AMD Piledriver rumours ... and expert conjecture

Page 170 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
Status
Not open for further replies.
Reynod

Reynod

Administrator
Mar 26, 2007
14,921
16
42,215
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 ...
 
Sort by date Sort by votes


2.gif


your words always rocks 😀
great sense of humor

btw
you were dating someone (few weeks ago, posted in old posts, you were prepraing some knowledge for that date)
so how was that?
 
Upvote 0 Downvote
This article sums up the PD improvements and issues pretty well.

http://semiaccurate.com/2012/05/25/trinity-is-more-than-the-sum-of-its-parts/

Only being able to decode 4 instructions per clock, when it can actually execute up to 12 instructions per clock. It's perpetually stalled for instructions.

PD has a bunch of small improvements that will help IPC slightly, but this big discrepancy is the kick in the nuts. Makes you wonder why they went to a module in the first place. The power reduction should help the overclocking headroom though.


 
Upvote 0 Downvote
I think we need to turn this into an AMD undervolting thread or at least let's actually talk about undervolting on AMD units since
voltage wise they run so, so much higher than Intel.

just a 'side' topic suggestion..
Click to expand...

AMD is like a a royal enfield bike no average consume lot of resources not fast pick up but when it stress tht big fat goes way beyond to handle same like enfield bikes hehe
 
Upvote 0 Downvote
Your spot on, Also don't forget Cincebench! When i test single core performance i also use the Fritz chess Benchmark but only run it on one core.

For fun here is some benchmarks that i did on my Phenom II x6 OC 3.9Ghz


CineBench
Single core: 1.18
All cores: 6.91
Fritz Chess Benchmark
Single Core
Relative Speed: 5.06
Kilo nodes per second: 2430
All Cores:
Relative Speed: 28.28
Kilo nodes per second: 13573
wPrime
32M: 7.02
1024M: 219.167


Be cool if Piledriver could beat that at stock! Whats sad is my friends Laptop with a I7(sandy) which is clocked at 2.2Ghz gets a 1.21 on Cincebench on just one core.
Click to expand...

Here is a thought. Are all those programs designed to take full advantage of advancing code (sse, avx, ect) or are they generic 386 only?

Also, should ipc also be taken in consideration with memory speeds or timings? Cinebench for one is affected by the memory so ite not purely ipc.
 
Upvote 0 Downvote
Someone earlier on the forum posted benchmarks of his/her FX-8120 with 4 cores @2.3GHz (same clocks as Trinity) enabled and put them up against Tom's Trinity review. Those showed that Trinity was ~10% faster than a similar BD. Factor in L3 cache, and whatever the tweaks AMD will do between Trinity and PD, and I think this statement may be rather believable.

It is interesting to note that the Donanimhaber indicates that instruction per clock (IPC) performance of Vishera will be 15% higher compared to that of current-generation Zambezi thanks to Piledriver micro-architecture as well as some other tweaks. Earlier it was widely believed that FX "Vishera" chips will only bring 10% speed improvement at the same clock-speed compared to the currently available chips.
Click to expand...

I may be wrong, but there is no "best case scenario" for IPC. There's no specific test that will show more IPC improvement than others. (at least by much). So if AMD gets their 15% (that big "if" is important) along with higher clocks, and much better power usage (Trinity's showed well), we may have a good set of chips on the way.
Click to expand...

Toms did pretty good with their x86 core comparison. A single core per CPU, no SMT or Turbod all clocked to the same speed. Overall if a CPU performs better than the IPC of that CPU is better than the one(s) it outperformed by such percentage.

Of course it never scales linearly with cores or clock speed but its a good base point to go from.

This article sums up the PD improvements and issues pretty well.

http://semiaccurate.com/2012/05/25/trinity-is-more-than-the-sum-of-its-parts/

Only being able to decode 4 instructions per clock, when it can actually execute up to 12 instructions per clock. It's perpetually stalled for instructions.

PD has a bunch of small improvements that will help IPC slightly, but this big discrepancy is the kick in the nuts. Makes you wonder why they went to a module in the first place. The power reduction should help the overclocking headroom though.
Click to expand...

This was the issue with BD as well, and its not suprising if PD has the same setup as its not a major overhaul of the arch, much like Deneb was not a major overhaul of the arch but a die shrink with much better clocks and some IPC improvements. Sort of like a stepping + I would say.

Here is a thought. Are all those programs designed to take full advantage of advancing code (sse, avx, ect) or are they generic 386 only?

Also, should ipc also be taken in consideration with memory speeds or timings? Cinebench for one is affected by the memory so ite not purely ipc.
Click to expand...

Memory speed is something. In synthetic tests, it can help. In the real world, most programs do not use a fraction of the memory bandwidth available to it (22GB/s from my DDR3-1600 is more than enough).

Even sadder is people who buy the double the price 2133MHz RAM for a system and there is not 2x the increase in performance. Had a custom build at work who got a IB CPU with some GSkill DDR3-2133MHz RAM. I didn't clock it to 2133 because, well it voids the CPU warranty as it needs 1.65v.

For what the guy was doing (photos) he probably wouldn't notice the difference anyways. Same goes with a AMD CPU, except Llano/Trinity which the GPU benefits from faster RAM.
 
Upvote 0 Downvote
Hello guys..
Just joining this forum,after a long time silent reader..hehe

Im just want to share my thought about AMD plan..

I think that AMD has already know everything about this arch..
But they just dont release it all to the its best performance..

They decided to release it little by little..
An improvement over improvement..

First they released the Buldozer,which all we know that didnt do much..
And they make us think that they developed Piledriver to revising Buldoz to make an improvement.. I think they already have Piledriver an even Steamroller before.. They just start to pull it out slowly..

I didnt have any evidence, i just reader all the forum that talk about this..

CMIIW

Thank you..
 
Upvote 0 Downvote
Here is a thought. Are all those programs designed to take full advantage of advancing code (sse, avx, ect) or are they generic 386 only?

Also, should ipc also be taken in consideration with memory speeds or timings? Cinebench for one is affected by the memory so ite not purely ipc.
Click to expand...

I was going to add something in those lines as well... PD is supposed to be more in-line with Intel's newer instructions, so compilers should take advantage from them as well. It was proven in Linux that you DO get a 10% increase with re-compiled code to support the new instruction set from BD, so for PD it should be a better landscape for free in Windows.

So you have to put that as well in the mix.

Cheers!
 
Upvote 0 Downvote
Of course they have piledriver and steamroller, even the mighty excavator, same as intel has technology like for or more years ahead from now. The companies usually hold back technology because it will be to risky to release what they have developed just recently, and it is a very good way to make money
 
Upvote 0 Downvote
Of course they have piledriver and steamroller, even the mighty excavator, same as intel has technology like for or more years ahead from now. The companies usually hold back technology because it will be to risky to release what they have developed just recently, and it is a very good way to make money
Click to expand...

If they do that, then it's mostly because of a lack of competition. And if 2 companies fix the tech advancement in an underground agreement, we can sue them to hell. It has happened before (ATI and nVidia were sued once for that reason IIRC).

We need to keep in check the big corporations to stop them from doing that or at the end of the day, we (the regular Joe consumer) will be screwed.

Cheers!
 
Upvote 0 Downvote
Hello guys..
Just joining this forum,after a long time silent reader..hehe

Im just want to share my thought about AMD plan..

I think that AMD has already know everything about this arch..
But they just dont release it all to the its best performance..

They decided to release it little by little..
An improvement over improvement..

First they released the Buldozer,which all we know that didnt do much..
And they make us think that they developed Piledriver to revising Buldoz to make an improvement.. I think they already have Piledriver an even Steamroller before.. They just start to pull it out slowly..

I didnt have any evidence, i just reader all the forum that talk about this..

CMIIW

Thank you..
Click to expand...

Or BD was only so bad to artificially set the bar low for CPUs, making subsequent launches look better by comparison. But now we're just talking conspiracy theories... Seriously though, I would be surprised if they didn't have plans for Steamroller already on the drawing board.
 
Upvote 0 Downvote
Of course they have piledriver and steamroller, even the mighty excavator, same as intel has technology like for or more years ahead from now. The companies usually hold back technology because it will be to risky to release what they have developed just recently, and it is a very good way to make money
Click to expand...

I think so.. Im feel something not fine in AMD.. Dunno? 😀 😀

If they do that, then it's mostly because of a lack of competition. And if 2 companies fix the tech advancement in an underground agreement, we can sue them to hell. It has happened before (ATI and nVidia were sued once for that reason IIRC).

We need to keep in check the big corporations to stop them from doing that or at the end of the day, we (the regular Joe consumer) will be screwed.

Cheers!
Click to expand...

What?? ATI and NVIDIA do that?? Nice info!

Or BD was only so bad to artificially set the bar low for CPUs, making subsequent launches look better by comparison. But now we're just talking conspiracy theories... Seriously though, I would be surprised if they didn't have plans for Steamroller already on the drawing board.
Click to expand...

I just think that The BD is the base arch and the other is the only revision.. AMD maybe hiding their "Nuke" from everyone until the day.. *toomuchsci-fi 😀

Plans yes, actual chips ready to go no. Intel isn't sitting on Haswell waiting for X months to pass, nor is AMD sitting on PD. I also doubt AMD made BD bad on purpose so they could release better looking chips.
Click to expand...

So what is the reason AMD launch BD? Why dont they just launch PD,which more reliable?

#sorry for my bad english.. 😀
 
Upvote 0 Downvote
Because PD wasn't ready. The final CPU needs to be ready months before it is launched. The foundries also need time to make enough of them to have on hand to launch them. (toss in a few more weeks for packaging them.) Why did they launch BD? Because they hadn't finished with PD yet and BD is what they had 3-5mo before they launched.
 
Upvote 0 Downvote
Memory speed is something. In synthetic tests, it can help. In the real world, most programs do not use a fraction of the memory bandwidth available to it (22GB/s from my DDR3-1600 is more than enough).

Even sadder is people who buy the double the price 2133MHz RAM for a system and there is not 2x the increase in performance. Had a custom build at work who got a IB CPU with some GSkill DDR3-2133MHz RAM. I didn't clock it to 2133 because, well it voids the CPU warranty as it needs 1.65v.

For what the guy was doing (photos) he probably wouldn't notice the difference anyways. Same goes with a AMD CPU, except Llano/Trinity which the GPU benefits from faster RAM.
Click to expand...
All programs respond differently. There is no true IPC test because of these differences. as far as saying faster ram is 2x the price, well, look at overall system perfromance vs overall cost.

~$1000 build, 1600 memory is $40 on newegg, and 2133 is $60 -15% discount, just call it $60. so you spent 2% of your total system cost for a ~0-8% or more (civ V increased 20% tested on BD) increase in performance.

you can't look at memory in direct price against itself, its a skewed result. when factored into the total cost is where you get the actual result.

I will say in a ~$300 build, maybe $20 is a bit more than justified (7% of the total cost, and chances are the particular cpu isn't going to respond as well). IMO it all depends on total system cost.

Killer deals are something to keep an eye out for, I got my 2133 memory when it was on sale, $45 for 8gb. Do I regret it in any way shape or form? not a chance.

You also have to take into consideration sse 4.2, AVX, and fma over PII. If those are used, then IPC is increased for that particular application as well.
 
Upvote 0 Downvote
Hello guys..
Just joining this forum,after a long time silent reader..hehe

Im just want to share my thought about AMD plan..

I think that AMD has already know everything about this arch..
But they just dont release it all to the its best performance..

They decided to release it little by little..
An improvement over improvement..

First they released the Buldozer,which all we know that didnt do much..
And they make us think that they developed Piledriver to revising Buldoz to make an improvement.. I think they already have Piledriver an even Steamroller before.. They just start to pull it out slowly..

I didnt have any evidence, i just reader all the forum that talk about this..

CMIIW

Thank you..
Click to expand...


Welcome to our forums.

:)
 
Upvote 0 Downvote
Toms did pretty good with their x86 core comparison. A single core per CPU, no SMT or Turbod all clocked to the same speed. Overall if a CPU performs better than the IPC of that CPU is better than the one(s) it outperformed by such percentage.

Of course it never scales linearly with cores or clock speed but its a good base point to go from.



This was the issue with BD as well, and its not suprising if PD has the same setup as its not a major overhaul of the arch, much like Deneb was not a major overhaul of the arch but a die shrink with much better clocks and some IPC improvements. Sort of like a stepping + I would say.



Memory speed is something. In synthetic tests, it can help. In the real world, most programs do not use a fraction of the memory bandwidth available to it (22GB/s from my DDR3-1600 is more than enough).

Even sadder is people who buy the double the price 2133MHz RAM for a system and there is not 2x the increase in performance. Had a custom build at work who got a IB CPU with some GSkill DDR3-2133MHz RAM. I didn't clock it to 2133 because, well it voids the CPU warranty as it needs 1.65v.

For what the guy was doing (photos) he probably wouldn't notice the difference anyways. Same goes with a AMD CPU, except Llano/Trinity which the GPU benefits from faster RAM.
Click to expand...

If I paid for a Corvette and drove within 0 - 60 mph 99% of the time, should I be angry if my Corvette did not go faster than 60 mph? I think so. Would you tell me it makes no sense for me to buy a Corvette or would you just say that I had too much money to spend?

This was the issue with BD as well, and its not suprising if PD has the same setup as its not a major overhaul of the arch
Click to expand...
Do you really know this? Why does the "voice of reason" always so anti-AMD with you?

 
Upvote 0 Downvote
All programs respond differently. There is no true IPC test because of these differences. as far as saying faster ram is 2x the price, well, look at overall system perfromance vs overall cost.

~$1000 build, 1600 memory is $40 on newegg, and 2133 is $60 -15% discount, just call it $60. so you spent 2% of your total system cost for a ~0-8% or more (civ V increased 20% tested on BD) increase in performance.

you can't look at memory in direct price against itself, its a skewed result. when factored into the total cost is where you get the actual result.

I will say in a ~$300 build, maybe $20 is a bit more than justified (7% of the total cost, and chances are the particular cpu isn't going to respond as well). IMO it all depends on total system cost.

Killer deals are something to keep an eye out for, I got my 2133 memory when it was on sale, $45 for 8gb. Do I regret it in any way shape or form? not a chance.

You also have to take into consideration sse 4.2, AVX, and fma over PII. If those are used, then IPC is increased for that particular application as well.
Click to expand...


I doubt buying 2133 ram would even make a difference(except for APU's). Not to mention Intel gets more bandwidth from 1333 then Amd does from 1866.

Amd really needs to work on this going forward. Graphics cards need high bandwidth Look at the difference in performance between a 6670 DDR3 and a DDR5 6670. At times its as high as 20%. Another thing Amd needs to do is include a small buffer for their graphics. They should look at these items when they shrink their APU to 28nm.

Noob2222 i'm sure you can agree and admit Amd needs to improve their performance per core. When their 12 core Opteron beats their 16 core Opteron their is problems.

Rory read is in idiot and should not be in control of a technology company when he says things like "Good enough performance". Its like Apple when they made a CEO in charge on their company when all he did is sell Pepsi a product that never changes. Technology changes so fast and we can't have someone who says good enough. I miss Dirk Meyer.

Meyer joined AMD in 1996, where he personally led the team that designed and developed the Athlon processor.
Click to expand...

Just saying :sol:
 
Upvote 0 Downvote
Hello guys..
Just joining this forum,after a long time silent reader..hehe

Im just want to share my thought about AMD plan..

I think that AMD has already know everything about this arch..
But they just dont release it all to the its best performance..

They decided to release it little by little..
An improvement over improvement..

First they released the Buldozer,which all we know that didnt do much..
And they make us think that they developed Piledriver to revising Buldoz to make an improvement.. I think they already have Piledriver an even Steamroller before.. They just start to pull it out slowly..

I didnt have any evidence, i just reader all the forum that talk about this..

CMIIW

Thank you..
Click to expand...
I agree that AMD probably has plans already in place up through Excavator even. Putting those plans in place is what takes time. Holding out to release a product makes sense, if yours is better than the competition's. AMD is behind already, so waiting to make progress isn't a good business choice.

Putting things into perspective, IF AMD can get 10-15% increases per generation, and the generations come out yearly, things will be looking pretty good not to far in the future.
 
Upvote 0 Downvote
Status
Not open for further replies.

TRENDING THREADS

Latest posts

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom