AMD to skip Piledriver and place hopes on Steamroller

[Andrew LB]

If this is a legit rumor, then call i can say is "ouch"...

AMD to skip Piledriver and place hopes on Steamroller?


By Hilbert Hagedoorn, August 21, 2012 - 8:05 PM N/A


Guru3D.com ImageVR Zone heard rumors (and really it's just that) that AMD's Piledriver "Vishera" processors will not be replacing the "Zambezi" Bulldozer chips because there's simply not enough improvement in per core performance from Bulldozer to Piledriver to justify a new product release. The site heard speculation that AMD is now placing its hope on Steamroller, a 28nm core planned for early 2013.

Only the "Abu Dhabi' dual die lower clock 16-core total server chip should be out next month. Why?

Well, one reason could be that, maybe, there is not enough improvement in per core performance from "Bulldozer" to "Piledriver" to justify the new product release, as it wouldn't improve the competitive position against Intel in the desktop segment at all. Whether it is 'instruction per clock' IPC or pure clock frequency, both have to contribute to the end performance to justify the product.

The other reason, linked to the first one, can be - likely - that AMD now has to place its hopes to the upcoming 'Steamroller' core early next year, coupled with the 28 nm process. I guess we can accept that and give the AMD CPU team more time to do that migration, as long as it is real soon now. Else, another excuse this time next year, to maybe wait for 'Excavator' to fix things, would be really unacceptable...
On another note, it's also unclear what's going on with AMD's desktop Trinity APUs.

http://www.guru3d.com/news/amd-to-sk...n-steamroller/

[EniGmA1987]

Someplace else mentioned this a few days ago as well. I hope it is wrong because Zambezi needs to be replaced, even if it is just with a 10% total improvement. Better than nothing.
If true, I hope it is because the Steamroller shrink is going much better than expected and will ship ahead of schedule. That is the only reason I could see for a company not to put out a new product to make money with.

[Greg83]

actually a great plan imho.
its only early next year. why release new 8 cores with nil real benefits. this way intel just remains dormant and has a safe feeling inside.
not like anyone with an 8 core amd is gonna upgrade to another 8 core , unless they have tons of disposable income.

i need a min a 50% improvement in perf to upgrade. which likely means, 10 cores, quad channel memory , pci-e on chip and 20% IPC increase.

[Gunslinger]

Maybe "Excavator" will dig their desktop market grave.

[st0ned]

Maybe "Excavator" will dig their desktop market grave.

Stop steamrolling AMD :O

[Greg83]

actually, thats the consumers, putting such high value in intels, they're behind in number of cores, but ppl buy them up anyway, knowing full well if amd goes down, that 2600K is $1000 and there won't be any real progress. the intel way.
n then that 3-10% better in gaming, for comparable priced processors will cost everything

an example being. i went 8 core amd. my buddies uncle went amd. my buddies cousin , who reads "reviews" went intel socket 2011 quad core. is about equal in games to me. :P losses in some games wins by marginal amounts in others and paid 3 times what we did.

even though i proved to him how he wasn't right about intel being better in the game he wanted it for. he still put all that money in , so he can upgrade to ivy-e and then , delayed and likely never gonna be in his budget.
and by the time it does, he'll have a 25% at best gains over me in games, after upgrading to a 6 core for $600 more dollars. making his pc upgrade cost 6 times what the rest of us who went amd's did. for 25% better perf lol in games running more then 60FPS on lcd's.

[st0ned]

Sorry but you are wrong specially for gaming, and not looking at the price intel Ivy & sandt steamrolles AMD FX CPU's.

[Greg83]

http://images.anandtech.com/graphs/graph5276/43293.png
http://images.anandtech.com/graphs/graph5276/43295.png


http://media.bestofmicro.com/I/B/318...202%201920.png
http://media.bestofmicro.com/I/C/318...202%202560.png
http://media.bestofmicro.com/I/E/318...203%201920.png
http://media.bestofmicro.com/P/R/318...rt3%202560.png

anything under 1080p has nothing to do with gamers. and that is why reviews of amd 8 core vs intel socket 2011 4 core is rare in the reviews
when i see intel winning in games, its rare and selective games of which they do not play or outside the scope of any real world use for that system.
prob due to single core perf. which has nothing to do with new games, where the gap between the oc'd 8120 and the oc'd 3180 is gonna be none existent in maxed out 1080p games.

[informal]

Is vrzone and guru3d for real with this "news article"?? This is total BS I'm afraid. AMD won't skip Piledriver,server parts are launching next month and desktop is following soon after that. Another thing to add is that SR core is slightly delayed and Kaver is not Q1 2013 any more. So Q2 for APUs and Q3 for server and desktop SR variants is most probable scenario.

I had no idea VRrzone sank so low now that they have to fabricate "news" in order to get traffic...

[st0ned]

http://images.anandtech.com/graphs/graph5276/43293.png
http://images.anandtech.com/graphs/graph5276/43295.png


http://media.bestofmicro.com/I/B/318...202%201920.png
http://media.bestofmicro.com/I/C/318...202%202560.png
http://media.bestofmicro.com/I/E/318...203%201920.png
http://media.bestofmicro.com/P/R/318...rt3%202560.png

anything under 1080p has nothing to do with gamers. and that is why reviews of amd 8 core vs intel socket 2011 4 core is rare in the reviews
when i see intel winning in games, its rare and selective games of which they do not play or outside the scope of any real world use for that system.

Are you for real ? Does AMD show up in anything but last at any of those links you gave... I rest my case then. And I don't know how you would be wanting to see bigger differences then that when you go for 1080p / 1600p resolutions where the graphic card is the weakest link, go for lower res where you are CPU bound and see what's going on.

[Greg83]

i didn't say they won , when did i say that, i said he paid 3 times for 3-10% gains. all those show 1-3% gains. he doesn't play single core games ....

all you proved is you were just trolling and didn't read anything

plus anything over 60FPS is m00t on lcd's

as i was talking ABOUT REAL WORLD, not theories of future possibilities, which might put amd on the top in the games he'd play, when games have 8 threads and in windows 8 , with the better support for the bulldoser. since in all these cases, it was a 1-3% dif.

hes only got crossfire 5770's still. he shoulda listened to me and budgeted with sense. instead of listening to reviews. and got a real upgrade for that $900 he dropped.

[SKYMTL]

Let's not get into fanboy territory on this one.

It makes sense from many standpoints to skip a delayed architecture and put additional resources into one that's launching in the not-so-distant future. However, should Steamroller be delayed in any way, things will get very, very ugly.

[informal]

They are not skipping Piledriver, the author of the text at VRzone is making things up.

[lutjens]

However, should Steamroller be delayed in any way, things will get very, very ugly.

Given AMD's abysmal track record on execution over the past few years, this statement doesn't give me a warm fuzzy feeling.

[Greg83]

it wasn't very clear. but he was only talking about on the am3+ desktop platform.

[FlanK3r]

not so true...New FX will be with higher clocks and few better performance per clock. Not much, but this is only refresh as SB to IB (look at performance with the same clocks SB and IB - IB turboost working differently than at SB-DT)

[MattiasNYC]

i didn't say they won , when did i say that, i said he paid 3 times for 3-10% gains. all those show 1-3% gains. he doesn't play single core games ....

I don't really have a dog in this fight, but I figured I'd just quote the above for truth.

Many people seem to forget or ignore the bigger picture. The bottom line is that if you have a budget to stick to and a specific use for the CPU you're buying then price/performance is king.

So I think your point was well made.

[Greg83]

couldn't they just release binned higher Zambezi and avoid the switchup and still release the new FX like 8170 http://www.cpu-world.com/CPUs/Bulldo...20FX-8170.html
if the IPC different isn't worth the release for a single processor for the market for the 4-5 months until the next one might be ready.

[Brice MJ]

Is vrzone and guru3d for real with this "news article"?? This is total BS I'm afraid. AMD won't skip Piledriver,server parts are launching next month and desktop is following soon after that. Another thing to add is that SR core is slightly delayed and Kaver is not Q1 2013 any more. So Q2 for APUs and Q3 for server and desktop SR variants is most probable scenario.

I had no idea VRrzone sank so low now that they have to fabricate "news" in order to get traffic...

There is NO Steamroller-based FX in any AMD client 2013 roadmap. And according to Charlie Kaveri is now H2/2013, with a Trinity 2.0 shipping in late Q1/early Q2 (aka Haswell launch date).

[Ket]

i didn't say they won , when did i say that, i said he paid 3 times for 3-10% gains. all those show 1-3% gains. he doesn't play single core games ....

all you proved is you were just trolling and didn't read anything

plus anything over 60FPS is m00t on lcd's

as i was talking ABOUT REAL WORLD, not theories of future possibilities, which might put amd on the top in the games he'd play, when games have 8 threads and in windows 8 , with the better support for the bulldoser. since in all these cases, it was a 1-3% dif.

hes only got crossfire 5770's still. he shoulda listened to me and budgeted with sense. instead of listening to reviews. and got a real upgrade for that $900 he dropped.

Performance numbers aside theres still one very strong reason to buy a intel CPU - power usage. Have you seen the insane power usage figures for Bulldozers? They are absolutely horrible in every possible way.

http://www.techspot.com/review/452-a...us/page11.html

In that link you can see the FX 8150 drawing 252w @ load and the FX 8120 drawing 244w @ load. Even the FX 6100 eats a considerable 225w @ load. By comparison, the very popular 2600k is only pulling 166w @ load. Unfortunately that article doesn't show how much power the AMD chips suck up when OCd, but the stock figures alone give you a good insight into the nightmare electricity bills that await those Bulldozer owners brave enough to OC their CPU. That article also shows that at more CPU limited resolutions the Bulldozer chips get comfortably beaten by the 2500k and 2600k CPUs, which shows that intel do indeed have a more efficient design all round.

[Greg83]

only idle power usage has any relevance though to a gamer. you only have 2 hours a day to game if you can afford the intel, cause you have a job.
which doesn't really use the same power as prime95 would.
now if you're a folder, you also do not care about power usage or the cost of hydro, where it might matter.
web browsing is still the main usage of all our pc's followed by movies. some cases games next.
its also twice the cores ... and only reason intels 4 cores win so much. is cause we had quad cores for so long and games often got coded upto 4 core.
so when comparing core for core at Mhz. instead of the perfs for the power usage. the intel is the power hog , just like the Pentium 4's with HT its mimicking

and this one painted a diff picture
http://www.bit-tech.net/hardware/cpu...-8120-review/7
since my buddy went the 2011 route, the power is higher there.

n http://www.hardwareheaven.com/review...rclocking.html

[Ket]

only idle power usage has any relevance though to a gamer. you only have 2 hours a day to game if you can afford the intel, cause you have a job.
which doesn't really use the same power as prime95 would.
now if you're a folder, you also do not care about power usage or the cost of hydro, where it might matter.
web browsing is still the main usage of all our pc's followed by movies. some cases games next.
its also twice the cores ... and only reason intels 4 cores win so much. is cause we had quad cores for so long and games often got coded upto 4 core.
so when comparing core for core at Mhz. instead of the perfs for the power usage. the intel is the power hog , just like the Pentium 4's with HT its mimicking

Did you not see I also mentioned the FX 6100? That eats a considerable 225w @ load. Load power figures are obviously more of interest than idle figures. Folders especially will care about how much power is being used @ load because their systems are always @ load. For folders, gamers, or casual internet browsers Bulldozer eats up more power at idle or load which will lead to somewhat higher electricity bills vs intel.

[Greg83]

yes, but the gap gets smaller for the socket 2011 3820 quad core. and amd is behind intel a bit with its tick tock. so it won't it be better to compare it to the socket 1366 ones. amd is behind yes. that i get. but if you think about it. thats a light bulb.

[demonkevy666]

Performance numbers aside theres still one very strong reason to buy a intel CPU - power usage. Have you seen the insane power usage figures for Bulldozers? They are absolutely horrible in every possible way.

http://www.techspot.com/review/452-a...us/page11.html

In that link you can see the FX 8150 drawing 252w @ load and the FX 8120 drawing 244w @ load. Even the FX 6100 eats a considerable 225w @ load. By comparison, the very popular 2600k is only pulling 166w @ load. Unfortunately that article doesn't show how much power the AMD chips suck up when OCd, but the stock figures alone give you a good insight into the nightmare electricity bills that await those Bulldozer owners brave enough to OC their CPU. That article also shows that at more CPU limited resolutions the Bulldozer chips get comfortably beaten by the 2500k and 2600k CPUs, which shows that intel do indeed have a more efficient design all round.

I'm not sure if power saving features where on in this Review.
things like, APM, C1E, C6, Cool and quite. All that is their to help save power.

However Care to explain to how a chip with such low IPC mange to get the OpenGL test to only show 9% slower FPS vs I5/I7 and 6% faster then phenom II X6 and 4% faster then X4
They're all using the Same GPU too "Gigabyte GeForce GTX 580 SOC (1536MB)"

back on topic for now
I think skipping it would be a good idea.
If the leaks shown are truly pile-driver then something has gone wrong with the L3 adding on the chips and works better with out the L3 cache as seen in trinity, which has pile-driver cores.

[highoctane]

To me it seems it would be odd for them to skip PD considering the resources/expense to get it ready to launch up until now, just throwing time and money away.

As for power consumption, with modern hardware a idle computer should be going S3 not sitting there sucking power and putting out heat doing nothing, with a ssd wake is near instant or as fast as the screen can power up.