Rumor: AMD "Piledriver" FX CPU production to begin Q3 2012

Page 8 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

ctsoth

Member
Feb 6, 2011
148
0
0
Saying that AMD has somehow advanced progress with Bulldozer is a joke, as is your nonsense callout.

I'd say it was more of an observation than a callout. Consider ... reading.

Bulldozer architecture info.

http://semiaccurate.com/2011/10/10/exclusive-a-look-at-the-bulldozer-architecture/
http://semiaccurate.com/2011/10/17/why-did-bulldozer-underwhelm/
http://semiaccurate.com/2011/10/17/bulldozer-doesnt-have-just-a-single-problem/

Trinity / Piledriver improvements

http://semiaccurate.com/2012/05/25/trinity-is-more-than-the-sum-of-its-parts/
http://semiaccurate.com/2012/05/28/trinity-has-a-brain-and-a-queue/

The current argument that bulldozer is a 4c + ht is patently absurd. I'll simply make two points. One, HT is far different than an integer unit. Two, such an argument establishes that to bee a 'core' a 'core' must have a floating point unit. This should elicit objections by anyone fit to carry on a conversation on this matter.

What I really look forward to, is in another 1 or 2 apu generations, AMD apus having a massive floating point unit/gpu paired with however many integer units. By the 4coar argument, this will bee a one core cpu.
 

pelov

Diamond Member
Dec 6, 2011
3,510
6
0
Are you a politician ??? you cant even answer a simple question ???

The part "you are willing/able" wasn't clear enough for you to answer that simple question ??? We are talking about $200-300 segment CPUs here, not what the majority of the market will buy.

Because I don't have to. If I had a chance to do it over, I'd buy an Ivy-Bridge/Trinity-based laptop or back then a Llano. If you want to get into hypotheticals in a fantasy world where only you live, you be my guest. You need a serious dose of reality.
 

SickBeast

Lifer
Jul 21, 2000
14,377
19
81
I'd say it was more of an observation than a callout.

It's glorified HyperThreading. Why split the integer unit into 2 units rather than making a single more powerful one?

Bulldozer is a failed CPU architecture and is a lesson in what not to do in terms of CPU design IMO. It is a failure on many levels (performance, heat, performance-per-watt, performance-per-dollar).
 

AtenRa

Lifer
Feb 2, 2009
14,003
3,361
136
Because I don't have to. If I had a chance to do it over, I'd buy an Ivy-Bridge/Trinity-based laptop or back then a Llano. If you want to get into hypotheticals in a fantasy world where only you live, you be my guest. You need a serious dose of reality.

I believe Core i7 3770K is as real as it can be. Anyway, you dont want to have an honest conversation and answer a simple question.

Carry on lads
 

pelov

Diamond Member
Dec 6, 2011
3,510
6
0
I believe Core i7 3770K is as real as it can be. Anyway, you dont want to have an honest conversation and answer a simple question.

Cary on lads

So I MUST choose between two chips, one real and another completely unreal in a hypothetical scenario that's limited by you and you're upset because I'm saying it doesn't represent reality...



I'll do quite well with an APU and a small discrete GPU on a laptop while you're spending too much money on idle cores that you nor anybody else can find a use for. With the saved cash I'll buy myself an ice cream sandwich! Or hundreds of them!

edit - actually, I will answer you.

I'd take the 3770K in a heartbeat. The 2 added cores will come at a lower stock clock speed (take a look at the 2011 platform as evidence. Higher core count = lower clock speeds given the same core architecture. Given different core architectures you'll always have lower clock speeds and lower IPC with a higher core count chip [Bulldozer]) as well as high power consumption (this one is obvious) and larger die meaning higher price. The 2 added cores and 4 threads wouldn't benefit me at all for my workloads and the power consumption, quicksync and toggling between IGP + discrete GPU would.

So, yea, the 3770K. Your hypothetical 6-core chip would cost too much, use too much power and be unnecessary for an overwhelming majority of users.
 
Last edited:

Vesku

Diamond Member
Aug 25, 2005
3,743
28
86
Hyperthreading - adds multithreading to a single logical unit, increases throughput of the single unit by roughly 20-30%.

CMT (Bulldozer) - shares resources between two logical units, provides roughly 80% of the throughput of 2 fully separate units.

It's glorified HyperThreading. Why split the integer unit into 2 units rather than making a single more powerful one?

Bulldozer is a failed CPU architecture and is a lesson in what not to do in terms of CPU design IMO. It is a failure on many levels (performance, heat, performance-per-watt, performance-per-dollar).
 

nismotigerwvu

Golden Member
May 13, 2004
1,568
33
91
AtenRa you're saying AMD is advancing progress when really they're just releasing HyperThreading 5+ years after Intel invented it.

Bulldozer has 4 cores. Only their marketing people say it has 8.

Only Intel didn't invent hyperthreading (unless you are referring only to the name, not the technology). SMT has been around since the 60's.
 

Vesku

Diamond Member
Aug 25, 2005
3,743
28
86
Also, Hyperthreading took quite a while to reach it's current usefulness. Kind of stunk on the P4. So judging CMT based on Bulldozer would be very similar to the reaction to HT on P4s.
 

SickBeast

Lifer
Jul 21, 2000
14,377
19
81
Hyperthreading still sucks. Most gamers turn it off. It only benefits some applications and in others it degrades performance.
 

pelov

Diamond Member
Dec 6, 2011
3,510
6
0
Also, Hyperthreading took quite a while to reach it's current usefulness. Kind of stunk on the P4. So judging CMT based on Bulldozer would be very similar to the reaction to HT on P4s.

They share more than that
 

Arzachel

Senior member
Apr 7, 2011
903
76
91
Hyperthreading still sucks. Most gamers turn it off. It only benefits some applications and in others it degrades performance.

Do you know of a real world scenario where Hyperthreading on SB/IB CPUs actually degrades performance? I've only seen it happen on very optimized benchmarks.
 

Ferzerp

Diamond Member
Oct 12, 1999
6,438
107
106
not that CMT is really better

With software today (and for the forseeable future), I'd rather have a single core that on one thread works at 130% and on 2 threads works at 160% (of an arbitrary baseline) than 2 cores that each work at 80% for a total of the same 160%.
 

guskline

Diamond Member
Apr 17, 2006
5,338
476
126
Is it possible that the improvements to the core of the Vishera Piledriver may exceed those of Trinity? I say this because Trinity is effectively out the door and there is still time over the summer to observe and make further "tweeks" to the Vishera Piledriver before release to production.
 

Arzachel

Senior member
Apr 7, 2011
903
76
91
http://www.overclock.net/t/671977/hyperthreading-in-games

In GTA 4 and in Crysis it lowers the performance. I will admit that it's a bit of a wash, but even if it is, it shows that it really does nothing to improve gaming performance at all.

Thanks! The impact is pretty small both when it helps and when it doesn't, I guess games don't take so well to HT.

Is it possible that the improvements to the core of the Vishera Piledriver may exceed those of Trinity? I say this because Trinity is effectively out the door and there is still time over the summer to observe and make further "tweeks" to the Vishera Piledriver before release to production.

The optimization guide refers differently to the Piledriver cores in Trinity and Vishera iirc, so it's possible.
 

Idontcare

Elite Member
Oct 10, 1999
21,118
59
91
With software today (and for the forseeable future), I'd rather have a single core that on one thread works at 130% and on 2 threads works at 160% (of an arbitrary baseline) than 2 cores that each work at 80% for a total of the same 160%.

What you want, and me too, is a more aggressive turbo-clocking situation.

Don't sell me a 3.5GHz chip that only clocks one core to a max of 3.9GHz when turbo-core kicks in.

Sell me a 3.5GHz chip that clocks one core at 5GHz or 6GHz when it is the only core being heavily utilized.

Remember, this was the dream:


Only they don't do it that way in reality. On my CPU they don't clock up on core such that it operates at high enough clockspeed to consume 95W (or the TDP limit) when the other cores are idle.

I want the turbo-core they implied we'd be getting. That's what I want.
 

Olikan

Platinum Member
Sep 23, 2011
2,023
275
126
Is it possible that the improvements to the core of the Vishera Piledriver may exceed those of Trinity? I say this because Trinity is effectively out the door and there is still time over the summer to observe and make further "tweeks" to the Vishera Piledriver before release to production.

yes, there will be more tweaks...

the most important ones are:
4 MOVs per clock ( only 2 in piledriver/trinity)
L3 cache

....the movs is the most important string today,
vishera can teoricaly have a IPC increase higher than 30% just by that!
( but it won't, there is other issues that won't let this happen )

bulldozer L3 cache is kinda useless for desktop, good for servers...piledriver is no diferent
 

TekDemon

Platinum Member
Mar 12, 2001
2,297
1
81
Thanks! The impact is pretty small both when it helps and when it doesn't, I guess games don't take so well to HT.
But it doesn't really hurt to leave it turned on in most situations since you're very unlikely to be CPU bound with a highly clocked Sandy or Ivy chip. In the linked benchmarks even the minimum frame rate was over 60FPS at all times and it's not even clear that the minimum was even CPU related. In the real world you would most likely not notice the difference between HT on and HT off since most games wouldn't be CPU bound by a 4Ghz+ sandy or ivy chip.
 

Ferzerp

Diamond Member
Oct 12, 1999
6,438
107
106
The only thing I'm aware of that HT gives a penalty on is Linpack, and that's only if you let it use 8 cores. If you restrict it to 4, it works just the same as it would with non-ht cores. There's really no reason to turn it off, and if you don't need/want it, get the i5 version.
 

Ferzerp

Diamond Member
Oct 12, 1999
6,438
107
106
What you want, and me too, is a more aggressive turbo-clocking situation.

Don't sell me a 3.5GHz chip that only clocks one core to a max of 3.9GHz when turbo-core kicks in.

Sell me a 3.5GHz chip that clocks one core at 5GHz or 6GHz when it is the only core being heavily utilized.

Remember, this was the dream:


Only they don't do it that way in reality. On my CPU they don't clock up on core such that it operates at high enough clockspeed to consume 95W (or the TDP limit) when the other cores are idle.

I want the turbo-core they implied we'd be getting. That's what I want.


That would be nice. As I've always said, single threaded performance is of much more benefit than throwing more cores at it. I'd take a quad at double the clock speed any day over an octo of the same architecture. I don't like going below 4 cores, but that's my work self leaking in to my personal systems since we're talking thousands of threads running on a single system, and you start to lose a lot to context switching if you don't have enough cores (but, unlike what one might intuit, single threaded performance is also very important there as well once enough cores are present, as only a couple of threads may really be heavily active at any one moment, so completing their work quickly provides a smoother experience)
 

AtenRa

Lifer
Feb 2, 2009
14,003
3,361
136
edit - actually, I will answer you.

I'd take the 3770K in a heartbeat. The 2 added cores will come at a lower stock clock speed (take a look at the 2011 platform as evidence. Higher core count = lower clock speeds given the same core architecture. Given different core architectures you'll always have lower clock speeds and lower IPC with a higher core count chip [Bulldozer]) as well as high power consumption (this one is obvious) and larger die meaning higher price. The 2 added cores and 4 threads wouldn't benefit me at all for my workloads and the power consumption, quicksync and toggling between IGP + discrete GPU would.

So, yea, the 3770K. Your hypothetical 6-core chip would cost too much, use too much power and be unnecessary for an overwhelming majority of users.

I believe you should read again Anand's SandyBridge E review. The 3960X at 3.3GHz humiliates Sandy 2600K that is clocked at 3.4GHz even in lower resolution gaming. Not to mention the huge performance difference in multithreaded applications. (before anyone say that 3960X is a $1000 cpu, continue reading)

A 6-core IB would have the ~same die size as current IB quad cores + iGPU like the 3770K. That means that a 6-core IB could cost the same as current Core i7 3770K. Both SBs and IBs are powergated, they shutdown idle cores plus lower both frequencies and Voltages in idle states. You will not have more power consumption with two extra cores if those are Idle. But you will get 30% to 50% more performance when you would need it because of those two extra cores.

Again we are talking about desktops here but you have to bring Laptops in to the conversation. You know im right but you trying to divert the conversation to a different domain.
Users that buy $200-300 Desktop CPUs dont need the iGPU. It is a dead die space for them. All im saying is that it is better to have two more cores sitting idle in that space and use them when you would need them than have the iGPU that they will never use it.
 

Cerb

Elite Member
Aug 26, 2000
17,484
33
86
Users that buy $200-300 Desktop CPUs dont need the iGPU. It is a dead die space for them. All im saying is that it is better to have two more cores sitting idle in that space and use them when you would need them than have the iGPU that they will never use it.
But, people that buy $150-200 CPUs do use the IGP. Why should they make separate chips, which will inevitably eat into their profit margins more than not offering a core or two more per chip will?
 

grimpr

Golden Member
Aug 21, 2007
1,095
7
81
I believe you should read again Anand's SandyBridge E review. The 3960X at 3.3GHz humiliates Sandy 2600K that is clocked at 3.4GHz even in lower resolution gaming. Not to mention the huge performance difference in multithreaded applications. (before anyone say that 3960X is a $1000 cpu, continue reading)

A 6-core IB would have the ~same die size as current IB quad cores + iGPU like the 3770K. That means that a 6-core IB could cost the same as current Core i7 3770K. Both SBs and IBs are powergated, they shutdown idle cores plus lower both frequencies and Voltages in idle states. You will not have more power consumption with two extra cores if those are Idle. But you will get 30% to 50% more performance when you would need it because of those two extra cores.

Again we are talking about desktops here but you have to bring Laptops in to the conversation. You know im right but you trying to divert the conversation to a different domain.
Users that buy $200-300 Desktop CPUs dont need the iGPU. It is a dead die space for them. All im saying is that it is better to have two more cores sitting idle in that space and use them when you would need them than have the iGPU that they will never use it.

Funny thing is that they will bring out the quicksync excuse and how fast and worth to have it is, they eat the fixed function transcoding hardware instead of 2 more full fledged x86 cores..D:
 
sale-70-410-exam    | Exam-200-125-pdf    | we-sale-70-410-exam    | hot-sale-70-410-exam    | Latest-exam-700-603-Dumps    | Dumps-98-363-exams-date    | Certs-200-125-date    | Dumps-300-075-exams-date    | hot-sale-book-C8010-726-book    | Hot-Sale-200-310-Exam    | Exam-Description-200-310-dumps?    | hot-sale-book-200-125-book    | Latest-Updated-300-209-Exam    | Dumps-210-260-exams-date    | Download-200-125-Exam-PDF    | Exam-Description-300-101-dumps    | Certs-300-101-date    | Hot-Sale-300-075-Exam    | Latest-exam-200-125-Dumps    | Exam-Description-200-125-dumps    | Latest-Updated-300-075-Exam    | hot-sale-book-210-260-book    | Dumps-200-901-exams-date    | Certs-200-901-date    | Latest-exam-1Z0-062-Dumps    | Hot-Sale-1Z0-062-Exam    | Certs-CSSLP-date    | 100%-Pass-70-383-Exams    | Latest-JN0-360-real-exam-questions    | 100%-Pass-4A0-100-Real-Exam-Questions    | Dumps-300-135-exams-date    | Passed-200-105-Tech-Exams    | Latest-Updated-200-310-Exam    | Download-300-070-Exam-PDF    | Hot-Sale-JN0-360-Exam    | 100%-Pass-JN0-360-Exams    | 100%-Pass-JN0-360-Real-Exam-Questions    | Dumps-JN0-360-exams-date    | Exam-Description-1Z0-876-dumps    | Latest-exam-1Z0-876-Dumps    | Dumps-HPE0-Y53-exams-date    | 2017-Latest-HPE0-Y53-Exam    | 100%-Pass-HPE0-Y53-Real-Exam-Questions    | Pass-4A0-100-Exam    | Latest-4A0-100-Questions    | Dumps-98-365-exams-date    | 2017-Latest-98-365-Exam    | 100%-Pass-VCS-254-Exams    | 2017-Latest-VCS-273-Exam    | Dumps-200-355-exams-date    | 2017-Latest-300-320-Exam    | Pass-300-101-Exam    | 100%-Pass-300-115-Exams    |
http://www.portvapes.co.uk/    | http://www.portvapes.co.uk/    |