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 ...

 

[Cazalan]

I was expecting a bit more on the graphics front but not bad overall.
They'll tweak some more performance with driver optimizations.

Any benches with Dual Graphics (7660G+7670M) yet?

 

[-Fran-]

Oh, I have to correct myself in one thing... It's not around 300-400 Mhz of difference; it's around 700Mhz of difference in the same process.

That's quite a big difference. Maybe the IPC will be still lower than K10 at the end and jdwii seems more on the mark for Zambezi's.

Cheers!

 

[truegenius]

Does anyone else have any estimates i would love to here them! Try to be realistic.

i will give mine but after some reviews of desktop trinity

but i think 20-30% overall performance increase with less heat

 

[noob2222]

The battery life could be many factors such as larger battery or possibly better power management features which I would expect out of Trinity.

I would have expected that intel's 22nm idle power would be able to keep up ... thats the same size battery. oddly enough ivy is near the bottom of the list all the time. I thought Ivy was supposed to be great for mobile, guess not.

 

[jdwii]

http://images.anandtech.com/graphs/graph5831/46668.png

I would have expected that intel's 22nm idle power would be able to keep up ... thats the same size battery. oddly enough ivy is near the bottom of the list all the time. I thought Ivy was supposed to be great for mobile, guess not.

Looks like Intel had to push their 22nm Processes to improve their Graphics. Since they don't have that much of a improvement on Power consumption or CPU performance so the GPU is the power hog.

 

[esrever]

http://images.anandtech.com/graphs/graph5831/46668.png

I would have expected that intel's 22nm idle power would be able to keep up ... thats the same size battery. oddly enough ivy is near the bottom of the list all the time. I thought Ivy was supposed to be great for mobile, guess not.

the ivy was a 45w chip. and Intel quad core i7s always push their thermals/power greatly.

 

[fazers_on_stun]

I disagree, the cpu is quite nice in improvements.
Those lappys are coming, so they wont miss, DT however is another matter.
Im not sure who said they expected alot more out of the cpu, but I havnt seen it, except from the usual suspects

You mean AMD?? 😀

IIRC it was probably DT and not mobile, but one of the AMD bloggers was saying 30% CPU improvement over Llano and I forget how much for the igp..

 

[JAYDEEJOHN]

AMD bloggers atre more often sloggers, as they tend to bend reality, and cause more unacceptable hopes.
As for hopes and dreams, maybe Obama should invest in AMD 😉

 

[noob2222]

Did my own tests

8120 @ 4 cores, 2.3 ghz.

30.27 & 20.39 clearly an improvement, ~10% not taking into account whatever benefit l3 cache has on this test

84.3, 31.5, 58, 42.67

still improved, question still remains, how much does l3 help this test?

Cinebench 11.5, 1.59
trinity cinebench 2.05

quite an improvement there.

 

[triny]

You mean AMD?? 😀

IIRC it was probably DT and not mobile, but one of the AMD bloggers was saying 30% CPU improvement over Llano and I forget how much for the igp..

I said 25-30% cpu overall better than Lano comparing A8 3850 to A10 5800 as well as 60-80 % + igp performance
Since the A10 5800 has not arrived yet you'll have to wait and see.All good things come to those who wait.
I own an I5 so to categorize me as an AMD blogger is far from the mark.

 

[guskline]

I do hope the Piledriver turns out to be a good bump up from the Thuban. I got high hopes that the 8150FX would go toe to toe with the 2500K in gaming but that was dashed. I already have a 1100 Thuban running on an Asus Sabertooth 990 FX mb but it won't be moved unless the Piledriver shows an improvement. We shall see.The Trinity review gives me a glimmer of hope that the Bulldozer core was tweeked. The question for the PileDriver is "how much".

 

[jdwii]

We will see who is closer in estimating Piledrivers performance Me or triny i say 15-20% boost OVERALL! Not 1 benchmark but 20+ benchmarks. And guskline to even think Piledriver will go toe to toe with a I5 ivy in terms of gaming means you have a LOT of hope! 😉

 

[viridiancrystal]

Did my own tests

http://media.bestofmicro.com/A/V/337351/original/Sandra arith.png

8120 @ 4 cores, 2.3 ghz.

30.27 & 20.39 clearly an improvement, ~10% not taking into account whatever benefit l3 cache has on this test

http://media.bestofmicro.com/B/0/337356/original/Sandra MM.png

84.3, 31.5, 58, 42.67

still improved, question still remains, how much does l3 help this test?

Cinebench 11.5, 1.59
trinity cinebench 2.05

quite an improvement there.

Thanks for the tests. I'm am shocked at how poorly the 8120 performs when down-clocked like that., but shows well for Trinity and PD.

I'm also quite surprised with the battery life Trinity pulled out. Llano has terrific battery life, but Trinity does even better in the same process with more transistors. That's quite a feat from AMD's engies.

Also, notice in the performance ballpark where Trinity lands... It's right in the doorstep of USD800+ Intel notebooks. If Trinity comes under USD700 that's going to be quite sweet across the board for notebooks and ultra-thins/portables.

Also, the IPC increase is not that noticeable, but you can tell it's there. They're not clocked that far from each other, yet Trinity is actually pulling ahead. Maybe not by a wide margin, but when you think it's using all the TDP it can against Llano, there should be like a 300-400Mhz difference only, which could put them in the same IPC ballpark.

IMO, TH's review will put some more perspective for all of us, but so far so good.

Cheers!

Most of Tom's tests put Trinity cpu about 10-20% faster than Llano cpu. How much was from clocks, we can only guess. However, in the mobile space, Trinity performs better than Llano in every aspect, while using less power. PD may not have a much improved IPC, but Trinity is quite a step up from Llano.

My own thoughts about Trinity. I'm impressed. Marginally better than Llano in every way, what more could you ask for?

Talking about Ivy's HD4K graphics being this so called "good enough" brings up a good point. Even though AMD and Intel's graphics are fairly close, AMD has a massive gpgpu advantage. Even more so when GCN is implemented. AMD needs to push heterogeneous computing, because now the only thing that a better gpu will do for the average person is take up die space. If GPGPU becomes widespread, then better igp will result in better performance in software. (don't say better cpu performance, because that isn't true)

Good job AMD, (i'm sure most of you AMD employees are reading this) keep up the good work.

 

[noob2222]

wasn't a full 8120, i cut it in half so it was equivalent to a downclocked 4100, trying to get apples to apples on the trinity.

I also couldn't run turbo enabled for most tests as T1 state is 17x multiplier on 4 cores. Had to attempt to mimick it by running a 135mhz fsb with a single core turbo at 3.2 ghz.

also ran it on 2.3ghz 200 mhz bus with turbo off to make sure the numbers were similar, they were nearly identical.

 

[esrever]

vishera looks to be a 20% improvement on bulldozer with performance and probably over 100% in efficiency.

Im very pleased with the power consumption of trinity and I will definately be looking at a 14" ultrathin with trinity.

 

[fazers_on_stun]

I said 25-30% cpu overall better than Lano comparing A8 3850 to A10 5800 as well as 60-80 % + igp performance
Since the A10 5800 has not arrived yet you'll have to wait and see.All good things come to those who wait.
I own an I5 so to categorize me as an AMD blogger is far from the mark.

?? Who was referring to you as an "AMD Blogger"? Not me.. AMD bloggers are generally understood to write blogs on AMD.com, and to use their real names as well..

IIRC it was jdwii and you making predictions about Trinity's performance - he predicted something like 30% gpu and 10% cpu. But you went much further and said "I also believe those are not benchmarks but an extrapolation of information ie: amd's initial release 20%cpu 50%gpu
Since then AMD has said that it would be more powerful than expected
In any case either 25to30% cpu 60-80%Gpu or AMD is full of poop which is entirely possible"

According to AT's review, gpu is about 20% faster averaged across 15 games and cpu was 20% faster in PCMark and 15% faster in Cinebench...

BTW, you also predicted a few months ago "I can't see trinity not out selling lano and BD combined as well as grabbing 30% of mobile market.
It's that good ,an A10 25 watt chip will bring unheard of video to ulv mobile formats
That's where Intel is weakest.
will it be the best of the best ,most fastest, no ,but it will be what lots of people will want during this recession."

So in other words, you're predicting Trinity will about double AMD's current mobile marketshare of 16%?

 

[viridiancrystal]

wasn't a full 8120, i cut it in half so it was equivalent to a downclocked 4100, trying to get apples to apples on the trinity.

I also couldn't run turbo enabled for most tests as T1 state is 17x multiplier on 4 cores. Had to attempt to mimick it by running a 135mhz fsb with a single core turbo at 3.2 ghz.

also ran it on 2.3ghz 200 mhz bus with turbo off to make sure the numbers were similar, they were nearly identical.

I know. Even so, it is surprising. My stock 8120 scores 4x as much. Doesn't make much sense to me.

 

[-Fran-]

?? Who was referring to you as an "AMD Blogger"? Not me.. AMD bloggers are generally understood to write blogs on AMD.com, and to use their real names as well..

IIRC it was jdwii and you making predictions about Trinity's performance - he predicted something like 30% gpu and 10% cpu. But you went much further and said "I also believe those are not benchmarks but an extrapolation of information ie: amd's initial release 20%cpu 50%gpu
Since then AMD has said that it would be more powerful than expected
In any case either 25to30% cpu 60-80%Gpu or AMD is full of poop which is entirely possible"

According to AT's review, gpu is about 20% faster averaged across 15 games and cpu was 20% faster in PCMark and 15% faster in Cinebench...

BTW, you also predicted a few months ago "I can't see trinity not out selling lano and BD combined as well as grabbing 30% of mobile market.
It's that good ,an A10 25 watt chip will bring unheard of video to ulv mobile formats
That's where Intel is weakest.
will it be the best of the best ,most fastest, no ,but it will be what lots of people will want during this recession."

So in other words, you're predicting Trinity will about double AMD's current mobile marketshare of 16%?

Actually, if you start thinking about how Trinity turned out, it's not THAT impossible to at least get near the 30% mark in a few years when the APU/Fusion thingy gets more adepts (Intel included). I'd say even that 20% is totally feasible in a few months.

Cheers!

 

[fazers_on_stun]

Actually, if you start thinking about how Trinity turned out, it's not THAT impossible to at least get near the 30% mark in a few years when the APU/Fusion thingy gets more adepts (Intel included). I'd say even that 20% is totally feasible in a few months.

Cheers!

For the low-end mobile market, perhaps - depends on how much Intel charges for the dual-core i3 HD4K IB's, which is probably where the quad-core Trinitys such as the A10 tested here and on AT will be competing most directly, pricewise anyway. I'd say also wait & see how the 17-W versions fare against each other - THG didn't compare to an IB at all, and AT used a 45-W part against the 35-W Trinity.

 

[jdwii]

?? Who was referring to you as an "AMD Blogger"? Not me.. AMD bloggers are generally understood to write blogs on AMD.com, and to use their real names as well..

IIRC it was jdwii and you making predictions about Trinity's performance - he predicted something like 30% gpu and 10% cpu. But you went much further and said "I also believe those are not benchmarks but an extrapolation of information ie: amd's initial release 20%cpu 50%gpu
Since then AMD has said that it would be more powerful than expected
In any case either 25to30% cpu 60-80%Gpu or AMD is full of poop which is entirely possible"

According to AT's review, gpu is about 20% faster averaged across 15 games and cpu was 20% faster in PCMark and 15% faster in Cinebench...

BTW, you also predicted a few months ago "I can't see trinity not out selling lano and BD combined as well as grabbing 30% of mobile market.
It's that good ,an A10 25 watt chip will bring unheard of video to ulv mobile formats
That's where Intel is weakest.
will it be the best of the best ,most fastest, no ,but it will be what lots of people will want during this recession."

So in other words, you're predicting Trinity will about double AMD's current mobile marketshare of 16%?

I always predict, but i never talk like that at least not in the last 6+ months. So please improve what your saying if you hinting at me. Not trying to cause somthing but i at least try to prove what i predict and with PD its not really that hard to see the performance its going to bring along with the power consumption unless the L3 cache is that power hungry.

I must say i expected a little more from Trinity's GPU but less on the CPU.

 

[fazers_on_stun]

I always predict, but i never talk like that at least not in the last 6+ months. So please improve what your saying if you hinting at me. Not trying to cause somthing but i at least try to prove what i predict and with PD its not really that hard to see the performance its going to bring along with the power consumption unless the L3 cache is that power hungry.

I must say i expected a little more from Trinity's GPU but less on the CPU.

Actually I was saying your predictions were much closer to the mark than Triny's 😀. That's a compliment!

For the record, here are some of Triny's more unrestrained predictions, both here and in what is now the Haswell thread:

Intel will have to scramble to improve igp
with the gains in performance of as much as 113% coming to the apu

hd4000 will be something like having a hd 6670 with beer bottle blinders and library of preset bitmaps popping in and out

IB delayed until q4 or 2013
DX 11 cracked one of their brass balls
dx 11.1 will crack the other one

exactly like I said IB small improvement
67% over hd 3000 is a small step forward lano is 250% ahead of hd3000 trinity is 500% ahead of hd3000.
haswell hopefully can catch lano at least

I am going by what Ive seen ces video lano game performance ,what I have read what AMD released the Tukiskish leak video and common sense .
my numbers may not be Brass balls accurate but neither are they that far off and may in fact be conservative.
Which is to say I don't know anymore than anyone else.
would you care to guess the numbers maybe you have more insight I'm an Intel guy so I am not up to snuff on AMD

 

[JAYDEEJOHN]

If you look, games that require more cpu did well with IVB, but again, it wasnt at the same power envelope, so, somethings got to give somewhere

 

[jdwii]

Actually I was saying your predictions were much closer to the mark than Triny's 😀. That's a compliment!

For the record, here are some of Triny's more unrestrained predictions, both here and in what is now the Haswell thread:

Intel will have to scramble to improve igp
with the gains in performance of as much as 113% coming to the apu

hd4000 will be something like having a hd 6670 with beer bottle blinders and library of preset bitmaps popping in and out

IB delayed until q4 or 2013
DX 11 cracked one of their brass balls
dx 11.1 will crack the other one

exactly like I said IB small improvement
67% over hd 3000 is a small step forward lano is 250% ahead of hd3000 trinity is 500% ahead of hd3000.
haswell hopefully can catch lano at least

I am going by what Ive seen ces video lano game performance ,what I have read what AMD released the Tukiskish leak video and common sense .
my numbers may not be Brass balls accurate but neither are they that far off and may in fact be conservative.
Which is to say I don't know anymore than anyone else.
would you care to guess the numbers maybe you have more insight I'm an Intel guy so I am not up to snuff on AMD

Oh lol yeah except for my Bulldozer prediction i'm usually close.

But that's what a prediction is your sometimes right and sometimes wrong. Just don't let emotions(fanboy in you) Take over and manipulate your prediction like i did with Bulldozer.

 

[palladin9479]

Can any of the mod's here contact the THG crew and see if they'll answer a few questions.

#1 Was any of those tests done while using processor affinity to force Turbo Boost?

Doing all sorts of tests on my Llano I can say with absolute authority that TB will not activate if you let Windows handle your process assignments. Windows will keep moving a thread around to different "under utilized" CPUs which prevents turboboost from kicking in. Putting my CPU at factory settings, I was able to force 1~2 cores to max speed using processor affinity, setting affinity to auto it would never hit max speed.

#2 Was there any tool provided from AMD or off the intarwebs that allows them to play with P / B states?

This would be good for analysis and seeing how much headroom there is on each CPU.

#3 Did someone check with a different set of memory?

The A8-3500M only supports DDR3-1333, that is why the DD3-1600 memory would only run at 1333 speeds. The fact that your getting IDENTICLE bandwidth measurements on a 3500M and a Trinity tells me something is wrong in the memory subsystem. The A8-35xxMX series was the Sabine line that supported DDR3-1600 memory, and it's more bandwidth then a 3500M. Memory bandwidth is incredibly important for the graphics system that your benchmarking.

Finially a comment, a 2.05 is still horrible for CB11.5. Can a quick test be done with CB11.5 on Single Threaded, then affinity lock the CB process and do another Single Threaded test and check the results. Will let you know what kind of improvement to expect in single / dual threaded applications.

 

[amdfangirl]

http://www.anandtech.com/show/5835/testing-opencl-accelerated-handbrakex264-with-amds-trinity-apu

In a nutshell: GPGPU is finally useful for me! <3

While video transcoding is significantly slower on Trinity compared to Intel's Sandy Bridge on the traditional x86 path, the OpenCL version of Handbrake narrows the gap considerably. A quad-core Sandy Bridge goes from being 73% faster down to 7% faster than Trinity. Ivy Bridge on the other hand goes from being 2.15x the speed of Trinity to a smaller but still pronounced 29.6% lead. Image quality appeared to be comparable between all OpenCL outputs, although we did get higher bitrate files from the x86 transcode path. The bottom line is that AMD goes from a position of not really competitive, to easily holding its own against similarly priced Intel parts.

This truly is the holy grail for what AMD is hoping to deliver with heterogeneous compute in the short term. The Sandy Bridge comparison is particularly telling. What once was a significant performance advantage for Intel, shrinks to something unnoticeable. If AMD could achieve similar gains in other key applications, I think more users would be just fine in ignoring the CPU deficit and would treat Trinity as a balanced alternative to Intel. The Ivy Bridge gap is still more significant but it's also a much more expensive chip, and likely won't appear at the same price points as AMD's A10 for a while.

AMD's VCE tested (but nobody cares)

😀