AMD Piledriver rumours ... and expert conjecture

[Reynod]

We have had several requests for a sticky on AMD's yet to be released Piledriver architecture ... so here it is.

I want to make a few things clear though.

Post a question relevant to the topic, or information about the topic, or it will be deleted.

Post any negative personal comments about another user ... and they will be deleted.

Post flame baiting comments about the blue, red and green team and they will be deleted.

Enjoy ...

 

[jdwii]

As for defending anything that is retarded as i own the Phenom. If i had a Phenom II x4 i would easily upgrade to the 8350 for its superior multithreaded performance and clearly better gaming performance.

 

[proffet]

Easy to say things with no proof

http://www.guru3d.com/articles_pages/amd_fx_8350_8320_6300_processor_4300_performance_review,10.html



At stock its about even

thanks for the link...

 

[cgner]

For gaming, which is primary use for most of us, Phenom II is still a better choice 3 years after its release. Tiny difference made by an 8 core monster is just disappointing, given that its a whole new archie that already got an update. :/ Maybe workstations will enjoy the performance....

 

[Cazalan]

PD is better long term than a Phenom II, simply for the FMA3 support. Intel did a switcharoo and did FMA3 instead of FMA4 which BD has. This means more apps are already coming out and in the queue that support FMA3.

If I were to build a new budget gaming rig today I probably would opt for the FX-6300. At $135 it's a steal vs the i3.

 

[-Fran-]

PD is better long term than a Phenom II, simply for the FMA3 support. Intel did a switcharoo and did FMA3 instead of FMA4 which BD has. This means more apps are already coming out and in the queue that support FMA3.

If I were to build a new budget gaming rig today I probably would opt for the FX-6300. At $135 it's a steal vs the i3.

And, well, not to stir Intel crowd's wasp nest... That FMA3 support point is valid as long as they (Intel) don't keep AMD out of the game when using ICC.

Also, since AMD's implementation of FMA3 is basically untested yet, when Intel actually supports it, they'll prolly have a better implementation or something. Like the "tesselation" support from the 4800 series AMD boasted about before DX11.

I know this is somewhat bitter, but not far from the truth.

Cheers!

 

[king smp]

Which I again stress can't be done for most workloads.

Look, we did experiments like this back in the 80's. MIT built a supercomputer with several THOUSAND CPU's inside it. And they discovered that most S/W loads do NOT scale well. And even those that do tend to drop off after 16-32, due to various reasons (I/O bottlenecks, locks, the OS scheduler, etc).

Sure, SOME applications can scale well close to infinity. SQL databases are largely independent on a per cell basis, so you can scale at close to 100%. Rendering is another example, hence GPU's. But most workloads do NOT scale in this way. Simple examples: You can't do 3d sound propagation without knowing the current geometry, so sound and rendering touch. That overlap hinders scaling. You can't do AI without inputs from the geometry and sound, so more overlap there. And so on. You don't have totally independent threads, and every time a thread touches another, you have a performance bottleneck.

Finally, you have the scheduler to consider. Developers do NOT use hardcoded thread logic (put thread "x" on core "y"), because there is no way to know current and future workloads on that core (EG: you are not the only thread running). So the cores threads are assigned to are left up to the OS scheduler. In theory, you put threads on the core with the least amount of current work, dynamically adjusting every time a thread gets run. But then you have to consider the CPU architecture (shared cache or dedicated cache per core? This matters if you share a lot of data between threads) as well, so even that gets sticky.

My point being simple: For the majority of workloads, applications can not, and will not, scale beyond a couple of CPU cores. No amount of coding by developers is going to change this very simple fact.

I also note that trying to determine scaling based on Task Managers 1 second sample rate is kind of silly when you think about it, considering most threads will run, finish their work, and be replaced with some other thread LONG before Task Manager gets its next workload sample.

+1 From my limited knowledge that makes alot of sense to me

I do want to take into consideration is the fact of how many people just run only one application at a time?

FYI I own a Phenom II x 6 1100T at 4ghz

Windows will easily run over 500 threads at one time so while most programs cannot take advantage of all six cores the fact remains that the OS managing many programs at a time will take full advantage of all cores available when faced with many apps running at a time.

Remember it is a multitasking world

I run dual display and I can easily have Netflix streaming,Handbrake,Cyberlink PowerDirector,MS Outlook,DVD Flick,multiple browser windows etc all open and running at a time

Gaming is about the only thing where I dont have too much running at a time but actually quite feasible to game and run background apps when running 4 cores or more

So I will state again
MOAH COARS is the future
even some smartphones are running quadcores now if I am not mistaken

 

[cgner]

It may be the future in 5 years, today is 2-3 cores.

 

[bawchicawawa]

It may be the future in 5 years, today is 2-3 cores.

How does it feel living in 2007?

 

[truegenius]

Numbers don't lie but people do

aha, modified statement 😛

even some smartphones are running quadcores now if I am not mistaken

for example samsung galaxy s3
or qualcomm s4 chips 😉
but maybe android os use them in a different manner

 

[cgner]

How does it feel living in 2007?

Sadly not much has changed since games are ported over from consoles like before.

 

[sarinaide]

Intel themselves have gaming problems, namely high res and maximum settings intels FPS fall off can be as high as 15-20% while AMD's is very minimal often around 5 FPS, this is also a trend seen in Nvidia cards compared to AMD cards. This means if a person is a FPS junky that wants 100FPS at ultra res and settings they then need the most expensive CPU with most expensive GPU solutions which is a tedious and pointless affair what with technology going forward so fast.

My other issue is "enthusiast market" what is an enthusiast nobody seems to really give a good answer to this, I have been on overclocking teams and gaming teams that did very well internationally yet I don't regard myself an enthusiast. Is an enthusiast a person that buys the most epeen equipment to play BF 3 against me with uber highest megatron settings of pwnage which when the game has hardware stutter which it does even on a intel epeen rig or is it me the dude running around with the lowest uber ugly mother of Cher settings pwning said person because I mitigate the hardware limitations for streamline performance yet it may look ugly as hell.......which is the enthusiast? (I care to point out this while using X79/58 multi card rigs we play on ugly settings)

Enthusiast is as open to debate as whether intel is any better or worse for anything PC related than AMD. If all a person cares about is maximum FPS then you are a pixel pusher with OCD for frames when I can happily game along at ultra presets and eye candy with a A10 with 7850 in the same manner as I can with my i3 itx build and similar setup.............and its cheap enough for me to throw out the chips next year and replace them with their replacement (ironic that the intel has no replacement on the same socket 😛)

 

[Cazalan]

Yep, 2 cores for the game, 1 core for the OS. The 4th core just used to spread the thermal load.

The NVidia Tegra 3 actually has 5 cores. The 5th one is slow speed for an ultra low power standby, where the 4 main cores get powered off.

 

[esrever]

Most high profile games made this year supports 4 active threads pretty well.

 

[cgner]

um no.

 

[jdwii]

For gaming, which is primary use for most of us, Phenom II is still a better choice 3 years after its release. Tiny difference made by an 8 core monster is just disappointing, given that its a whole new archie that already got an update. :/ Maybe workstations will enjoy the performance....

What i don't understand is you keep saying games don't support beyond 2-3 cores which is true for most games made today, but then you say it takes a 8 core Piledriver to beat a Phenom II and then you say the Phenom II is still a better choice Those extra cores are not even being used for gaming(as you just pointed out) just 4 or less 90% of the time

I'm not saying its worth upgrading to piledriver for gaming if you're on phenom II but a Phenom II by no means is still a better choice if you gave me a Phenom II x4 955 and a 6300 and a H100 or 212+ i could achieve higher FPS with a 6300 same thing with a 1100T vs a 8350.

What i'm trying to do is find all the things about steamroller and try to make a honest estimate on its performance improvements over Piledriver.

Cache is easily one thing that hurts BD/PD and it seems that issue isn't being touched on

The L2 cache accounts for a large chunk of AMD’s core leakage, so shutting half or more of it down can definitely help with battery life. The resized cache is no faster (same access latency); it just consumes less power.

Steamroller brings no significant reduction in L2/L3 cache latencies. According to AMD, they’ve isolated the reason for the unusually high L3 latency in the Bulldozer architecture

http://www.anandtech.com/show/6201/amd-details-its-3rd-gen-steamroller-architecture/2

Power consumption will definitely go down by the looks of it.

 

[proffet]

Intel themselves have gaming problems, namely high res and maximum settings intels FPS fall off can be as high as 15-20% while AMD's is very minimal often around 5 FPS, this is also a trend seen in Nvidia cards compared to AMD cards. This means if a person is a FPS junky that wants 100FPS at ultra res and settings they then need the most expensive CPU with most expensive GPU solutions which is a tedious and pointless affair what with technology going forward so fast.

My other issue is "enthusiast market" what is an enthusiast nobody seems to really give a good answer to this, I have been on overclocking teams and gaming teams that did very well internationally yet I don't regard myself an enthusiast. Is an enthusiast a person that buys the most epeen equipment to play BF 3 against me with uber highest megatron settings of pwnage which when the game has hardware stutter which it does even on a intel epeen rig or is it me the dude running around with the lowest uber ugly mother of Cher settings pwning said person because I mitigate the hardware limitations for streamline performance yet it may look ugly as hell.......which is the enthusiast? (I care to point out this while using X79/58 multi card rigs we play on ugly settings)

Enthusiast is as open to debate as whether intel is any better or worse for anything PC related than AMD. If all a person cares about is maximum FPS then you are a pixel pusher with OCD for frames when I can happily game along at ultra presets and eye candy with a A10 with 7850 in the same manner as I can with my i3 itx build and similar setup.............and its cheap enough for me to throw out the chips next year and replace them with their replacement (ironic that the intel has no replacement on the same socket 😛)

Intel having gaming problems....
where, link it..

 

[gamerk316]

Intel themselves have gaming problems, namely high res and maximum settings intels FPS fall off can be as high as 15-20% while AMD's is very minimal often around 5 FPS, this is also a trend seen in Nvidia cards compared to AMD cards. This means if a person is a FPS junky that wants 100FPS at ultra res and settings they then need the most expensive CPU with most expensive GPU solutions which is a tedious and pointless affair what with technology going forward so fast.

Easy to explain: As you bump res, you move from a CPU bottleneck to a GPU bottleneck. And as Intel has higher baseline FPS, they lose a higher percentage of FPS when a GPU bottleneck occurs.

 

[fazers_on_stun]

And, well, not to stir Intel crowd's wasp nest... That FMA3 support point is valid as long as they (Intel) don't keep AMD out of the game when using ICC.

Also, since AMD's implementation of FMA3 is basically untested yet, when Intel actually supports it, they'll prolly have a better implementation or something. Like the "tesselation" support from the 4800 series AMD boasted about before DX11.

I know this is somewhat bitter, but not far from the truth.

Cheers!

Couple years ago apps using Intel's compiler amounted to maybe 5% of the total. Has that number gone up since then? If not, worrying about Intel crippling AMD seems like a red herring..

 

[fazers_on_stun]

Easy to explain: As you bump res, you move from a CPU bottleneck to a GPU bottleneck. And as Intel has higher baseline FPS, they lose a higher percentage of FPS when a GPU bottleneck occurs.

Methinks Sarinaide is still trying to argue TR's article showing Intel as having the better gaming CPUs

 

[fazers_on_stun]

Off-topic, but does anybody else see these annoying ads that pop up over any images that are linked here on THG?? They have to be from BOM as it doesn't matter where you link the images from..

Really, Tom's - you annoy far more people than your advertisers sell to... Almost as bad as the popups when first entering the home page here..

 

[earl45]

Off-topic, but does anybody else see these annoying ads that pop up over any images that are linked here on THG?? They have to be from BOM as it doesn't matter where you link the images from..

Really, Tom's - you annoy far more people than your advertisers sell to... Almost as bad as the popups when first entering the home page here..

+1

 

[sarinaide]

Methinks Sarinaide is still trying to argue TR's article showing Intel as having the better gaming CPUs

http://techreport.com/r.x/cpu-gaming-2012/bf3-beyond-16.gif

http://techreport.com/review/23750/amd-fx-8350-processor-reviewed/7

Why is that one so different?

Anyways its all immaterial, as I said I game on uber low mega ugly settings that way I don't get pwned by a guy using the title mr GTX690, instead I take his doggie tags.

And no I was just arguing the HARDOCP article which showed intels frame fall off at extreme res and extreme settings, to the point it equals a bulldozer (at the time) fps at the exact same settings.

 

[gamerk316]

http://techreport.com/review/23750/amd-fx-8350-processor-reviewed/7

Why is that one so different?

Because BF3 is GPU limited:

So CPU choice isn't going to make a huge deal in SP BF3, as long as you are above a certain threshold. Skyrim however, tends to be CPU limited:

Leading to the results you would expect, with all AMD CPU's trailing by a significant margin.


Hence why its important to know what you are benchmarking.

 

[proffet]

^
good posting..
explains a lot.

 

[gamerk316]

^^ BF3 should be removed from all CPU tests, because I can guarantee most every CPU is going to get the same exact FPS count, +-5% in either direction. Good GPU benchmark, horrid CPU benchmark.