Both AMD and Nvidia have tended to launch higher end flagships before moving to fill the rest of the lineup. The same is expected for Nvidia’s Pascal launch, with the “flagship” GP104 based cards set to arrive before the midrange GP106 does. As we know GP104 will arrive with the GTX 1080 and 1070 at Computex and reports are surfacing that we won’t see GP106 till the autumn of this year.
By holding back the release, Nvidia balances out their sales and gives a chance for last gen cards to get be cleared out of stock first. This is especially true of the low-end as GM206, the current Maxwell midrange chip, is still in production. Many of Nvidia’s AiB partners also recently launched new 75W GTX 950 cards, raising suggestions that Nvidia is preparing cards to counter the more power efficiency Polaris 11 which should arrive before GP106.
With GP104 reaching down to the 1060Ti, the GP106 will likely power the GTX 1060, 1050Ti and maybe even the GTX 1050 as well. According to the source, the chip will also not require a PCIe power connector, with a TDP of at least under 75W. This would put is TDP near that of Polaris 11 which is also under 75W. With GP106 already spotted on the Drive PX2, it’s only a matter of time though till Nvidia launches the cards,
After the recent Taiwanese earthquake, many Nvidia and AMD watchers may have worried about their upcoming Pascal and Polaris GPUs. While TSMC did eventually reveal that there would be a hit to their chip production, especially 16nm, it seems like things should be fine. According to the latest reports, TSMC is planning to double their 16nm wafer production from 40,000 per month up to 80,000 per month.
While this number may still be slightly depressed due to the earthquake, it does mean TSMC is taking in more 16nm orders and is able to supply them. Nvidia is relying on TSMC to supply them with 16nmFF+ GPUs for use with Pascal which is set to launch later this year. A ramp up now would mean the a mid-2016 launch for the earliest Pascal chips, right in line with rumours. For AMD, TSMC will play a lesser role as Polaris may be using GlobalFoundries 14nmLPP exclusively.
One snag in the above analysis is that these maybe Apple A10 SoCs. Apple has been moving away from Samsung as their main chip supplier and Apple may be starting to ramp up iPhone SoC orders. Either way, the fact that 16nmFF+ is doing well means the earthquake likely won’t affect chip supply and prices.
After introducing their latest SSD two months ago in China and Japan, Samsung is bringing their latest drive to the rest of the world. Based off of the long-standing TLC tradition at Samsung, the 750 EVO uses the latest planar 16nm TLC NAND to bring a more budget solution to the masses. Coming in the more budget oriented 120 and 250GB capacities, the latest drive will hit lower price points than it’s faster 850 EVO cousin. While the industry is starting to transition to 3D or V-NAND.
While the industry is starting to transition to 3D or V-NAND, planar NAND is still very much relevant and Samsung’s 16nm process is their latest. As the 850 and 850 EVO demonstrated, V-NAND carries a strong speed benefit over planar NAND, leading to better performance. Plain old TLC drives like the 840 tended to be slower than their MLC counterparts. Despite this, the use of the familiar SLC cache allows the 750 EVO manages to mask most of the performance differences compared to V-NAND, allowing it to match the 850 EVO in short, bursty workloads.
Still, a heavier workload would likely overwhelm the SLC cache and cause performance to drop to TLC levels. Other noteworthy specifications include the dual core MGX controller, 256MB of DDR3 cache and improved LDPC ECC and 256Gbit dies. Overall, by releasing a 16nm planar NAND based SSD will help Samsung develop and prepare for the day when 3D-NAND scales down to 16nm or even lower. Hopefully, the 750 EVO won’t inherit the read speed issue faced by the 840 EVO and the still unfixed 840.
In the hours after the Taiwanese earthquake, TSMC was one of the major tech firms affected by the magnitude 6.4 quake. While the semiconductor manufacturer has initially stated that only minor damaged has occurred, it looks like effects will be longer lasting than expected. In fact, the damage will cause a greater than 1% disruption in Q1 2016 wafer shipments.
Taiwan Semiconductor Manufacturing Company or TSMC has 9 fabs in operation in Taiwan, with Fabs 2, 3, 5, 6, 8, 12A, 12B, 14 and 15 located in the island country. While it’s unknown which fabs were impacted, Fab 6 and 14B is already back up and running while. The main damage is due to a reassessment of the damage done to Fab 14 which is expected to take longer to restore. The original estimate was that 95% of foundry machines would be back up and running within 3 days.
Despite the damage and disruption, TSMC is still expected to hit its revenue targets for the quarter. The firm does have a $100 million range for its target so I would presume that the disruption will likely remain under 2%. Hopefully, this means that Nvidia and AMD won’t face any delays and it’s still much better than what happened to HDDs after the Thai floods in 2011 which saw a massive price spike.
Last month we received word that the first Pascal chips would be launching a bit sooner than expected. At that time, GP100 was expected to drop in April and GP104 in June. According to the latest rumors, it looks like that timetable was accurate, with the GP100 based Tesla chip coming in April around GTC. What’s more, we’re getting more details about when the rest of Nvidia’s Pascal lineup will launch.
As with the previous report, GP104 will arrive in June and it looks like the GTX lineup will be based off that, with both the GTX 1080 and 1070 being GP104 chips. Near the end of the year in Q4, we can expect GP106 and GP107. These will be longer end chips and likely power the GTX 1060 and 1050. Finally, we have the Titan which will use GP100 and a GP108 in early 2017. This follow’s Nvidia’s new trend of releasing a GTX x80 first, followed by the Titan, then finally a GTX x80Ti. While it’s good for Nvidia’s sales, it moves high-end users into a quicker upgrade schedule than if all the cards launched at the same time.
The Tesla launching first makes a whole lot of sense as enterprise users can pay the high premiums for early HBM2 and 16nm. A June launch for GP104 may point to them using HBM2 as the timeline is a bit tight for GDDR5X though it is doable. The biggest question is how well Pascal will perform as it is a stopgap architecture between Maxwell and Volta, like a Maxwellv2 though the die shrink to 16nm should make some great gains in and of itself.
Remember the $600 and $700 Polaris GPUs we spotted last month in a shipping manifest? It turns out that those chips may not be the top end Polaris chip launching this year. According to the shipping tracker, a ‘Video Graphic Card’ labelled C993 shipped from Canada on February 1st and was valued at 110,603 INR. That’s $1600 or nearly 3x the value of the C981 and C980 that we spotted earlier.
According to AMD, they plan to launch both a Polaris 10 and Polaris 11, with one being a high-end card while the other would service the budget/mainstream market. We had expected the $600-700 C981/980 to be the top end flagship while the $200-$300 C924/913 would be the budget model. With the C993 at $1600, this card completely blows the other cards out of the water and makes it seem like that might be the flagship Polaris card.
At $1600 though, this card would be one of the most expensive cards AMD has released in a long time. At $1600, it would be beyond what Nvidia has charged for the Titan (X, Black), though it would make sense as a dual-GPU utilizing 2 Polaris chips like the C981 and C980. Maybe this is why there are only 2 Polaris chips technically launching but they may have multiple variants and different configurations.
For those waiting on Nvidia’s next-generation GPUs, the wait may not be as long as expected. Last week, we found out that Pascal would be arriving a bit sooner than expected, in the earlier parts of 2H 2016 rather than late in 2016. Now, it looks like Nvidia may be moving even faster than those rumours, with GP100 to arrive in April with GP104 to follow 2 months later in June. What’s more, the GTX 1080 will also debut in June and reportedly be based off GP104; perhaps we’ll see it at Computex 2016?
While GP100 or Big Pascal will launch first in April, that is only for the Titan and various enterprise models. This is in line with what Nvidia has done in the past by launching models with higher margins in order to reduce their risk and grab as much of the early adopter crowd cash. Later on, the more mainstream GP104 will follow up with gaming oriented GeForce models with the compute units cut out. The biggest change is that the Titan will be launching before the GeForce this time.
If the GTX 1080 is based off GP104 as rumoured, this would suggest a GTX 1080Ti based off of GP100 would arrive later on, just like what happened with the 9xx and 6xx series. For those looking to get the very best gaming card for the next generation, waiting may be a smart move. AMD is also set to launch their own Polaris GPUs around the same time though it looks like Nvidia may beat them to the punch with GP100.
More and more information is pointing to the state of readiness for AMD’s upcoming Polaris GPUs. According to information spotted in AIDA64 and HWiNFO Changelog, support for 3 new GPUs, Ellesmere, Baffin and Greenland has appeared. As these nomenclatures predate AMD’s Polaris announcement, we can assume that Greenland is Vega 10 while Baffin and Ellesmere are one of Polaris 10 and 11. With AMD going around and handing out the PCI-e ID for Polaris, this means engineering samples are just around the corner.
Switching away from the Islands based noncom lecture, the new architecture maintains it’s GCN roots but is otherwise heavily improved. Graphics guru Raja Koduri noted that the new GCN 4.0 is built purposefully for use with the new 14/16nm process and FinFETs. Combined other redesigned blocks and units on the GPU, the new cards will offer a revolutionary improvement over the past. The use of HBM2 and GDDR5X also means these cards will be both more power efficient and able to push higher resolutions easier.
With Polaris set to arrive in mid-2016, it’s only a matter of time before we get more information and leaks out of AMD. Once launched, the new architecture will usher in a new era for GPUs and hopefully for AMD as well.
Even as this generation’s GPUs are continuing to fly off the shelves, Nvidia is already gearing up for their Pascal launch. Despite being quieter than AMD, it looks like Nvidia will launch their Pascal cards around the same time, in 2H 2016 as AMD’s Polaris will. What’s more, 2H 2016 will see Nvidia’s flagship Pascal GPU based on TSMC’s 16nmFF+ process and utilizing HBM2. This is still a rumour right now but it does fit the time frame since 1H 2016 would be too soon and 2017 too late.
The biggest question is what does”flagship” mean exactly. Ever since GTX 680 was launched, Nvidia has been playing around with the word flagship. Traditionally, the big dies like GF110 would launch first with the smaller mainstream GF104 launching after. Kepler and Maxwell saw that switch with GK104 and GM104 launching ahead of GK110 and GM200 respectively. This suggests that the so-called “flagship” may only be GP104 and not GP100. Even if it is GP100, it may well be a cut-down version, similar to how the GTX 780 was the cut-down variant of the later GTX 780Ti. This strategy does maximize sales for Nvidia but isn’t that great for consumers.
Whatever the card is, be it GP104 or GP100, it is going to use HBM2, giving it at least 512GB/s with 8GB of VRAM but potentially much higher at 16-32GB with 1TB/s+ of bandwidth. With AMD set to launch Polaris around the same time, Q3 2016 should make for exciting times as a slew of new GPUs arrive.
Ever since HBM1 was revealed and launched with Fury X, many have been looking forwards to what HBM2 would bring along in 2016. While HBM1 brought large power savings and a major boost in memory bandwidth, it was largely limited to a relatively low 4GB capacity. HBM2, however, is set to provide a boost in capacity and bandwidth by increasing the number of stackable dies. We’re now getting reports that AMD’s upcoming Polaris chips will utilize HBM2.
As a major revamp of the GCN architecture, a Polaris flagship GPU would be the natural product to debut HBM2. A flagship GPU much more powerful than current generation chips due to the new architecture and process node would likely require more memory bandwidth to feed it and a high memory capacity as it would be meant for VR and 4K gaming. Being the largest chip in the lineup, the flagship would also benefit from the major power savings, helping offset its core power consumption. The confirmation of HBM2 also suggests that we will be getting high-end Polaris chips this year.
At the same time, AMD is also confirming that they will continue to use GDDR5 and likely GDDR5X as well. At CES, AMD showed off a low powered Polaris chip using GDDR5 that was able to provide the same performance as Nvidia’s GTX 950 but with a significantly lower power consumption. With such a leap in efficiency, the HBM2 chips will likely be light years ahead of current cards in terms of efficiency if GDDR5 already shows such massive gains.
Nvidia has been mum about their upcoming Pascal architecture, even as rival AMD has been showing off their new Polaris chips at CES. Set to launch in 2016 as well, we’re now getting a hint that Pascal may soon be ready. According to shipment tracking site Zabua, several chips likely to be GP104 were sent out on December 29th last year. With engineering samples out, Nvidia may have GP104 ready in about 6 months.
Being a 37.5×37.5mm BGA package, the chip is most likely the GP104, replacing the current GM204 which comes in at about 40x40mm. Since GP104 is “small Pascal”, it will likely come as a GTX 970/980 replacement for the GM204. As with the Kepler and Maxwell releases, Nvidia probably plans to lead with their smaller dies first and later launch their top end GP110 chips in order to maximise sales.
With Pascal engineering samples just shipped, it looks like AMD will have several months lead on Nvidia for the upcoming process node. Given the gains Polaris has shown off, Pascal may do the same. With Maxwell as efficient as it is though, Pascal will probably find most of its efficiency gains from the new 16nm node. It will be interesting to see which direction Pascal will take Nvidia.
When AMD unveiled their new Polaris architecture last week, details about the timeline were a bit scarce. Other than previous confirms about a mid -2016 launch for one of the Polaris cards, AMD had been mum about when Polaris would actually arrive. Now information has surfaced about Polaris shipments has been spotted. According to a shipment tracking site, 4 Polaris cards have already been shipped around, with the earliest from Jan 2015.
The more recent models were the C981 and C980 which shipped in late December and early January. Valued at 47,578 and 40,790 INR respectively, these should fall into the product lineup around the R9 390/390X. Surprisingly, the earliest Polaris shipment, the 913, was in January 2015, nearly a year ago. this means AMD has had working Polaris silicon for nearly a year. A 924 was also shipped in August. Given their reported value, these 2 chips will likely be mid-tier about where the R7 370 and R9 380 fall.
If AMD had Polaris working for them for nearly a year, it suggests that a launch may come sooner than expected. This bolstered by the fact that AMD already has development advanced enough to show off a working demo at CES. Whatever the case, Polaris is fast approaching and will likely launch within 6 months.
First leaked yesterday, we’re now able to bring you the full AMD presentation on their upcoming Polaris GPU architecture. Set to ship mid-2016, Polaris will be using a 14/16nm FinFET process and bring massive power consumption and efficiency improvements. According to RTG SVP Raja Koduri, AMD says the massive potential that would come out of moving to FinFETs and a lower process node at the same time and decided to design a new architecture just for that.
Not surprisingly, the biggest focus is on the efficiency side of things. AMD is claiming a historic leap in performance per watt that any Radeon GPU has ever seen. The key to this is the new compute units for Polaris, or GCN 4.0. While we can expect much remains similar to past GCN revisions, the new release will be adding more parts to make the chip more efficient. A more efficient hardware scheduler, primitive discard accelerator and improved memory compression are expected to help as well. While HBM(2) will help reduce power consumption, GDDR5X models will be launched first.
To show off their improved efficiency, AMD showed off working Polaris silicon with presumably GDDR5X. Facing off against a GTX 950 in an identical system, the Polaris part managed to pull 1.65x less power from the wall as a whole. Using only 86W total compared to the 140W on the Nvidia card. This was while both cards were running Star Wars Battlefront at 1080p 60fps. Even if some of the gains are coming from the FinFETs and die shrink, that is still pretty impressive as AMD has been lagging behind on this front. The demo GPU was made using Globalfoundries 14nm FinFET process fo what it’s worth but expect some 16nm parts from TSMC as well.
Just as AMD has done with previous GCN releases, Polaris will also see other parts of the modular system upgraded. These include the Command Processor, Geometry Processor, Multimedia Cores, Display Engine, L2 Cache and Memory Controller. For the Multimedia Cores, the biggest additions are support for 4K HEVC (h.265) encoding and decoding at 60 FPS which should be welcome as well as AMD continues to push HDR. On the connectivity side, DisplayPort 1.3, and, at long last, HDMI 2.0a are supported.
For AMD, 2016 will be a critical year as both their GPU and CPU get a major architectural overhaul at the same time they get a major process upgrade. If executed well, this may finally pull AMD out of it’s slow decline and bring the firmly back into the black. Whether that happens remains to be seen.
AMD has largely been building off of GCN since its initial debut in 2011, with incremental improvements made via GCn 1.1 and 1.2. Combined with the limitations of having to use the 28nm process, the GPU space hasn’t seen any great gains of late. However, that is set to change soon with the arrival of AMD’s new Greenland and Nvidia’s Pascal architectures. We’re now getting some information that Greenland will launch in Summer 2016 and what’s more, jump straight to 14nm.
Greenland was rumoured to use the 14nm process from Samsung/Globalfoundries, eschewing the 16nm from TSMC, their usual supplier. If this is the case, AMD and Nvidia’s cards will not only have different architectures but finally different process nodes. Greenland manufacturing is to begin in June with the launch to be held sometime in late summer, in time for the back to school and holiday sales.
AMD had previously confirmed that the 14nm process would be used for GPUs but this is a major change in strategy. Globalfoundries and Samsung are also expected to use the 14nm process to make AMD’s next-gen Zen CPU as well. From an integration standpoint, it does make it easier for AMD to produce APUs since both the CPU and GPU will target the same node. Combined with HBM2, next years GPU and CPU launches should be pretty eventful.
With both AMD and Nvidia set to debut new GPU architectures next year, many are expecting both greater performance and improved performance per watt. While Nvidia has recently taken the lead in this area, it looks like AMD is making big gains themselves. According to CEO Lisa Su, Artic Islands will double performance per watt over the current offerings. This will place Artic Islands well past Maxwell though Pascal is still an unknown at this point.
In order to get such a large improvement, AMD is counting on their architectural improvements, continued use of HBM, as well as the process advantage moving from 28nm to TSMC’s 16nm FF+. At this point, it’s hard to determine what current offerings means, with calculations based off Tonga coming in ahead of Maxwell by a good margin but using Fiji as a baseline will give even better gains.
With improved performance per watt, we can also expect raw performance to improve as well. With lower TDP, we can expect larger dies which obviously give more performance if done right. A new architecture also opens up room for good gains as does the increased die sizes available from the new process. 2016 will be pretty interesting as AMD and Nvidia battle it out, with AMD given a chance to take back some much need marketshare.
Right after announcing their 16nm TLC plans, Micron has acquired an SSD controller firm. Tidal Systems, made up of engineers from Link-A-Media Devices (LAMD) and SandForce was founded last year and has managed to create 2 of their own SDD controllers already. With in-house controllers, Micron will be able to iterate more quickly and provide a more unified SSD platform.
For the longest time, Micron has had to rely on third-party controllers to power their SSDs. While Silicon Motion does provide a somewhat complete controller and firmware package, Micron’s main supplier Marvell does not. This has meant that Micron has long had to write their own firmware for most of their SSDs. With the controller now in-house as well, Micron has room to reduce costs and also build on their long expertise in writing firmware.
Tidal Systems also brings their expertise around Low-Density Parity Check (LDPC) to the table. As NAND transitions to use TLC (Triple-Level Cell) and QLC (Quad-Level Cell) on ever smaller lithographies, the number of errors is sure to go up. With LDPC, SSD makers will be able to offset some of the error and endurance issues.
Even with Tidal Systems in-house now though, don’t expect any Micron drives to ship with the new controllers anytime soon. Toshiba took quite a while to integrate their OCZ controllers with their drives. With this purchase, it only leaves SanDisk as a major NAND and SSD producer without their own controller. It is critical though that SanDisk has been able to do the most with Marvell controllers, leveraging their in-house firmware to eek out performance that rival’s Samsung’s.
Micron Technologies is finally moving into the TLC (Triple-Level Cell) NAND market, with shipments of consumer SDDs starting in Q4 this year. TLC is generally cheaper than MLC, holding up to 3 bits compared to the 2 with the older technology. This allows for cheaper SSDs as it requires fewer NAND dies to reach a certain capacity compared to MLC. By Q3 2016, Micron is expecting about 50% of their SSDs will be using TLC.
Along with Intel, Micron’s joint venture IMFT has largely focused on shrinking the process with MLC in order to gain die savings. IMFT was one of the earliest NAND producers to reach 20 and 16nm. On the other hand, SanDisk/Toshiba and Samsung have long been using TLC in their SSDs, preferring to move to new processes slower. Samsung for instance, released their first TLC drive back in 2013 and SanDisk their’s in 2014.
Combining their leading 16nm process with TLC should offer great savings for SSD buyers. TLC does come with a number of drawbacks though, most notably lower endurance and performance. TLC generally only can last 1,000 P/E (program/erase) cycles while MLC dos much better at 3,000. Even with only 1,000 cycles though, TLC should be more than enough for most consumers, especially if a bit of extra NAND is set aside. On the performance front, a caching system like those used in the Samsung 840/850 EVO or SanDisk Ultra II can mitigate most of the issues. Micron first trialed their caching system with the MX200 which should be a great starting point to work from for the TLC drives.
Micron has not yet revealed any details for their TLC drives. Intel too, given their joint IMFT venture may be trying out TLC drives in the near future. Combined with 3D XPoint and 3D NAND from SanDisk/Toshiba, 2016 should be a pretty good year for SSDs.
TSMC has just scored a major CPU customer as AMD is allegedly moving their Zen CPUs over to the fab. Originally meant for GlobalFoundries 14nm process, delays at the once AMD owned fab have led to a change to use the 16nm process at TSMC, the same one used for AMD and Nvidia’s next-gen GPUs. Zen is AMD’s next CPU architecture, aimed at improving IPC by 40% over current Excavator products.
According to the source, GF has been facing issues with getting their 14nm production ramped up. The fab’s main owner, the government of Abu Dhabi, has been cutting expenses due to low oil prices. Due to that and difficulties in retooling the 28nm equipment to 14nm, volumes and yields on the new process are below expectations. It was also the delays for the 32nm process at GF that caused Bulldozer to launch later than expected back in 2011.
With both TSMC and GF offering FinFET processes, AMD should see good efficiency gains on top of moving to a new process. While AMD had previously been mum about which FinFET process it would use, most had expected GF to win out due to the long relationship between the two firms. With TSMC now confirmed, the biggest question is whether or not the fab can handle all the CPUs, GPUs and SoCs planned for next year. Hopefully, TSMC 16nmFF+ process will be able to hit the clocks speeds required of desktop CPUs.
Thank you WCCFTech for providing us with this information
It’s no secret that Nvidia are working their next-generation graphics hardware, even with their flagship 980Ti and Titan series of cards doing well, there’s always going to be something better on the horizon. The next Nvidia hardware will be powered by the new GP100 silicon, which has reportedly entered its testing phase, putting it through its paces to judge its viability for future consumer hardware.
This means that Nvidia now has a few completed chips which may one day make it into their next-gen graphics cards. GP100 is based on the Pascal architecture, which is set to feature more than 17 billion transistors, made possible by TSMC’s 16nm FinFET+ node construction. Of course, the excitement doesn’t end there, as the new hardware an HBM2 memory interface, which leaves room for up to 32GB of extremely fast memory, although it’s more likely this next-generation will not go above 16GB.
Early predictions, which are usually quite accurate, expect that Pascal GP100 hardware, with HBM2 memory could outperform the current Titan-X hardware but 60-90%, and I honestly don’t doubt that fact.
With AMD and Nvidia cooking up HBM2 cards, the next generation of the GPU wars is going to be a lot of fun!
Thank you TechPowerUp for providing us with this information.
While there had been some rumours that Nvidia would turn to Samsung’s 14nm process for GPUs, it appears those were wrong. For the longest time, Nvidia has relied on TSMC to manufacture their chips and it appears this relationship is continuing. Set to launch next year, Nvidia’s Pascal architecture will reportedly use TSMC’s latest 16nm process. This will be the same process used for AMD’s upcoming Greenland GPUs.
As with AMD’s Greenland, Pascal will be a new architecture with new features and other improvements. Most notably, Pascal will be paired with HBM2, allowing for up to 16GB of VRAM and 1TB/s of memory bandwidth. Other additions include support for NVLINK, Nvidia’s GPU interconnect and mixed precision support. With Kepler and later Maxwell, Nvidia had been stripping out compute power, leading to better power efficiency but at the cost of compute performance. Pascal is set to fix this and bring Nvidia’s compute power back on par with AMD’s, though likely at the cost of efficiency.
Even though Samsung lost out this time, the simple fact that they were in competition with TSMC speaks volumes. TSMC has been falling slightly behind in terms of process technology and trying to meet Apple’s insatiable demand. In some ways, using Samsung would have made sense as Samsung is also set to be a major HBM2 supplier as well, simplifying the production for Nvidia. In the end though, it seems that TSMC’s long experience with Nvidia and GPU’s won out.
According to multiple sources, AMD is working on 3 new GPUs as part of their 2016 lineup. Likely part of the Arctic Islands release, the 3 chips all fit the theme being called Greenland, Baffin and Ellesmere. Greenland is set to be the flagship product while Baffin and Ellesmere will probably target other segments like performance and mainstream.
While we’ve already gotten some information about what Greenland will look like, we’re also learning that the ISA (Instruction Set Architecture) is undergoing major changes. Since 2011, AMD has been working with GCN, or Graphics Core Next. There have been several iterations from the original GCN 1.0 like 1.1 which powers the R9 290/290X and GCN 1.2 which is found in the R9 285 and Fury/Fiji. With this latest ISA change, AMD is moving beyond GCN to “post-GCN” on a radically new architecture. In some ways, the time spent with GCN highlighted how AMD has a slower cycle than Nvidia but also showcased the staying power of the architectures, which is finally seeing AMD’s early invest pay off with DX12.
This new ISA is also set to net AMD massive gains in the power efficiency department. While Nvidia showed strong gains moving from Kepler to Maxwell, AMD is reporting that their new ISA will double power efficiency. This should leapfrog Maxwell and will likely be a strong competitor to Pascal. Given the use of either 14nm or 16nm FinFETs as well, we may well see even more massive power savings. Combined with up to 32GB of HBM2 which is also pretty power efficient, 2016 looks to be a pretty big year.
Both Baffin and Ellesmere are also expected to be entirely new cards as well. This will bring a much-needed refresh to the AMD lineup which largely consists of rebrands at this point. Hopefully, these new cards and Zen will bring enough to the table and turn things around for the beleaguered firm.
Thank you WCCFTech for providing us with this information
After being stuck for what seems like forever on 28nm, we’re finally getting a glimpse of the monsters set to arrive with TSMC’s 16nm process. Code-named Pascal, Nvidia’s top end 16nm GPU is reportedly pushing 17 billion transistors, set to replace the current GM200.
To put that number in context, the current Titan X only clocks in at about 8 billion transistors, making the “GP100” Pascal more than twice as complex and likely twice as dense. Even AMD’s monster Fury X only pushes 8.9 billion transistors, which is still far and behind Pascal. Combined with a reported 32Gb of HBM VRAM at the highest SKU, Pascal may show a massive jump in performance compared to our current chips.
These gains are only possible with the new 16nm FinFET process from TSMC. Being nearly twice as dense, 16nm would allow Nvidia and AMD to double transistors in only a slightly larger die size. Combined with better power efficiency from being a lower process, FinFETs and HBM, efficiency should also improve despite having more transistors. Despite being called 16nm, TSMC’s process is closer to Intel’s 22nm or Samsung’s 20nm design, so there is certainly even more room to shrink in the future.
While CPUs have not benefitted as much from increased transistor counts, GPUs are relatively less complex and easier to make full use of the extra transistors. With DX12 and Vulcan in line as well as the new architectures from Nvidia and AMD, these new technologies should create a perfect storm to push GPU performance and gaming forward.
Thank you Fudzilla for providing us with this information
We have known for a while now that Nvidia wanted to get onto the HBM wagon too and that is pretty much a given. It’s the next generation video memory and it looks extremely promising so far. The next gen GPU should be the GP100 and it will reportedly rock somewhere between 4500 and 6000 CUDA cores and it will be coupled with HBM2 memory.
The latest news now points towards two new graphics cards: one for the consumer market and one for the server and workstations, again this is almost a logical assumption anyway. The first new GPU will use 4-Hi stacks of HBM2 memory and be aimed at the consumer market while the other business-oriented GPU will feature 8-Hi stacks of HBM2 memory. Effectively that means HBM powered graphics cards from Nvidia with up to 32GB of very fast VRAM.
Whether the new consumer model will be named the GeForce GTX 1080 is currently unknown. It could very well be that Nvidia is working on a new naming too, not only to avoid the thousand digits, but also to underline that is truly is a next generation graphics card; much like AMD has done with their Fury branding.
We’ll make sure to keep you updated as more news and leaks emerge on these new Nvidia 16nm based graphics cards.
Thank You TweakTown for providing us with this information
In a move that is sure to please AMD and Nvidia fans, TSMC has started volume production of 16nm FinFET chips. According to TSMC president and co-CEO Mark Liu, the ramp up for 16nm will be more aggressive than their 20nm process, leading to improved market share for TSMC. Both AMD and Nvidia rely heavily on TSMC to deliver their chips, with the mobile focus of 20nm leading to an extended cycle on the 28nm for PC GPUs.
With the reveal that AMD has taped out their first FinFET chip, its looks like the chip was for TSMC 16nm FinFET, not Global Foundries/Samsung 14nm FinFET. With shipments for 16nm already started, AMD and Nvidia may have new GPUs set for Q2/Q3 2016, offering improved power efficiency and lower power consumption.
TSMC also revealed that their 10nm process, based heavily on 16nm, will also begin production in 2016. The 10nm process will improve frequencies by 15% and power consumption by 35%, which is understandable given that FinFETs tend to reduce clock speeds when they are introduced. 7nm is also planned for early 2018, which is a pretty aggressive schedule for TSMC.
With TSMC back on track, the future for GPUs looks brighter. Given the differing standards for semiconductors, Intel 14/10nm and Samsung 14nm are not comparable with TCMC’s. So while TSMC may reach 10nm before Intel does, Intel will actually still have the smaller process.
While much of the focus from yesterday’s financial call was the poor state of finances, another small tidbit about future plans was mentioned. CEO Lisa Su revealed that AMD has taped out their first FinFET chips back in June this year. While the exact process node has not been revealed, we can assume it is either TSMC 16nm or Samsung/Global Foundries 14nm.
From when the chip first gets taped out to initial production, it takes about a year for the chip to get to market. It takes about 3 months for the chip to get produced, and about 1 month to implement all the fixes. Then the cycle happens once more, taking another 4 months, then finally, production is ramped up, taking a grand total of about 12 months. This means we can expect the FinFET chips to arrive about June/July in 2016, assuming AMD doesn’t hit any major obstacles.
FinFETs will help AMD claw back power efficiency from Intel, who already moved to FinFETs with their mainstream 22nm process back in 2013. Given the timeframe, the chip is likely either a GCN Arctic Islands chip or a Summit Ridge Zen based processor. Both, but especially on the CPU side, can do with better efficiency as Nvidia and Intel are ahead on that front. FinFETs can also reduce overclocking headroom so AMD will have to focus on improving instruction efficiency to compensate. In all, 2016 looks to be a good year for AMD if they can make it there.
AMD has revealed what the company plans to do with its GPUs and CPUs in the next 5 years at the PC Cluster Consortium event in Osaka Japan, where AMD’s Junji Hayashi revealed the company’s roadmap.
During the event, AMD has focused on its graphics IP and the products that involved it, including discrete Radeon graphics cards and Radeon powered Accelerated Processing Units. There have been talks about AMD’s upcoming K12 ARM as well as the x86 Zen CPU core, including a strategy of how the company plans to introduce both x86 and ARM powered SOCs to the market in a pin for pin compatible platform code-named SkyBridge.
It is said that both CPUs are 64-bit capable parts coming in a 14nm FinFET ‘shell’, but one is based on the ARMv8 architecture while the other is based on the more traditional x86 AMD64 architecture, having them target the server, embedded, semi-custom and client markets.
AMD has also talked about “many threads” revealing that the K12 will come with Simultaneous Multi-Threading (SMT) technology in contrast to the company’s Clustered Multi-Thread (CMT) technology we are able to see in the Bulldozer family. SMT essentially takes advantage of the various resources in the core which are underutilized and dedicate to an additional, slower, execution thread for added throughput. In contrast, CMT is looking for opportunities to share resources between two different CPU cores, instead of doing it inside a single CPU core.
Hayashi also revealed AMD’s GPU roadmap, which reveals that the company is employing a two-year cadence to updating its GPU architecture inside APUs. It looks like the company will add Accelerated Processing Units with updated GPU architectures once every two years. The roadmap also reveals that AMD plans to introduce what it described as a High Performance Computing APU which carries a 200 – 300 watts TDP, having the company stating that the APU in question will excel in HPC applications.
AMD apparently did not attempt to go with newer APUs in the future because it was not viable in terms of memory bandwidth. Instead, the company’s stacked High Bandwidth Memory will be used as an alternative, making the design extremely effective. The second generation of HBM is said to be 9 times faster than GDDR5 memory and 128 times faster than DDR3 memory.
The company has not revealed any code names for the GPU architectures, but a previous leak pointed out that the architecture will debut on 16nm FinFET and will be code-named Arctic Islands. More specific details about AMD’s products will be revealed in May at the Financial Analyst Day event.
Thank you WCCF for providing us with this information
Some people just have a keen eye when being at the right place at the right time, and as such the Patriot Torch SSD has been spotted as listed on Amazon ahead of its launch. While the drive isn’t even listed on the product pages yet, we do get some information about the drive from this listing. It actually already has one user review dated the 4th November 2014.
The Patriot Torch is built around the Phison SSD controller and paired with qualified 16nm Synchronous MLC NAND flash chips to deliver a good performance and reliability while staying cost-effective. The sequential read and write speeds are listed up to 545MB/s and 430MB/s for the 120GB model while the 240GB model can achieve 555MB/s reading and 535MB/s writing.
The only two features mentioned are TRIM and Advanced wear-leveling, but we can assume that SMART is present just as well, and DevSleep is a good bet. It’s backed by a 3-year warranty and supports Windows XP, Vista, 7, 8, 8.1, Mac OS, and Linux.
It’s listed as available and ready for dispatch at a price of $106.01 (MSRP $109.99) for the 240GB model while the 120GB capacity will cost you $66.63 (MSRP $79.99).
Thanks to TechPowerUp for providing us with this information