Sorry to burst your thought bubble, there's no way high performance GPUs that spit out 250W of heat is going to be viable on a LPP node. I honestly don't even think its viable on the TSMC 20nm SOC node either. That's why NV isn't jumping on board. Why would AMD do it?
Exactly. Even NV owners such as tviceman and toyota have already calculated that GM204 is about 70% more efficient per watt in avg. gaming performance than GK204, but when this is pointed out, people just ignore facts and mathematics.
NV's measurement of efficiency was tied to GFLOPs/Watt, not in terms of gaming performance/watt, but as usual marketing won over logic and facts! NV also used 195W as indicative of GK104's power usage and 165W for the 980. This is crazy misleading since they used the lowest possible power usage of the reference 980 despite 99% of after-market 980's using much more than 165W of power at max load and they used the higher measurement for 680 when reference 680 uses less than 195W.
---
Anyways, back to Fiji. The interesting part is that Fiji XT is supposed to be R9 380/380X, while Bermuda XT is R9 390/390X.
R9 390 series = Bermuda XT
http://www.pc-specs.com/gpu/AMD/R-300_Series/Radeon_R9_390/2264
"Finally the performance of the card in OpenGL is approximately 50% higher than the R9 290x (63.6 vs 42.4 GB / s) so it seems that AMD will offer a good performance increase with Pirate Islands."
http://www.profesionalreview.com/2014/11/12/primeros-detalles-de-la-radeon-r9-390x/
"The bleeding edge R9 390X is next and will be based on Bermuda. You can expect the GPU to be targeting both the 980 Ti and the Titan X."
http://www.redgamingtech.com/amd-r9-380x-february-r9-390x-370x-announced/
"Previously there was a news that AMD Radeon 300 series would feature the Bermuda core and on R9 390X and R9 390 Bermuda XT would fuse with them. But last week the news from Zauba send everyone in the new void of thinking, as according to their latest entry AMD Next Generation of Radeon R9 300 Series will feature the Fiji XT GPU. Previous rumors claimed that, Bermuda GPU would be the part of AMD Next Generation of Radeon R9 300 Series."
http://tech4gamers.com/amd-r9-radeo...ns-4gb-3d-stacked-hbm-4096-stream-processors/
Right now, it's difficult to say if Fiji XT is R9 380/380X and Bermuda XT is R9 390/390X, if the codenames are reversed now, or if AMD will hold back the real Bermuda XT for the R9 490/490X, or someone incorrectly assumed that Fiji XT that leaked is R9 390 series, when in fact it's only R9 380 series.
If you guys look at AMD's strategy, their next gen mid-range cards roughly equal their last gen flagship cards in performance (similar to NV):
1. HD6950 (2nd fastest) < HD7850 (2nd fastest mid-range); HD6970 (fastest last gen) < HD7870 (next gen top mid-range)
2. HD7950 (2nd fastest high-end) < HD7950 V2 ~ R9 280 (2nd fastest mid-range); HD7970/7970Ghz (fastest last gen) ~ R9 280X (next gen top mid-range)
3. HD7970 was about 25-30% faster than HD7870 (2nd tier level card of that gen)
4. R9 290X was about 24-30% faster than HD7970Ghz/R9 280X (2nd tier card of that gen)
Based on the above, we are starting to see some trends:
Trend #1 - AMD's next gen mid-range is about as fast as last gen flagship. Therefore, we can estimate that R9 380/380X will be about as fast as R9 290/290X.
Trend #2 - AMD's flagship card tends to be about 25-30% faster than its comparable mid-range card of the same generation. Therefore, we can estimate that if R9 380/380X ~ R9 290/290X, then R9 390/390X will be at least 25-30% faster.
However, the rumours are suggesting that R9 380X is aimed at GTX970/980 cards, which means R9 390/390X should be much faster than 25-30% than R9 290/290X for this to be true because 980 is 20% faster than R9 290X. For R9 380X to compete with a 980, it has to be faster than R9 290X, which means if we apply 25-30% faster over R9 380X, we'll be way faster than 30% over the 290X.
"The R9 380X (based on the Pirate Islands architecture) is said to be aimed squarely at Nvidias recently released Nvidia Maxwell GTX 970 and 980."
http://www.redgamingtech.com/amd-r9-380x-february-r9-390x-370x-announced/
It's possible based on connecting these dots that R9 390X will actually be faster by more than 30% than R9 290X.
Note that there are sources for RS's post...
Got any sources, please?
Once the next drivers come out I can show everyone.
14.4xx(RC2) or 15.xxx(Beta/Alpha).
Last edited:
most reviews I have seen show the GTX 980 is around 15% faster than R9 290X at 1440p/1600p. The gap is closer to 10% at 4k.
http://www.pcper.com/reviews/Graphi...70-GM204-Review-Power-and-Efficiency/Conclusi
" Benchmarks and real-world game play proved that to be true: the GTX 980 was as much as 15% faster than the R9 290X and was only beaten by the AMD flagship card in one of our six games"
http://www.techpowerup.com/reviews/NVIDIA/GeForce_GTX_980/26.html
http://www.hardware.fr/articles/928-20/recapitulatif-performances.html
http://www.computerbase.de/2014-09/geforce-gtx-980-970-test-sli-nvidia/6/
http://www.sweclockers.com/recension/19332-nvidia-geforce-gtx-980-och-gtx-970/19#pagehead
Last edited:
RussianSensation
Elite Member
- Sep 5, 2003
- 19,458
- 765
- 126
^ It'll be interesting to see the response of NV users if 390X beats 989 by 20% at $549 in Spring 2015. Since right now a $550 980 is 20% faster and that shrinks at 1600p and above against a 290X, then a $550 390X would mean 980 shouldn't cost more than $349. After all, since people think a 20%, $250 more expensive 980 is worth buying, then surely a $550 390X should sell by truck loads at $550 with 20% more performance. Something tells me though the same story will repeat itself with people saying "too late" for 390X and others will either wait for GM200. I think 390 non-X might shake things up, if it beats 980 and AMD prices it at say $399-429, then it could be the sweet-spot card for CF.
Last edited:
380X preview
FPS
power consumption
Edited_1
Performance:
390X = 295X2
Power consumption:
390X = 290
FPS
power consumption
Edited_1
Performance:
390X = 295X2
Power consumption:
390X = 290
Last edited:
I'll be very interested to see how the next release goes with Su in charge. Reed ruined every release while he was there.
I am pretty sure AMD has significant changes to its architecture for GCN 2.0 based R9 390X. I am betting that AMD will really shake up things with the world's first HBM based GPU just as it did with the world's first GDDR5 based GPU - Radeon HD 4870. AMD already has huge fill rate improvements and tesselation improvements with GCN 1.2 aka Tonga. With 512 GB/s HBM and Tonga's color compression memory bandwidth per shader is going to be up by more than 50%. AMD needs to work out a few things - improved perf/shader, better perf/transistor and better perf/watt. Rearchitecting a few core elements like the compute Unit (maybe 128 stream processors) and the related architectural plumbing for higher efficiency and performance will help AMD get there. HBM also brings significant power reduction compared to GDDR5 for a given bandwidth. Add to it techniques like voltage adaptive operation, more fine grained power gating and a significantly better GF 28SHP (vs TSMC 28HPM) and they should all make it possible for R9 390X to be much more power efficient than R9 290X. I am looking forward to R9 390X being around 50% faster than R9 290X ( and even higher in perf/watt) and 30% faster than GTX 980 (atleast at 1440p and 4k). :thumbsup:
http://www.anandtech.com/show/8742/amd-announces-carrizo-and-carrizol-next-gen-apus-for-h1-2015
But I don't see AMD going very cheap. I think AMD will bring out a true GM200 competitor and the prices will be competitive. I am guessing R9 390 at USD 450 - 500 and R9 390X at USD 600 - 650.
Last edited:
flash-gordon
Member
- May 3, 2014
- 123
- 34
- 101
Great analysis.
There are two rumours really new to us this generation:
- A hi-end GPU being manufactured by GloFo, so we don't really know where TDP, clocks and leakage may go.
- HBM
If the process and HBM works as advised I think we should see the whole board inside 200W.
The same way, I don't see the HBM out of the 1Ghz specification as that SiSoft leak suggests, what's the need for that? In gaming it won't be the bottleneck some time...
SlickR12345
Senior member
They will come out very strong and destroy Nvidia is order to take market share. Like even if they only have giant lead in performance of 1 month before Nvidia responds, just the press and word of mouth info about how better AMD cards are will be enough for people to go and buy AMD cards for the next 6-9 months.
So AMD will come out strong with the latest and fastest technology and crush Nvidia significantly.
I mean Nvidia is working on a 980 upgrade and its going to be about 20% faster according to all the reliable sources and AMD's flagship the 390x will be about 40% faster than Nvidia's 980TI.
So AMD will come out strong with the latest and fastest technology and crush Nvidia significantly.
I mean Nvidia is working on a 980 upgrade and its going to be about 20% faster according to all the reliable sources and AMD's flagship the 390x will be about 40% faster than Nvidia's 980TI.
RussianSensation
Elite Member
- Sep 5, 2003
- 19,458
- 765
- 126
As in 40% faster than GM200 powered 980Ti, or did you mean 980Ti as the card 20% faster than a 980?
Based on your estimates then:
980 = 100%
980Ti = 120%
390X = another 40% faster than 980Ti or 1.2x1.4 = 168%?
No way.Like I said if 390X beats 980 by 20-25%, it will be a MAJOR accomplishment given the time span and how overpriced 980 is. Looking at the benchmarks above, it's not quiet there yet at 15-16% but I see the power usage is only 200W. I really want AMD to design a card 50-75% faster than 290X at 250-300W TDP but it looks like the performance/watt marketing is going to win over this one given how much AMD was criticized for it with 7970/290 series.
The leaked data above was very accurate before 970/980 launch so it seems believable that some 300 series card can achieve 16% over the 980.
I don't think AMD will "crush NV" because you are forgetting 3 things:
1) GM204 overclocks really well, which means NV can respin GM204 and add another 10-15% more performance similar to GTX770 vs. 680. On top of that they can market GM204B with 8GB which will be a nice bonus for 4K gaming.
2) NV could easily introduce a less cut down GTX970Ti and/or drop the price of the 980 to $349-399 to combat AMD's card in that range. Look at 980's price of $550-600 vs. 970 and it's clear that NV has a lot of headroom to drop prices and still be very profitable. NV basically priced 980 close to a 780 replacement with most after-market cards going for $580-600 but it's just a mid-range 256-bit card! Don't underestimate their ability to lower prices on the 980.
3) You are forgetting that NV will have GM200/210 with probably 50% more performance over 780Ti! If 390X beats 980 by 15-20% but GM200 beats 390X by another 15-20%, A LOT of gamers will think: "Hmmm, if I am already spending $550, why not just spend $700 and get the fastest possible GPU", especially with 6GB+ of VRAM that GM200 will have vs. likely HBM limited 4GB of the 390X.
While 390X should beat a 980 or it's a total disappoitnment given that it's going to be late by 6 months or so (unless AMD prices it at $399 or something), NV hasn't showed off the Big Maxwell yet. Therefore, it's too premature to write off NV. Given that NV has had the fastest single GPU crown since 8800GTX (!), and high-end gamers gravitate towards NV 2:3 in the $300+ segment, I wouldn't bet against them. NV has also locked in some major games for GW including, MGS, Batman Arkham Knight, Project CARS and the Witcher 3. What games does AMD have in 2015 that are part of GE? So far I haven't seen much.
What's good is that we should finally see competition since NV definitely won't be able to get with with pricing 980 at $550-600, which is good for gamers looking to finally upgrade from 670/680/7950/7970 style cards. :thumbsup:
*Note* if the ~ 200W power usage and 16% performance increase over 980 is true, I don't see how AMD was able to pull that off on 28nm.
Last edited:
Good post, RS
I think ungrounded expectations stand to devastate and can easily turn a respectable accomplishment into something much less. I think most everyone hopes AMD can come out with a gm200 crusher for $499, but already preaching crazy figures steals away the chances of the chip really impressing. See, if everyone expects the 390x to be 40% faster than the gm200, if it actually does achieve that, its merely meeting its expectations. Anything less will be deemed a disappointment by the ones expecteing much much more.
If people set their expectations so extremely high, it would have to be even that much better to really really impress.
There is really no good in that, as far as i see it.
There is one gripe i have with your last few post. Its your time frames. You post as if AMD waited till the gm204 launched before they ever started working on their next flagship. That is incredibly misleading. You are doing the same when you refer to the gm204 launching 1 year after the 290x and saying, "only beating it by 20%"
It is completely wrong and a backwards way of looking at things.
It is simply.... not how it works
Yes, the gm204 launched after the 290x but not on AMDs time scale. That is where you are seriously mixing things up. These chips are developed in-house and on their own time frame. I know, from a consumer stand point things may appear in a certain order but bunching them up together is usually not correct.
The GM204 came down a linage internal at Nvidia. From the Gf1XX to GK1XX to GMXXX. Just as the 390x will come out of AMDs own progression. From concept to launch, 90% of that time both companies are totally in the dark and dont know exactly what it will be launching against. Your observations are completely over simplified and are based on imagination more than reality. That is not meant as stab towards you and i dont mean for it to be taken harshly. It is just really really wrong to look at the progression of GPUs in such an oversimplified way. But I guess.. you have every right, if you choose to.
I just wanted to let others know that, for the record
That is just not how it works
I think ungrounded expectations stand to devastate and can easily turn a respectable accomplishment into something much less. I think most everyone hopes AMD can come out with a gm200 crusher for $499, but already preaching crazy figures steals away the chances of the chip really impressing. See, if everyone expects the 390x to be 40% faster than the gm200, if it actually does achieve that, its merely meeting its expectations. Anything less will be deemed a disappointment by the ones expecteing much much more.
If people set their expectations so extremely high, it would have to be even that much better to really really impress.
There is really no good in that, as far as i see it.
There is one gripe i have with your last few post. Its your time frames. You post as if AMD waited till the gm204 launched before they ever started working on their next flagship. That is incredibly misleading. You are doing the same when you refer to the gm204 launching 1 year after the 290x and saying, "only beating it by 20%"
It is completely wrong and a backwards way of looking at things.
It is simply.... not how it works
Yes, the gm204 launched after the 290x but not on AMDs time scale. That is where you are seriously mixing things up. These chips are developed in-house and on their own time frame. I know, from a consumer stand point things may appear in a certain order but bunching them up together is usually not correct.
The GM204 came down a linage internal at Nvidia. From the Gf1XX to GK1XX to GMXXX. Just as the 390x will come out of AMDs own progression. From concept to launch, 90% of that time both companies are totally in the dark and dont know exactly what it will be launching against. Your observations are completely over simplified and are based on imagination more than reality. That is not meant as stab towards you and i dont mean for it to be taken harshly. It is just really really wrong to look at the progression of GPUs in such an oversimplified way. But I guess.. you have every right, if you choose to.
I just wanted to let others know that, for the record
That is just not how it works
- Feb 2, 2009
- 14,003
- 3,362
- 136
If 390x is 20% faster than 980 then it should be priced higher than 980. I dont see why AMD should price a faster/better product lower than the competition. NVIDIA does exactly that every single time, AMD can do the same. And it is the right think to do, after all they deserve it if they can pull out a product better than 980.
2015 Mantle titles
Star Citizen
Battlefield : Hardline
Star Wars : Battlefront (Dec 2015 ??)
I believe there are more GE games coming 2015 with or without Mantle.
I believe it is 28nm
RussianSensation
Elite Member
- Sep 5, 2003
- 19,458
- 765
- 126
Ya, I know that. These cards are 2-4 years in development. My point is from a gamer's point of view, the timeline/time frame matters. For example, I got my 7970s in early 2012 which is nearly 3 years ago. If I waited 3 years without upgrades, for me another 3-6 months for 20-30% more performance or much lower prices with 980's performance is not much wait time left. There are plenty of gamers that don't upgrade very often which means for them waiting 2-3 years to upgrade is not uncommon. From that perspective, 970/980 way under delivered from a performance point of view if looking as 1 year next gen cards vs. 290X/780Ti.
Think about it another way, on average GPU speed increased about 35% per annum from Sept 2009 (HD5870) to Nov 2013 (780Ti).
http://www.computerbase.de/2013-12/grafikkarten-2013-vergleich/10/
Therefore, by roughly December 2014 we should have a card from NV/AMD that's 35% faster than 780TI but it's not here yet. I always talk about the price/technology curve because historically all GPUs have reverted to it in the last 20 years. Right now the 980 misses that curve completely and if 390X is 35% faster than 780TI by March 2015, it's only late by about 3-4 months on that curve using historical performance increase data of about 35% per year (averaged). If 390X fails, then GM200 should meet the curve.
Note: Another way to look at it is opportunity cost of waiting. If you wait 3 years to buy a new card, you should expect about 2.5X the performance increase, roughly (1.35x1.35x1.35 = 2.46X). Surely enough 290X delivers 2.3X over HD6970 and the time difference between their launches is slightly less than 3 years. Therefore, given how long I've waited to upgrade from HD7970, in 3-3.5 years I should be able to buy a single GPU card with 2.3-2.5X the performance but 980 again misses that mark by a bit. From a technology curve point of view, 980's 20% over 290X is not enough, it should be 35%. Alternatively, if the performance doesn't meet the technology GPU curve, then price/performance has to compensate and here again the 980 falls short. That's why I am looking forward to 390/390X/GM200.
For one, the 390X is late so to compensate for the 6 months that technology should have moved, it should offer either superior features by miles than 980 OR superior performance per $ (like 290 vs. 780) OR be way faster to command a much higher price. AMD doesn't have the same brand loyal users as NV. 480 launched 6 months late with only 15-18% more performance over 5870 and NV users waited 6 months for it. People aren't going to be waiting for the next AMD card that's only 15-18% faster than a 980 for $600-650. If AMD priced it at $550 and delivered 15-20% more performance, they are moving the performance/$ metric and bringing a lot more competition. Also, if you look from 970/980's point of view bringing lower price and faster performance is good for your brand value.
Remember every day any new card is out, it drops in value since tech "ages" even if the price on the market stays the same. Sooner or later the real price is reflected in the market, such as 295X2 now going for just a hair over $650 vs. $1499 debut price! Therefore, by March 2015, 980 isn't worth $550. Its real value at that point is probably $475 but the market just hasn't priced it in yet because the VGA market isn't an actively traded commodity/stock. By end of 2015, the real value of 980 will probably be around $399 or so, and usually the GPU markers refresh it with a major price drop (aka 680--> 770). Therefore, in 6 months from 980's launch, you probably should be able to get 15-20% more performance for $550 anyway. Of course this may have little to do with reality as you have seen AMD/NV raise prices significantly and bifurcate generations in the last 3 years. This makes it tough to say what 390X or GM200 would be priced at. For example, if 390X is $599 but GM200 is 20% faster at $699, then 390X is a horrible deal. If NV prices GM200 at $899 or $999, well that's totally different.
Last edited:
Wild Thing
Member
- Apr 9, 2014
- 155
- 0
- 0
I believe you'll be pleasantly surprised to find it is a 20nm production process....hence the stellar power use/perf numbers.:thumbsup:
I am not sure if there will be a GM200/GM210 anytime soon - people forget Nvidia just released a power consumption optimised GK110,the GK210. Since the Nvidia large die GPUs are subsidised by higher margin professional sales,I doubt the GM200/GM210 will around for a while,unless the GK210 is one of the shortest life GPU products Nvidia ever made. Hardly any customers would want to buy any GK210 based products if a GM200/GM210 based one was around the corner and that's the problem - in constant load compute operations Maxwell is not as big an improvement over Kepler when compared to more peaky gaming workloads. Hence,this is why Nvidia just refined the GK110 - I expect large Maxwell will be waiting for a new process node IMHO,unless it is mahoosive(over 600MM2) 28MM chip.
I have a feeling Nvidia will just relaunch the GTX980 with higher boost clocks,and they can produce a dual GPU card much easier this time too.
I have a feeling Nvidia will just relaunch the GTX980 with higher boost clocks,and they can produce a dual GPU card much easier this time too.
Last edited:
- Feb 2, 2009
- 14,003
- 3,362
- 136
My bad, i wanted to say higher priced than 970 not 980. If 390X is 20% faster than 980 i wouldnt be surprised if it would released at $499-550 MSRP.
It will not make sense to price it lower if you are the performance champion at the time. When GM200 will be released 4-6 months later and if it is faster and priced at the same price, then you can lower your price but not when you have the fastest GPU on the market.
- Feb 2, 2009
- 14,003
- 3,362
- 136
I think the gk210 is just a special design that was born out of a specific case or need. There is possibly a contract that could use the extra cache and buffers. This is also why its 2gpus on one pcb. And since it was already designed, they released it as a new tesla card.
Most of these big contracts have a specific task and special code designed around the gpu it will be using. A gm200 would need tweaked code to be taken full advantage of. I am just saying, the gk210 doesnt have to mean there is no gm200. I think there will be one anyway. Just like we saw shipping info on the gm204 and the gk210, we have also seen a gm200. I fully expect all 3 to come to market in due time
Most of these big contracts have a specific task and special code designed around the gpu it will be using. A gm200 would need tweaked code to be taken full advantage of. I am just saying, the gk210 doesnt have to mean there is no gm200. I think there will be one anyway. Just like we saw shipping info on the gm204 and the gk210, we have also seen a gm200. I fully expect all 3 to come to market in due time
RussianSensation
Elite Member
- Sep 5, 2003
- 19,458
- 765
- 126
But NV's research paper shows that GM200 will have more than double the GFLOPs/Watt of GK110.
http://videocardz.com/51195/nvidia-maxwell-gm200-pascal-gp100-confirmed-research-paper
Nothing precludes NV from releasing an even faster professsional card and raise prices even more. After all, we didn't exactly have $1000 gaming Titans and $3000 dual-GPU Titan Z cards 10 years ago. We have seen NV and AMD both raise prices on tech, with Titan Z and 690 priced well above cards like GTX590/5970/6990 and so on. NV can easily create another class of professional cards priced even higher than GK210.
Based on the research paper, the timelines he provides are about 6-months off from wide launch availability. It says Summer 2014 for GK210 but it just launched recently. Since the paper specifies Dec 2014 for GM200, it should launch within 6 months of that.
ShintaiDK
Lifer
- Apr 22, 2012
- 20,378
- 146
- 106
Still a lot cheaper than the days of 40000$ professional graphcis cards. And thats even without adding inflation on top since.
Also even looking 10 years back. Professional FireGL And Quadro GPUs from ATI and nVidia was easily 1000-3000$ back then. (That would be over 4000$ in todays money.).
But selective memory is selective memory.
Please reread the toms hardware test everyone seems to be taking this from.
This simply isn't true.
SirPauly
Diamond Member
- Apr 28, 2009
- 5,187
- 1
- 0
I would like to see this come to fruition and good to see rumors on architectural efficiency than strong vocal advocacy for water cooling for reference.
We are yet to see which card these perf nos are for. Without that detail its impossible to tell. I do not think that this is R9 390X. There is no way AMD is releasing a 200W GPU as its flagship when it clearly knows a GM200 is due for launch in Q1 2015.
dude this overclocking business applies to every card. So don't go in that direction. the biggest factor which limited Hawaii overclocking was heat and power. Given the 200w power draw of this card is closer to power draw of HD 7970(925 mhz) there is going to be no shortage of overclocking headroom.
RS it all depends on AMD's R9 390X and R9 380X. If they are competitive with GM200 and GM204 then NV cannot do anything. AMD will hit back hard for the loss of marketshare and take back what they lost and maybe even more.
GTX 980 is 5% faster in this chart wrt GTX 780 Ti. According to you GM 200 with 50% more resources than GM204 will be 45% faster. Sorry but you just have to look at GTX 770 to GTX 780 Ti in the same chart to see that perf never scales linearly on the same architecture. For 87.5% more shaders the perf lead is < 50% (around 45%). Most reviews I have seen put the 780 Ti at 50% faster than GTX 770. So what you are saying is not going to happen on the same architecture and same node. You will be lucky if Nvidia gets 30% from GM200 wrt GM204. Do not make the mistake of comparing R9 390X and R9 290X similarly because there are too many variables at play - Improved Architecture, new high bandwidth memory system, improved process (GF 28SHP).
I agree its too premature to tell anything. But the odds are looking good for AMD if this chart is true. Obviously its still a rumour. btw just because you don't know how it does not mean its not possible. You were one of the most vocal supporters saying that R9 290X would definitely be slower than Titan and even GTX 780. We all know how that ended up. So don't repeat the mistake. I think people are underestimating AMD and thats good as that leaves room for surprises. :thumbsup:
Given that AMD has now confirmed that Carrizo-L is 28nm, I am now pretty much sure that AMD has skipped 20nm. There is no way AMD is making 300 - 350 sq mm high performance GPUs when they chose to skip doing a <= 100 sq mm low power mobile SOC. I am now sure that other than Apple and Qualcomm the rest of the industry will go straight to FINFET.
Samsung, GF and TSMC will provide enough 16/14 nm FINFET capacity starting from 2015 and you will see a rapid adoption from low power to high performance chips. I can now see the first 16nm GPUs in early 2016. Samsung expects to have 30% of their 12 inch wafer capacity at 14nm by end 2015. GF will be ready too to help Apple, Qualcomm and AMD ramp their FINFET products.
http://aod.teletogether.com/sec/20141030/2014Q3_script.pdf
page 12
"In terms of the ramp-up, already the wafers for mass production are fed in for the target date of going mass production in -- or at the end of this year. And we believe that the ramp-up will be going on as scheduled in 2015. Our plan is to ramp up very quickly during 2015 so that by the end of 2015 this would account for 30% of our 12-inch capacity"
TSMC also will be ramping in Q3 2015 (10% of Q4 2015 revenue at 16 nm FINFET) with a steep Q1 2016 ramp (20% of wafer revenue at 16nm FINFET).
http://www.eetimes.com/document.asp?doc_id=1324617
"We believe 20nm yields are approaching 80%, and 16 nanometer FF+ SRAM yields exceed 90%, improving confidence that FinFET will reach 1% of TSMC's third-quarter 2015 sales, 10% of fourth-quarter 2015 sales, and approaching 20% by the first quarter of 2016," Credit Suisse analyst Randy Abrams said in a Nov. 12 report following the TSMC announcement. TSMC "also may secure Apple for the next iPad in the fourth quarter of 2015 and iPhone 7 orders in 2016."
By early 2016 there will be no shortage of FINFET capacity with TSMC, Samsung and GF providing enough wafers for the entire fabless industry. In fact few analysts say there could be oversupply and that could lead to foundry price wars.
Last edited:
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 24K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter