Bulldozer 50% Faster than Core i7 and Phenom II

[TheMailMan78]

Firstly, what guy above me said is true.

Secondly more and more games/apps will support multi threading as time goes on, I've already got a whole bunch of programs that use all 6 of my cores.

The rate that multithreading is supported will increase and actually only get faster and faster.

Eventually every app that comes out will be multithreaded by default ( free potential performance boost after all)

Also, bare in mind a some games you can fiddle with settings and force them to multithread

Sometimes it doesn't make a difference but other times it will reduce minimum fps/frame drops in general.
Cool!

I own a few games that utilize all 6 of my cores.

 

[WarEagleAU]

Being an Avid AMD Fanboy and lover of their procs, I take this with less than a grain of salt. IF true, well done and about time, but you have to look at this within reason of what Intel has been doing since AMD whooped its ass and it came back strong and continues to do so. I will wait to see performance numbers but hopefully the new architecture will be amazing.

 

[oli_ramsay]

 

[Wrigleyvillain]

I own a few games that utilize all 6 of my cores.

What else besides BF3?

 

[Jstn7477]

Funny that this thread has been dug back up. 10 months later and now we know the answers.

 

[15th Warlock]

This thread is old... No sense in beating a dead horse it should be locked

 

[TheMailMan78]

What else besides BF3?

Shogun, F1, Any source game, Supreme Commander and a dozen others.

 

[nt300]

AMD also shared with us that Windows 7 isn't really all that optimized for Bulldozer. Given AMD's unique multi-core module architecture, the OS scheduler needs to know when to place threads on a single module (with shared caches) vs. on separate modules with dedicated caches. Windows 7's scheduler isn't aware of Bulldozer's architecture and as a result sort of places threads wherever it sees fit, regardless of optimal placement. Windows 8 is expected to correct this,
LINK:
http://www.anandtech.com/show/4955/the-bulldozer-review-amd-fx8150-tested/11

 

[de.das.dude]

AMD also shared with us that Windows 7 isn't really all that optimized for Bulldozer. Given AMD's unique multi-core module architecture, the OS scheduler needs to know when to place threads on a single module (with shared caches) vs. on separate modules with dedicated caches. Windows 7's scheduler isn't aware of Bulldozer's architecture and as a result sort of places threads wherever it sees fit, regardless of optimal placement. Windows 8 is expected to correct this,
LINK:
http://www.anandtech.com/show/4955/the-bulldozer-review-amd-fx8150-tested/11

some said ms was working on patch for win7 to fix this.

 

[Super XP]

Interesting....

We can hope that Windows 8 and upgraded applications and utilities that use the new FX instructions will make it more competitive, and I'd expect these things right about the time Ivy Bridge become available.

Pros:
+ First consumer eight-core processor
+ Officially supports 4GHz-plus turbo speeds and DDR3-1866 memory
+ An FX system has 42 PCI-E lanes as opposed to the 24 lanes of a Sandy Bridge system
+ 990FX chipset supports NVIDIA SLI. Finally.
+ AMD finally has a 32nm processor with good overclocking

Cons:
- Requires a new Socket AM3+ motherboard
- Single core performance has remained static
- Full performance requires Windows 8 system and applications that use its new instructions
- Overall similar performance to Core i5 2500K, but at a higher price

Ratings:
•Performance: 8.00
•Construction: 9.00
•Overclock: 9.50
•Functionality: 8.50
• Value: 8.00
Final Score: 8.60 out of 10.

http://benchmarkreviews.com/index.p...k=view&id=831&Itemid=63&limit=1&limitstart=17

 

[Mussels]

the extra PCI-E lanes make me think it may be worth it for crossfire/SLI


edit: really, windows 8? thats an interesting comment

 

[Deleted member 67555]

the extra PCI-E lanes make me think it may be worth it for crossfire/SLI


edit: really, windows 8? thats an interesting comment

Oh you caught that too...Yeah...kinda like spitting in someones wound to get the salt to spread better.

 

[kid41212003]

the extra PCI-E lanes make me think it may be worth it for crossfire/SLI


edit: really, windows 8? thats an interesting comment

Let me clarify that for you...

It's nothing fancy really.


Bulldozer has 4 "packages". Each of them contains 2 cores.

For Turbo mode to work correctly, workload must be on a single package. If workload splits between 2 cores between 2 packages, the cpu won't be able to turn off cores that are not in use.

Bulldozer can turn off "package" but not individual core.

 

[specks]

If its true then

 

[wolf]

the Bulldozer architecture focuses on greater inter-core communication and reconfigured ALU/FPU to achieve higher instructions per clock cycle (IPC) compared to the previous generation K10.5, on which its current Phenom II series processors are based.

seems like what they wanted but not what they achieved... especially when you compare it to PII X6 1100T or X4 980, let alone the last 2 generations of core i5/i7....

I am so idssapointed in Bulldozer, almost heartbroken... I wanted it to succeed and compete against SB quite badly, but it's almost like they are still in the water at the moment, they added 2 cores but made each core do less than a PII... ?

 

[Nick89]

Woa, what happened between January an October?

 

[Super XP]

We were trying to kill off this thread, and now you resurrected it

 

[Frick]

Only mods can truly kill threads!

 

[de.das.dude]

i dont think they can kill news threads. anyhoo keep to topic or leave this thread.

 

[cadaveca]

Sure, but you must understand what started the conversation is that this bit of info that started this thread was some guerilla marketing, seemingly more aimed at creating disappointment when the actual product arrives. Alot of people are going to be screaming about how AMD lied about the 50% performance they indicated now, when the product launches. Of course, the info didn't come from AMD, seemingly, so one must wonder why it was even posted in the first place...and I don't mean psoted here on TPU...I mean posted on the orginal website/source.

that point alone, of course, leads to us discussing proper marketing practices. Logical way for the conversation to progess, it seems to me.

I find it funny that I was warning about the poor marketing creating false expectations for months before the launch, as seen in the quote above, from back in January.

Nothing really changed over those months, per se...at least for myself.


No need to hide this thread!

 

[CDdude55]

We know you said that, you've mentioned how you've said that multiple times. Really the best strategy that i think most people should take is to not expect things will go a certain way no matter what people are saying, as the only people who were really mad or sad about the FX chips were people who saw ''new architecture'' written everywhere and hence proceeded to wet their pants and expect more then what they got.

When it comes to this thread though, even a blind man could of easily seen that this threads prediction was not going to come true, 50% faster then a Core i7 is significant and felt mroe like wishful thinking as opposed to anything factual.

 

[erocker]

Aren't you people tired of this conversation yet? Don't get me wrong, I find most of you amusing.

 

[cadaveca]

We know you said that, you've mentioned how you've said that multiple times. Really the best strategy that i think most people should take is to not expect things will go a certain way no matter what people are saying, as the only people who were really mad or sad about the FX chips were people who saw ''new architecture'' written everywhere and hence proceeded to wet their pants and expect more then what they got.

When it comes to this thread though, even a blind man could of easily seen that this threads prediction was not going to come true, 50% faster then a Core i7 is significant and felt mroe like wishful thinking as opposed to anything factual.

Sure. But it seems that I cannot stress this point enough. I'm not some boy-genius or anything..as you say, it's obvious, but many have ignored the obvious.

I'll be using this same example for some time, not because I'm tooting my own horn...but because really, I don't know jack. It's pretty painful to see peopel ignore the obvious, and then be upset because of it.

Aren't you people tired of this conversation yet? Don't get me wrong, I find most of you amusing.

meh. Bulldozer will be talked about for years and years. I still cannot get FX-8120 or FX-8150 in lcoal shops, and I was at them all yesterday. Not ONE chip has come into town yet. Until it does, this is the best and closest I can get to playing with my own.

 

[Super XP]

Well since we are on topic, what about this post about the Bulldozer's L1, L2 and L3 cache. What I am reading is Bulldozer may need a complete cache overhaul. Am I correct? Or AMD can fix this with minor process revisions and tweaks.

Total cache can be deceptive.
As I indicated in earlier speculation thread scrypt is VERY L1 cache dependent.

While the Bulldozer has more total cache (L1+L2+L3) it has less L1 data cache (L1 cahce is divided into discrete data & instruction caches).

Phenom II has 64KB of L1 data cache per core.
Bulldozer has 16KB of L1 data cache per integer core.

The hypothesis I proposed in the speculation thread was that Bulldozer would do better (8 cores vs 6 cores) if scrypt lookup table would fit in the L1 cache of Bulldozer. Your benchmark just answered that question.

The larger L2 & L3 cache of Bulldozer is immaterial. There is a 3 clock cycle latency to L2 cache and 20 (IIRC) clock cycle latency to L3 cache. It would appear that the scrypt lookup tables can't fit in 16KB thus the CPU is being idled thousands of times per hash waiting for data to SLLLLLLLOOOOOOOOOWWWWWWWWWWWWLLLLLLLLLYYYYYYYYY make it way from L2 -> L1. L3 cache is likely completely unused for the datasets used by scrypt.

The nice thing is you have shown 16KB of L1 cache is likely insufficient. We know 64KB is sufficient. That gives us an upper and lower bounds.


The i5 series CPU have 32KB of L1 cache. Clock for clock they tend to underperform the Phenom II but still do ok. My guess is that there may be some cache misses but not too many which allows decent performance.

On edit: looks like I was incorrect. Clock for clock i3/5/7 series outperforms Pheom II. Phenom II has higher overall performance but that is due to more cores & higher overclock. That would indicate 32KB is sufficient.

Do you guys think a revision or stepping will make any difference besides make the power consumption better?

If there's a L1 cache bug, absolutely; could be 10% right there. If other efficiencies altogether added another 10% (include BIOS revisions in there too), we could see a 20%-25% improvement, which pretty much puts this chip where it needed to be; if not a thriller, certainly no disappointment.

 

[Damn_Smooth]

I find it funny that I was warning about the poor marketing creating false expectations for months before the launch, as seen in the quote above, from back in January.

Nothing really changed over those months, per se...at least for myself.


No need to hide this thread!

I honestly thought that they had a chance though. AMD's marketing team did a great job of leading me to believe that they were on to something. Oh well, lesson learned.

We know you said that, you've mentioned how you've said that multiple times. Really the best strategy that i think most people should take is to not expect things will go a certain way no matter what people are saying, as the only people who were really mad or sad about the FX chips were people who saw ''new architecture'' written everywhere and hence proceeded to wet their pants and expect more then what they got.

When it comes to this thread though, even a blind man could of easily seen that this threads prediction was not going to come true, 50% faster then a Core i7 is significant and felt mroe like wishful thinking as opposed to anything factual.

Truer words I have not heard spoken in a long time.