Rumour: Bulldozer 50% Faster than Core i7 and Phenom II.

[Cogman]

If you want to show benches in multi threading show recent benches with retail chips not cherryt ones like the above. And that decoder is from the CS4 days. Outdated and dont make use of the multi core systems so well

http://ppbm5.com/Benchmark5.html
^thats recent with retail chips with clocks the systems run 24/7 on

Ummm. Even in the very earliest of days, x264 has made excellent use of multi-cored systems. In fact, a 47% increase with 50% more cores is pretty dang close to the theoretical maximum speed increase one could expect to see.

x264 is anything but cherry picking. The developers do their best to make it run as fast as possible on whatever architecture you have. They don't favor intel or AMD in the development of x264.

x264 is an ENCODER not a DECODER. Big difference. Even this 2-3 year old version of x264 runs circles around the crappy encoders published by adobe. Newer versions are only going to be faster.

Pretty much every unbiased review of x264 (read, a review NOT published by someone making a competitor) places it squarely at the top in pretty much every category.

 

[HW2050Plus]

As I said in the first page of this thread:

Hey, an eight core CPU is faster than a four core CPU! Who woulda thought??

Really, would you compare that 2600K to an Athlon X2??

It was said already, but again for you:
There is nothing wrong with that to compare two processors showing 8 cores to the operating system. Intel and AMD use different terms and that may confuse. 1 AMD module is equal to 1 Intel core and 1 AMD core is equal to 1 Intel Thread.

They did not say it was a 4 module Zambezi. They said it was a quad core.

Not saying it's legit (because I definitely think it isn't). Just pointing out that a quad-core Zambezi means exactly that: 4 cores (only 2 modules). It helps to just forget the modules because the OS supposedly never sees them, and just reports the number of int cores as cores.

This indication is what definitely makes the entire leak fake to me. If the leak was anywhere near legit, there would have been no rookie mistake regarding what a quad core BD means, and therefore the ridiculous comparison to a 2600K (which the quad core zambezi is shown to dominate) would not have been made in the first place. A 4-core/4-thread Zambezi spanking a 4-core/8-thread 2600K in a multi-threaded benchmark, and with less GHz? Absolutely rubbish.

This, like most other leaks, are just page-hit materials. JFAMD already spelled it out: Under his watch, no leaks, no benchies, nothing, until launch. Everything else is just fiction.

There are some more obstacles in this "leaked benchmark". However I would not put too much attention in this 4 core thing. So the mentioning of "4 core" does not irritate me. However other obstacles which Dresdenboy pointed out.

I am afraid we will get some more bullshit fakes as leaks. However samples of Bulldozer are out now. We could see some real leaks.

Currently we do not have anything reliable.

 

[Mopetar]

Company A markets a CPU with a base frequency of 2Ghz that will turbo to 3Ghz.

Company B markets a CPU with a base frequency of 2.5Ghz that will turbo to 3Ghz.

Oh, by the way, under light loads they both also underclock below their base frequency.

Which CPU has more headroom?

If I decide to market CPU B as a 1.8Ghz CPU does it now have more headroom?

There's not enough information to answer that question, but either way, it entirely misses the point of what I was trying to say. If both companies have techniques to get their chip to maximum TDP, it really doesn't matter how TDP is measured or who has more headroom.

 

[Phynaz]

especially if the thread counts are the same.

What in the world does thread counts have to do with anything??

Sun has a 24 thread CPU. Does that mean it should be compared to a 24 core cpu?

 

[Phynaz]

1 AMD module is equal to 1 Intel core and 1 AMD core is equal to 1 Intel Thread.

Read what JF has posted and quit spreading false information. 1 AMD module = 2 cores.

AMD is not going to market modules, they are going to market cores.
Cores = Cores, get it?

 

[sawtx]

What in the world does thread counts have to do with anything??

Sun has a 24 thread CPU. Does that mean it should be compared to a 24 core cpu?

Yet core count is important? Price and performance should be the main factors for us, but if you want to compare CPUs without knowing those thread count is more important than core count.

 

[Phynaz]

Yet core count is important? Price and performance should be the main factors for us, but if you want to compare CPUs without knowing those thread count is more important than core count.

Ok.

In that case the Sun CPU I just mentioned blows away BD to no end, right? After all, you don't know it's price or performance, so therefore more threads makes it better than BD or SB.

Threads don't do processing, cores do.

 

[JFAMD]

Core counts and thread counts are what the loyalists will use to argue superiority over each other.

Total performance per socket
Power consumed per processor
Price per processor

If we all stick to these, then architectural differences are neutralized and we can actually have a decent conversation.

That is, amazingly, how probably 95%+ of the world would evaluate the processors.

 

[piesquared]

Ok.

In that case the Sun CPU I just mentioned blows away BD to no end, right? After all, you don't know it's price or performance, so therefore more threads makes it better than BD or SB.

Threads don't do processing, cores do.

So for ZLib tests, for example, performance doesn't increase with increased thread count? Answer: yes it does. Anyway, if if costs me the same amount for a 100,000 core chip that outperforms a 4 core whatever chip, I couldn't care less what propaganda intel wants to spin, i'll choose the 100,000 core chip. And I suspect the overwhelming majority of consumers(>90&#37 would agree...

 

[sawtx]

Ok.

In that case the Sun CPU I just mentioned blows away BD to no end, right? After all, you don't know it's price or performance, so therefore more threads makes it better than BD or SB.

Threads don't do processing, cores do.

Do you think that the Xbox 360 CPU is faster than an dual core SB? It has more cores so it has to be much faster, right?

If what you say is true then by disabling hyper-threading on an i7 you don't lose any performance. Wow I need to go tell Intel to redo their architectures and to not improve performance by increasing thread count. The whole "you're comparing 4 core to 8 core" thing seriously sounds like a fanboy trying to convince people that their stuff is still better than the competition.

 

[LoneNinja]

Core counts and thread counts are what the loyalists will use to argue superiority over each other.

Total performance per socket
Power consumed per processor
Price per processor

If we all stick to these, then architectural differences are neutralized and we can actually have a decent conversation.

That is, amazingly, how probably 95%+ of the world would evaluate the processors.

I know that's how I look at it, I don't care if 1 processor takes more cores to beat another as long as power consumption and price are similar.

 

[Mopetar]

I know that's how I look at it, I don't care if 1 processor takes more cores to beat another as long as power consumption and price are similar.

Well, there's the other ~ 5% that are enthusiasts and want the best performance regardless of price or performance per watt. That's why Intel can charge them $1000 for a processor that isn't $700 better than a 2600K in performance. They also make a ridiculous amount of margin off of these chips as well so being able to own that market segment isn't a bad thing.

 

[AtenRa]

If i remember correctly, the 1366 socket CPUs was only 1% of the total Intel CPUs and i would bet that 6-core Gulftown CPUs were less than 1% of the total 1366 CPU Segment.

That is not a lot of money.

Intels 6-Core Gulftown was only about prestige and PR to drive the sales of other Intel CPUs.

 

[Mopetar]

Yes, but it's a profitable 1%. Most sales are going to be for their mid-range chips that end up in crappy $600 Dell boxes, but if that 1% of sales generates an insane amount of profit, it's probably worth as much as their bottom 10% of CPU sales.

Look at Apple as an example. They probably only control around 15% of the smartphone market, but about 50% of the profits. Sales volume is nice, but it doesn't always translate into profitability.

 

[Phynaz]

Do you think that the Xbox 360 CPU is faster than an dual core SB? It has more cores so it has to be much faster, right?

If what you say is true then by disabling hyper-threading on an i7 you don't lose any performance. Wow I need to go tell Intel to redo their architectures and to not improve performance by increasing thread count. The whole "you're comparing 4 core to 8 core" thing seriously sounds like a fanboy trying to convince people that their stuff is still better than the competition.

I notice that none of the BD advocates mention the Xeon 7560. I wonder why they are afraid to make comparisons to that CPU?

If calling me a Intel fanboy helps you sleep better at night have at it. I'll introduce you to the folks in the video forum that call me an AMD fanboy and you can argue which one I am with them.

In the mean time I will continue to call out making comparisons between two CPU's, when one has twice the number of processing units than the other as stupid. Especially when you bring into play the extremely limited amount of software that can take advantage of eight cores.

 

[Arkadrel]

Core counts and thread counts are what the loyalists will use to argue superiority over each other.

Total performance per socket
Power consumed per processor
Price per processor

If we all stick to these, then architectural differences are neutralized and we can actually have a decent conversation.

That is, amazingly, how probably 95%+ of the world would evaluate the processors.

^ this, anything else just does not make any sense.

In the mean time I will continue to call out making comparisons between two CPU's, when one has twice the number of processing units than the other as stupid. Especially when you bring into play the extremely limited amount of software that can take advantage of eight cores.

Only 3 things matter:

performance + price + power used.

*IF*

2 core Intel cpu costs same as 4 core AMD cpu
2 core Intel cpu performance same as 4 coro AMD cpu
2 core Intel cpu power same as amd cpu

why not compaire them? they cost the same, they give same performance, they use same power.

your the one thats being silly about core vs cores haveing to be equal to compair cpus.....

for me its price/performance.... thats what I look for in a cpu, I dont care if it has 1 core or 16 cores, if its price is cheap, and its performance is fast! Thats the cpu Im buying, its the same for the rest of the world, except for fanboys.

 

[itsmydamnation]

I notice that none of the BD advocates mention the Xeon 7560. I wonder why they are afraid to make comparisons to that CPU?

If calling me a Intel fanboy helps you sleep better at night have at it. I'll introduce you to the folks in the video forum that call me an AMD fanboy and you can argue which one I am with them.

well to be fair we are talking about the consumer space, there has been very little talk about the 16core/8mod bulldozer either so your point is hardly relevant or meaningful.

In the mean time I will continue to call out making comparisons between two CPU's, when one has twice the number of processing units than the other as stupid. Especially when you bring into play the extremely limited amount of software that can take advantage of eight cores

then your not doing a very good job a lot of different things like:
the differences in the uarchs
the difference in execution units
the difference in peak vs average utilisation
the impact issue,fetch,retirement have on performance
the fact software is only becoming more threaded
that fact that users do lots more multi tasking then they used to

. if intel cores are twice as wide as AMD's and they have HT, how isn't it a fair comparison? "turbos" also make your point far less relevant because low threaded apps will get big turbo's.

Then there is also understand what chews power in a chip, here's a hint go look at the size of an int add or mul unit compared to FP add or mul. Then you need to look what kinds of software or tasks use what.

you seem to think only the metric you define are important, unfortunately for your metrics are no longer important and are becoming increasingly more unimportant.

 

[Phynaz]

if its price is cheap, and its performance is fast! Thats the cpu Im buying, its the same for the rest of the world, except for fanboys.

So if somebody has different purchasing criteria than you they are a fanboy.

I'll just let that one hang out and simmer for while.

BTW, I'm still waiting on that link to the 175Mhz IBM CPU that is faster than a Sandy Bridge.

You havnt seen the 175mhz supercomputers IBM made have you?
you dont need 4GHz CPUs.... not if you have very high ICP, even a 175mhz cpu can be MUCH MUCH more powerfull than a 4ghz sandybridge cpu.

 

[sawtx]

I notice that none of the BD advocates mention the Xeon 7560. I wonder why they are afraid to make comparisons to that CPU?

We can if you want but the comparison would be against it's competitor which would be the 16 core BD.

If calling me a Intel fanboy helps you sleep better at night have at it. I'll introduce you to the folks in the video forum that call me an AMD fanboy and you can argue which one I am with them.

I didn't mean to call you a fanboy, just that your argument is one a fanboy would make. Also there is no reason you can't be a fanboy for Intel CPUs and AMD GPUs.

In the mean time I will continue to call out making comparisons between two CPU's, when one has twice the number of processing units than the other as stupid. Especially when you bring into play the extremely limited amount of software that can take advantage of eight cores.

Again, core count doesn't matter, unless you can prove that hyper-threading doesn't add performance to Intel chips. Price and performance should matter the most, and if the "50% better" is real that probably gives us both as the performance comparison should be made against similarly priced current CPUs.

Also there are many pieces of software that will use as many threads as you can feed them with, I use several basically everyday. That isn't even bringing up multi-tasking.

 

[maddie]

As I said in the first page of this thread:

Hey, an eight core CPU is faster than a four core CPU! Who woulda thought??

Really, would you compare that 2600K to an Athlon X2??

You're definitely on a roll.

Intel has hyperthreading that allows 1 core to execute 2 threads if one of them stalls. You should realize that in cases of no stalls, only 1 thread executes. This is not a free lunch as there is a small increase in transistors.

AMD has developed a module that allows 2 threads at all times. You cannot separate the cores in a module. This also has a transistor penalty, but a VERY small one compared to 2 distinct cores.

It would seem obvious to any neutral observer that this is a very big break from traditional design layouts.

The old definition of cores is changing and we will have to think about them differently.

If a 4 core hyperthreading CPU is close in transistor budget to a 4 module, 8 core CPU, then what is the problem from comparing them directly?

We should be more creative in our thinking, don't you agree?

 

[AtenRa]

AMD has developed a module that allows 2 threads at all times. You cannot separate the cores in a module. This also has a transistor penalty, but a VERY small one compared to 2 distinct cores.

We have no idea how big the Shared frond end of the Module is and we have no idea how big the shared FPU 2 x 128-bit FMACs are, dont take only the Integer core.

Yes they save die size and transistor count and many other things with shared units in the design but we have no idea how much they have managed to save compared to 2 full independent cores yet.

 

[maddie]

We have no idea how big the Shared frond end of the Module is and we have no idea how big the shared FPU 2 x 128-bit FMACs are, dont take only the Integer core.

Yes they save die size and transistor count and many other things with shared units in the design but we have no idea how much they have managed to save compared to 2 full independent cores yet.

I'm saying this based on this info.

"Two dedicated integer cores
- each consist of 2 ALU and 2 AGU which are capable for total of 4 independent arithmetic or memory operations per clock per core
- duplicating integer schedulers and execution pipelines offers dedicated hardware to each of two threads which significantly increase performance in multithreaded integer applications
- second integer core increases Bulldozer module die by around 12%, which at chip level adds about 5% of total die space

 

[Phynaz]

Intel has hyperthreading that allows 1 core to execute 2 threads if one of them stalls. You should realize that in cases of no stalls, only 1 thread executes.

Incorrect.

The two threads execute simultaneously. SMT - Simultaneous MultiThreading

http://www.realworldtech.com/page.cfm?ArticleID=RWT040208182719

 

[Phynaz]

Price and performance should matter the most, and if the "50% better" is real that probably gives us both as the performance comparison should be made against similarly priced current CPUs.

I agree it's fair to compare similarly priced CPUs.

How much does does Bulldozer cost?

 

[maddie]

Incorrect.

The two threads execute simultaneously. SMT - Simultaneous MultiThreading

http://www.realworldtech.com/page.cfm?ArticleID=RWT040208182719

Is it truly possible for 2 threads to use the same part of the pipeline at once? Really asking.

Can anyone explain this?