AMD Bulldozer Finally!

LeJosh · 2 Oct 2011 at 21:51

Not in today's world no but once developers develop engines that are designed to run on multiple threads (i.e. X: Rebirth) then we may see good bumps in performance from either team.

There aren't many games out just now that use more cores/threads.

Martini1991 · 2 Oct 2011 at 21:54

LeJosh said:
Not in today's world no but once developers develop engines that are designed to run on multiple threads (i.e. X: Rebirth) then we may see good bumps in performance from either team.

There aren't many games out just now that use more cores/threads.

And I've said previously that BD's modular design, with the ability to "Add cores easily" means that in the future, it'll be very good. Right now? Not so much.

grandmaster · 2 Oct 2011 at 22:20

The 8170 is out in January.
The ones before that are out in October.
8150 is performance wise an Intel 990X or something. According to the benchmarks.
The rest is nigh on par with an i7 2600k
I just hope prices will come down for the i7 as that is what i will be going for.
Always found Intel to be more stable than AMD in the long run.

CAT-THE-FIFTH · 2 Oct 2011 at 23:25

It seems the Passmark scores for the Core i7 2600K were for overclocked CPUs.

Here is a comparison with a Core i7 2600K at stock speeds:

http://www.overclock.net/15137769-post6861.html

The stepping of the Bulldozer CPU is not indicated in the chart.

This post mentioned Bulldozer will have a reasonable IPC increase:

http://semiaccurate.com/forums/showpost.php?p=136387&postcount=2208

Martini1991 · 2 Oct 2011 at 23:28

CAT-THE-FIFTH said:
It seems the Passmark scores for the Core i7 2600K were for overclocked CPUs.

Here is a comparison with a Core i7 2600K at stock speeds:

http://www.overclock.net/15137769-post6861.html

This post mentioned Bulldozer will have a reasonable IPC increase:

http://semiaccurate.com/forums/showpost.php?p=136387&postcount=2208

EDIT : Sounds like a very solid piece of evidence

, I hope it's true, but I can't see it.

And, you didn't read did you?

All @ 3.6GHz
All have 570GTXs and....AMD FX loses both in 2d graphics and 3d graphics

i7 950 has hyper-threading turned off

Also, all of them have 4GB and again AMD FX loses(in mem test)

CAT-THE-FIFTH · 2 Oct 2011 at 23:34

Martini1991 said:
EDIT : Sounds like a very solid piece of evidence , I hope it's true, but I can't see it.

And, you didn't read did you?

Yes I did read the whole thread.

The same chap who post the chart mentioned something else:

http://www.overclock.net/amd-cpus/791495-bulldozer-blog-live-687.html#post15138181

"Originally Posted by flashtest:
If they pass i5, in rendering/gaming (3d studio specially) i have two less motherboards to sale and two bulldozers to preorder "

"Seronx:It does"

Martini1991 · 2 Oct 2011 at 23:36

CAT-THE-FIFTH said:
Yes I did read the whole thread.

The same chap who post the chart mentioned something else:

http://www.overclock.net/amd-cpus/791495-bulldozer-blog-live-687.html#post15138181

"Originally Posted by flashtest:
If they pass i5, in rendering/gaming (3d studio specially) i have two less motherboards to sale and two bulldozers to preorder "

"Seronx:It does"

This is his evidence

3D Graphics test then extrapolating 1024x768x32 to 1920x1080x32

Ergo, pointless crap, we may's well wait for real benchmarks, we don't all sit around playing pass mark. I want to see how it compares to current AMD CPU's with multi-gpu configurations etc.

FrenchTart · 2 Oct 2011 at 23:40

Martini1991 said:
Not really, it proves that throwing cores isn't a solution to lower IPC.

My point was that 4.6ghz vs 3.8ghz wasn't a fair comparison.

IPC is of course part of the issue but the main reason that quad core isn't suffering compared to hex core is that apps aren't making use of the extra cores.

The way things currently work you're of course right. Sadly, this isn't likely to change in the near future either.

Turning my comment into something other than what it was is a bit tiresome though.

Martini1991 · 2 Oct 2011 at 23:42

FrenchTart said:
My point was that 4.6ghz vs 3.8ghz wasn't a fair comparison.

IPC is of course part of the issue but the main reason that quad core isn't suffering compared to hex core is that apps aren't making use of the extra cores.

The way things currently work you're of course right. Sadly, this isn't likely to change in the near future either.

Turning my comment into something other than what it was is a bit tiresome though.

I didn't intend to do that.
Of course your 2500k's at a higher clock, ergo better etc.

But, AMD have went the "Add more cores way", if you're noticing gains with less core, albeit a higher clock, then it doesn't bode well for adding more cores at present.

Although it's not an exact science.
4 x 4.6 = 18.4
6 x 3.8 = 22.8

With less theoretical processing power, you see gains, clocks and cores don't amount to much if the IPC isn't there, that's what I'm getting at. You're going to argue, even though you know it's right

.

grandmaster · 2 Oct 2011 at 23:46

Guys if you google, an extensive list of benchmarks and performance conclusions have come out for the bulldozer

Martini1991 · 2 Oct 2011 at 23:47

grandmaster said:
Guys if you google, an extensive list of benchmarks and performance conclusions have come out for the bulldozer

Likely fake, or used to hype it up, same as with any launch (lol).

sunama · 2 Oct 2011 at 23:57

Have AMD issued an actual (official) release date for Bulldozer?

Martini1991 · 2 Oct 2011 at 23:57

sunama said:
Have AMD issued an actual (official) release date for Bulldozer?

Nope.
There's a slated date which has been leaked, but it's not official announced.
There's a Youtube video where they confirm October for some region or another, I forget.

tommybhoy · 3 Oct 2011 at 02:28

Martini1991 said:
There's a Youtube video where they confirm October for some region or another

Asia.

grandmaster · 3 Oct 2011 at 03:00

A. I doubt the benchmarks are faked. Check them out.
B. Bulldozer is being shipped due for release between Oct 12-24
c. 8150 Available for launch.
D. 8170 - Release date Q1 2012

CAT-THE-FIFTH · 3 Oct 2011 at 09:51

It seems Piledriver might be on socket AM3+ which is good news if true:

http://translate.google.fr/translat...esil-Bulldozer-mimarisi-icin-ilk-detaylar.htm

Martini1991 · 3 Oct 2011 at 09:54

CAT-THE-FIFTH said:
It seems Piledriver might be on socket AM3+ which is good news if true:

http://translate.google.fr/translat...esil-Bulldozer-mimarisi-icin-ilk-detaylar.htm

With the majority of specs being the same as Zambezi and on the same chipset? Generally AMD release new chipsets for releases.

If it's true, then fair enough, but we're already at a point where AMD's backwards compatibility is affecting progression.

gareth170 · 3 Oct 2011 at 10:17

well i did say am3+ might be here longer than people may think.

nitram100 · 3 Oct 2011 at 11:54

I'm happy that my AM3 board supports AM3+ now due to a bios update from gigabyte, but part of me wishes AMD would just design a completely new socket and really compete with Intel in raw speed and performance.

nightmare99 · 3 Oct 2011 at 11:54

Pah only PCI Express Gen2