AMD "Piledriver" refresh of Zambezi - info, speculations, test, fans

[FlanK3r]

I think, there will be more...Its not all about high clocks...New instruction set, heterogenous computing, more better OC potential than Llano (example A8-3870K, 3 GHz stock, max stable average is 3.6 GHz. A10-xxxx will have nice OC potential, more than 600 MHz)

[Oese]

base clocks is 1.9ghz vs. 2ghz. Turbo may work better then on Llano and it may have some advantage due to that. But that advantage then at least seems to (over)compensate for the 2m/4t design and former ipc losses so i think one can say that it matches Llano on same clocks at lower tdp, which includes better IGP also. I think thats a big step forward in any case compared to bulldozer at launch...

[fakhrain]

http://amdfx.blogspot.de/2012/05/amd...a10-4655m.html

dunno if it has been posted before somewhere.. looks good. 2m/4t trinity @ 2.0 ghz 2.3 turbo matches llano 4c @ 1.9 ghz 2.3 turbo (which doesnt work) at 10w less tdp..

The Llano A6-3400M is actually a 1.4ghz part with 2.3ghz turbo.... not 1.9ghz.

[tbone8ty]

http://amdfx.blogspot.de/2012/05/mor...43-better.html

posted over in trinity thread in the xtremenews section

[zhadoom]

http://amdfx.blogspot.de/2012/05/mor...43-better.html

posted over in trinity thread in the xtremenews section

Good news. Not fantastic but recovering wasted time ( bulldozer ).
If this performance is real to final product Trinity may become very profitable to entry level desktop and Notebooks.

[haylui]

http://www.chiphell.com/thread-470447-1-1.html

A10-4600M VS A8-3500M

seems not too bad

[GenTarkin]

man, 3.2ghz mobile part, thats a mighty fine clock speed and still maintain a 35w tdp

[tbone8ty]

need a FX-4100 vs A10-4600m vs Phenom II x4 comparison

[Oese]

The Llano A6-3400M is actually a 1.4ghz part with 2.3ghz turbo.... not 1.9ghz.

yes that one is odd in there.. however, now we dont need these speculations anymore

[Oese]

The Llano A6-3400M is actually a 1.4ghz part with 2.3ghz turbo.... not 1.9ghz.

yes that one is odd in there.. however, now we dont need these speculations anymore

[Oese]

http://www.expertreviews.co.uk/lapto...ototype-laptop

The Intel Core i5-2450M is therefore 8 per cent faster than the AMD Trinity A10-4600M, and seeing as the Intel processor has an 8 per cent faster clock speed this would seem about right. AMD has finally produced a processor that can match Intel's chips in 2D tasks - at least the last generation

Trinity matches a Sandy Bridge i5 clock per clock - you ever dreamed that? :O

http://photos.expertreviews.co.uk/im..._166819_52.jpg

at better battery life whilst having smaller battery :O

[BeepBeep2]

http://www.expertreviews.co.uk/lapto...ototype-laptop



Trinity matches a Sandy Bridge i5 clock per clock - you ever dreamed that? :O

http://photos.expertreviews.co.uk/im..._166819_52.jpg

at better battery life whilst having smaller battery :O

I apologize but the 2450M is a dual-core part with HT. Looks more like 50% behind to me...
Also, gaming performance is rather disappointing when compared to Ivy Bridge, it is better, but not as much as one would expect...most likely due to CPU limitation.

Though, I guess if it is clearly better than Llano CPU-wise I can't complain, though they are now a year further into the game.

[xdan]

Yes, i5 mobile cpu's aren't quad like desktops, they are just dual core i3's with TURBO.
Only i7's mobile are quads.
Multithread performance is pretty worse.
Just chech the video editing performance, it's lower than Llano's.
The average media is better due the image editing media which is probably single thread where Trinity has 3.2ghz TURBO compared with 2.4Llano(1.7ghz stock).

[Oese]

aha thanks for clarifying. whatever 4t vs. 4t comparison to me seems legid, and the i5 is direct competitor, not the i7 AMD just dont offer a high end part here.. And more then 2m/4t in mobile sector to me is only worth it for mobile workstations, nothing i'd need.

6% better then llano IPC like in the other thread is valid? Turbo makes it hard to clearly compare them..

[BeepBeep2]

aha thanks for clarifying. whatever 4t vs. 4t comparison to me seems legid, and the i5 is direct competitor, not the i7 AMD just dont offer a high end part here.. And more then 2m/4t in mobile sector to me is only worth it for mobile workstations, nothing i'd need.

6% better then llano IPC like in the other thread is valid? Turbo makes it hard to clearly compare them..

no, that claim was made for the Phenom II X4 940, DDR2 and all...

[informal]

Why so much debate about IPC in mobile platforms? Trinity is NOT a slow chip ,you get ~3.2Ghz for single threaded tasks and anyone who will tell me this is not enough for 90% of laptop buyers is a fool in my eyes. 3.2Ghz Trnity is probably in the same league as 2.8Ghz C2D in single thread tasks and anyone who knows how fast that is knows this is MORE than enough for 90% of laptop buyers. AMD knows this and they finally produced a chip that excels in perf./watt and baterry life while offering best in class iGPU and very good CPU performance. How many potential buyers need i7 with 8 threads in their laptops? Minority ,a large one. Majority will be better of with well rounded perf./watt efficient chip that can provide decent frames with decent IQ for much less money.

[BeepBeep2]

How many consumers are doing much more than streaming HD video on their laptops anyway? How many are utilizing GPGPU or playing games? A minority...
A simple 160 shader iGPU @ 400 MHz can cover the majority and lower idle AND load consumption very considerably.

I like APUs, don't get me wrong...and I like the way Trinity is coming along...but CPU performance is a little disappointing when compared to 2 core intel solutions with SMT.

Because I don't play games, I can't wait to see the 18w Trinity lineup.

I can't wait to see how they overclock too though, to be honest

[informal]

Let's make some things clear here. 2 core intel CPUs with SMT have very aggressive Turbo and base clocks(yeah,"blame" intel's superior process tech). Apart from that, 4 "core" Trinity has 2 floating point units(2 threads capable each),same as 2 core+SMT intel solutions. So essentially,what you get is a rough parity in SSE/fp benchmarks if clocks were always equal (which they are not of course).
This is why 8 "core" bulldozer on desktop cannot beat 4 core SB/IB in floating point/SSE workloads. It essentially has the same number of floating point units and in workloads that are not FMA optimized it can only reach parity(or around that level roughly). That is AMD's choice,they targeted the design for server integer workloads leaving FP/SSE to heavy optimizations(FMA) to make up for reduced number of execution units per chip.

Numbers that Trinity shows at least give us clue about solid IPC boost for a "quick fix" (of around 5-10%) and a solid power draw reduction at similar clock as bdver1. If this gives us 4-4.6Ghz Vishera with 5-10% higher IPC then AMD may challenge 8T IB in quite a few workloads.

[SEA]

How many consumers are doing much more than streaming HD video on their laptops anyway? How many are utilizing GPGPU or playing games? A minority...
A simple 160 shader iGPU @ 400 MHz can cover the majority and lower idle AND load consumption very considerably.

Khm... you forgot hardware accelerated youtube (HD). are there many people who never used it? Probably everybody uses it...
Almost everybody these days uses gpu just for web browsing. And, the more shaders, the faster page rendering. In many cases only GPU can help to browse comfortably, say in netbooks, tablets, and many desktops.
And it is not just wideo accelerated, many other things.
opera_hw_accelerated.jpgchrome_hw_accelerated.jpg

Also, don't forget, starting from Windows Vista, all windows also hardware accelerated whenever you just move some window.

[BeepBeep2]

Let's make some things clear here. 2 core intel CPUs with SMT have very aggressive Turbo and base clocks(yeah,"blame" intel's superior process tech). Apart from that, 4 "core" Trinity has 2 floating point units(2 threads capable each),same as 2 core+SMT intel solutions. So essentially,what you get is a rough parity in SSE/fp benchmarks if clocks were always equal (which they are not of course).
This is why 8 "core" bulldozer on desktop cannot beat 4 core SB/IB in floating point/SSE workloads. It essentially has the same number of floating point units and in workloads that are not FMA optimized it can only reach parity(or around that level roughly). That is AMD's choice,they targeted the design for server integer workloads leaving FP/SSE to heavy optimizations(FMA) to make up for reduced number of execution units per chip.

Numbers that Trinity shows at least give us clue about solid IPC boost for a "quick fix" (of around 5-10%) and a solid power draw reduction at similar clock as bdver1. If this gives us 4-4.6Ghz Vishera with 5-10% higher IPC then AMD may challenge 8T IB in quite a few workloads.

When was that unclear? Was I wrong in what I said

Khm... you forgot hardware accelerated youtube (HD). are there many people who never used it? Probably everybody uses it...
Almost everybody these days uses gpu just for web browsing. And, the more shaders, the faster page rendering. In many cases only GPU can help to browse comfortably, say in netbooks, tablets, and many desktops.
And it is not just wideo accelerated, many other things.
opera_hw_accelerated.jpgchrome_hw_accelerated.jpg

Also, don't forget, starting from Windows Vista, all windows also hardware accelerated whenever you just move some window.

...

How many consumers are doing much more than streaming HD video on their laptops anyway? How many are utilizing GPGPU or playing games? A minority...
A simple 160 shader iGPU @ 400 MHz can cover the majority and lower idle AND load consumption very considerably.

I like APUs, don't get me wrong...and I like the way Trinity is coming along...but CPU performance is a little disappointing when compared to 2 core intel solutions with SMT.

Because I don't play games, I can't wait to see the 18w Trinity lineup.

I can't wait to see how they overclock too though, to be honest

...

[Electroneng]

When are the desktop versions due? I have not seen a lot of information on that lately!

[SEA]

CPU performance is a little disappointing when compared to 2 core intel solutions with SMT.
...

Sorry for unclear answer...
That was the point of my answer - everybody uses gpu all the time, even not realizing that, and even not using GPGPU or games.
Majority is browsing and streaming HD so they need gpu to have their laptops' batteries last and last.
They have enough cpu performance for those uses anyway....

[Oese]

no, that claim was made for the Phenom II X4 940, DDR2 and all...

aha lolz

Well anyway, it seems like AMD really succeeded in their strategy. For me, a a10-4600 or a10-4655 seems to be best buy. I hope for a quality laptop (non-glare panel, enough brightness and contrast for use outside, good keyboard, good case, good battery) with these processors.

I like the large iGPU cause i play mostly older games with laptops, but where athlon ii x2 with 4650 gpu and i5/HD3000 (Dell Latitude) dont really deliver a good performance. Trinity would fit much much better for me, but i dont like cheap entertainment laptops or ultrathins since i work on them also and they are too slow for my liking.

And well, I am interested in IPC since its not only trinity but vishera I am interested in, of course. My BD runs nice but I'll drop in next gen FX instantly it seems

[muzz]

I'll be happy to see some real Piledriver reviews....APU means squat to me.

[tbone8ty]

http://www.xbitlabs.com/news/cpu/dis...ips_in_Q3.html


so Trinity has "early" piledriver cores and Vishera will have "full" piledriver cores... can somebody explain this?

so basically vishera piledriver is even better than trinity piledriver?

interesting since trinity piledriver cores showed some nice improvements over bulldozer.