Some Preliminary Results of AMD Thuban @ 2.8GHz (OC'd too)

Page 2 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.

Do you think the final results will mirror these preliminaries or do you still think that Phenom II

  • Final results will mirror preliminary results

    Votes: 29 63.0%
  • Equally clocked Phenom II X6 will outperform Core i7s by a considerable margin

    Votes: 17 37.0%

  • Total voters
    46


Notice no temps posted. Would like to see those to see what its running at. I highly doubt it will run cooler than a quad core or use less power. If it was 32nm then maybe.

Would also like to know what typr of air cooling he is using and what the stock voltage is. Since Thuban is a new core, it is possible that AMD tweaked their 45nm and lowered the stock voltage but I doubt hes on 4.3GHz on air on stock voltage.



In single to quad core, I expect it to be behind Core i7. In 5+ cores I expect it to be on par or slightly beat Core i7 due to it having 2 more real cores vs 4 fake cores. I do not expect it to kill current Core i7s and I expect Gulfton to beat it in highly multithreaded apps.

But of course certain people want me to be amazed and drool all over it. Guess I need to put on my green kool-aide glasses for that.... whoops I broke em. Damn....

BTW, did anyone else see this guys stuff:

http://hwbot.org/community/user/nickshih?tab=submissions

7.1GHz Core i5 670 (almost as fast as the 8GHz Pentium 4....)

And a 6.8GHz I7 980X..... damn............ thats almost as fast as the fastest clocked Phenom II..... but with 6 cores.....
 

No need to upgrade if you have a tri/quad core proc. already, but this might be a nice upgrade from my Athlon X2 3800 Socket 939. My 5000 series graphics card might get some breathing room.
 
I haven't been around for a while but the incredible stench of bull coming from the intel fanboys was too hard to resist.

Wait...what is going on here? Elmo is trying to say that he always said Thuban would match i7 clock for clock? :pt1cable:

jimmysmitty has moved on to comparisons with intel's 32nm chips, after months of claiming Thuban wouldn't beat an i7 920? :na:

Badtrip is probably having a REALLY badtrip atm. :lol:

----

Here are the results -

i5 = finished. If you bought one of these, you already know what is coming. You are now looking at what your i5 cost, and doesn't give. Better luck next time guys, remember AMD is good too - and you might be able to plug bulldozer into the same socket. :lol:

45nm i7's = probably looking at a 40%-250% reduction in price depending on model.

All of this done by AMD on an older process with NO MONEY. Wake up kids, you've been backing the wrong side for too long. :)
 


Hello :)

It is a fact that I have stated that a 6 core Phenom II would likely perform close to a Core i7 CPU (at the same clocks). I've stated this for a while now. I haven't changed that position. Preliminary results indicate that what I speculated would occur... is occurring.

And please stop it with the diminutive terms ("kids"). I consider those insults and if you continue on that path I will report you to admins and mods. Just giving you a fair warning.

Oh and you ask for proof?
http://www.tomshardware.com/forum/page-266071_28_50.html

Since it's going to be a 45nm processor. It's safe to assume that it won't beat a any mildly overclocked Core i7 920 Processors.

6 cores and it still can't get the job done. I know this because my friend has a Dual Shanghai rig (8 total Shanghai cores) and I slaughter his rig in terms of CPU performance... it's not at all a fair fight.

(We also know how badly Istanbul compares to Corei7s today as well):
http://techreport.com/articles.x/17005/11

That's right.. your eyes are not deceiving you. An Intel Xeon X5550 (2.667GHz) is besting an AMD Istanbul based Opteron 2.6GHz.

4 cores vs. 6 cores.

Add to the fact that Corei7s use far faster DDR-3 memory than their Xeon based counterparts (non-registered non-ecc often exceeding 1600MHz with tighter timings) and you've got quite the ass whooping on the horizon.

Some will claim that the Phenom II X6 processor will be using DDR-3... so what. Have you seen the level of "benefits" DDR-3 has over DDR-2 on the Phenom II architecture?

AMD needs Bulldozer... and they need it now. 2010 looks to be a snooze for team AMD in terms of CPUs (but the ATi/AMD evergreen refresh with the 384-bit GDDR-5 memory controller looks promising.

That's right... look at the date.


Here are some of the things you've said on the subject (same thread):

So much clueless guesswork in this thread now it is barely worth correcting.

Thuban will be out soon and it will defeat all current i7's...intel know this already and that is why gulftown has been renamed to an i7 part.

If you weren't so damned closed minded you'd be able to figure out simple stuff like that yourselves btw. Keep dreaming.

Oh btw - http://www.top500.org/lists/2009/11

Where are the nehalems in the top 5 again?
 
Lol I can't visit this site without Mousemonkey following every word I type so what makes you think they don't already know? :lol:
 



Jenny, what you fail to understand is unimpressive results are nothing but unimpressive results. In Sigumund Freud's own words "Sometimes a cigar is just a cigar. ”
 
Christ elmo fail a bit less please. You realise you linked pretty damning evidence I'd been searching for? 😀

Since it's going to be a 45nm processor. It's safe to assume that it won't beat a any mildly overclocked Core i7 920 Processors.

Clock for clock, mildly overclocked?

Backtrack more Elmo.
 


Yeah.. mildly overclocked to match the Phenom II X6's processing speed (so far I've seen a 2.8GHz model in action a Core i7 920 is clocked at 2.66GHz).

No fail there.

Also worth noting (and taking into account) how long ago the Core i7 920 was released vs. the fact that AMD is needing 6 cores to match the performance (see equivalent) of the Nehalem architecture.

I think an admin should step in and take care of you. You are one mean person that's for sure. Your obsession over AMD has you insulting people left and right. I'm a tad worried for your psychological health to be quite honest (not an insult.. I truly am worried).

It's only a freaking CPU... no reason to go to war over it.
 
I think Intel has produced some fabulous processors the last few years (albeit for some big $), but I still think Thuban will be faster clock-for-clock than my Athlon X2 939. :wahoo:
 



without a doubt
 


Wake me up when I can SLI and Crossfire on the same AMD board. That has always been the main benefit and isn't possible now. Sure others said differently, but they were just delusional.

Besides the Phenom II 925 and Athlon II X4 620 are still the best processors out there, nothing can justify the extra cost for 99% of people. Thuban is good, but its just another useless enthusiast part. I know this is an odd side to take, but come on, software is so far behind modern CPUs it is hard to care anymore.

Thuban is a good chip if the initial impressions are correct and should be one of the top contenders for the enthusiasts, but remember there are somethings that have little to do with straight processor speed. In the end a product needs to fit a persons needs.

So is the 800 series chipset going to have SLI and Crossfire support like the P55 and X58? Is there any news of this for AMD at all?
 


Um... what? I am still comparing it to Core i7. I said it will not beat Core i7 with 4 cores in anything that uses 4 cores or less. And in anything that uss 5 cores or more it will be on par with or slightly better than a quad core Core i7. And for the record, it will not beat a 920 in a clock per clock basis.

I did say it will NOT beat a Core i7 980x but thats because its a 6 core vs 6 core. Is it wrong to compare 6 vs 6? What is it with you AMD fans that get all chapped when someone tries to compare CPUs that are available? When a review site benched Phenom against Core i7, you said it wasn't fair since Phenom was ment to take on Core 2 Quad. Yet you are more than happy to compare the HD5K t the GTX200 series and no to Fermi ONLY because the HD5K series is better.

Wake up and reread my post. Then again I forget, I am a Intel only fan. When in reality I take everything into comparison, instead of just whatever makes one company look better (price for example). Or have you not noticed I have clearly stated both Thuban and Gulftown are nearly useless for most people right now and its best to wait for Sandy Bridge/Bulldozer since both of those will actually give MORE performance in EVERYTHING instead of certain highly threaded apps? Probably not since your green kool-aide glasses filters it out....



Anything would. Even a low end Sempron II or Celeron (newer that is) probably would.

Man how did you survive this long with that?



As far as we know, so far no. I highly doubt it will come either since getting SLI from nVidia is like trying to get a liberal to lower taxes. Near impossible. Intel only got it by playing a dangerous game which they won. Not sure AMD has it in them to do the same..... or the funds for that matter.
 


While that is true, how hard would it be for AMD to allow and make nVidia support SLI and Crossfire on their boards. They can do a dual partnership so that the boards are actually decent, and nVidia get's their boards sold to all the enthusiast gamers buying AMD.

The lack of SLI and Crossfire support is what is really hurting AMD in my mind, at least when it comes to gamers.
 


SLI and Crossfire are on the AMD platform. You mean both on the same board?
 


I doubt they would. AMD is focused on nVidia in their ATI division. They are wanting to get more people to buy Radeons and use CrossFire. I doubt they would pratner up and even if they did, I doubt the FTC would allow them since they are the only 2 discrete GPU makers out there.



Yes. Both on the same mobo.
 

What I find hilarous about this is who is trying to antagonize whom? Jennyh used the term "kids" ... what else fits there ... morons, idiots, fanboys, nvidiots, spintel, ... Since when is this open forum supposed to be cesnored for being politially correct ... and what term should be used that won't offend anyone at all?

Everyone picks on Jennyh because of the known fact that she has been banned before, and now everyone just wants to keep fuelling the fire. who is the one really at war?

As far as the phenom x6, its a tweaked phenom core, you can't expect miracles from a company that Intel put nearly bankrupt (still not out of the question.) How long did it take Intel to answer the athlon and take over the fastest cpu battle,(how many billions of dollars were spent trying to do so) and now AMD is supposed to reclaim it over night otherwise they are just a piss poor cpu company. Quite frankly I find it piss poor to support a company thats hell bent on corruption at every level of the game, you know who I am talking about.

Also, exactly how do we compare clock to clock? do we use the vendor's specified will never run this slow clock speed or do we disable turbo to get an accurate and precice comparison? If we disable it then all the Intel fanboys cry fowl even if they are just trying to see clock for clock comparisons. Clock for clock comparisons died with the P4/A64 days imo. But now that Intel is faster clock for clock its all important again because it shows how vastly superior Intel is.

I will keep my comparisons on a price/performance basis for any purchases I make.
 
It is not that hard to guess how the Thuban's going to perform. It obviously going to perform just like quad core Phenom IIs at everything except some encoding/decoding situations where it can make use of those extra 2 cores. These are same old Phenom II cores, despite bunch of hypes from AMD fanboys. Unless games (or most of other programs) can start using those extra cores, 6-cores are basically useless for me. We would be talking about Sandy Bridges and Bulldozers when 6+ cores actually mean something.
 
My gosh, you Tom's people are still pathetic.

Thuban introduces new lower-capacitance gating. You do know what 'gates' are, don't you?

The new gating process means less power consumption, less heat and less leakage within the same thermal design.

Do I have to explain to you what that means?



 


Um, presuming you're speaking of low-K dielectric, that's fine & dandy but the bottom line still remains how it performs. And Thuban still basically uses the 3-issue K8 design. While I'll wait for the NDA to lift and the reviews in English to come out, preliminary indications are that Thuban is not going to be anything special with the still-low IPC. It's apparently just another product slotted into an already overcrowded price point segment, which is probably dominated by gaming, and it remains to be seen how well Thuban will do in that area.

IMO, Thuban selling at the same or lower price as P2-965 last summer, yet being 50% larger in die size, means AMD's margins will be low again.

Besides, Thuban still lacks the magic HKMG sauce :kaola:
 


Actually, IMO this thread was fairly respectful & peaceful, until you unloaded the turd bomb above 😀...

Here are the results -

i5 = finished. If you bought one of these, you already know what is coming. You are now looking at what your i5 cost, and doesn't give. Better luck next time guys, remember AMD is good too - and you might be able to plug bulldozer into the same socket. :lol:

45nm i7's = probably looking at a 40%-250% reduction in price depending on model.

All of this done by AMD on an older process with NO MONEY. Wake up kids, you've been backing the wrong side for too long. :)

Hmm, weren't you saying i5 was "finished" last summer before it even released?? :kaola: Not to mention your stating that P2-965 would not be affected at all by i5, and then to strangely go silent as AMD reduced the price from $270 or whatever it was, down to what - $180?

Also, what did AMD do with that $1.25B from Intel again?? 😀