turbo = win
Hyperthread = meh

IIRC hyperthreading can provide up to ~30% performance boost in the case of multithreaded but lightly-loaded apps, where a core spends a significant amount of time waiting around for data. Why not switch to another thread and do work during the otherwise wasted clock cycles? Esp. when HT only adds something like 5% extra transistors?

IMHO, Turbo & HT are the same performance-boosting philosophy, but at opposite ends of the threadedness spectrum. But if your apps are mainly single or low-threaded ones with few idle clock cycles, like some games, then you would probably be better off disabling HT.

 

[Chad Boga]

IIRC hyperthreading can provide up to ~30% performance boost in the case of multithreaded but lightly-loaded apps, where a core spends a significant amount of time waiting around for data. Why not switch to another thread and do work during the otherwise wasted clock cycles? Esp. when HT only adds something like 5% extra transistors?

IMHO, Turbo & HT are the same performance-boosting philosophy, but at opposite ends of the threadedness spectrum. But if your apps are mainly single or low-threaded ones with few idle clock cycles, like some games, then you would probably be better off disabling HT.

fazers,
Even though I know what you said about hyperthreading's potential performance gain was correct, I think that figure of 30% though is such a best case scenario, that I would probably mention that gains of 10 to 15% are a more realistic measure, as this is also being mindful that there are instances were there is no improvement and in some instances slight degradation.

 

[werxen]

IIRC hyperthreading can provide up to ~30% performance boost in the case of multithreaded but lightly-loaded apps, where a core spends a significant amount of time waiting around for data. Why not switch to another thread and do work during the otherwise wasted clock cycles? Esp. when HT only adds something like 5% extra transistors?

IMHO, Turbo & HT are the same performance-boosting philosophy, but at opposite ends of the threadedness spectrum. But if your apps are mainly single or low-threaded ones with few idle clock cycles, like some games, then you would probably be better off disabling HT.

Maybe in a couple more years when multithread is mainstream I would agree. Till then, rock on turbo.

 

[keithlm]

When was the last time a new stepping or an errata fix(outside Barfelona's TLB bug) ever gave an IPC improvement that anyone noticed in any reviews?

AH... but you are conveniently forgetting that is not just a stepping... there is also a new die. Are you going to pretend that the last time they did a stepping and die together that nothing changed? Is that your final prediction?

So I think my prediction there will be proven right again, no matter how desperate you are to pretend it won't be the case.

Because you are the almighty Chad Bogas we must accept your predictions as being more important than anybody else's predictions. Good luck with that one. Especially since you already failed.

:lol: Keith, not only don't you understand the technology, you can't follow when someone talks about it all.

The point I was making which whooshed over your head, is that you are massively wrong when you infer/state that the i5-750 is running most of the time at 3.2Ghz in the benchmarks and that is why it is kicking AMD's arse.

I see you are falling back to personal insults and pretending that somebody doesn't understand your less than intelligent ranting. You are not revealing unusual behavior for somebody that is desperate.

I am saying that as the i5-750 is likely not to be hitting it's turbo max that often, neither will the turbo feature on Thuban hit it either, so thinking that now AMD has turbo, it is going to catch up significantly to Intel is wishful thinking.

VERY HUMOROUS: When people pointed out that the average home user would not really gain much benefit from Turbo, they were denigrated. Now you are going to support that viewpoint which the Intel fans previously ridiculed people for. Obvous case of double standards.

But sadly for you (and your argument) the benchmarking platforms we are talking about are not normal systems operating in average conditions. Most of the benchmarking reviews were done under optimal conditions that generally guarantee that the i5-750 was hitting max as much as possible. Sorry to point that out and completely ruin your argument.

So after all the jumping around pretending how hyperthreading was a great benefit that allowed Intel to "win" you are now going to start to belittle multi-threaded applications also.

Nope, and even allowing for how poor your interpretation of just about everything is, I can't work out how you managed to confuse yourself and ended up drawing that conclusion.

Not surprising that you can't work it out, you have already shown yourself to be as terribly confused as you pretend to claim that others are.

You mentioned that the 3.2Ghz PhenomII with 6 cores and Turbo won't be as good as the 3.4Ghz Quad core without the turbo-gimmick. (DUH.) And yet in the past so many people pretended that the obscure multi-threaded benchmarks showed how much superior their i7 chips because of hyperthreading. Now you want to discount those same benchmarks. It seems you want to have it both ways. Another example of double standards. Then you want to pretend that adding additional cores is not in any way similar to hyperthreading.

Keith,
People are laughing right now at how you can't even follow mainstream technology and keep making one crazy assumption after the other.

As I said before, keep plucking that chicken.

What part of your ranting did I not follow? The part where you became Captain of the Fail Boat by resorting to insults and making assumptions? Or the part where you attempt to obscure the facts by attempting to present ridiculous non-facts?

 

[keithlm]

Well keith i have to admit Hyperthreading and turbo actually are kinda useful and are therefore not just gimmicks.

Also, chad, what did you do to gain such a superior knowledge of tech compared to that Kieth fellow who (if he was being serious) has been 'programing in assembly since the 70's?

You are correct in that Turbo-Boost has the potential to be more useful than just a marketing gimmick. But with the way Intel has currently implemented the feature it really is only useful for one thing: Benchmarks. A serious user will manually overclock, and an average Joe won't care. After AMD adds this same type of feature the real truth of this viewpoint will become even more apparent.

Hyperthreading is pretty much a shortcut taken when you don't want to add real cores. Hyperthreading will pretty much disappear in the future. (And what AMD is putting into their newest design is not hyperthreading no matter how often some posters use it an attempt to get a flame war started.)

BTW: I haven't really done any assembly since the beginning of the 80's. But it is still fun to come out and play with the kiddies. It makes them crazy pointing out obvious minor things that they want to pretend aren't true. It is even more fun because they don't realize that no matter how much they cry about these things they can't change some things just because they don't like them.

 

[Chad Boga]

AH... but you are conveniently forgetting that is not just a stepping... there is also a new die. Are you going to pretend that the last time they did a stepping and die together that nothing changed? Is that your final prediction?



Because you are the almighty Chad Bogas we must accept your predictions as being more important than anybody else's predictions. Good luck with that one. Especially since you already failed.

Keith,
Even though you should be a master of recognising failure on a daily basis(I presume you have a mirror), what have I supposedly failed at?

I am stating that the per core performance of Thuban will only show improvements due to Turbo, if you run a benchmark on Thuban that doesn't use more than 4 threads, then clock for clock, it will be no faster than the PhII.

What is your prediction?

I see you are falling back to personal insults and pretending that somebody doesn't understand your less than intelligent ranting. You are not revealing unusual behavior for somebody that is desperate.

:lol: You are a funny and confused man Keith. Suggesting I am falling back to insults, when that is your stock in trade.

And what am I supposedly desperate about? This seems to be your deepest desire, that people who think you are a fool, are desperate about something.

VERY HUMOROUS: When people pointed out that the average home user would not really gain much benefit from Turbo, they were denigrated. Now you are going to support that viewpoint which the Intel fans previously ridiculed people for. No, because y Obvous case of double standards.

But sadly for you (and your argument) the benchmarking platforms we are talking about are not normal systems operating in average conditions. Most of the benchmarking reviews were done under optimal conditions that generally guarantee that the i5-750 was hitting max as much as possible. Sorry to point that out and completely ruin your argument.

Well Keith we don't know if the benchmarking conditions are optimal and thus the Windows Scheduler doesn't inhibit Turbo in benchmarking runs.

But the point you are skipping over is that if AMD's turbo works as well or better than Intel's turbo in the benchmarks, then in no way will I seek to diminish that, and would still continue to see the wide range of benchmarks as being representative of the performance difference between each CPU maker.

You mentioned that the 3.2Ghz PhenomII with 6 cores and Turbo won't be as good as the 3.4Ghz Quad core without the turbo-gimmick. (DUH.)

You shouldn't keep sticking to the conversations you imagine you have in your head, but rather read what has been written.

I was comparing the 2.8Ghz Thuban with a top Turbo of 3.2Ghz to the PhII 3.4Ghz. I made that comparison because they look like they will be priced the same.

And yet in the past so many people pretended that the obscure multi-threaded benchmarks showed how much superior their i7 chips because of hyperthreading. Now you want to discount those same benchmarks. It seems you want to have it both ways. Another example of double standards. Then you want to pretend that adding additional cores is not in any way similar to hyperthreading.

Keith you are just desperate and confused. I am not doing what you are suggesting I am doing.

The top of the line Thuban will be AMD's most compelling desktop part on performance alone, but it will be 50% dearer than a PhII 3.4Ghz, so whilst it will beat the PhII in most, if not all benchmarks, it's margin of victory in a good many of those benchmarks will not be so great as to make the 50% price premium seem worth it. Just as I would never buy a i7 more expensive than the 860 or 930, the small performance advantage isn't worth the extra cost.

The Thuban that is priced the same as the PhII 3.4Ghz, will probably lose more benchmarks than it wins against its sibling.

It is for those above reasons that I have stated that I don't see Thuban shaking things up that much, but I doubt you will understand what I have written and instead will imagine something different, so that you can respond once more with utter nonsense.

 

[werxen]

Hey Chad, do us all a favor and stop posting!
- Toms

 

[Chad Boga]

Hey Chad, do us all a favor and stop posting!
- Toms

Hey Werxen, do us all a favour and stop breathing!
- Toms

 

[someguy7]

Oh boy.

 

[yannifb]

Hey yannifb, do us all a favor and never stop posting!
-Everyone. Ever.

 

[jimmysmitty]

The last steppings that AMD had didn't improve IPC, much like the last stppings that Intel had didn't improve IPC.

Phenom I to Phenom I 50 was mainly to kill the TLB bug. Phemon I to Phenom II was not a stepping, complete new processor.

Phenom II C3 was just like most stepping improvements, lower the TDP and increase overclocking potential.

I can see Thuban having a small IPC improvement, but nothing major. Its not a rework of Phenom II but simply a 6 core Phenom II.

I still say it wont be impressive due to its low price. AMD is not in the charity business.