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 ...

 

[-Fran-]

Trinity uses the piledriver core

Zambezi uses the bulldozer core.

Reynod uses the 48 year old organic core.

Dual cores are outdated, bro.

Cheers! 😛

 

[mayankleoboy1]

Dishonored released. Overwhelmingly positive reviews.

 

[Chad Boga]

then again, maybe its taking them 2 years to fix the 22nm problem. Rght now its a vaoporware cpu slated for q3 2013, about the time they plan on releasing 14nm Atom cpus.

Haswell will be released to market before they start to ship their 22nm server cpus based on Ivy bridge? doesn't sound too exciting.

If 22nm was above expectations, they would show up this year.

I see your mastery of manufacturing processes in CPU design also extends to the Economics of the Industry and Server verification processes. :lol:

 

[sarinaide]

A A10 5800K with a discrete 7850 games within 3FPS of a i3 3220 and 5FPS of a i5 3450, so it dispels the notion of invalidating the necessity for APU's on the market. Factor in Lucid which one can switch between Discrete and Integrated mode for rendering and encoding performance the Trinity was said to be a major player and it surely delivered.

 

[Reynod]

I see your mastery of manufacturing processes in CPU design also extends to the Economics of the Industry and Server verification processes. :lol:

Chad ... where are you going with this discussion?

 

[Chad Boga]

Chad ... where are you going with this discussion?

It's called addressing fanciful claims that don't stack up.

That's part of a discussion forum, well at least normal discussion forums.

 

[tonync_01]

From AMD: 7% IPC increase and 7% frequency increase.

http://www.youtube.com/watch?v=o5hxOsrukPY

 

[sarinaide]

From AMD: 7% IPC increase and 7% frequency increase.

http://www.youtube.com/watch?v=o5hxOsrukPY

Yep, it was clear from the testing that we got that the Vishera chips were going to hit the 15% performance gains intended, really good effort made.

 

[gamerk316]

Yep, it was clear from the testing that we got that the Vishera chips were going to hit the 15% performance gains intended, really good effort made.

*facepalm*

A 7% increase in IPC does not equate to a 7% performance gain.
A 7% increase in clockspeed does not equate to a 7% performance gain.

So yeah, marketing.

 

[Cazalan]

What are you smoking dude ? 😀

It's not a native quad core chip without an iGPU. The iGPU is just disabled; other than that, its exactly the same die as what is used to make the 5800K's. How the hell can you except better single/multi threaded performance from that?!!

Probably not a significant IPC advantage but the CPU wouldn't have to respond to any interrupts of the iGPU asking for memory/resources.

CPU should run cooler and clock higher with the iGPU disabled.

If you have a decent video card already it's the best way to get Trinity without waiting for Vishera. Although it's rumored to be out later this month. 23rd perhaps.

 

[blackkstar]

*facepalm*

A 7% increase in IPC does not equate to a 7% performance gain.
A 7% increase in clockspeed does not equate to a 7% performance gain.

So yeah, marketing.

IPC is such a BS measurement. Better IPC in what? And, correct me if I'm wrong, but 3.6ghz to 4ghz is an 11% increase in clock speed, right?

4 - 3.6 = .4
.4 / 3.6 = 11.1%

Unless they are counting turbo too

4.2 = 4 = .2
.2 / 4 = 5%

Even then that's 8%. I wouldn't take those 7% figures seriously at all. I just hope AMD is going to try and not over-hype PD so everyone isn't super disappointed.

 

[Cazalan]

hexus.net tests trinity's ram scaling

If AMD were smart they'd do some joint marketing with memory vendors to boost Trinity sales/performance.

 

[sarinaide]

*facepalm*

A 7% increase in IPC does not equate to a 7% performance gain.
A 7% increase in clockspeed does not equate to a 7% performance gain.

So yeah, marketing.

I am so pleased you are in a jovial mood. I never equated 1:1 how that scales to real world performance could be substantially more.

 

[viridiancrystal]

If AMD were smart they'd do some joint marketing with memory vendors to boost Trinity sales/performance.

AMD is selling their own memory sticks now, so they could do it themselves :lol: Its almost like they might have planned it that way 😱

 

[gamerk316]

IPC is such a BS measurement. Better IPC in what? And, correct me if I'm wrong, but 3.6ghz to 4ghz is an 11% increase in clock speed, right?

4 - 3.6 = .4
.4 / 3.6 = 11.1%

Unless they are counting turbo too

4.2 = 4 = .2
.2 / 4 = 5%

Even then that's 8%. I wouldn't take those 7% figures seriously at all. I just hope AMD is going to try and not over-hype PD so everyone isn't super disappointed.

Technically, IPC is application dependent. The simplified equation for time to complete a given benchmark (T) is:

T = IPC * Clockspeed * Number of cores * scaling factor

Since the clockspeed and number of cores are known, and the scaling can be estimated based on Task Manager, so you can compute a reasonable figure for IPC for a given application after we know how long it takes the benchmark in question to complete. Its not a flat number, and will vary by workload.

But we can usually assume IPC will be in the same ballpark, so we can guestimate that a 7% IPC increase would be 7% across the board for most workloads.

 

[Cazalan]

AMD is selling their own memory sticks now, so they could do it themselves :lol: Its almost like they might have planned it that way 😱

I didn't understand that move. They could have made more money offering AMD certification or something. Let the memory vendors deal with the overstock.

The performance edition sticks are just too slow right now. They need 2133/1866 at least.

http://www.newegg.com/Product/Product.aspx?Item=N82E16820103006

They max out at 1600.

 

[Cazalan]

Even then that's 8%. I wouldn't take those 7% figures seriously at all. I just hope AMD is going to try and not over-hype PD so everyone isn't super disappointed.

Doesn't look like they're marketing it at all really.
For the channel sales you're lucky to find a "2nd generation" label on the products with Trinity chips.

 

[fazers_on_stun]

then again, maybe its taking them 2 years to fix the 22nm problem. Rght now its a vaoporware cpu slated for q3 2013, about the time they plan on releasing 14nm Atom cpus.

Last I heard it was April 2013 for the desktop parts, which is early Q2. And considering the ton of sku's for Ivy parts, in plentiful supply, and the 4th fab nowhere near capacity, all that indicates their 22nm is better than expected, perhaps better than their 32nm node which was the best in the industry..

Haswell will be released to market before they start to ship their 22nm server cpus based on Ivy bridge? doesn't sound too exciting.

If 22nm was above expectations, they would show up this year.

No, server parts have to be extensively tested and validated which is why they show up later, at least for Intel.

 

[Cazalan]

Taken straight from AMD's website.

http://www.amd.com/us/products/desktop/processors/a-series/Pages/a-series-pib.aspx

"AMD A10 Processors include all of the above features, plus:

45% more graphics performance than 1st Gen AMD A8 APUs '11
26% better system performance than 1st Gen AMD A8 APUs '12
"


With note 11 being a singular benchmark, 3DMark Vantage.
With note 12 being a singular benchmark, PCMark7

Not sure I'd call that hype but very selective for sure.

 

[-Fran-]

Taken straight from AMD's website.

http://www.amd.com/us/products/desktop/processors/a-series/Pages/a-series-pib.aspx

"AMD A10 Processors include all of the above features, plus:

45% more graphics performance than 1st Gen AMD A8 APUs '11
26% better system performance than 1st Gen AMD A8 APUs '12
"


With note 11 being a singular benchmark, 3DMark Vantage.
With note 12 being a singular benchmark, PCMark7

Not sure I'd call that hype but very selective for sure.

I call it "marketing 101", hahaha.

I hate that kind of picky showing and every brand is at fault doing it. They could at least make a funny remark against the competition or something 😛

Anyway, so now the show will begin on the 23rd of October, right? Let's see if my estimation is correct in the 7% to 9% overall improvement.

Cheers!

 

[jdwii]

Technically, IPC is application dependent. The simplified equation for time to complete a given benchmark (T) is:

T = IPC * Clockspeed * Number of cores * scaling factor

Since the clockspeed and number of cores are known, and the scaling can be estimated based on Task Manager, so you can compute a reasonable figure for IPC for a given application after we know how long it takes the benchmark in question to complete. Its not a flat number, and will vary by workload.

But we can usually assume IPC will be in the same ballpark, so we can guestimate that a 7% IPC increase would be 7% across the board for most workloads.

I wonder how much Intel is going to improve IPC with their new design i hear 15-20% if so that will show Amd up even more which actually makes me sad.

 

[mayankleoboy1]

15-20% IPC improvement for Intel is almost impossible. Mainly because all their branch predictors and cache coherency algorithms are already at the saturation point.

 

[amdfangirl]

^ Then we are missing is a little corporate espionage and we'll have a competitive AMD then, no? 😛

 

[blackkstar]

15-20% IPC improvement for Intel is almost impossible. Mainly because all their branch predictors and cache coherency algorithms are already at the saturation point.

I agree with this as well. Intel knows it too, it's why Haswell talk is mostly about their iGPU. If AMD was serious, they could definitely catch up to Intel. Intel isn't playing the performance game anymore either.

Everyone talks about how the desktop is a dying breed, but it will never die. People need it to render, edit photos, do 3d modeling, create music, compile software, etc. Even if some of these tasks move to the cloud, not everyone is going to want to move to the cloud. There will always be a desktop market.

What I would really like to see AMD do is get rid of FX entirely and leave APUs and Opterons, and then create an enthusiast platform around Opterons that use the same sockets and allow overclocking. I don't care how much slower a Bulldozer or Piledriver core is compared to an Intel core. When you get two 16 core Piledriver Opterons in a system and give it a moderate overclock, it doesn't matter unless you're playing games. All AMD would need to do is create a non-server chipset with the things servers need (ECC, SAS, etc) removed, add some PCIe lanes, and make sure you can raise the bus speed.

 

[amdfangirl]

Well socket unification is one thing AMD needs with FM1, FM2, AM3+ all selling in the mainstream market at the same time.