AMD "Zembezi

Page 8 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Status
Not open for further replies.

illfindu

Distinguished
Nov 30, 2009
370
0
18,810
Hey im looking towards the future and I'm eyeing the AMD 8 core bulldozer CPU'S coming down the line. I'v seen some source say there going to use a AM3+ Socket and im wondering if that means youll be able to toss one in a current AM3+ compatible board?
Will my current http://www.newegg.com/Product/Product.aspx?Item=N82E16813130297&nm_mc=OTC-Froogle&cm_mmc=OTC-Froogle-_-Motherboards+-+AMD-_-MSI-_-13130297 msi 870A Fuzion work i noticed just now that its a AM3 not a AM3+ I'm guessing that means ill need a new mother board cause the sockets are comparable?
 
Solution


That "market" is no different from any other market. Not all software makes the best use of 24 cores - some of the tests in that link didn't even make use of 12 cores. How is a lower clocked 24 core server supposed to perform against a higher clocked 12 core server when the workloads are only optimised for 12 cores?

22156.png


The result of this scaling is that for once, you can notice which CPUs have real cores vs. ones that have virtual (Hyper...
Gulftown also has 25% more transistors and is on a process ahead of Phenom II X6. You can't compare them core for core.

You can't even compare the bloomfield/lynnfield cores to Phenom II, as again the intel cores are bigger. The whole core argument is a crock of ***.
 
I'll admit I'm a bit AMD leaning at times, but saying you can't compare the performance of two products that are out at the same time because one has 'bigger cores' or is on a smaller process makes no sense to me. Just when can you compare AMD to Intel then? Only back in the old days when they were copies?
 


They are the best 6 cores vs 6 cores from both camps that are available for sale RIGHT NOW.

When BD gets released, it will be compared to what ever Intel has available at that time.

Ivy Bridge and its 8 cores won't be out 'till later this year, so is it going to be fair to compare 8 core BD vs 4 core SB?

Is it Intel's fault that it has taken AMD this long from rehashing its architecture time and time again for the past few years?

All the talk in here about real cores, who has more cores, the speed, blah blah blah

I just want to know what BD will do real world, period. I've already gone through a lot of the architecture, and I am subscribed to a few threads like this one in a few forums.

Are 4 BD cores ( 2 modules ) going to compete with Intel's current Quad Core offerings ( 2500k, 2600k )

Or are the 8 core BD's ( 4 module ) going to compete with the Quads?

I'm interested because I am considering picking up a 980x, but I'm also waiting for BD to get released to see how it will live up to the hype. The architecture looks good on paper, that's about it, and that's all we know.

So yea, I'm impatiently waiting, as I can get a good deal on a 980x right now, which makes it worth it. 😉
 

If you want to really be picky, that 980x is probably using turbo boost to get to 3.46ghz in most of the workloads. If it's anything like the quad i7. At stock settings the 920 would be 2.8ghz while folding, prime 95, etc. with turbo boost on. It would also be 2.8ghz while surfing the internet, so pretty much any workload would make it the 21x multiplier instead of 20x.
 


And will that really make a difference in the whole grand scheme of things?

1 is obviously faster regardless, especially when you take overclocking into account.

I've seen people's rigs with a 1090T @ 4.4 bottlenecking a SINGLE 470, a 980x doesn't see a bottleneck 'till you do Tri SLI 580's

Can anyone here make an educated guess whether BD's 8 core will be better than a 980X?

In Video Editing, Encoding, Rendering, Highly Threaded Apps - On paper it's supposed to...
 


You can since its all AMD has out.

One thing I hated when Core 2 45nm hit is people cried foul when comparing it to 65nm Phenoms when thats what AMD was offering.

AMD will probably always be behind in process. Its normal. Compare whats offered, not what will be.
 


Sadly, that's the issue, you can't fault Intel for being ahead in the Fab process, they've had 32nm processors for close to a year, AMD is just now getting there.

Intel will be at 22nm maybe at the end of this year, early next year.

They'll be at 14nm in 2014 perhaps?

Intel to spend $5 billion on new 14nm fab in Arizona

When Paul Otellini isn't too busy talking about being jilted by Nokia, he spends his time hosting presidents and splashing billions of dollars on new manufacturing facilities. Intel's CEO is wrapping his tumultuous week on a high note, having welcomed Barack Obama to Chipzilla's Oregon facility and treated the president to the happy news that Intel will invest $5 billion back into the US economy by building its most advanced fab yet -- which will introduce an impossibly small 14nm production process -- in Arizona, to begin operation in 2013.
Source: http://www.engadget.com/2011/02/19/intel-to-spend-5-billion-on-new-14nm-fab-in-arizona-creating-4/

When is AMD expecting to hit 22nm and below?

I really hope BD lives up to the hype.
 



Research Quantum Tunneling. When that happens, it will give AMD a chance to catch up.
 

The 1100t also has turbo.
 


Going to save that for a rainy day...
 

Only a fool would make that statement in all seriousness.

Intel are concentrating in order on 3 things

1. Increasing the strength of individual cores
2. Adding more cores over time
3. Increasing clockspeed

With the Bulldozer Approach it looks like AMD are concentrating in order on these 3 things

1. Adding more cores
2. Increasing clockspeed
3. Increasing the strength of individual cores

With the difference in approach, Intel's CPU's will be far more suited to 90+% of desktop users, but with respective pricing from AMD and Intel & AMD's marketing, AMD may be able to get 30+% of the desktop market.


Here's the difference: Cores are not constrained like clock speed is. Look at the last 5 years and ask yourself, how have cores progressed? (3X) How has clock speed progressed?

If you had to lay a bet on where performance in 2013 was going to come from, how many of you would place your bet on clock speed and how many would bet on core counts?
You seem to have conveniently left out the IPC of cores in your comments.

Intel's advantage over AMD isn't now or in the near future going to be based on clocked speed, or number of cores, it is going to be based on the IPC of their cores.

The reality is that clock speed isn't going anywhere. Nice to see that Intel has cornered the market on that, but where is is going?
Intel have cornered the market on the strength of their individual cores and where is that going? Well it certainly ain't going away like you are hoping people believe, so that AMD's flock of chickens will be able to plow the fields better than Intel's two oxen. :kaola: :lol:

 


No, that is not true.



You won't see benchmarks. I don't know what they are showing, but I caution against people getting too excited about what will be shown there. I am guessing that after cebit is over there will be 100 threads saying "AMD was supposed to launch and...."



Price matters far more than raw performance. Your sig says that you are running an I7 930. Why not a 980? Price I am guessing. Price matters more because price is a driver in probably 99%+ of the purchases, the 980 has less than 1% of the shipments. If performance was the #1 driver, that share would be higher. Obviously price matters. A lot.
 

1100T = 3.3GHz, 3.7GHz Turbo

You're right, price does matter, but the discussion here turned into who has the better cores. I am merely stating that Intel has by far better cores than AMD does ( core for core, and clock for clock ). 980x vs 1100T 'nuff said.

http://www.anandtech.com/bench/Product/142?vs=203

I highly doubt that if the 1100T was on par with the 980x, that it would cost $239, and the 980x certainly wouldn't be $1000.

If AMD had better performing products right NOW, Intel wouldn't be able to get away with charging $1000 for those processors, just like the 580 wouldn't cost $500 if the 6970 gave it a run for its money, but let's not get off topic.

We all know that the price of the processors aren't based on speed as an incremental value. My 930 cost $300 and the 980x certainly isn't 3 times faster, but it's just the way it is. AMD charged $1000 for their FX series dual cores just 4 years ago, why? Simply because they could, as Intel had nothing that came close to that performance at the time. ( Intel isn't the only one that has charged $1000 for processors simply because their processor are the top of the line at the time. )

When I put my rig together, my budget was $2000, and it didn't make sense for me to go 980x, but ever since Intel price dropped the 970, and released the 990x... I have a friend who has a 980x and is upgrading to the 990x ( don't ask, dude has like $7000 worth in that rig, and wants the 990x )

He's willing to sell me his 980x for the price of the 970, after selling my CPU, the upgrade doesn't seem so costly @ just $400 - $500, besides, I was also considering upgrading to the 970 before my buddy said he would rather me take the 980x.

I also hinted at this earlier...


Can anyone here make an educated guess whether BD's 8 core will be better than a 980X?

In Video Editing, Encoding, Rendering, Highly Threaded Apps - On paper it's supposed to...

My buddy also has a 2600K, and loves it, and has compared them both head to head, he said that for many people who are looking to spend $300 on a processor, the 2600K is the way to go, as it trades blows with the 980x. But for those who use multi-threaded apps, and do a lot of rendering and encoding, the 980x is hands down the better processor.

The sheer performance of the 2600K is trading blows with the performance of a $1000 processor for $700 LESS, Intel must be doing something right.

http://www.anandtech.com/bench/Product/287?vs=142

I'm wondering if it's really worth it for me to wait for BD or jump on this 980x deal.
 


I doubt very much whether a BD 8 core model will out perform the 980x across the board in the situations you outlined, but their 8 module/16 core CPU should, and it will be interesting to see what they price it at and when it is available.

 


You're right, that's why I'm not even considering the 2600K.

Since the 16 core Opteron is a server processor, I can't see that being under $1000.
 


You are confusing things. First you say price matters. Then you go on to say that intel has better cores while ignoring the price factor.

Phenom 1100T = ~$250
i7 980 = ~$1000

Yes, the 980 is faster, but is it 4X faster?

If you are allowed to look only at performance and claim a winner, then I could look only at price to claim a winner. We would both be right, given our differing views.

But if you look at price/performance, AMD is the winner.

And if you believe that performance trumps price performance, then why aren't more people buying the 980?

 
Does anybody really care if they are "cores" or if they are "not cores"? If a BD CPU outperforms a SB CPU at the same price point, thermal envelope, and preferably with a similar OC headroom, i dont care what you call them, and nobody should. If AMD is calling them cores, they are cores, if AMD is calling them Frosted Flakes, they are Frosted Flakes.

@JF, i wouldnt say that Intel is throwing more clock speed at the problem. Intel is currently about one node ahead of AMD. Smaller the node, higher clock speed potential. Intel is not throwing higher clock speeds into their CPU's to necessarily solve a problem, just because its the natural progression of things. If Intel went from 90nm to 32nm and kept the clock speeds at 1.8 GHz, that would be very dumb. AMD on the other hand has nothing else to do but add cores. They cant raise the clock speed too much, its already at 3.5 GHz for the quads. And trust me with my OC experience with my X4 955, it doesnt scale too well over 3.7 GHz. So clock speed is out. AMD cant just roll out a new arch to fix Phenom, that takes time, hence BD. The only real thing left is to add cores. Intel raises clock speed because its just the natural progression of a shrinking node. AMD add cores because theres nothing else to add, and it does make AMD competitive in some niche markets where high core count is important.

@kg2010 vs. JF You're both kinda right. As ive stated many times before, clock for clock, core for core, Intels current cores leave AMD in its dust. However, when you consider the fact you can buy 4 1100T's for the price of one 980x, i think the winner is obvious. I mean, you could buy 2 different sets of the 1100T+RAM+Mobo for the price of just 1 980x individually.

@Price vs Performance, one thing we all have to realize is that we enthusiasts make up a fraction of the market. I actually walk into bestbuy saying they need more RAM because they ran out of room for their music :lol: If you tell 100 people walking into bestbuy, "This computer can do 100,000,000 computations per second with 2 GB of VRAM and a total of 10 GB of RAM with a 32nm CPU, and only costs $1,500" and "This computer has a big screen and can store all your music and videos and cost $750" which one do you think they would buy? Not to mention if they are mildly educated, and they know that more is better, would they rather have a 1100T 6 Core at 3.3 GHz with 8GB of RAM for $1000 or an i7 Quad core at 2.8 GHz with 6 GB of RAM for $1250? 6>4, 3.3>2.8, 8>6, which one are they going to think is faster?
 


Isn't BD the new architecture to carry AMD forward after the K10?

@kg2010 vs. JF You're both kinda right. As ive stated many times before, clock for clock, core for core, Intels current cores leave AMD in its dust. However, when you consider the fact you can buy 4 1100T's for the price of one 980x, i think the winner is obvious. I mean, you could buy 2 different sets of the 1100T+RAM+Mobo for the price of just 1 980x individually.
Yes but there are other choices one can make to get performance that exceeds the 1100T, without having to get the 980x and thus the crazy price. These choices will be even more plentiful once the Sandy Bridge chipset fiasco is behind us. 😀
 
The problem I forsee is this:

1: The majority of users aren't going to be seeing the performance benifits from BDs Intiger units [which seem more optimized for the server market...]

2: Scaling as you add more cores declines over time [specifically, it gets harder to keep all the cores working.] As such, as you add more cores, you get less benifit. I know in studies, 16 cores is about the limit where you would expect scaling to essentially stop being effective...

I expect BD to pull ahead in most benchmarks that scale well, simply due to its extra theoretical horsepower. In everyday use, I expect SB to be faster. The Floating point bottleneck and issues with scaling may well be BD's achillies heel.
 
price/perormance matters to me... to some users it doesn't becasue money doesn't matter to them


news flash every consumer is different.

I started with a good msi mobo and a athlonII x4 640 because price/perf was great there, i then picked up a used 955 and a cm 212+ clocked it to 3.6 and i think it is an excellent machine for the money.

if i had 4-5 k 980 would be the way to go ... btu truthfully if i had that much money i'd probably buy somehtign more useful... like a newer mroe efficient truck 😉

my small point is this bd will come, it will preform how it performs, it will be priced accordingly, and i'll still be around here on the forums telling peopel to buy whichever preocesser performs the best in thier price range, get the system that fits your budget... halo products are a reality for a very small % of the cpu aware market so why cater to them? thats my $0.02
 

I personally hate turbo when doing benchmarks... it only is supposed to work in certain situations(like 3 cores getting boosted with turbo core).

I guess since the X6 gets a higher boost when it can boost then the benchmarks are a little more even, but it's still not X clockspeed compared to X clockspeed with both turbos enabled.
 


I got off track talking about 2 separate points. A product's performance vs the competition also determines its price, though when you have the best processor available, it's not always incremental based on speed.

I agree wholeheartedly that price matters, thus I went with a 930 over a 980x, I didn't see the point at the time in spending $700 extra for it. More importantly, I also chose a 930 over a 1090T, and it was only a $30 - $50 difference or less can't remember, they were both close in price back then. In that scenario I chose a Quad core over a 6 core based on performance, as price was hardly an issue between the 2.

Right Now, Intel's cores are better than AMD cores, especially when you take overclocking into account.

The only way AMD continues to compete is by pricing their products cheaper, and making the price / performance, this strategy can't be that profitable long term.



Agreed, hey, if the 8 core BD is priced at $329 for example, then it's fair game to compare it to the 2600K regardless of core count. I don't think anyone can say but it's not fair, it's 8 cores vs 4 cores, in this case, it doesn't matter since they are priced the same. I just wonder what type of pricing BD will have if it spanks a 2600K? $500 wouldn't be unreasonable.



Yea true - I was looking at both scenarios. Price / performance AMD wins comparing a 1090T to a 980x.

But what happens when we now compare a 2500K to a 1090T?
http://www.anandtech.com/bench/Product/288?vs=146

The whole price / performance argument can be seen differently with that comparison, for 99% of performance / gaming rigs today the 2500K is an excellent choice. For Multi-Threaded apps, the 2600K is trading blows with a 980x.

Intel's MAINSTREAM processors are trading blows with previous gen Enthusiast line processors.

I personally know a few people that switched from a 1090T to a 2500K and have seen significant improvements in overall performance.

Plus for new rigs, even AMD guys were recommending the 2500K, and that speaks volumes.

So, we'll see what BD brings to the table, I sure hope it delivers as the competition will be good for everybody.
 
^If it spanks the 2600K it will probably be a $1K FX chip. AMD will use their better performance to price higher like Intel does right now. I mean if it spanks a 2600K it probably will spank a 980X since the 2600K easily competes with the 980X.



Then which is it? We keep hearing one thing and some people say something else.

I would like for somethings to be clearer so we can easily understand what AMD is attempting to do. Just saying "no, that is not true" doesn't help at all.

The confusing thing is that while we heard from early rumors this was CMT we now hear that they are real cores although they are sharing a few more resources than the older K10 or current Intel archs do.

So which is it?
 
Status
Not open for further replies.