AMD to launch desktop Bulldozer "Zambezi" at E3 Show in L.A. (June 7-9)

[Postmodum]

Interestring graphic right here...
( My guess is that the big-toys-for-big-boys cpu's will be only released for Q4 )

[Hornet331]

hm, I think, this is not about permormance against SB 2600k, kts only overall price position for desktop pc.

Price is based on performance... amd always priced its cpus accordingly to intels, they where a bit cheaper but offered same performance.

I call bollocks on that chart, even if its overall system price, the only differene in price will be the cpu and the mainboard. And that mainboads with same features are usually only 30-40€ more expensive on the intel side, I doubt AMD will sell the top of the line unlocked 8core BD for sub 300€.

It makes no sense to artificially cut your margines, when you have a superior product.

[FlanK3r]

here is not dount about performance Zambezi in rendering. In rendering will be more close to Gulftown than SB...Some theoretical calculations exists at XS from Informal and others.

[Postmodum]

Well... it's the market they need to win ... Let's see ...
( it's hard to believe yes ... )

[Vinas]

Boxes
http://www.amdforum.se/artikel/exklu...och-fx-4-1025/

Those boxes look really nice in the way that it incorporates ATi's lineage into the product. Personally, I'm holding off on SB to see what the big bad bulldozer will do. Which ever is better I will buy, but having used intel ever since I upgraded from my AMD K6 II, I kind of miss AMD these days. Here's to hoping they release a real alternative to SB!

[Particle]

I call bollocks on that chart, even if its overall system price, the only differene in price will be the cpu and the mainboard.

I don't know if the chart is legit or not, but I don't think your reasoning here is sound. More expensive systems from OEMs tend to include things like bigger hard drives, larger monitors, and faster video cards by default. I'd be surprised if we see a great deal of configurations with 8-core BD processors using a GX chipset for graphics, for instance. It's not to say there won't be any, but hardware of that class is very likely to be bundled with mainstream discrete graphics.

If anything, I think the pricing (should it be true) is a great indicator that the CPUs will be affordable instead of carrying FX pricing of yesteryear.

[Manicdan]

i notice that the numbers are very low for the FX series, its core count followed by 110 with the exception for 81x0, so it seems that AMD is showing its lowest end of each BD cpu vs SB. (65/95w versions maybe?)

also being that its system price, if its 800$ total for a 2500k or a 6110 system, chances are that could mean its a 200$ intel cpu vs a 250$ amd cpu since intel boards are generally more expensive.

[Oliverda]

AMD Aims to Fight Core i7 “Sandy Bridge” with Bulldozer.
http://www.xbitlabs.com/news/cpu/dis...Bulldozer.html

This chart was made by X-bit labs if I'm not mistaken.

[Hornet331]

No its just bad business if you sell your product under value. If 8 core BD can deliver nearly the same performance or better as a hexacore form intel and selling it for less then a mainstream quadcore that would make your investors more then mad... you just waste 200€ of margine...

I can see the 8core BD somewhere around the 500€ mark, cause thats where intels hexacores start. So if we see the lower clocked octocores starting with 400€ and then gradualy increasing its price with performance, you get a nice margine, and offer better performance for the money you spent.

[demonkevy666]

Interestring graphic right here...
( My guess is that the big-toys-for-big-boys cpu's will be only released for Q4 )

looks like it's most of the system price.

Price is based on performance... amd always priced its cpus accordingly to intels, they where a bit cheaper but offered same performance.

I call bollocks on that chart, even if its overall system price, the only differene in price will be the cpu and the mainboard. And that mainboads with same features are usually only 30-40€ more expensive on the intel side, I doubt AMD will sell the top of the line unlocked 8core BD for sub 300€.

It makes no sense to artificially cut your margines, when you have a superior product.

well that's true for amd, but intel has a huge manufacturing advantage.

No its just bad business if you sell your product under value. If 8 core BD can deliver nearly the same performance or better as a hexacore form intel and selling it for less then a mainstream quadcore that would make your investors more then mad... you just waste 200€ of margine...

I can see the 8core BD somewhere around the 500€ mark, cause thats where intels hexacores start. So if we see the lower clocked octocores starting with 400€ and then gradualy increasing its price with performance, you get a nice margine, and offer better performance for the money you spent.

really everyone is hoping $300 from what I've seen.

[informal]

Here is an interview with Macci And John Taylor at cebit11. Judging by their tone in the interview,to me it looks like AMD may price FX parts aggressively against intel SB and Gulftown parts even if FX performs somewhat better. Folks need to remember that in client space more cores don't have a proportional impact in performance numbers as in server space-much less impact in average sum of workloads as can be seen here. SB 2600K has ~8% higher IPC,3/5% higher clock(def/Turbo) and 50% less cores and is only around 4.7% slower than 980x. It is also more than 3x cheaper than 980x. Which is a better value? SB of course,unless you do some insanely parallel workloads on your PC when 980x is much faster. Which is a faster chip? 980x of course,but due to nature of workloads SB hold very well and even outperforms 980x in many applications!

Same will go for FX BD parts(relative to 980x).Notice what Mr Taylor said in the interview @ 2:18 and onwards : " BD is designed to offer 30-50% better performance in the same TDP and roughly the same die area as AMD's previous designs". This probably goes for single and MT workloads,respectably. So for AMD it's important that they hit 40+% range of better performance Vs Thuban.Why? Well,if you look carefully at the chart I linked above,you will see that AMD precisely needs around 34% better performance overall in order to match 980x in desktop workloads. They need 28% overall better performance to match SB 2600K. If they hit north of 40% range,they will be combating and maybe even beating 990x Gulftown part and be very close to SB 6 core products. All this within 125W TDP range. They can push for 140W "XE" part and go for 5% more performance (from 3.5Ghz to 3.7-3.8Ghz?) and have a "halo" product,but this depends on their ability to hit these speeds and their marketing department.

[Tomasis]

True, but my point is that Bulldozer might be an underwhelming flop.
I just fear with so much hype going on for so long that AMD might be setting themselves up for a big fall.
John

cmon!

hype created from internet..

loads of people 90% customers PC dont know about hype.. they buy when it is available

you judge amd future by hype created by some computer typing people behind screen, lol

[Manicdan]

snip

so i think what your saying is that clock speeds are becoming the main price factor for client systems instead of core count. which does make sense.

sure back in the day 1 -> 2 cores was very well noticed
2 -> 4 cores is quite noticeable
4 -> x cores is not so noticeable unless you have specific tasks for it.

however i hope that the perf potential of BD is able to keep up with SB in perf/mm2 since larger sized SB chips are coming, and could drive down the price of BD, or just inflate the prices of higher core count SBs

[-Boris-]

Here is an interview with Macci And John Taylor at cebit11. Judging by their tone in the interview,to me it looks like AMD may price FX parts aggressively against intel SB and Gulftown parts even if FX performs somewhat better. Folks need to remember that in client space more cores don't have a proportional impact in performance numbers as in server space-much less impact in average sum of workloads as can be seen here. SB 2600K has ~8% higher IPC,3/5% higher clock(def/Turbo) and 50% less cores and is only around 4.7% slower than 980x. It is also more than 3x cheaper than 980x. Which is a better value? SB of course,unless you do some insanely parallel workloads on your PC when 980x is much faster. Which is a faster chip? 980x of course,but due to nature of workloads SB hold very well and even outperforms 980x in many applications!

Same will go for FX BD parts(relative to 980x).Notice what Mr Taylor said in the interview @ 2:18 and onwards : " BD is designed to offer 30-50% better performance in the same TDP and roughly the same die area as AMD's previous designs". This probably goes for single and MT workloads,respectably. So for AMD it's important that they hit 40+% range of better performance Vs Thuban.Why? Well,if you look carefully at the chart I linked above,you will see that AMD precisely needs around 34% better performance overall in order to match 980x in desktop workloads. They need 28% overall better performance to match SB 2600K. If they hit north of 40% range,they will be combating and maybe even beating 990x Gulftown part and be very close to SB 6 core products. All this within 125W TDP range. They can push for 140W "XE" part and go for 5% more performance (from 3.5Ghz to 3.7-3.8Ghz?) and have a "halo" product,but this depends on their ability to hit these speeds and their marketing department.

Just one thing. 40% better performance is not the same thing as 40% better bench results. A 5GHz processor don't get twice as good results as a 2.5GHz processor. So if 980X has 40% better results than Phenom X6, it's probably has much higher performance advantage.
BD needs much more than 50% better performance to compete with 990X.

People will once again accuse a company of lying about performance just because thay don't realize these things. I don't say you don't understand this, but there are many others out there.

[informal]

Just one thing. 40% better performance is not the same thing as 40% better bench results. A 5GHz processor don't get twice as good results as a 2.5GHz processor. So if 980X has 40% better results than Phenom X6, it's probably has much higher performance advantage.
BD needs much more than 50% better performance to compete with 990X.

People will once again accuse a company of lying about performance just because thay don't realize these things. I don't say you don't understand this, but there are many others out there.

There are a variety of benchmarks for desktop out there.Many of them are biased one way or the other,many of them have little to do with real world workloads. But at the end of the day,when an Average Joe user reads TR,AT or any other HW website and reads BD review ,he will probably look for the one page that says summary of results.There he will see an overall average score relative to other competing parts. Even if BD trounces other chips in say several heavy fp workloads(rumored CB11.5 score of ~11 ?) it may still be a bit slower in some non-thread integer heavy workload(still to be seen though). This may bring down the average score and put it just a bit over ,or maybe even under, 980x.

Users will have to look at particular tests and what they mean for them. If the test is relevant for you in some way and chip performs as you expected and the pricing fits,there is no reason not to pick it up. I really do expect Zambezi to be a very good desktop chip. Even if it is not overall the fastest thing on desktop,it just needs to be competitive in performance/high end segment. People will buy it.

[-Boris-]

There are a variety of benchmarks for desktop out there.Many of them are biased one way or the other,many of them have little to do with real world workloads. But at the end of the day,when an Average Joe user reads TR,AT or any other HW website and reads BD review ,he will probably look for the one page that says summary of results.There he will see an overall average score relative to other competing parts. Even if BD trounces other chips in say several heavy fp workloads(rumored CB11.5 score of ~11 ?) it may still be a bit slower in some non-thread integer heavy workload(still to be seen though). This may bring down the average score and put it just a bit over ,or maybe even under, 980x.

Users will have to look at particular tests and what they mean for them. If the test is relevant for you in some way and chip performs as you expected and the pricing fits,there is no reason not to pick it up. I really do expect Zambezi to be a very good desktop chip. Even if it is not overall the fastest thing on desktop,it just needs to be competitive in performance/high end segment. People will buy it.

BD will still need more than 50% higher performance than Phenom II to compete with 980X. Of course there are differences between loads. But if 980X score 50% higher in benches than Phenom II, then BD needs maybe 70%+ higher performance to get these 50% higher results.

A 4GHz chip has twice the performance of a 2GHz chip, but might only have 40% higher FPS. You can't just look at benches and think that they accurately show the differences in performance for a single part of the system.

[informal]

By "FPS" I believe you think of gaming? If so games are and will be mostly GPU bound.Those that still depend on CPU will probably see a big boost from improved uarchitecture BD brings. I think BD will have no problems being among top dogs in gaming.
Also,BD probably will be 70+% faster in some tests than Phenom II. The rumored score of 11 in CB11.5 ,versus 6 core Thuban(not 4 core Deneb based Phenom II), hints at this. People these days mostly look at benchmarks. When AMD states that Zambezi was designed to offer 30-50% higher performance than Thuban then it will offer that much performance in real world applications.How much better it can be in theory,if properly optimized for,is anyone's guess. With proper FMA optimization or if ie. targeted for AES,it may be even 2x faster Vs non-optimized results.

[-Boris-]

By "FPS" I believe you think of gaming? If so games are and will be mostly GPU bound.Those that still depend on CPU will probably see a big boost from improved uarchitecture BD brings. I think BD will have no problems being among top dogs in gaming.
Also,BD probably will be 70+% faster in some tests than Phenom II. The rumored score of 11 in CB11.5 ,versus 6 core Thuban(not 4 core Deneb based Phenom II), hints at this. People these days mostly look at benchmarks. When AMD states that Zambezi was designed to offer 30-50% higher performance than Thuban then it will offer that much performance in real world applications.How much better it can be in theory,if properly optimized for,is anyone's guess. With proper FMA optimization or if ie. targeted for AES,it may be even 2x faster Vs non-optimized results.

When AMD says 50% higher performance they are talking about total throughput, that's what they have said all the time. And I doubt that they mean some average benches. If I know AMD right they are talking about numbers pretty close to best case scenario. And a rumored benches is not something we should base performance estimates on.

I bet that AMD will have a very nice processor and I will by it, but I dont think it will compete with the best six core i7s. Especially not the SB ones.

[informal]

Since we have no real Zambezi benchmark number it's unknown how it will perform.It may be a total flop for all we know . But AMD knew a 1+ year how i7 Gulftown performs and they still call Zambezi their high performance part aimed at enthusiasts .They also revived FX branding.This at least suggests the performance should be close to what Gulftown offers.It may not be able to beat it completely but as I said before AMD just needs to be competitive. Intel will have 3.46Ghz Gulftown when Zambezi launches so it will be a hard task to equal this monster,let alone beat it.

[FlanK3r]

BORIS: in rendering cant SB better than Zambezi...Sometimes is 1100T better than 2600k in render, example 3DsMax scenarios, POV Ray.

[AbortRetryFail?]

Price is based on performance... amd always priced its cpus accordingly to intels, they where a bit cheaper but offered same performance.

I call bollocks on that chart, even if its overall system price, the only differene in price will be the cpu and the mainboard. And that mainboads with same features are usually only 30-40€ more expensive on the intel side, I doubt AMD will sell the top of the line unlocked 8core BD for sub 300€.

It makes no sense to artificially cut your margines, when you have a superior product.

Sure it does. Welcome to Capitalism 101: May I introduce you to ... The Demand Shift ?



By setting a price point (in the case of the graph - Price at $3 instead of $4) at a 25% reduction, Demand is increased 60% from 15m units to 24m units.

In semiconductors, the more of something you make, the more efficient your production. Yields increase, and your margins will actually grow at higher production levels (even with the higher cost of base materials) with increased demand.

This is how AMD gathered market share on the graphics side (by fighting at price points) with nVidia.

From what I have seen, AMD also feels that 'gate-first' is much more efficient, simple and cost-effective than 'gate-last' in their process. One would hope that greater production would only extend process efficiency and build higher yields with better margins.

[ichime]

I hope that the product is as exciting as the box it comes in, cuz that box art kinda rekindled my enthusiasm for AMD a little bit...

[Hornet331]

Sure it does. Welcome to Capitalism 101: May I introduce you to ... The Demand Shift ?



By setting a price point (in the case of the graph - Price at $3 instead of $4) at a 25% reduction, Demand is increased 60% from 15m units to 24m units.

In semiconductors, the more of something you make, the more efficient your production. Yields increase, and your margins will actually grow at higher production levels (even with the higher cost of base materials) with increased demand.

This is how AMD gathered market share on the graphics side (by fighting at price points) with nVidia.

From what I have seen, AMD also feels that 'gate-first' is much more efficient, simple and cost-effective than 'gate-last' in their process. One would hope that greater production would only extend process efficiency and build higher yields with better margins.

Price is only one factor... again you just dont read what i have posted.... if you have a product that has better performance then your competitors product, you already generate increased demand.

So tell how this will make economical sense to price a bigger chip that will perform better @ the same price point as a smaller chip that performs worse... oonly to get more marketshare...?

oh yeah right lets us play the "marketshare at all cost" game that hector played back in 2006 again.... turned out pretty go.. oh wait...

AMD at that time was manufacturing at full capacity and still made a loose... now we have a total different situation, amd doesn't has to be concerned anymore with making the production more efficent, since they are only customers at a external fab. Its there problem to make the process as efficent as possible.

ATI gathering marketshare was mostly due to the fact that NV was late to the game and had no card that could compare to ATIs offerings in the midrange for 5 months. Plus there chip was smaller and offered nearly the same performance for nearly the same price (a bit cheaper)... see where that goes again...

Economics 101 is nice in theory, but fails at real markets where much more then price matters to sell a product.

[Kristers Bensin]

When AMD says 50% higher performance they are talking about total throughput, that's what they have said all the time. And I doubt that they mean some average benches. If I know AMD right they are talking about numbers pretty close to best case scenario. And a rumored benches is not something we should base performance estimates on.

I bet that AMD will have a very nice processor and I will by it, but I dont think it will compete with the best six core i7s. Especially not the SB ones.

wow so much nonsense in your latest posts. u should read your own posts again and u might see what i point at You should also read more on bulldozer design if you like the concept.

[SMTB1963]

...It makes no sense to artificially cut your margines, when you have a superior product...

...No its just bad business if you sell your product under value. If 8 core BD can deliver nearly the same performance or better as a hexacore form intel and selling it for less then a mainstream quadcore that would make your investors more then mad... you just waste 200€ of margine...

You know, I've always wondered why I've been so angry with AMD since buying their stock many years ago. Turns out that all this time, they could have charged more for their processors - but they didn't! What a bunch of idiots. Instead of keeping their margins artificially low, AMD could've been paying out a big fat dividend to keep me happy...right?

I can see the 8core BD somewhere around the 500€ mark, cause thats where intels hexacores start.

So...your position is that AMD should price its chips the same as intel's chips of equivalent performance? That's your idea of a smart business move for AMD?!? Really?!?!