AMD FX 8 Core and 4 Core Processor Systems Seen Running at E3

I'm really excited about finally building an AMD system, and I think it will be right around the time BD comes out. Haven't bought AMD since like 10 years ago, when I got burned with a string of bad Athlon chips (probably like 4 in a row, ouch!). I finally want to give them another shot. I hope they come correct with this release.
 
Mackowitz said:
Does voltage matter? Serious question, not a troll or fanboi. If a 140w cpu ran at 1.4v@100amps or 0.7v@200amps or 2.8v@50amps, they all draw the same power, is there a difference?
Click to expand...
Because that's not the issue or how it works in this case.

It's a sign of what it takes to even get the CPU up to 3.6GHz (Turbo Core?) on the stepping shown. It's likely that particular CPU shown was not a "125W" model, which is the real problem. Unless GF is unique in its position of having a high performance 32nm process designed to use higher voltages than where it left off at 45nm, then either the process or processor stepping are not working too well at this point.

This is an example of why AMD is not happy with the performance of the current stepping, recently calling it uncompetitive. A 125W CPU would ship at even lower speeds than what is essentially an over-volted, who-cares-what-power-it-uses CPU shown at E3.

Example shmoo plot: http://www.realworldtech.com/includes/images/articles/cell3-6.gif Notice 2 things: the edge of where frequency reaches maximums at a given voltage and how power rises quickly at each speed as voltage is increased (but the upper range of frequency is increased).

Seriously, even when AMD gives bad news directly (and demonstrates it with a screenshot), does anyone here take it at face value and not try to hand wave the problem(s) away? LOL Hopefully for AMD, the B2 revision does a lot better than what was shown at E3. Next month!
 
Corvette said:
For me, BD could offer an overclockable budget chip. I already have a Corsair H50 on my C2D 6300. What I'd like in my next computer is 4 cores which the low cost Intels don't have and the option to OC. This is especially nice if the 4 core AMD performs similar to the 2 core i3 before the OC and at about the same price. Throw in a nice motherboard and a bunch of RAM and I'll have a nice upgrade along with some options for upgrades in a few years.
Click to expand...

Here's the hitch though - the 4 core BD is rumored to debut at 220, which is roughly the price of the 2500k. In order to lure people in that direction I think they would need to have an IPC advantage (which as far as I can tell they likely won't) or they would need to be able to OC past 4.5ghz without a hassle. Maybe just me but I know I'd have a hard time buying a processor I know is inferior when I could've gotten a 10-20% improvement for the same price. Granted BD was designed for high clocks so hopefully we'll be seeing 5+ghz oc's but who knows. So much about BD performance and capability is still up in the air that we may as well just dream that the 4 core will debut at 120 instead of 220 anyway.
 
Devilpup said:
Here's the hitch though - the 4 core BD is rumored to debut at 220, which is roughly the price of the 2500k.
Click to expand...
Where did you here that the 4 core was 220? Last I heard the 4 core was 190, 6 core was 240, and two 8 core's at 290 and 320.

http://wccftech.com/2011/05/20/amd-bulldozer-llano-fusion-pricing-information-leaked/
http://www.brightsideofnews.com/new...ricing-mid-24300-for-an-8-core-processor.aspx

amdbulldozerlianofiyat_dh_fx57.jpg
 
From here:

http://www.techpowerup.com/146236/AMD-Bulldozer-Llano-Pricing-Surface.html

Has the two 8 core models at 320 and 290, 6 core at 240, and 4 core at 220. For the most part it looks consistent except for the 4 core model.

Given that the rumored clocks for the top model are something like 3.8ghz with 4.2ghz turbo, I can't imagine the 4 core being clocked much higher out of the box. Even if the 4 core will do 4.5ghz without a fuss, it will need an IPC advantage to put it ahead of an intel pick. Otherwise I dare say it would be worth the possible price premium for the performance.

If we take some of the [H] tests on SB here and assume AMD's claims of a 20% IPC improvement over PII, that still puts it behind the intel offerings by a not insignificant margin. Although in that case BD will be looking a lot better than previous offerings as far as IPC, and might actually be competetive as long as the clocks can keep up.
 
Good points. At some point it will be easier to compare each platform directly. I think my cooling will work with either sockets.
 
Devilpup said:
From here:

http://www.techpowerup.com/146236/AMD-Bulldozer-Llano-Pricing-Surface.html

Has the two 8 core models at 320 and 290, 6 core at 240, and 4 core at 220. For the most part it looks consistent except for the 4 core model.
Click to expand...
There is a $50 difference between the lowest 8 core and the 6 core, yet there is only a $20 difference between the 4 core and 6 core. Ya, I would say it looks inconsistent.

Considering AMD has already confirmed the $320 price for the top end 8 core, I would say putting $50 difference between each model makes more sense, especially to the consumer. Otherwise the consumer has zero incentive to buy the lower end model when only $20 more will get you 50% more cores. My bet is the $190 price for the 4 core is closer to the truth than $220.

In a straight talk, AMD confirmed that the pricing for FX series is going to be extremely aggressive, with their king of the hill processor, the FX-8130P coming with a street price of $320.
Click to expand...
 
It will be nice when we are lucky enough to see

PII X4 same clocks benched against FX X4 at same clock, and same with the PII X6's/FX X6's.

The fact that you can get an X4 BE for $105.00 on newegg, vs their newer FX X4 and I think I read it should cost $180.00? To make that a worthwhile upgrade it will really have to perform. Same with the X6->FX X6.

If their IPC only goes up 10 % lol AMD is _ _ _ _ _ _ .:D

They need to be like 25% or better than PII's IMO, across the board.
 
If their IPC only goes up 10 % lol AMD is _ _ _ _ _ _ .

They need to be like 25% or better than PII's IMO, across the board.
Click to expand...

They will have higher clocks @ lower power draw coupled with an IPC increase. I am optimistically hopeful that it will be around 25% so that clock for clock, core for core it will be about the same as the Intel's gen 1 i series processors (Nehalem architecture) at a lower power draw + additional instruction sets.
 
Anyone want my board and processor in my signature cha cha ... if the chips are that badass mines for sale! I know some of you hardcore Intel only fans will want it.
 
Just Start calling the Bulldozer chips...

AMD CONAN

Because it is going to go barbarian on Intel! WOOT!!!
 
tangoseal said:
Just Start calling the Bulldozer chips...

AMD CONAN

Because it is going to go barbarian on Intel! WOOT!!!
Click to expand...

I'd like to know what you're basing this on? It's certainly not performance based on the numbers posted above.
 
RamonGTP said:
I'd like to know what you're basing this on? It's certainly not performance based on the numbers posted above.
Click to expand...

Which were engineering samples and not the same thing you can buy in stores when the time comes ....
 
stefanm said:
Useless numbers, no way retail Bulldozer will score that low.
Click to expand...
The difference between a random site with a bare screenshot and the link I posted is that the persons running the benchmarks are known and the hardware used (motherboard and CPU) is pictured.

From the softpedia link, since you probably didn't click through to see what a real benchmark looks like when the user has the hardware:
http://www.chiphell.com/thread-210890-1-1.html
(XS is down right now) http://www.xtremesystems.org/forums/showthread.php?272022-Amd-Bulldozer-tested-youtube

Sorry. Looks like a total dud when the actual hardware was tested instead of what's shown in fantasy blog "screenshots".
 
That's no "fantasy" blog, OBR has a history of breaking nda and posting numerous leaks, furthermore he is completely biased against AMD so there is no chance of him faking these results just to make BD look better and they aren't that good anyway for an 8 core CPU. Again the results you posted are completely irrelevant they are not representative at all of Zambezis performance.
 
Last edited:
stefanm said:
Again the results you posted are completely irrelevant they are not representative at all of Zambezis performance.
Click to expand...
Benchmarks from known users with proof that they were run are "completely irrelevant"? LOL The benchmarks are not redacted, and both are apparently B0 stepping chips, the same ones motherboard makers had 2 weeks ago at Computex.

Enjoy the fantasy while it lasts. Signs are not pointing to good things for Zambezi.

But wait a second...
I mean, seriously? AMD has called the current state of Zambezi "uncompetitive" and some random blog using unpictured hardware and redacted :rolleyes: leaks show fantastic performance which completely contradict AMD's own assessment, and that's somehow more believable?

/shakes head... it's hopeless
 
Believe what you want dude but BD will be that bad only in your dreams. Top 8c FX will beat the 2600K in heavily MT programs like CB 11.5 and lose in most if not all low thread scenarios.
 
pxc said:
Actual benchmark testing is showing BD performance to be quite disappointing. Enjoy your fantasy.
Click to expand...
Aren't these tests being done on B0 silicon? Is there anyone testing on B1/B2 revisions? The Softpedia link kinda proves the story about B0 not meeting AMD's expectations on performance...
 
JMccovery said:
Aren't these tests being done on B0 silicon? Is there anyone testing on B1/B2 revisions? The Softpedia link kinda proves the story about B0 not meeting AMD's expectations on performance...
Click to expand...

Yes. AMD is still tweaking timings, p-states, power gating and turbo probably.
 
AMD_Gamer said:
It is pretty obvious that the samples that are being tested are sabotaged in some way.
Click to expand...

Of course they are. There is no way AMD could have yet another processor that offers sub-par performance. It's not like they've been doing that for the better part of the last 30 years. Oh wait................ :rolleyes:
 
Dan_D said:
Of course they are. There is no way AMD could have yet another processor that offers sub-par performance. It's not like they've been doing that for the better part of the last 30 years. Oh wait................ :rolleyes:
Click to expand...

"sub-par performance" ?

Thunderbird, Athlon64, Athlon64 X2, FX, Zacate, Llano. Were all better than what Intel had to offer at the time they were released.
 
sirmo said:
"sub-par performance" ?

Thunderbird, Athlon64, Athlon64 X2, FX, Zacate, Llano. Were all better than what Intel had to offer at the time they were released.
Click to expand...

You are talking about a just a few of the releases in the last 30 years.

To me the biggest problem AMD has is transistors/process. AMD can on design brilliant processors but it does not have the money to invest in improving its process that Intel does. At the moment AMD is 16 months behind at 32nm. Yes AMD has SOI but that does not appear to be benefiting them anymore. I mean AMD's 45nm process was no better than Intel's 45nm process. In 8 to 10 months Intel will release 22nm using 3D transistors. Intel spent over 3 billion dollars developing this technology. AMD/GF does not have the cash flow to do that. I do not believe AMD will have 3D transistors for at least 2.5 years.
 
drescherjm said:
You are talking about a just a few of the releases in the last 30 years.

To me the biggest problem AMD has is transistors/process. AMD can on design brilliant processors but it does not have the money to invest in improving its process that Intel does. At the moment AMD is 16 months behind at 32nm. Yes AMD has SOI but that does not appear to be benefiting them anymore. I mean AMD's 45nm process was no better than Intel's 45nm process. In 8 to 10 months Intel will release 22nm using 3D transistors. Intel spent over 3 billion dollars developing this technology. AMD/GF does not have the cash flow to do that. I do not believe AMD will have 3D transistors for at least 2.5 years.
Click to expand...

Well the entire world is behind Intel on the Fab process. Monopoly has been quite nice to Intel. AMD is currently on the same fab process. This is why I time my upgrades to fab process switches. Phenom I was actually the only time I skipped.

GloFlo and TSMC will always be behind on this. I was responding more to the sentiment that AMD has always had "sub par performing" products, which is obviously not true.

God knows Intel even after dumping 3 billion into Larrabee still can't make a decent GPU to save their life.
 
AMD_Gamer said:
It is pretty obvious that the samples that are being tested are sabotaged in some way.
Click to expand...
Yeah, by design and production. ;)

Whatever is wrong with those was apparently not fixed in B1 since those samples were not given to motherboard manufacturers after AMD got those back from GF (according to mobo makers at Computex... lesson here: keep your direct customers happy and don't string them along like end users).

Power consumption/clock speed improvements are supposedly coming in B2 next month, right? I don't see how that will change overclocked benchmarks that are available now but we can wait and see.

Still, it is a bit of wishful thinking to believe that performance is going to radically change within a few weeks vs overclocked samples.

Lemme tell ya how the July preview is going to go: "gaming benchmarks" using GPU limited settings will show that an overclocked BD FX is on par with some SB "X" model. People will dance with joy. Oops, spoilers!
 
sirmo said:
Well the entire world is behind Intel on the Fab process. Monopoly has been quite nice to Intel. AMD is currently on the same fab process. This is why I time my upgrades to fab process switches. Phenom I was actually the only time I skipped.

GloFlo and TSMC will always be behind on this. I was responding more to the sentiment that AMD has always had "sub par performing" products, which is obviously not true.

God knows Intel even after dumping 3 billion into Larrabee still can't make a decent GPU to save their life.
Click to expand...

Couple things wrong about your argument.

1) This is the year 2011 not 2005
2) We're talking about CPU's not GPU's

I guess if you talk about completely different products or a brief history in time more than half a decade ago, you might find some notches in AMD's belt. Too bad it's completely irrelevant here.

I'll bet you thought barcelona was gonna walk all over intel becuase it was a native quad core too huh?
 
RamonGTP said:
I'll bet you thought barcelona was gonna walk all over intel becuase it was a native quad core too huh?
Click to expand...

As an AMD fanboy I was so frustrated when AMD converted their second Dresden Fab from the 90nm process to the 65nm process when Intel had already begin using the 45nm process. I though this was the best chance to catch up.
 
sirmo said:
"sub-par performance" ?

Thunderbird, Athlon64, Athlon64 X2, FX, Zacate, Llano. Were all better than what Intel had to offer at the time they were released.
Click to expand...

That's just a small sampling of products produced over a relatively short time span out of the last 30 years. My point is that AMD has a history of making promises and falling short on delivery. They also have a history of making sub-par performing processors. Note I didn't say that they weren't a good value at certain price points because that isn't generally the case. However AMD is forced to sell certain processors as loss leaders just to pay the bills and they can't compete in the same price points Intel does at present. They've been unable to do so during most of the last 30 years. Anyone who fails to see this is blind. AMD bringing back the FX name sounded like a good sign until all evidence pointed to them selling these processors in the sub-$350 price point. That tells us all we need to know about the performance of Zambezi. If Zambezi were competitive enough with Sandy Bridge (and could be fixed with a minor revision of the silicon B0 > B1 > B2) then it would likely cost more than Sandy Bridge with some parts costing around the same and offering more performance, or slightly less expensive with comparable performance. The fact that we are seeing pricing around that of the Core i7 2600K isn't really all that good for an 8 core processor in my mind. If AMD were able to deliver on their promises of matching Nehalem than they'd be offering four or six core performance at that price point and 8 core CPUs would be priced in line with Gulftown's lesser priced siblings. (IE: Core i7 970.) I'm sorry but 8 "real cores" being roughly competitive with 4 real cores and 4 logical cores is kind of sad.

And I've got to say that after looking at those benchmarks and comparing them against Sandy Bridge (The one's from Chiphell which is the most reliable source I've seen quoted thus far) I have to say that Bulldozer / Zambezi are shaping up to be the next Phenom (Again you can evidently spell fail with a "PH".) all over again. The fact that they are revising the silicon so much so close to "launch" is relatively scary too. I don't think a B2 / B3 revision is going to get it done. But we'll see. AMD has been full of surprises here and there throughout their history, but they've also been relatively predictable most of the time as well.

drescherjm said:
You are talking about a just a few of the releases in the last 30 years.

To me the biggest problem AMD has is transistors/process. AMD can on design brilliant processors but it does not have the money to invest in improving its process that Intel does. At the moment AMD is 16 months behind at 32nm. Yes AMD has SOI but that does not appear to be benefiting them anymore. I mean AMD's 45nm process was no better than Intel's 45nm process. In 8 to 10 months Intel will release 22nm using 3D transistors. Intel spent over 3 billion dollars developing this technology. AMD/GF does not have the cash flow to do that. I do not believe AMD will have 3D transistors for at least 2.5 years.
Click to expand...

Exactly.

drescherjm said:
As an AMD fanboy I was so frustrated when AMD converted their second Dresden Fab from the 90nm process to the 65nm process when Intel had already begin using the 45nm process. I though this was the best chance to catch up.
Click to expand...

Well going to 45nm would have cost AMD a lot more than going to 65nm and they probably hadn't designed their products around 45nm at that time. I just don't think it would have ultimately benefited them enough at that time to justify the cost.
 
Last edited:
sirmo said:
Considering it was just released and reviewed (today), shouldn't be too long before you can get your hands on one.
Click to expand...
Llano has been "shipping" since March according to AMD. I do have an idea for the reason of mobile parts shipping months ago and lack of laptops for sale on launch day. Think filling up a large bucket a drop at a time. ;)
 
pxc said:
Llano has been "shipping" since March according to AMD. I do have an idea for the reason of mobile parts shipping months ago and lack of laptops for sale on launch day. Think filling up a large bucket a drop at a time. ;)
Click to expand...

~150 OEM products are using Llano. There was a story around few weeks ago stating they were sold out. Limited availability due to high demand.. oh well. :D
 
Dan_D said:
And I've got to say that after looking at those benchmarks and comparing them against Sandy Bridge (The one's from Chiphell which is the most reliable source I've seen quoted thus far) I have to say that Bulldozer / Zambezi are shaping up to be the next Phenom (Again you can evidently spell fail with a "PH".) all over again. The fact that they are revising the silicon so much so close to "launch" is relatively scary too. I don't think a B2 / B3 revision is going to get it done. But we'll see. AMD has been full of surprises here and there throughout their history, but they've also been relatively predictable most of the time as well.
Click to expand...

http://crazyworldofchips.blogspot.com/2011/06/zambezi-es-performance-weirdness.html

http://crazyworldofchips.blogspot.com/2011/06/revisiting-bulldozer-es-weirdness-story.html

Not sure if he's right but his speculation is as good a guess as any of the speculation here. I doubt it'll be Phenom I. I think people just like predicting doom like the world ending dude.
 
InternationalHat said:
http://crazyworldofchips.blogspot.com/2011/06/zambezi-es-performance-weirdness.html

http://crazyworldofchips.blogspot.com/2011/06/revisiting-bulldozer-es-weirdness-story.html

Not sure if he's right but his speculation is as good a guess as any of the speculation here. I doubt it'll be Phenom I. I think people just like predicting doom like the world ending dude.
Click to expand...

Well we hope it won't be another Phenom 1. And that speculation seems relatively plausible from just what I glanced over.
 
You must log in or register to reply here.
Back
Top