AMD CPU speculation... and expert conjecture - Page 9
Tags:
-
AMD
-
CPUs
Last response: in CPUs
Related resources
- Tek Syndicate: Expert Conjecture and Speculation - Forum
- AMD Steamroller rumours ... and expert conjecture - Forum
- Will the 4+4 pin CPU power on the Corsair RM850 Gold reach to 8 pin header on the Asus Z87- Expert in the Corsair Obsidian 750 - Tech Support
- I need a CPU expert's opinion on this - Tech Support
- 5/5 on CPU expert, but no badge? - Forum
dozerman
December 6, 2012 3:02:09 PM
fazers_on_stun said:
Heh, we heard ya the first time
..Anyway AMD has been stuck around the 5% mark in server marketshare for over a year now - if PD and/or the SeaMicro effort pays off, we might get a hint during the Q4 earnings report due out in about 6 or 7 weeks from now. But AMD undergoing another 15% layoff a few months after the last one, plus selling their Austin campus, plus seeking out Goldman Sachs and getting another cash bailout from ATIC does not bode well for any possible sudden increase in their cash flow from the most lucrative segment.
Only thing I've seen recently is Rory Read stating something about black Friday not sucking too much, which would seem to be about the consumer segment as I doubt IT buyers honor the black Friday tradition too much and line up at night to be first in line to buy a server farm at 50% off prices
..Sorry about that. My phone kept telling me that the data didn't go through, so I kept hitting the submit button. XD
-
Reply to dozerman
esrever said:
the chips were designed to run higher. Just looking at the power consumption and the OC numbers. If you can cool bulldozer, it OCed. The leaking 32nm process made it impossible to run the chips at high speeds. They wished it would have run faster, but there is a limit to every 32nm process. Intel, IBM, TSMC, Samsung, GF, none went past 4Ghz without the power consumption going through the roof.
If Intel or TSMC were shipping 5ghz standard CPUs then you might have an argument against GF.
-
Reply to Cazalan
Cazalan said:
They wished it would have run faster, but there is a limit to every 32nm process. Intel, IBM, TSMC, Samsung, GF, none went past 4Ghz without the power consumption going through the roof. If Intel or TSMC were shipping 5ghz standard CPUs then you might have an argument against GF.
ibm ran at 5ghz at 32nm. its not only the process that dictates how high a cpu can clock.
-
Reply to esrever
esrever said:
ibm ran at 5ghz at 32nm. its not only the process that dictates how high a cpu can clock.You missed the part "without the power consumption going through the roof".
In all fairness, AMD has been consistent with PD in the 125W envelope using 4Ghz.
Intel as well, if you think about OC. They've gone past 4Ghz since a long time ago if they wanted to keep the 125W TDP mark.
Now, depending on the way to use Silicon to get things out in a given process node, you have different thermal behaviors. AMD has always been more resistant to heat, but that doesn't mean it's a bad process because it heats up more (half agree with myself on this, haha).
I like to use the 4850 single slot using 65nm (IIRC) as a very good example; that thing could handle 80°c + with no issues XD
Cheers!
EDIT: Typos and missing text
-
Reply to Yuka
It all depends on pipe length and other factors, as well as usage.
Why dont GPUs do 3-4 Ghz?
Waaaaaay too wide, so they dont need to, as width used in parallism does what its designed to do, as we see Knights corner/LRB etc go with many cores as well.
A true Central Processor as used in DT can and does use higher clocks, where specialized chips on same node can go higher
Why dont GPUs do 3-4 Ghz?
Waaaaaay too wide, so they dont need to, as width used in parallism does what its designed to do, as we see Knights corner/LRB etc go with many cores as well.
A true Central Processor as used in DT can and does use higher clocks, where specialized chips on same node can go higher
-
Reply to JAYDEEJOHN
http://money.msn.com/business-news/article.aspx?feed=OB...
Advanced Micro Devices will scale back on purchases from manufacturing partner GlobalFoundries Inc in 2013, allowing the struggling microprocessor firm to conserve cash and reflecting a slowdown in the personal computer market.
A distant runner-up to Intel Corp in PC microprocessors, AMD is trying to raise capital and arrest steep declines in revenue as the PC market sputters. It divulged plans last month to sell and lease back its campus in Austin, Texas, to raise cash and fund its chipmaking business.
On Thursday, AMD said it had struck a deal with GlobalFoundries, under which it lowered wafer purchase commitments for the fourth quarter of 2012 to about $115 million, and to $1.15 billion in fiscal 2013.
Along with lowered payments for research and development, the agreement with the foundry partner should allow AMD to return to free cash flow generation by the second half of 2013, the company said in a statement.
Quote:
AMD cuts back on GlobalFoundries purchases in 2013Advanced Micro Devices will scale back on purchases from manufacturing partner GlobalFoundries Inc in 2013, allowing the struggling microprocessor firm to conserve cash and reflecting a slowdown in the personal computer market.
A distant runner-up to Intel Corp in PC microprocessors, AMD is trying to raise capital and arrest steep declines in revenue as the PC market sputters. It divulged plans last month to sell and lease back its campus in Austin, Texas, to raise cash and fund its chipmaking business.
On Thursday, AMD said it had struck a deal with GlobalFoundries, under which it lowered wafer purchase commitments for the fourth quarter of 2012 to about $115 million, and to $1.15 billion in fiscal 2013.
Along with lowered payments for research and development, the agreement with the foundry partner should allow AMD to return to free cash flow generation by the second half of 2013, the company said in a statement.
-
Reply to fazers_on_stun
earl45
December 6, 2012 9:43:10 PM
fazers_on_stun said:
http://money.msn.com/business-news/article.aspx?feed=OB...Quote:
AMD cuts back on GlobalFoundries purchases in 2013Advanced Micro Devices will scale back on purchases from manufacturing partner GlobalFoundries Inc in 2013, allowing the struggling microprocessor firm to conserve cash and reflecting a slowdown in the personal computer market.
A distant runner-up to Intel Corp in PC microprocessors, AMD is trying to raise capital and arrest steep declines in revenue as the PC market sputters. It divulged plans last month to sell and lease back its campus in Austin, Texas, to raise cash and fund its chipmaking business.
On Thursday, AMD said it had struck a deal with GlobalFoundries, under which it lowered wafer purchase commitments for the fourth quarter of 2012 to about $115 million, and to $1.15 billion in fiscal 2013.
Along with lowered payments for research and development, the agreement with the foundry partner should allow AMD to return to free cash flow generation by the second half of 2013, the company said in a statement.
When is there going to be good news for AMD?
-
Reply to earl45
esrever said:
ibm ran at 5ghz at 32nm. its not only the process that dictates how high a cpu can clock.I recall that it needed about 1200W just to power the CPU. Not in the same ballpark.
It also costs over $20,000 for the CPU the way IBM charges for "activated" cores.
For that kind of money I'm sure you can get all kinds of overclocked and water cooled Intel/AMD kit running over 5Ghz.
-
Reply to Cazalan
Chad Boga
December 7, 2012 2:58:16 AM
earl45 said:
When is there going to be good news for AMD?R Read may be the solution
Obviously, the current business model wasnt working, so shifting priorities was essential.
Tighten your focus, head towards future markets, dump those that arent heading that way, at least within the parameters within the current assets within the company.
Its going to hurt alot, but hopefully will get them going again.
-
Reply to JAYDEEJOHN
Samsung’s share of world silicon foundry market remains low. According to market research organization Gartner, in 2011 TSMC had 48.8% of the market, far leading United Microelectronics Corp.’s 12.1%, GlobalFoundries Inc.’s 12% and Samsung’s 1.6%.
http://news.cens.com/cens/html/en/news/news_inner_42252...
Not sure if this reflects AMDs slowing of GF or not.
http://news.cens.com/cens/html/en/news/news_inner_42252...
Not sure if this reflects AMDs slowing of GF or not.
-
Reply to JAYDEEJOHN
Chad Boga said:
Well if your speculation about it being virtually impossible for Intel to produce 14nm desktop CPU's proves to be true, then all these businesses will be saved.
production is the easy part, keeping it from turning into a smoldering pile of ash when its turned on is something different.
making 14nm into a bga makes sense as you don't have the socket storing heat or insulating the bottom of the chip, instead it helps transfer heat to the motherboard itself.
-
Reply to noob2222
earl45 said:
When is there going to be good news for AMD?IMO that is good news for AMD - they are cutting their dependency on GF. By getting away from their special dependency on GF, the latter will now have to compete on a level playing field for AMD's business, against TSMC for now anyway. It's about time AMD recognized the problem and took steps to remedy the situation.
And, this is evidence that not all the problems lie with AMD - as almost everybody here recognizes, GF has a pretty bad track record at least with fabbing AMD's products. In this lousy economy, AMD needs to execute to perfection, from upper management providing direction and leadership, to the engineers designing the products, to the fab(s) making the product timely, competitively and low enough costs for AMD to get its margins back up, and to marketing.
-
Reply to fazers_on_stun
earl45
December 7, 2012 12:40:49 PM
fazers_on_stun said:
IMO that is good news for AMD - they are cutting their dependency on GF. By getting away from their special dependency on GF, the latter will now have to compete on a level playing field for AMD's business, against TSMC for now anyway. It's about time AMD recognized the problem and took steps to remedy the situation.And, this is evidence that not all the problems lie with AMD - as almost everybody here recognizes, GF has a pretty bad track record at least with fabbing AMD's products. In this lousy economy, AMD needs to execute to perfection, from upper management providing direction and leadership, to the engineers designing the products, to the fab(s) making the product timely, competitively and low enough costs for AMD to get its margins back up, and to marketing.
I can agree with you on AMD getting away from GF is pretty good news.
-
Reply to earl45
Cazalan said:
I recall that it needed about 1200W just to power the CPU. Not in the same ballpark. It also costs over $20,000 for the CPU the way IBM charges for "activated" cores.
For that kind of money I'm sure you can get all kinds of overclocked and water cooled Intel/AMD kit running over 5Ghz.
Lol Intel could probably make a 5Ghz CPU with a TDP of 125-145 watts.
Chad Boga said:
Well if your speculation about it being virtually impossible for Intel to produce 14nm desktop CPU's proves to be true, then all these businesses will be saved.
For just a second i though that was someone else i was going to say your nuts Intel will not only produce 14nm before Global disasters but they will also be able to produce high-end CPU's unlike GF.
earl45 said:
I can agree with you on AMD getting away from GF is pretty good news.I obviously agree as well, Amd just cant keep making these mistakes over and over again.
Llano- Low demand, Bulldozer delayed 1 year when they came out power hungry and hot(Amd's fault to) Now we have Steamroller delayed and are we even going to have a Steamroller APU in 2013 or just a slightly tweaked PD?
Solution Get away from GF but still use them start to use TSMC more. Be cool if they could use even more fabs samsung maybe? Start cutting management left and right start bringing in some new engineers, be better to get a few good engineers(ones back up with experience and a good rep) vs a lot of average ones.
^ Do that and in 3-5 years they will be back, times will be bad for them next couple of years but as long as they start now they might be able to stay in business without needing to team up with Arm or Samsung.
-
Reply to jdwii
AMD makes less sense by the day. They've paid a small fortune to sever their ties with GF, and it still hasn't ended.
The latest report that they cut back orders on $500mil to $115mil, gives a short term savings of 385mil. But it adds a whopping $320 mil severance fee. That means every chip they've paid for in Q4 actually costs them 3x as much.
To cover that loss they have to sell their Texas campus for 250mil, which doesn't even cover the FEE.
Why are there not more firings at AMD. The execs making these arrangements have sunk the company.
The latest report that they cut back orders on $500mil to $115mil, gives a short term savings of 385mil. But it adds a whopping $320 mil severance fee. That means every chip they've paid for in Q4 actually costs them 3x as much.
To cover that loss they have to sell their Texas campus for 250mil, which doesn't even cover the FEE.
Why are there not more firings at AMD. The execs making these arrangements have sunk the company.
-
Reply to Cazalan
-
Reply to JAYDEEJOHN
mayankleoboy1
December 7, 2012 11:54:08 PM
^^
seems like intel is ahead only in terms of fabrication. they haven't made power efficient chips like the ones based on arm instruction set. makes me think if intel's x86 itself is power inefficient which is why intel is competing with die shrinks instead of aggressively tweaking/stripping down their own instruction set and/or architecture.
seems like intel is ahead only in terms of fabrication. they haven't made power efficient chips like the ones based on arm instruction set. makes me think if intel's x86 itself is power inefficient which is why intel is competing with die shrinks instead of aggressively tweaking/stripping down their own instruction set and/or architecture.
-
Reply to de5_Roy
Cazalan said:
AMD makes less sense by the day. They've paid a small fortune to sever their ties with GF, and it still hasn't ended. The latest report that they cut back orders on $500mil to $115mil, gives a short term savings of 385mil. But it adds a whopping $320 mil severance fee. That means every chip they've paid for in Q4 actually costs them 3x as much.
To cover that loss they have to sell their Texas campus for 250mil, which doesn't even cover the FEE.
Why are there not more firings at AMD. The execs making these arrangements have sunk the company.
You can probably thank Hector Ruiz for that deal.. The good news is that he got fired from GF for his role in the insider-trading scandal. Well OK he resigned before getting his arse handed to him, but IMO the writing was on the wall already..
I know there are commonly penalties for severing a contract purchase, as fabs have to be able to plan ahead to keep their expensive equipment somewhat busy, but where are the penalties GF should have paid to AMD for all those delays and low yield ramps for Llano and BD? This penalty business seems pretty one-sided IMO. Plus all the R&D payments from AMD to GF for SOI fabrication, I suppose. How about GF cough up some $$ to AMD for GF being stupid enough to go with gate-first HKMG on 32nm?? Or IBM for that matter, since it was IBM urging all their fab consortium members including GF, that gate-first was the way to go?
But the top execs at AMD giving themselves IIRC a $68M pay & bennie increase this year is just sad. Whatever happened to reducing your pay to $1 a year during times of trouble, as the CEOs of other companies have done in the past when falling on hard times?? Just goes to show these guys are more interested in lining their pockets than much of anything else.
-
Reply to fazers_on_stun
mayankleoboy1 said:
^ What i understood from the previous links you posted was 1. Intel is far ahead of the competition
2. More and more fabs are looking to skip a node
3. move from 28nm planar to 20nm planar is not worth RoI.
4. FinFET is the future
5. Intel is far ahead of the competition
IIRC it was about 18 months ago that GF was proclaiming "we don't need no stinkin' FinFETs!"
.Personally I think all these fabs, TSMC included, are going to find lots of surprises when trying to skip nodes and at the same time going to 2.5D or 3D transistor structures. IOW, their roadmaps are way too optimistic IMO. And seeing as how I'm batting nearly 1000 in the prognostication dept. lately, my opinion is worth its weight in gold the photons emanating from your display!
But, as always we shall see in a year or two.
-
Reply to fazers_on_stun
mayankleoboy1
December 8, 2012 1:51:21 PM
fazers_on_stun said:
IIRC it was about 18 months ago that GF was proclaiming "we don't need no stinkin' FinFETs!"
.Personally I think all these fabs, TSMC included, are going to find lots of surprises when trying to skip nodes and at the same time going to 2.5D or 3D transistor structures. IOW, their roadmaps are way too optimistic IMO. And seeing as how I'm batting nearly 1000 in the prognostication dept. lately, my opinion is worth its weight in gold the photons emanating from your display!
But, as always we shall see in a year or two.
The level of BS execs are willing to spew to appease the investors keeps rising.
TSMC hasn't even fully ramped their 32nm production and that's close to 3 years after Intel started shipping 32nm. Looking at the roadmaps for Nvidia/AMD we can see nothing new will be viable through 2013 and likely through 2014.
The foundries are having trouble at the cutting edge because the vendor nor customer want to admit any fault. The engineers are always right, let the physics be damned. AMD blames GF, GF blames AMD. Neither wins.
Intel gets the job done because they're both in the same boat. The process engineers and the ASIC design engineers have to work closely together. They're less concerned with who's going to get the bill for each mistake.
-
Reply to Cazalan
AMD will continue to support socketed processors well into the future
Subject: Processors | December 5, 2012 - 05:33 PM | Ryan Shrout
Tagged: socket, Intel, BGA, amd
Over the past week or more we have been seeing a lot of news about Intel's rumored move to leave the world of socket-based processors behind after the pending Broadwell parts are released as BGA - ball grid array - and are soldered to motherboards directly. I would highly encourage everyone to read Josh's thoughts on the subject that are not nearly as damning as others might have you believe.
However, we got this official note from AMD earlier in the week that I thought I would share:
AMD has a long history of supporting the DIY and enthusiast desktop market with socketed CPUs & APUs that are compatible with a wide range of motherboard products from our partners. That will continue through 2013 and 2014 with the “Kaveri” APU and FX CPU lines. We have no plans at this time to move to BGA only packaging and look forward to continuing to support this critical segment of the market.
As the company that introduced new types of BGA packages in ultrathin platforms several years ago, and today offers BGA-packaged processors for everything from ultrathin notebooks to all-in-one desktops, to embedded applications and tablets, we certainly understand Intel’s enthusiasm for the approach. But for the desktop market, and the enthusiasts with whom AMD has built its brand, we understand what matters to them and how we can continue to bring better value and a better experience.
Obviously AMD is trying to persuade PC builders that not only is its path the safest in the future but maybe that supporting AMD today might help make sure it can arrive to the future well enough to continue the enthusiast path.
If Intel even starts to heavily side with BGA processors, is a move to AMD in your future again?
Subject: Processors | December 5, 2012 - 05:33 PM | Ryan Shrout
Tagged: socket, Intel, BGA, amd
Over the past week or more we have been seeing a lot of news about Intel's rumored move to leave the world of socket-based processors behind after the pending Broadwell parts are released as BGA - ball grid array - and are soldered to motherboards directly. I would highly encourage everyone to read Josh's thoughts on the subject that are not nearly as damning as others might have you believe.
However, we got this official note from AMD earlier in the week that I thought I would share:
AMD has a long history of supporting the DIY and enthusiast desktop market with socketed CPUs & APUs that are compatible with a wide range of motherboard products from our partners. That will continue through 2013 and 2014 with the “Kaveri” APU and FX CPU lines. We have no plans at this time to move to BGA only packaging and look forward to continuing to support this critical segment of the market.
As the company that introduced new types of BGA packages in ultrathin platforms several years ago, and today offers BGA-packaged processors for everything from ultrathin notebooks to all-in-one desktops, to embedded applications and tablets, we certainly understand Intel’s enthusiasm for the approach. But for the desktop market, and the enthusiasts with whom AMD has built its brand, we understand what matters to them and how we can continue to bring better value and a better experience.
Obviously AMD is trying to persuade PC builders that not only is its path the safest in the future but maybe that supporting AMD today might help make sure it can arrive to the future well enough to continue the enthusiast path.
If Intel even starts to heavily side with BGA processors, is a move to AMD in your future again?
-
Reply to skitz9417
mayankleoboy1 said:
^ Well since every fab (samsung, TSMC, GF) is skipping a node, i think they know what they are doing.Well 2014 is only 13 months away, so we shall see..
There's a reason why the industry leader (Intel) minimizes the risks inherent in new nodes/new designs with their tick-tock model. And as far as TSMC and GF go, their track record at 40nm and 32nm and 28nm is not exactly inspiring a lot of confidence
.-
Reply to fazers_on_stun
TSMCs 28 was fine.
You need to go every other node for problems.
40 wasnt good.
Earliest adopter AMD gpus were is short supply for early ramp up, as other customers came along, each was alloted so many wafers, keeping supplies lower, but as for a decent node is was very good, and is fully ramped now.
Understand, this isnt Intel where 2-3 months can be moved very easily, and stockpiling is OK.
These customers dont want stockpiling, they want new node ASAP
You need to go every other node for problems.
40 wasnt good.
Earliest adopter AMD gpus were is short supply for early ramp up, as other customers came along, each was alloted so many wafers, keeping supplies lower, but as for a decent node is was very good, and is fully ramped now.
Understand, this isnt Intel where 2-3 months can be moved very easily, and stockpiling is OK.
These customers dont want stockpiling, they want new node ASAP
-
Reply to JAYDEEJOHN
http://www.overclock.net/t/1299828/lr-amd-shows-steamro... their is a another fourm people can leran more on steamroller
-
Reply to skitz9417
mayankleoboy1
December 10, 2012 5:09:38 AM
when is the amd steamroller release 2013 or 2014
You've seen the slides everywhere and read through what Josh could observe and predict from those slides but at the end of Hot Chips will still know little more about the core everyone is waiting for. The slides show a core little changed from Bulldozer, which is exactly what we've been expecting as AMD has always described Steamroller as a refined Bulldozer design, improving the existing architecture as opposed to a complete redesign. SemiAccurate did pull out one little gem which might mean good news for both AMD and consumers which pertains to the high density libraries slide. The 30% decrease in size and power consumption seems to have been implemented by simply using the high density libraries that AMD uses for GPUs. As this library already exists, AMD didn't need to spend money to develop it, they essentially managed this 30% improvement with a button press, as SemiAccurate put it. This could well mean that Steamroller will either come out at a comparatively low price or will give AMD higher profit margins ... or a mix of both.
http://www.theinquirer.net/inquirer/news/2208525/amd-st...
The Bulldozer architecture within in AMD's current FX series CPUs, has generated no end of controversy (see pp. 1-123) since its release late 2011. For one sister site Atomic's own lab testing saw the top model FX-8150 come up short in a large number of areas. How then, will Bulldozer's successors fare?
In the first update for AM3+ systems, second generation processors (codenamed Piledriver) are very much inbound for desktops this year, and a few tweaks bring a welcome 10% performance bump and much lower power draw. In that vein we should see desktop A-Series branded (CPU + graphics parts) over the next few weeks and FX-Series (enthusiast CPU only processors) later in 2012.
Even more interesting is just how much next-year's further evolved 'Steamroller' processor (expected to also be compatible with AM3+) will shake things up. Bulldozer had the disctinction of being the first completely new x86 (PC) chip design since the VIA Cyrix III circa 2000 and chipmaking is a hugely complex task. Five year gestation or not, this first of a breed ended up rushed and has significant potential for optimisation - AMD have given us their story on 'how' and 'how much faster' Steamroller will be.
Where the second-gen Piledriver concentrates primarily on adding 'Resonant Clock Mesh' technology to its Bulldozer predecessor, the third gen Steamroller will be a major reworking of the underlying design that debuted in Bulldozer. Previous AMD roadmaps targeted a 10% per generation improvement in performance (which Piledriver met) however with the design being more fleshed out, AMD now imply a 30% performance increase. This is a pretty significant boost if true and more than a few sources suggest it is - then again 'more than a few' sources suggested Bulldozer would be good.
Chief among the design improvements is doubling the number of pre-processing 'decode' modules, bringing this part of the chip back to the traditional one-per-core level. Given the amount of time Bulldozer and Piledriver spend with inactive processing areas (indicating a design flaw in the pre-processing stage) this is a significant and straightforward win. The processing centres (Floating Point and to a lesser extent Integer units) also receive significant changes, as do the Prediction unit and Cache:
Steamroller was the subject of the recent keynote speech by AMD Chief Technology Officer Mark Papermaster, along with a fair bit of marketing speak on a 'Surround Computing Era'. Nonetheless along with a move to a slightly more advanced 28nm 'Bulk' (from 32nm 'SOI') manufacturing process, the significant redesigns outlined in Steamroller do hold the potential to yield equally significant performance gains. This in turn may close the gap with Intel, whose fourth-generation 'Haswell' architecture (due May-June 2013) is set to give roughly a 10-20% over its third generation 'Ivy Bridge' designs. We should see more concrete information from Intel on Haswell over the next week at its flagship IDF conference, which starts September 11th.
All in all it does seem certain AMD will deliver a significant performance jump with Steamroller, exactly how significant this will turn out to be and when the chips will reach shelves is not as clear yet. We'll keep you up to date as the details come into focus
"With that in mind, the HDL slide was rather interesting. AMD is claiming that if you rebuild Bulldozer with an HDL library, the resulting chip has a 30% decrease in size and power use. To AMD at least, this is worth a full shrink, but we only buy that claim if it is 30% smaller and 30% less power hungry, not 30% in aggregate. That said, it is a massive gain with just a button press.
AMD should be applauded, or it would have been, but during the keynote, the one thing that kept going through my mind was, “Why didn’t they do this 5 years ago?”. If you can get 30% from changing out a library to the ones you build your GPUs with, didn’t someone test this out before you decided on layout tools?"
http://www.tomshardware.com/news/AMD-Steamroller-Piledr...
http://www.techradar.com/au/news/computing-components/p...
http://www.extremetech.com/computing/135105-amd-details...
There was nothing catastrophically wrong with the design itself, but not all the bugs were ironed out and the design was not completely optimized.
At the time, Barcelona seemed rushed out, as the company was in a hurry to offer something different than the popular Athlon 64 X2 available in those days.
All that was needed was some grooming in the bug forest and a considerably larger level 3 cache. Then we had Phenom II that was and still is successfully competing against its respective Core 2 Duo counterparts.
The Bulldozer story was similar.
While the design differences between Phenom II and Bulldozer are quite significant and the number of floating point units (FPU) has been halved, many expected the design to literally bulldoze the competition out of the way.
We weren’t so sure about the FlexFP. Our opinion was that it would lead to some serious efficiency and throughput increases in the server sector, but on the desktop side we would have considered AMD lucky if it managed to get 4 FPUs to perform like the previous 6 FPUs.
To its credit, AMD has successfully managed exactly that, but that’s certainly not enough to fight Ivy Bridge.
The incoming Vishera silicon with Piledriver enhancements is what Bulldozer should have been in the first place.
We’re sure AMD’s Dirk Meyer was envisioning something like Steamroller. He should have made the initial Bulldozer more like today’s Piledriver, but at least we know that Bulldozer had some issues that have now been ironed out.
Many say that Steamroller will not be able to face Haswell, but the reality is that Steamroller is apparently a proper implementation of the FlexFP concept. The doubling of the cache, dispatch and fetching units will greatly increase the performance.
Many are estimating conservative values ranging between a 20% and a 30% performance improvement over the current Bulldozer processors, but sources inside the engineering department at AMD are reportedly expecting 45% performance improvements.
This puts AMD’s Steamroller beyond Ivy Bridge’s performance and right against Intel’s Haswell.
Some are wondering about the company’s statements that they won’t be fighting for the desktop market anymore.
We believe that’s just PR talk to divert attention from the fact than AMD’s current top desktop CPUs barely make it against Intel’s quad-cores with no fighting chance against any hexa-core from the CPU giant.
The most pressing question right now is whether AMD’s next desktop platform will keep the current socket or whether they will go for something with DDR4.
It would really be early for the company to go for DDR4, as even Intel is not planning DDR4 platforms until 2014, but it will also seem strange to see AMD change two sockets in two years.
A new Steamroller socket with support for DDR3 will allow AMD to get away from any drawbacks that the current AM3+ platform might have, but then, in 2014 they will be forced to put another socket out with DDR4 support.
One strategy would be to keep the current AM3+ design and use some level 4 cache to improve on the lack of a triple or quad channel memory controller, but that remains to be seen
Yesterday we showed some leaked information on AMDs upcoming Piledriver CPUs, today AMD officially announced information (check this presentation) on a new architecture; Steamroller. The new 28nm chips using this architecture should have lower power usage while performing better. The details were unveiled by AMDs CTO Mark Papermaster during the Hot Chips conference. In 2011 AMD introduced a new modular micro-architecture. The two cores of each module share components which should have resulted in better performance than Intel’s Hyperthreading but lower performance than two seperate cores. The architecture debuted in AMDs FX Bulldozer CPUs but was not able to deliver the performance as expected.
The Piledriver architecture should already be an improvement but in 2013 AMD hopes to introduce Steamroller which should be an even bigger improvement. These chips would build on Bulldozer and Piledriver architecture but should again be faster and consume less watts. The bottlenecks in Bulldozer and Piledriver are e.g. the shared fetch and decode hardware in the front-end. In Steamroller the decode hardware is doubled, which means the rest of the core doesn’t have to wait for instructions. The fetch hardware remains shared between cores. With all the changes, AMD expects a 30% increase in performance.
AMD also improved the shared floating point unit which each Steamroller module has. The FPU still has the same execution capabilities, but it has become smaller. Also the MMX unit and 128-bit FMAC pipes share some hardware. With the reduction of the pipeline resources AMD hopes to deliver the same throughput while consuming less power and smaller hardware, something that already should have been done in Bulldozer and Piledriver CPUs.
Also cache of the CPUs has been changed, L1 cache sizes have gone up and also the interface between L1 and L2 cache has been improved. Even better, L2 cache memory that isn’t used can be turned off which should result in less power usage.
The first products using the Steamroller architecture should appear in 2013, and should bring AMD less power consuming but more powerfull processorcores which can be used in AMDs upcoming APUs. In 2014 AMD will introduce another new architecture: Excavator with, you guessed it, even better performance
http://www.bit-tech.net/news/hardware/2012/08/29/amd-st...
http://wccftech.com/amds-kaveri-based-28nm-richland-apu...
http://www.pcgamer.com/2012/08/29/amd-stokes-up-its-eng...
http://www.xbitlabs.com/news/cpu/display/20120829180358...
http://hothardware.com/News/AMDs-Next-Gen-Steamroller-C...\
http://forums.overclockers.co.uk/showthread.php?t=18465...
https://www.youtube.com/watch?v=iKW0CkTTBU0
https://www.youtube.com/watch?v=PMSWhPrT99g
and sorry for the long post guys
You've seen the slides everywhere and read through what Josh could observe and predict from those slides but at the end of Hot Chips will still know little more about the core everyone is waiting for. The slides show a core little changed from Bulldozer, which is exactly what we've been expecting as AMD has always described Steamroller as a refined Bulldozer design, improving the existing architecture as opposed to a complete redesign. SemiAccurate did pull out one little gem which might mean good news for both AMD and consumers which pertains to the high density libraries slide. The 30% decrease in size and power consumption seems to have been implemented by simply using the high density libraries that AMD uses for GPUs. As this library already exists, AMD didn't need to spend money to develop it, they essentially managed this 30% improvement with a button press, as SemiAccurate put it. This could well mean that Steamroller will either come out at a comparatively low price or will give AMD higher profit margins ... or a mix of both.
http://www.theinquirer.net/inquirer/news/2208525/amd-st...
The Bulldozer architecture within in AMD's current FX series CPUs, has generated no end of controversy (see pp. 1-123) since its release late 2011. For one sister site Atomic's own lab testing saw the top model FX-8150 come up short in a large number of areas. How then, will Bulldozer's successors fare?
In the first update for AM3+ systems, second generation processors (codenamed Piledriver) are very much inbound for desktops this year, and a few tweaks bring a welcome 10% performance bump and much lower power draw. In that vein we should see desktop A-Series branded (CPU + graphics parts) over the next few weeks and FX-Series (enthusiast CPU only processors) later in 2012.
Even more interesting is just how much next-year's further evolved 'Steamroller' processor (expected to also be compatible with AM3+) will shake things up. Bulldozer had the disctinction of being the first completely new x86 (PC) chip design since the VIA Cyrix III circa 2000 and chipmaking is a hugely complex task. Five year gestation or not, this first of a breed ended up rushed and has significant potential for optimisation - AMD have given us their story on 'how' and 'how much faster' Steamroller will be.
Where the second-gen Piledriver concentrates primarily on adding 'Resonant Clock Mesh' technology to its Bulldozer predecessor, the third gen Steamroller will be a major reworking of the underlying design that debuted in Bulldozer. Previous AMD roadmaps targeted a 10% per generation improvement in performance (which Piledriver met) however with the design being more fleshed out, AMD now imply a 30% performance increase. This is a pretty significant boost if true and more than a few sources suggest it is - then again 'more than a few' sources suggested Bulldozer would be good.
Chief among the design improvements is doubling the number of pre-processing 'decode' modules, bringing this part of the chip back to the traditional one-per-core level. Given the amount of time Bulldozer and Piledriver spend with inactive processing areas (indicating a design flaw in the pre-processing stage) this is a significant and straightforward win. The processing centres (Floating Point and to a lesser extent Integer units) also receive significant changes, as do the Prediction unit and Cache:
Steamroller was the subject of the recent keynote speech by AMD Chief Technology Officer Mark Papermaster, along with a fair bit of marketing speak on a 'Surround Computing Era'. Nonetheless along with a move to a slightly more advanced 28nm 'Bulk' (from 32nm 'SOI') manufacturing process, the significant redesigns outlined in Steamroller do hold the potential to yield equally significant performance gains. This in turn may close the gap with Intel, whose fourth-generation 'Haswell' architecture (due May-June 2013) is set to give roughly a 10-20% over its third generation 'Ivy Bridge' designs. We should see more concrete information from Intel on Haswell over the next week at its flagship IDF conference, which starts September 11th.
All in all it does seem certain AMD will deliver a significant performance jump with Steamroller, exactly how significant this will turn out to be and when the chips will reach shelves is not as clear yet. We'll keep you up to date as the details come into focus
"With that in mind, the HDL slide was rather interesting. AMD is claiming that if you rebuild Bulldozer with an HDL library, the resulting chip has a 30% decrease in size and power use. To AMD at least, this is worth a full shrink, but we only buy that claim if it is 30% smaller and 30% less power hungry, not 30% in aggregate. That said, it is a massive gain with just a button press.
AMD should be applauded, or it would have been, but during the keynote, the one thing that kept going through my mind was, “Why didn’t they do this 5 years ago?”. If you can get 30% from changing out a library to the ones you build your GPUs with, didn’t someone test this out before you decided on layout tools?"
http://www.tomshardware.com/news/AMD-Steamroller-Piledr...
http://www.techradar.com/au/news/computing-components/p...
http://www.extremetech.com/computing/135105-amd-details...
There was nothing catastrophically wrong with the design itself, but not all the bugs were ironed out and the design was not completely optimized.
At the time, Barcelona seemed rushed out, as the company was in a hurry to offer something different than the popular Athlon 64 X2 available in those days.
All that was needed was some grooming in the bug forest and a considerably larger level 3 cache. Then we had Phenom II that was and still is successfully competing against its respective Core 2 Duo counterparts.
The Bulldozer story was similar.
While the design differences between Phenom II and Bulldozer are quite significant and the number of floating point units (FPU) has been halved, many expected the design to literally bulldoze the competition out of the way.
We weren’t so sure about the FlexFP. Our opinion was that it would lead to some serious efficiency and throughput increases in the server sector, but on the desktop side we would have considered AMD lucky if it managed to get 4 FPUs to perform like the previous 6 FPUs.
To its credit, AMD has successfully managed exactly that, but that’s certainly not enough to fight Ivy Bridge.
The incoming Vishera silicon with Piledriver enhancements is what Bulldozer should have been in the first place.
We’re sure AMD’s Dirk Meyer was envisioning something like Steamroller. He should have made the initial Bulldozer more like today’s Piledriver, but at least we know that Bulldozer had some issues that have now been ironed out.
Many say that Steamroller will not be able to face Haswell, but the reality is that Steamroller is apparently a proper implementation of the FlexFP concept. The doubling of the cache, dispatch and fetching units will greatly increase the performance.
Many are estimating conservative values ranging between a 20% and a 30% performance improvement over the current Bulldozer processors, but sources inside the engineering department at AMD are reportedly expecting 45% performance improvements.
This puts AMD’s Steamroller beyond Ivy Bridge’s performance and right against Intel’s Haswell.
Some are wondering about the company’s statements that they won’t be fighting for the desktop market anymore.
We believe that’s just PR talk to divert attention from the fact than AMD’s current top desktop CPUs barely make it against Intel’s quad-cores with no fighting chance against any hexa-core from the CPU giant.
The most pressing question right now is whether AMD’s next desktop platform will keep the current socket or whether they will go for something with DDR4.
It would really be early for the company to go for DDR4, as even Intel is not planning DDR4 platforms until 2014, but it will also seem strange to see AMD change two sockets in two years.
A new Steamroller socket with support for DDR3 will allow AMD to get away from any drawbacks that the current AM3+ platform might have, but then, in 2014 they will be forced to put another socket out with DDR4 support.
One strategy would be to keep the current AM3+ design and use some level 4 cache to improve on the lack of a triple or quad channel memory controller, but that remains to be seen
Yesterday we showed some leaked information on AMDs upcoming Piledriver CPUs, today AMD officially announced information (check this presentation) on a new architecture; Steamroller. The new 28nm chips using this architecture should have lower power usage while performing better. The details were unveiled by AMDs CTO Mark Papermaster during the Hot Chips conference. In 2011 AMD introduced a new modular micro-architecture. The two cores of each module share components which should have resulted in better performance than Intel’s Hyperthreading but lower performance than two seperate cores. The architecture debuted in AMDs FX Bulldozer CPUs but was not able to deliver the performance as expected.
The Piledriver architecture should already be an improvement but in 2013 AMD hopes to introduce Steamroller which should be an even bigger improvement. These chips would build on Bulldozer and Piledriver architecture but should again be faster and consume less watts. The bottlenecks in Bulldozer and Piledriver are e.g. the shared fetch and decode hardware in the front-end. In Steamroller the decode hardware is doubled, which means the rest of the core doesn’t have to wait for instructions. The fetch hardware remains shared between cores. With all the changes, AMD expects a 30% increase in performance.
AMD also improved the shared floating point unit which each Steamroller module has. The FPU still has the same execution capabilities, but it has become smaller. Also the MMX unit and 128-bit FMAC pipes share some hardware. With the reduction of the pipeline resources AMD hopes to deliver the same throughput while consuming less power and smaller hardware, something that already should have been done in Bulldozer and Piledriver CPUs.
Also cache of the CPUs has been changed, L1 cache sizes have gone up and also the interface between L1 and L2 cache has been improved. Even better, L2 cache memory that isn’t used can be turned off which should result in less power usage.
The first products using the Steamroller architecture should appear in 2013, and should bring AMD less power consuming but more powerfull processorcores which can be used in AMDs upcoming APUs. In 2014 AMD will introduce another new architecture: Excavator with, you guessed it, even better performance
http://www.bit-tech.net/news/hardware/2012/08/29/amd-st...
http://wccftech.com/amds-kaveri-based-28nm-richland-apu...
http://www.pcgamer.com/2012/08/29/amd-stokes-up-its-eng...
http://www.xbitlabs.com/news/cpu/display/20120829180358...
http://hothardware.com/News/AMDs-Next-Gen-Steamroller-C...\
http://forums.overclockers.co.uk/showthread.php?t=18465...
https://www.youtube.com/watch?v=iKW0CkTTBU0
https://www.youtube.com/watch?v=PMSWhPrT99g
and sorry for the long post guys
-
Reply to skitz9417
Uh oh:
http://techreport.com/news/24026/amd-lowers-wafer-order...
http://www.fudzilla.com/home/item/29743-intel%E2%80%99s...
So now AMD is cutting wafer orders, and taking a $320 Million hit as a result.
http://techreport.com/news/24026/amd-lowers-wafer-order...
http://www.fudzilla.com/home/item/29743-intel%E2%80%99s...
So now AMD is cutting wafer orders, and taking a $320 Million hit as a result.
-
Reply to gamerk316
The funny thing is that AMD said they were going to cancel vishera pd for 2012 and release in 2013 instead, but they released in 2012, on schedule. So, It'll be interesting to see if they drop sr in 2013. I think they really need to. I really want one of these Steamy chips.
Another thing that comes to mind is, "Will AMD keep the AM3+ chipset and add DDR4 and PCIe 3.0 to the mix?"
I currently see no need for PCIe 3.0. But, DDR4 is coming out with intel's new chipset next year... I'm curious to see the performance gains vs DDR3.
Another thing that comes to mind is, "Will AMD keep the AM3+ chipset and add DDR4 and PCIe 3.0 to the mix?"
I currently see no need for PCIe 3.0. But, DDR4 is coming out with intel's new chipset next year... I'm curious to see the performance gains vs DDR3.
-
Reply to griptwister
Chief Executive of AMD: We Are Not Interested in Low-Volume Customers.
AMD Interested Only in High-Volume Design Wins
http://www.xbitlabs.com/news/cpu/display/20121210000002...
i chuckled at this bit:
so much for moar cores (in fx cpus).....
i guess he's emphasising low power apus like kabini.
AMD Interested Only in High-Volume Design Wins
http://www.xbitlabs.com/news/cpu/display/20121210000002...
i chuckled at this bit:
Quote:
“Not all design wins are created equal. Actually, more is not better. More equals more complexity, it means testing platforms, back-end costs that drive [company’s] costs. […] What I am going to do in our space is to focus on design wins that are going to drive volume. […] Every design has associated cost with testing, putting on the platform, etc.,” said Rory Read, chief executive of Advanced Mice Devices, at Credit Suisse technology conference in late November.so much for moar cores (in fx cpus).....
i guess he's emphasising low power apus like kabini.
-
Reply to de5_Roy
de5_Roy said:
Chief Executive of AMD: We Are Not Interested in Low-Volume Customers.AMD Interested Only in High-Volume Design Wins
http://www.xbitlabs.com/news/cpu/display/20121210000002...
i chuckled at this bit:
Quote:
“Not all design wins are created equal. Actually, more is not better. More equals more complexity, it means testing platforms, back-end costs that drive [company’s] costs. […] What I am going to do in our space is to focus on design wins that are going to drive volume. […] Every design has associated cost with testing, putting on the platform, etc.,” said Rory Read, chief executive of Advanced Mice Devices, at Credit Suisse technology conference in late November.so much for moar cores (in fx cpus).....
i guess he's emphasising low power apus like kabini.
The original approach was a good one, too bad it was a messy screw up of execution failure. You know BD on 32nm when Llano was first introduced would have been a way better product.
Anyway, that's the correct choice given the landscape. You can't win a drag race with an under-powered engine and lower grip tires, but you can go drifting and have a chance... I guess, hahaha.
I still want to know about DDR4 plans on AMD's side.
Cheers!
-
Reply to Yuka
-
Reply to skitz9417
JAYDEEJOHN said:
TSMCs 28 was fine.You need to go every other node for problems.
40 wasnt good.
Earliest adopter AMD gpus were is short supply for early ramp up, as other customers came along, each was alloted so many wafers, keeping supplies lower, but as for a decent node is was very good, and is fully ramped now.
Understand, this isnt Intel where 2-3 months can be moved very easily, and stockpiling is OK.
These customers dont want stockpiling, they want new node ASAP
Somehow I doubt nVidia would agree that TSMC's 28nm was just fine
.. -
Reply to fazers_on_stun
fazers_on_stun said:
Somehow I doubt nVidia would agree that TSMC's 28nm was just fine
..Higher demand from earlier adopters left nVidia somewhat behind the 8 ball early on.
As for costs, everyone had to pay those rates, as newer nodes are simply more expensive to make.
Perf wise, its been a very good node
-
Reply to JAYDEEJOHN
- First
- Previous
- 9 / 329
- 10
- 11
- 12
- 13
- … More pages
- Next
- Newest
- 1
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9 / 329
- 10
- 20
- 30
- 40
- 50
- 60
- 70
- 80
- 90
- 100
- 200
- 300
Related resources
- SolvedAMD CPU for gaming (fanboi question) solution
- Solvedwill a amd 750k cpu hold back my 280x solution
- SolvedAMD GPU and CPU speeds? solution
- SolvedAMD Athlon ii x4 640 CPU/GTX 480 temps good? solution
- SolvedUpgrading CPU from AMD solution
- SolvedIntel Core i3-4130 or AMD FX-6300 CPU? solution
- SolvedAmd CPU and 12 Gb of Ram solution
- SolvedCPU-Cooler (AMD FX-8350) solution
- SolvedAMD FX6350 CPU Gaming boost? Any suggestions? solution
- SolvedWhich CPU-Cooler? (AMD FX-8350) solution
- SolvedHelp identifying an AMD Athlon II CPU solution
- Solvedasus sabertooth 990fx rev 2 amd 8350 sudden crash and cpu led solution
- Solvedasus sabertooth 990fx rev 2 amd 8350 sudden crash and cpu led solution
- Solvedis the amd a10-7850k a good cpu solution
- SolvedAMD Athlon CPU heatsink fan silent solution
- More resources
!