First better info about Steamroller (SR) after Hotchips 2012
http://www.anandtech.com/show/6201/a...r-architecture
http://www.computerbase.de/news/2012...hwenk-zurueck/
http://techreport.com/articles.x/23485
First better info about Steamroller (SR) after Hotchips 2012
http://www.anandtech.com/show/6201/a...r-architecture
http://www.computerbase.de/news/2012...hwenk-zurueck/
http://techreport.com/articles.x/23485
nice one
Werd!
Excavator looks to be twice as high on Amd's chart ;)
30% increase over steamroller! Lol
What process node will be available by then? 20nm?
Maybe the increase will be big because it will be an APU by then!
Should we include jaguar/hondo? Eh maybe not lol
it's too quite on this subject matter ?!?!?!
has Two decoders 4 way to each core ? (unknown if true)
Has increased instruction cache size (size unknown)
most likely will still have 16Kbytes 4 way cores
Less L2 cache said to be 512Kbytes
Steamroller will be successful
I'm predicting an 8 core APU with all thr die spacr savings of 28nm and dynamic library's
Kaveri is already confirmed in AMD's slides from Analyst Day to have 4 cores and 8 Compute Units (8 x 64 = 512 Radeon Cores). And the move from 32nm SOI to 28nm Bulk won't show any major gains in die savings.Quote:
Originally Posted by tbone8ty
you're right in that we wont see any improvements from 32nmSOI->28Cmos in terms of die space or power consumption but where we will see gains is cost. If I remember correctly SOI is something like 20-30% more expensive, with luck AMD will be able to give us a good dual module apu to compete with the pentiums in both desktop and laptop. I really think that trinity would make an amazing ultrabook processor, sadly hp make the only AMD 'sleekbook' at 15"(!).Quote:
Originally Posted by Smartidiot89
It could very well result in power savings if that is true that GF 32nm is leaky, very leaky, and 28nm is better.Quote:
Originally Posted by bobbth
But just like any new process tech on a processor for AMD, it will probably take time to get right. So 28nm bulk may be just as leaky at first as 32nm SOI has been so far.
I post this in the other thread too.
http://www.anandtech.com/show/6201/a...architecture/2Quote:
Steamroller brings no significant reduction in L2/L3 cache latencies. According to AMD, they’ve isolated the reason for the unusually high L3 latency in the Bulldozer architecture, however fixing it isn’t a top priority. Given that most consumers (read: notebooks) will only see L3-less processors (e.g. Llano, Trinity), and many server workloads are less sensitive to latency, AMD’s stance makes sense.
so L3 cache has some problem ?
L3 has every time higher latencies than others caches, it depends at architecture.
seems like kaveri will be a nice step in both cpu and gpu performance
I heard on techreport podcast that amd is going to release a 19w quad core trinity for ultrathin market.
Steamroller "bdver3" Compiler Support:
http://www.phoronix.com/scan.php?pag...tem&px=MTIwNDY
Interesting thing like this:
Quote:
The bdver3 contains three pipelined FP units and two integer units. Fetching and decoding logic is different from previous fam15 processors...
this confirms that the MMX is a separate pipeline from the others then.Quote:
Originally Posted by cezar
Like the Steam roller diagrams shows.
No steamroller? No kaveri? R u serious? Amd is :banana::banana::banana::banana:ed
radeon 7,XXX and 8,XXX.Quote:
Originally Posted by tbone8ty
Game over AMD. Seems like next 5 quarters of big money loss. In their current financial situation there is no hope for them.
Tap tap tap - SGS 2
Kabini with Jaguar looks great, Im thinking. Vishera all next year? If AMD will be lucky, maybe at end of 2013 is SR possible. In worse case it could be sometimes in H1 2014. So...we will see maybe higher Vishera models in 2013.
possible with vishera on 22nm node?
No, i think no....Im meaning, we will see maybe 8320/8350 in 95W version later and new FX-8370 (maybe 8390) in 2013
Something is odd with that chart. I was under the impression that Richland was using Steamroller cores and is based on the 28nm manufacturing process.
you are wrong, Richland is in old roadmap still for FM2 and with Piledriver cores - it is "Trinity 2.0" platform in some roadmap. Everything on 28nm node (Kaveri/bdver3) is delayed to 2014, GloFo 28nm is not ready for mass production. Steamroller is 28nm architecture, is delayed too.Quote:
Originally Posted by SKYMTL
this slide remind me, "someone about whom we do not speak" blog :) was he right?Quote:
Originally Posted by Olivon
Yup it's gonna be a long long 2013 for amd
Proof on the delay to 2014? All I ever saw was speculation based on the upcoming layoffs.Quote:
Originally Posted by PedantOne
Yeroon
Why not? They've done dumb shrinks before. 22nm might help them hit 95w.Quote:
Originally Posted by FlanK3r
What's a "dumb shrink"? They don't simply press the shrink button when a process is finished and working chips comes out the oven. There are enormous challenges to it and lots of things needs to be remade - it's costly and extremely time consuming, think not of several months rather several quarters. Just shrinking to a new process without doing any sorts of improvements doesn't make sense.Quote:
Originally Posted by Darakian
I recall one of the AMD executives (i can't find the quote anymore) making a statement during an interview and saying something to the effect that the "piledriver core of 2013 will not be the same piledriver core of the initial release".Quote:
Originally Posted by Olivon
Maybe they have further enhancements that will occur in 2013 to piledriver before shifting to SR cores in 2014.
I think it's just a clock bump. With SR finished there is no point in tweaking PD core and I doubt they can get much more than what we have now anyway. PD is a solid bump in performance Vs Bulldozer,I just don't think it can hold up to Haswell next year,that's all.
Is Radeon 2.0 = GCN?
A dumb shrink is the term given to shrinking a core without giving a redesign, ex. newcastle to winchester in the athlon 64 days. As I recall the only benefit to doing a dumb shrink is to lower the power consumption.Quote:
Originally Posted by Smartidiot89
And to get things tested so the new architecture can be designed better. Its why new node + new arch atr the same time almost always have problems of some sort. Just look at Phenom I and Bulldozer. All fixed with subsequent core tweaks because of too many issues from trying too much at once. Steamroller looks great on paper, I just hope it executes well since it is a shink and somewhat of a core re-design again. Granted, at least its not a 100% architecture change so it has more hope than Bulldozer did.Quote:
Originally Posted by Darakian
everywhere I've looked says richland is 28nm APU of piledriver. it's trinity on half node (yes AMD is doing half node with APU's)
Kaveri is steamroller apu on 28nm. Steamroller was already stated to be in tapped out and sampling off 28nm.
richland it self looks to be taken off some older road maps it was on them with Krishna and Wichita. Those are no long their and Brazo 2.0 in both their places.
In the case of Phenom I, it was not their first 65nm processor. Remember Brisbane Athlon 64's?Quote:
Originally Posted by EniGmA1987
This makes no sense. How can AMD release bdver3 GCC support which implies massive changes to decoders and stuff, and then cancel it? Wouldn't GCC support for Steamroller mean that it's already done? Are they just going to throw the chip away?
Also, what about rumors of PS4 using steamroller? I don't understand. Does Sony need so many Steamroller CPUs that AMD can't provide enough for everyone and has to use a tweaked 32nm PD 2.0 for desktop?
hmm whats radeon 2.0?Quote:
Originally Posted by tbone8ty
If I well understand, Richland is Trinity + HSA feature and Bolton PCH. Graphic parts is maybe the same.
Radeon 2.0 = VLIW4Quote:
Originally Posted by tbone8ty
explains slide a little more maybe
http://wccftech.com/amd-richland-apu...d-8000-series/
hopefully kabini and tamesh on 28nm do well
i think Radeon 2.0 = GCN ;)Quote:
Originally Posted by vietthanhpro
I know Radeon 2.0 = VLIW4 ;) Seriously though, it has been stated quite clearly in earlier press materials that "Radeon 2.0" is in Trinity, which is VLIW4. Don't expect more then clock bumps with Richland...Quote:
Originally Posted by tbone8ty
in official AMD TRinity slides is Radeon Cores 2.0 = VLIW4, Radeon GCN 1.0 = GCNQuote:
Originally Posted by tbone8ty
Yeroon = proof is that leaked roadmap?
Compiler support would imply that the design is done. Fabbing a working chip and doing so en-mass is the next step. Do you have a link for the compiler support?Quote:
Originally Posted by sdlvx
Now, Trinity: 4 core x 3.8ghz x 8 + 6 SIMD x 64 x 0,8Ghz x 2 = 736 GFLOPSQuote:
Originally Posted by Slide #34 of Dr. Lisa Su's presentation
--> Richland 28nm: 819 GFLOPS
Richland: 4 core x 4Ghz x8 + 6 SIMD x 64 x 0.9Ghz x 2 = 819.2 GFLOPS
Yeah, slide is good enough for me. Havent seen that one earlier.Quote:
Originally Posted by PedantOne
It makes sense that no major changes will happen until 28nm, so richland being PD cores and vliw4 doesn't suprise me. I did read about trinity 2.0 in the summer, but was hardly mentioned again, but if trinity 2.0 = richland, that makes a lot of sense.
this is getting confusing, whats the actual changes for richland compared to trinity? im reading that its just trinity with CGN.
Well trinity with GCN would be great for a refresh....Some improvement in memory controller would be welcomed also :)
in very old roadmap was Richland APU, it was Dual Core + VLIW-4 196 SP?s, originally Trinity for notebooks!
http://www.abload.de/img/cdb3a587-856e-4043-87ojbqe.jpg
But this is "new Richland", There are two possibilies about it:
1. Piledriver CPU + VliW-4 GPU on 32nm node, higher clocks, nothing more
2. Whole Trinity (Piledriver/VLIW-4) on 28nm node
I bet first is right!
http://semiaccurate.com/2012/11/06/a...gain-2014-now/Quote:
More recently, SemiAccurate?s moles have come back and said that Kaveri has slipped yet again to 2014 if it is still alive. Big if. That would put the chip 18 months late best case, and then you have to ask yourself why not just skip it? Time will tell what happens, but one thing that is not in question is if there will be an updated APU in 2013. That answer is definitely no, Intel has the game all to itself next year.
Time will tell if any of the rumors are true or false.
if anything steamroller needs time for maturity.Quote:
Originally Posted by AMDforME
From x264dev irc
Thanks to cal_guy on B3DQuote:
2012-11-14 23:42:21 < Skyler_> amd just sent me a Steamroller CPU.
so they do have ES out already? :P
is this real?
that means in 2013 we have steamroller :)
no way, there are NO SR ES samples in the wildQuote:
Originally Posted by Evantaur
it's an FM2 sample, he says kaveri, could be richland.Quote:
Originally Posted by PedantOne
richland is 32nm, Kaveri 28nm, right?
Possibly cancelled:
http://semiaccurate.com/2012/11/19/a...and-excavator/
Richland is 28nm.Quote:
Originally Posted by FlanK3r
Do not we know for half a year, that is the last AMD High-end desktop cpu ever is Vishera? - Yes, we know already.
If you believe OrbQuote:
Originally Posted by PedantOne
Then you also believe a 5770 is faster then a 5870. =|
what?
OBR is "special" guy. Best avoided.
On the topic of "Kaveri cancelled" BS ,SA retracted the story since AMD denied the big cores are cancelled and stated they are still on track.
i wonder how many clicks they got from that
Of course AMD denied it, what else should they do? SA are obligated to print the AMDs denial comment - doesn't make it true.
So if Amd came out with an 8 core FX chip with steamroller what r the projected % gains from say an fx-8350?
If I recall steamroller is more of an ipc improvement
We won't be seeing FX based on SR core next year. Maybe 2014. Best case scenario is APU Kaveri,if it launches next year.
OBR may be an ass, he has been correct on quite a lot of things lately. Dismissing anything he says based on the fact that he's a bit "special" is kind of weird.Quote:
Originally Posted by informal
He is also banned here,but you know that :). Even a broken clock shows a correct time 2x a day,it doesn't mean it's OK to always check for time on it :).
He wasn't banned for misinformation, but for being an ass.
Has he been way off lately on AMD-related topics? Afaik, most of the number he posted were at least close to accurate.
He saw one roadmap that has no new FX in 2013 and proclaimed Vishera is the last high perf. desktop CPU. He reposted stuff(Vishera ES) from a Chinese forum/site and claimed he leaked it. And he has a pathological hatred for AMD. That's not a reliable source IMO,that's an attention whore :) (pardon my French).
He said that PD's (8350) power consumption is higher than BD (8150).Quote:
Originally Posted by massman
Oh I forgot to add: he claimed that HD48xx series won't have 800SPs but 480SP. He failed there too. And that blog post of his is gone,of course.
No, he said his sample had higher consumption. Which is absolutely plausible as some other sites measured this as well under certain load tests. People need to read more carefully before bashing the Nameless One :DQuote:
Originally Posted by repman
whats i know, in times of HD4800 his blog doesnot exists! You are trying to deceive us again as always, bro :)Quote:
Originally Posted by informal
Link? I checked quite a lot of reviews and haven't seen that.Quote:
Originally Posted by boxleitnerb
He posted it for sure,I cannot recall where but there are others who remember it I'm sure ;).Quote:
Originally Posted by PedantOne
Also you said:
Who are "us". Are you OBR? If so why did you register with this account here again?Quote:
You are trying to blame us again as always, bro
It doesnt matter if the blog existed or not. What matters is that all those misinformation came from OBR himself, be it from his blog, forum posts, etc.Quote:
Originally Posted by PedantOne
sorry for my english, i am german. i thought blame us is deceive us :) corrected
repman - http://www.3dcenter.org/artikel/laun...r-rechenkernen
From six FX-8350 reviews has FX-8350 higher power draw in four! table : Volllast (Syst.) down on page
So of all the things he wrote, there's now one or two that are incorrect. That's not very much of a basis to disregard everything he writes by principle.
his attitude towards amd just shows that he's a douchebag. He doesn't post anything but amd stuff. His posts about amd stock just shows you that his blog has only one motive.
regardless can we not talk about anymore?
this is derailing the thread
back on topic.Quote:
Originally Posted by tbone8ty
ah yes back to steamroller.
I was thinking about this module only has one fetch for two cores
A normal dual core would technically have two fetches for the two cores one for each core.
Can the front end even give enough data for two decoders from one fetch, it seems like it would be the major bottleneck?
SR core addresses this ;).Quote:
Originally Posted by demonkevy666
hopefully there is a amd fanboy on the JP Morgan finacial team and there showing him a steamroller chip and hes like daaaaaaaamn intel killa
lol
No. SR has two decodes, one fetch.Quote:
Originally Posted by informal
Page 12
Well what I meant it addresses the front end bottleneck(and fetch is in front end). And they optimized the fetch stage Vs BD/PD,it's more efficient. But yeah,there is one fetch per module as before.
I think the fetch is not the main culprit in BD/PD design, the numbers of decoders were.Quote:
Originally Posted by informal
It's obvious it's decoding stage since that's what AMD doubled ;). But they did optimize the fetching stage too,it's more power efficient now and has a micro-op queue -this is a new feature Vs BD/PD cores .
I feel like they could use a ring bus for steamroller.Quote:
Originally Posted by informal
wow 28nm is far out, well at least its on the road map now and not in lala land
Posted a while ago. Still true?
http://www.theinquirer.net/inquirer/...or-steamrollerQuote:
CHIP DESIGNER AMD has said that its upcoming Vishera chip will not be the last to use Socket AM3+.
Nobody knows,AMD included :D. They have changed their roadmaps so many times in the last few years that making any predictions on the socket longevity front is useless.
Yeah, holy trueQuote:
Originally Posted by informal
by the time they release a new desktop part they will be forced to have a new socket cause am3+ will be so old by then
did anyone read this ?Quote:
"Steamroller is not Bulldozer Enhanced. F*** no. The layout might look the same but our LEGO blocks are completely different. When all is said and done we should get 45% improvement and this goes to show how the Bulldozer was f***** design. This is all what Bulldozer was supposed to be."
Read more: http://vr-zone.com/articles/amd-push...#ixzz2GYA1FnAK
Yes it was posted already on this forum. Smells like bs to me.
http://www.fudzilla.com/home/item/29...an-28nm-bobcat
This is about kabini....looks promising