AMD Bulldozer Finally!

[LeJosh]

Stop comparing a BD core to a Phenom II core or an Intel core? A module is still two threads, two Phenom II cores = 2 threads, two SB 2500k cores = 2 threads.

Even that aside, a BD module compared to a 2600k core (Both execute 2 threads)
In single threaded app's, the 2600k is bound to beat it, and in two threaded apps the gap becomes a lot closer, possibly better than the 2600k's 1 core, although the slides show that isn't the case, and that it's slower.

It does strike me that a lot of people don't actually know or are misleaded by AMD's marketing strategy regarding modules vs cores.

 

[Final8y]

there was never any question of morality, just deluded ranting

From someone that does not know me personally you have nothing to back that up with.
But you cant read this so you have done nothing but talk without having to face up to your comments.

And it is over because its clear that people cant stop the continuation of making more comments so i will not respond any more to them in this thread.

 

[pmc25]

Scaling won't be linear. It's a totally pointless argument. That's about all there is to say.

I find it interesting that the higher clocked Bulldozers run the NB (not CPU/NB - does this even exist in its prior Phenom II state in BD?) at 2.2Ghz as opposed to 2Ghz in the lower clocked SKUs.

 

[CAT-THE-FIFTH]

What about this?

Doesn't look completely convincing... 4.8GHz with 8 cores the score is 7.8. With my 2600K with 4 cores (8 threads obviously) at 4.6GHz I get 9.04. Or is Cinebench Intel biased?

I know I said I didn't expect Bulldozer to be "that" much better, if at all, but I would expect the scores to be a lot closer, and at 4.8GHz it "should" be overtaking a 2600K...?

EDIT: Hang on... I have just noticed...

Why would an AMD slide showing the scores for Bulldozer chip have "Your System" as in Intel build... I call BS!

EDIT 2: Nevermind, it's the official screenshot from http://www.maxon.net/downloads/cinebench.html. It could be genuine I guess, but I would have thought in an AMD presentation they would have their own screenshot.

AMD would be quite aware of the Phenom II X6 scores too and reviews would have course compare Bulldozer to the Phenom II X6 too.

What if I said I've actually seen BD.

http://www.vortez.net/articles_pages/amd_fx_cpu_preview_event_and_overclocking,1.html


So yes I have seen it but that's all I'm gonna say. Like I've been saying to some people, just wait and see what happens in a few weeks rather than keep speculating.

So have a few other people too on xs who are quite well known.

 

[Martini1991]

It does strike me that a lot of people don't actually know or are misleaded by AMD's marketing strategy regarding modules vs cores.

It doesn't matter what it's called though does it? A Module still executes 2 threads.

 

[LeJosh]

It doesn't matter what it's called though does it? A Module still executes 2 threads.

Not really no but people are saying 8 cores instead of 4 modules and thinking that an 8 "core" processor should be doing better. Although, AMD haven't officially released benchmarks and it's getting very misty in here.

 

[Martini1991]

Not really no but people are saying 8 cores instead of 4 modules and thinking that an 8 "core" processor should be doing better. Although, AMD haven't officially released benchmarks and it's getting very misty in here.

I hate "official".
Ideally AMD should have just had the 8150 as a 4 "core" that does 8 threads.

People don't care for the specifics, just how it performs, and they had unrealistic opinions.

 

[CAT-THE-FIFTH]

This has been mentioned over at semi-accurate by the admin:

http://semiaccurate.com/forums/showpost.php?p=135619&postcount=1825

I would definitely wait until launch before making any decision.

Ideally AMD should have just had the 8150 as a 4 "core" that does 8 threads.

IIRC,that is what chew over on xs was saying too(guy was involved in the overclocking event).

 

[Hagar]

Yes lets have a weeks sabbatical.

restraint please gentlemen

 

[mmj_uk]

..then why is an 8 'core' bulldozer only as fast as a 6 'core' Phenom II..in Cinebench 11.5..?

Somethings not right here...and i hope for AMD's sake its another set of fake benchmarks...

Probably because a 1090T has 6 full-blown cores.

Until now CPU manufacturers have simply taken a single core design and bolted them together to come up with 2, 4 and 6 core chips (with usually only the L3 cache shared between all cores).

A so called '8 core' Bulldozer consists of 4 modules, each module appears to be a dual core chip with bits that AMD deem to be 'not used much' either reduced or shared (or both) between each pair of cores (pair of cores = module).

 

[gareth170]

for people that still don't understand, another way of thinking of it is, a module = a house, 2 core = 2 people in the house, one of them is a full adult, the other is a teenage, and they will have to share some of the food = some of the resources.

 

[Jokester]

I think the way AMD is wanting to go is to have smaller, lower performance cores, but more of them to increase the performance in a similar manner to their GPUs. Problem is at the moment, other than for software that is fully multithreaded it's going to cost them performance. They really need those high clock rates it seems.

 

[CAT-THE-FIFTH]

Hopefully,this means x264 performance on Bulldozer will improve:

http://forum.doom9.org/showthread.php?p=1528746#post1528746

 

[Martini1991]

I think the way AMD is wanting to go is to have smaller, lower performance cores, but more of them to increase the performance in a similar manner to their GPUs. Problem is at the moment, other than for software that is fully multithreaded it's going to cost them performance. They really need those high clock rates it seems.

And when this was suggested previous, people were ridiculed

 

[Martini1991]

for people that still don't understand, another way of thinking of it is, a module = a house, 2 core = 2 people in the house, one of them is a full adult, the other is a teenage, and they will have to share some of the food = some of the resources.

The two cores are the same.
But saying one = 1 and the second = 0.8 makes more sense as that's how their performance would come across.

 

[Jokester]

The two cores are the same.
But saying one = 1 and the second = 0.8 makes more sense as that's how their performance would come across.

Is it not more along the lines that if only one core is used = 1, both cores = 2 x 0.8?

 

[Martini1991]

Is it not more along the lines that if only one core is used = 1, both cores = 2 x 0.8?

I meant it as in, when a second thread is executed on the module, the BD core only adds 0.8.

So, total performance would be 1.8 in a double threaded application over a single BD core in a module.

 

[Jokester]

I meant it as in, when a second thread is executed on the module, the BD core only adds 0.8.

So, total performance would be 1.8 in a double threaded application over a single BD core in a module.

Ah ok, I get you.

 

[pmc25]

Hopefully,this means x264 performance on Bulldozer will improve:

http://forum.doom9.org/showthread.php?p=1528746#post1528746

I see that and raise you:

http://forum.doom9.org/showthread.php?p=1528813#post1528813

I have access to a 2x16 core Bulldozer server, plus AMD is mailing us a physical system.

Yes, there will be XOP optimizations. Yes, they will help, at least a % or 2.

 

[gareth170]

The two cores are the same.
But saying one = 1 and the second = 0.8 makes more sense as that's how their performance would come across.

i know

the meaning was the adult = 1, the teenager = 0.8