News Posts matching #28 nm

Return to Keyword Browsing

AMD Denies Hidden GCN CUs in Tahiti

Over the past few days, we were hearing rumors from many quarters that AMD's "Tahiti" high-performance GPU may have been a deviation from an older specification, and that it really has 2304 stream processors spread across 40 GCN compute units (CUs), instead of the 32 the Radeon HD 7970 ended up with. Both AMD and NVIDIA create more redundant components on their chips than their SKUs end up getting, so they could increase yields, it's a process commonly known as "harvesting".

On Tuesday, AMD quashed the rumor in an e-mail to Bright Side of News, in which it said that Tahiti XT (Radeon HD 7970) makes use of all the CUs there are, on the chip. The 40 CU / 2308 SP rumor gained some weight with the fact that since AMD is venturing into unknown territory (TSMC's 28 nm process, built after quite some delays and failures), it could do some heavy harvesting. Examples of harvesting in recent past include Intel Sandy Bridge-E Core i7 processors, which use only up to 6 out of 8 cores on the silicon, and only up to 15 MB out of 20 MB available on it; and GeForce GTX 480, which used only 480 out of 512 CUDA cores available on the GF100 GPU.

HD 7970 Overclocked to 1.26 GHz: 28 nm Tech Really Stretches Its Legs

Welcome to the first TechPowerUp news post of 2012! Read on for a couple of impressive overclocking feats with the HD 7970 graphics card.

It looks like the new AMD Radeon HD 7970 could be a bit of a dark horse and a lot more potent than its stock specifications would suggest - excellent for creating a competitive graphics card market. The reviews at stock speeds show the flagship HD 7970 to be around 10-15% faster than NVIDIA's flagship GTX 580, which doesn't seem all that impressive since the GTX 580 has been on the market for over a year now. However, what the reviews haven't really shown, is what kind of an overclocking monster the HD 7970 is. It definitely looks like AMD could have easily beaten the GTX 580 by a much bigger margin than they did, had they wanted to and it makes one wonder why they didn't.

VR-Zone have spent the New Year weekend overclocking this beast, having reached a whopping 1.26 GHz core clock speed with their HD 7970 - and decent benchmark improvements to go with it. Also, with the fan at 100%, the card never got above a very comfortable 68 degrees centigrade while running Furmark, which is amazing considering how this test is specifically designed to heat a graphics card to the max - but please see the update at the bottom of the article. The stock cooler may be noisy, but it's certainly very effective: an excellent result which will prolong the working life of the card.

Radeon HD 7950 Specifications Leaked

Last Thursday, AMD launched the Radeon HD 7970 graphics card based on its new 28 nm "Tahiti" silicon, but it remained tight-lipped about the specifications of the more important SKU that will be based on it, the Radeon HD 7950. The HD 7970 will carry a launch price of US $550, making the HD 7950 an SKU to watch out for. According to details released by XTReview, the "Tahiti Pro" or HD 7950 will be carved out this way:
  • 1792 stream processors, 28 GCN compute units
  • 112 TMUs, 32 ROPs (derived)
  • 384-bit GDDR5 memory interface
  • 3 GB memory, memory clock around 5.00 GHz
The core clock speed, the exact memory clock speed, and more importantly, the target price-point, remain unknown. The Radeon HD 7950 is expected to be launched on the 9th of January.

Christmas Special: The PC Technology of 2011

Welcome to the TechPowerUp 2011 PC technology Christmas special. We hope that you will enjoy reading it while tucking into your turkey, Christmas presents and a little too much wine... In this article, we go through the technology of 2011 that has had the most significance, the most impact and was generally the most talked about. It's not necessarily the best tech of 2011 which is the most significant though, since lemons can be just as significant as the ground-breakers in how they fail to deliver - and the backlash that goes with it.

January: Intel Sandy Bridge i5 & i7

Released on January 9th, the new Intel Core i5 & i7 processors were based on Intel's second generation Core architecture built on a 32 nm production process (HEXUS review). They included an IGP (Integrated Graphics Processor) physically on the same piece of silicon along with HyperThreading. These new dual and quad core processors soundly beat all previous generations of Intel processors in terms of processing performance, heat, power use, features and left AMD in the dust. Therefore, Intel badly needed some competition from AMD and unless you have been living under a rock, you will know how that turned out in October with the launch of Bulldozer. Sandy Bridge was a sound win and is generally considered to be the only architecture worth considering at this point. The i5-2500K is currently at the sweet spot of price/performance. It comes at a stock speed of 3.3 GHz, but typically overclocks to an amazing 4.5 - 5 GHz with a decent air cooler and without too much difficulty in getting there. Models in the budget i3 range were released at various times later. See this Wikipedia article for details.

AMD Dual-GPU Radeon HD 7990 to Launch in Q1 2012, Packs 6 GB Memory

Even 12 months ago, an Intel Nehalem-powered gaming PC with 6 GB of system memory was considered high-end. Now there's already talk of a graphics card taking shape, that has that much memory. On Thursday this week, AMD launched its Radeon HD 7970 graphics card, which features its newest 28 nm "Tahiti" GPU, and 3 GB of GDDR5 memory across a 384-bit wide memory interface. All along, it had plans of making a dual-GPU graphics card that made use of two of these GPUs to give you a Crossfire-on-a-stick solution. AMD codenamed this product "New Zealand". We are now getting to learn that codename "New Zealand" will carry the intuitive-sounding market name Radeon HD 7990, and that it is headed for a Q1 2012 launch.

This means that Radeon HD 7990 should arrive before April 2012. Tests show that Tahiti has superior energy-efficiency compared to previous-generation "Cayman" GPU, even as it has increased performance. From a technical standpoint, a graphics card featuring two of these Tahiti GPUs, running with specifications matching those of the single-GPU HD 7970, looks workable. Hence, there is talk of 6 GB of total graphics memory (3 GB per GPU system).

Club3D Radeon HD 7970 Graphics Card Pictured

Although AMD's Radeon add-in board partners (AIBs) are still "under oath" (read: NDA) not to disclose their products before the 9th of January, Club3D's Radeon HD 7970 press-shots still made it to the web. Pictured below are the Club3D Radeon HD 7970 graphics card and its box-art. The box art is typical Club3D, nothing exceptionally new there. The card, too, sticks to AMD's reference board design, till the point where it uses a red-colored PCB instead of a black one (found on AMD's HD 7970 press samples).

Based on the 28 nm "Tahiti" silicon, Radeon HD 7970 is the first DirectX 11.1 compliant graphics card, it packs 2048 GCN cores, and 3 GB of GDDR5 memory across a 384-bit wide memory interface. The core is clocked at 925 MHz, and the memory at 1375 MHz (5.50 GHz). It is expected that this card will stick to the price AMD disclosed, around US $550.

AMD Radeon HD 7900 Key Features Listed

We've already been through the specifications of HD 7970 "Tahiti" in some detail that matters to those who can draw a performance hunch looking at them. This latest slide shows you the feature-set this GPU comes with. To begin with, there are three main categories of feature updates: Graphics CoreNext, AMD Eyefinity 2.0, and AMD APP Acceleration. AMD claims CoreNext to be a "revolutionary" new architecture that changes the way the GPU crunches numbers.

For the past five generations (since Radeon HD 2000), AMD GPUs have used the VLIW (very-long instruction word) core arrangement. Even the latest VLIW4 introduced by Radeon HD 6900 series, was an evolution, than a revolution of that. CoreNext replaces VLIW stream processors with super-scalar Graphics Compute cores. This should translate to higher performance per mm² die-area, resulting in smaller GPUs, giving AMD room for greater cost-cutting if the competition from NVIDIA for this generation takes effect. The GPU itself is built on TSMC's new 28 nm silicon fabrication process. Next up, AMD confirmed support for PCI-Express 3.0 interface, that nearly doubles system bus bandwidth over the previous generation.

AMD Pulls Radeon HD 7970 Launch to December 22

In a surprising move, AMD pulled the launch date of Radeon HD 7970, a high-performance single-GPU graphics card based on the 28 nm Tahiti silicon, up to December 22, 2011; from its earlier launch date of January 09, 2012. The January date was a lot more than speculation, as older presentation slides from AMD to distributors and retailers talked specifically about it. The move to pull December 22 (next Thursday) spices things up in the run up for CES. First, it gives AIB partners full freedom to show off their custom-design graphics cards at the event, along with full details about GPU specifications and clock speeds.

According to a VR-Zone report, Radeon HD 7970 will launch on December 22, 2011, this will be the day you will be able to read reviews of the card (at least the AMD reference design board), online. It will be a limited launch (read: paper-launch), but one can expect "full" retail availability of the card by January 09. Another interesting bit of information is concerning the Radeon HD 7950. This card will be available in non-reference board designs from day one, it will however launch on January 09.

AMD Starts Shipping 28 nm GPUs for Revenue

AMD CEO Rory Read, speaking at the IT Supply Chain conference organized by Raymond James this Tuesday, said that his company had begun shipping 28 nm GPUs for revenue (meaning, in volumes big enough to fetch revenue). With it, AMD fulfilled its promise to be the first to the market with GPUs built on the 28 nm silicon fab process. AMD's foundry partner for these chips is TSMC. "We are ramping 28nm [products] with TSMC in Taiwan and shipping the products here and now. We are very excited about the products," said Read.

At the upcoming CES event, AMD will formally unveil a range of products that will use its 28 nm GPUs. CES will give AMD a good opportunity to bag design wins with large volume manufacturers of notebooks and PCs. What this means for the enthusiast community is that whenever AMD does launch its Radeon HD 7900 series, it won't be a "paper-launch".

GeForce Kepler 104 and 100 GPU Specifications Compiled

A quick stroll through our previous article about how the GeForce Kepler family of next-generation GPUs is laid out, would tell you that GeForce Kepler 104 (GK104), is going to be NVIDIA's answer to AMD's Tahiti. GK104 will be a high-performance (≠ high-end) GPU by NVIDIA that will have many of the features that were reserved for its previous high-end GPUs (such as a 384-bit wide GDDR5 memory interface), but will not be NVIDIA's most powerful GPU in the series. The throne will be kept empty for GK100, which will comply with NVIDIA's "go all in" design ideology for high-end GPUs.

3DCenter.org compiled a few specifications of the GK104 and GK100. They go like this:
GK104
  • 640 to 768 CUDA cores
  • 80 to 96 TMUs (depending on what the CUDA core count ends up being)
  • 384-bit GDDR5 memory interface, 48 ROPs
  • Built on the 28 nm TSMC process
  • Products based on this will launch in the first quarter of 2012

January 9 Launch Date for AMD Radeon HD 7900

Ladies and Gentlemen with graphics card upgrade plans, circle the date January 09, 2012, for this is going to be the day AMD will launch its next generation high-performance graphics cards in the Radeon HD 7900 series, according to reliable market sources DonanimHaber spoke with. On the 9th, AMD is expected to unveil at least two new SKUs in the HD 7900 series, most likely, HD 7970 and HD 7950. These will be based on the new 28 nm "Tahiti" silicon that will use completely redesigned number-crunching machinery, and a very wide memory bus.

AMD Tahiti GPU Specifications Compiled

If the word on the optical fibers is true, we are less than a month away from the launch of AMD's next high-end graphics card family based on its next high-performance GPU, codenamed "Tahiti". According to 3DCenter, AMD will launch new graphics card models based on this GPU around January 10, 2012. It is expected that we'll learn a lot more about these GPUs, maybe even come across AIB-branded graphics cards, at the upcoming CES event.

3DCenter compiled specifications of "Tahiti", based on bits and pieces of information from various sources. The specs can be listed out as:
  • 4.50 billion transistors, die-area of 380 mm², built on TSMC 28 nm process
  • Advanced GCN 1D architecture
  • 2048 1D processing cores
  • 128 TMUs, 48 ROPs
  • 384-bit wide GDDR5 memory interface, memory clock slightly below 1 GHz, target bandwidth of 240~264 GB/s
In Gandhi's words, salt is as free as the air.

HP Starts Listing Notebooks with Radeon HD 7600M Graphics

PC major HP made a slight slip, disclosing AMD's upcoming Radeon HD 7000M series GPUs as options available for some of its notebooks. These are the first public disclosures of a 28 nm GPU. The new options that had been made available were Radeon HD 7670M, and Radeon HD 7690M, both based on the "Thames-XT" GPU. The Thames-XT GPU is built on the 28 nm process, and features current-generation VLIW4 stream processors, a 128-bit wide GDDR5 memory interface, and is DirectX 11 compliant. AMD unveiled this GPU at a press event in London on the 5th.

NVIDIA GeForce Kepler Roadmap Compiled

2012-13 promises to be a period of big graphics product launches, centric to a new DirectX version, DirectX 11.1, which will ship with Microsoft's next major Windows version (currently referred to as Windows 8). Information compiled by ExpertsPC.com and 4Gamer.net tables what NVIDIA's next-generation graphics family could look like, and around what time it could be released to market. With its next-generation GeForce Kepler family of GPUs, NVIDIA will follow a sensible bottom-up product release model, to ensure that it isn't met with any technical hurdles with TSMC's new 28 nm manufacturing process, and so it could launch GPUs with increasingly higher transistor counts, till its top-of-the-line GPU is outed.

The first GPU in NVIDIA's pipeline is the GeForce Kepler 107 (GK107), on which will be based entry thru lower-mainstream SKUs. The data doesn't reveal things like core counts, but points out that GK107 will have a 128-bit wide GDDR5 memory interface, will use the current-generation PCI-Express 2.0 bus, will be built on the 28 nm process, and will support DirectX 11.1. This will be followed by the GK106, on which "sweet-spot" SKUs could be based. This will be NVIDIA's first PCI-Express 3.0 compliant GPU, it will have a 256-bit wide GDDR5 memory interface.

TSMC 28 nm Technology in Volume Production

TSMC today announced that its 28nm process is in volume production and production wafers have been shipped to customers. TSMC leads the foundry segment to achieve volume production at 28nm node.

TSMC's 28nm process offering includes 28nm High Performance (28HP), 28nm High Performance Low Power (28HPL), 28nm Low Power (28LP), and 28nm High Performance Mobile Computing (28HPM). Among these technology offerings, 28HP, 28HPL and 28LP are all in volume production and 28HPM will be ready for production by the end of this year. The production-version design collateral of 28HPM has been distributed to most mobile computing customers for their product-design use.

AMD's First 28 nm GPUs in December

It looks like AMD will have the symbolic achievement of launching its first GPUs built on the new 28 nanometer process in 2011 itself. Sources told Heise.de that AMD is working towards launching some of its planned 28 nm GPUs in the second week of December, 2011. One of these sources specifically named December 06. Details on whether the launched GPU will be for the mobile (notebook) or desktop (graphics card) platforms; or even whether it will use the VLIW4 or so-called 'NextGen' compute architecture, are not known at this point.

Another source reinforced the theory that the launch will be more about symbolism than volume manufacturing for sales. It's likely that a small number of these GPUs will be manufactured, just about enough to send to OEMs for their qualification, and perhaps even the media for published performance testing. We expect these GPUs to be lower-end or mid-range GPUs, and since AMD is reserving the NextGen compute architecture for only the high-end GPU part, these ones will most likely use VLIW4.

AMD to Turn to TSMC for ''Bulldozer'' Manufacturing

AMD is rumored to be seeking ties with TSMC, Taiwan's premier semiconductor manufacturing foundry, for future manufacturing of its "Bulldozer" architecture processors, according to a report by DonanimHaber. This has two very distinct implications: first, AMD could be facing issues with GlobalFoundries 32 nm HKMG node, its de facto foundry for CPU manufacturing, and second, this could just be an obvious development of future low-power APUs based on the new x86 architecture being manufactured at TSMC, much like how current E-series and C-series APUs are.

Then again, AMD doesn't exactly have any APUs in works that use "Bulldozer" architecture for the x86 cores, rather, its successor codenamed "Piledriver". Another couple of important things to note here are that TSMC does not have a 32 nm bulk node (it was scrapped with the transition to 28 nm bulk), and its HKMG (high-K metal gate transistor) manufacturing technology is deployed rather recently. It would be interesting to follow this development.

NVIDIA Begins Sampling First Kepler GPUs

Graphics major NVIDIA has begun sampling the first GPUs of its next-generation Kepler family. Kepler succeeds the present generation Fermi family of GPUs, that make up the GeForce 400 and 500 series. Members of the GeForce Kepler family will follow the codename nomenclature of GK1xx, like GeForce Fermi followed GF1xx. The first GPU being sampled is codenamed GK107. The codename suggests that this isn't the top-end part, it's more like a lower-mainstream or value segment offering.

It is reported that GK107 has a 128-bit GDDR5 memory bus, which also supports inexpensive DDR3 memory. It is built in small quantities on TSMC's 28 nanometer fab process (as the foundry isn't ready for 28 nm bulk manufacturing. NVIDIA will follow a "bottom-to-top" strategy, when dealing with a new fab process technology. It will first design the smallest, simplest GPUs in the lineup, and then gradually move on to larger ones. The first GK107-based SKUs will succeed the GeForce GT 500M series. NVIDIA will carve out four SKUs, internally, NVIDIA will refer to those as N13P-LP, N13P-GS, N13P-GT and N13E-GE, with N13E-GE being the "enthusiast" part. Its market SKU will likely succeed the GeForce GT 560M.

AMD Showcases its First 28 nm GPU

AMD showed off its first graphics processor (GPU) built on TSMC's cutting-edge 28 nanometer silicon fabrication process, the next foundry process standard for discrete GPUs. Bulk manufacturing at TSMC's Fab 15 facility at 28 nm is still taking shape, TSMC will take volume orders only later this year. For the moment, it can run small batches for designers to test their designs. The GPU was running on a mobile platform (pictured below, the red PCB), cooled by a compact copper-channel air cooler, leading us to believe that this is a mainstream segment GPU, if not lower. The demo platform was showcased running DirectX 11 title Dirt 3. Besides that, absolutely no other details were shared, not even a company codename for the GPU board.

AMD's Next-Generation Wichita and Krishna APUs Detailed

In its latest presentation to industry partners, AMD detailed its upcoming Deccan low-power computing platform, targeting the market Intel's Atom and VIA's Nano processors do. AMD is currently behind the "Zacate" and "Ontario" processors, which deliver high performance/watt x86 computing at low power draw and costs. The company's future platform will be called "Deccan," consisting of processors codenamed "Wichita" and "Krishna," targeting the ULV desktop and netbook markets, respectively. With the next generation, AMD is looking to take advantage of the 28 nanometer manufacturing process to put four x86-64 cores based on the Bobcat architecture on a single piece of silicon, with an integrated memory controller and AMD Radeon discrete-class graphics.

The biggest change here isn't the fact that there are four cores, or that it's built on 28 nm, but that Wichita and Krishna are completely single-chip. The FCH or Fusion Controller Hub has been completely fused into the APU silicon. Motherboards and notebook logic boards will have just one big chip, with no "chipset" of any form. This makes AMD's Wichita and Krishna the industry's very first true x86-based consumer SoC (system on chip). The integrated memory controller now supports DDR3-1600 MHz memory. The integrated AMD Radeon graphics is set to get a performance and SIMD boost, as well, including a Secure Asset Management Unit (SAMU). AMD's next generation APUs are slated for 2012.

AMD Radeon HD 7000 Series to be PCI-Express 3.0 Compliant

AMD's next generation of graphics processors (GPUs) that will be branded under the HD 7000 series, are reported to be PCI-Express Generation 3 compliant. The desktop discrete graphics cards will feature PCI-Express 3.0 x16 bus interfaces, and will be fully backwards-compatible with older versions of the bus, including Gen 1 and Gen 2. Motherboards sold today feature Gen 2 PCI-E slots, although some of the very latest motherboards launched by major vendors feature PCI-Express 3.0 slots.

The new bus doubles the bandwidth over PCI-E 2.0, with 1 GB/s of bandwidth per lane, per direction. PCI-Express 3.0 x16 would have 32 GB/s (256 Gbps) of bandwidth at its disposal, 16 GB/s per direction. AMD's next generation of GPUs, codenamed "Southern Islands" will be built on the new 28 nm process at TSMC, and will upscale VLIW4 stream processors. Some of the first PC platforms to fully support PCI-Express 3.0 will be Intel's Sandy Bridge-E. Whether AMD's GPUs have hit a bandwidth bottleneck with PCI-E Gen 2, or is AMD trying to just be standards-compliant, is a different question altogether.

AMD Southern Islands GPU Codenames Surface

It was discovered that an early version Catalyst 11.7 contains entries for new GPUs, spelling out the possible codenames of upcoming AMD Southern Islands GPUs. Southern Islands is the next generation of AMD Radeon GPUs, which will possibly see propagation of VLIW4 SIMD design to GPUs in all market segments, and possible adoption of 28 nm manufacturing process. Moving on to codenames, 3D Center observes that the top of the line GPU is codenamed "Tahiti", there is a possibility of two single-GPU variants.

A graphics card that makes use of two Tahiti chips is codenamed "New Zealand". "Thames" is the codenamed of a possible low cost mobile chip, it could also be a performance chip. There's some uncertainty here. Not much is known about another chip codenamed "Lombok", either. Perhaps it's a performance desktop chip.

TSMC Reiterates 28 nm Readiness by Q4 2011

TSMC reiterated that it will be ready with a 28 nanometer manufacturing process by Q4 2011. The semiconductor company handles manufacturing of graphics processors for both AMD and NVIDIA. After the current 40 nm process, 32 nm, the next milestone process, was canceled for GPU makers to leap to 28 nm, this caused the foundry transition to the next process to take longer than usual. The current 40 nm process already seems to be saturated by GPUs with over 3 billion transistors, which are barely able to maintain acceptable thermal specs without using some sort of power-load throttling mechanism.

TSMC Chairman and CEO, Morris Chang, confirmed that tape-outs will be starting as early as in Q3, and production of 28 nm chips will start in Q4. Chang expects that up to 3% of TSMC's revenues will be made from 28 nm chips by the end of the year. "We plan to have around 2% or 3% of our total revenue in the fourth quarter [to] be 28nm. The tape-outs of the 28-nanaometer will start to ramp in the second half, starting in the third quarter and then more in the fourth quarter. But the real momentum [for 28nm], we believe, will be next year," Chang said. Apart from GPUs, the 28 nm process will also benefit ARM processors, with multi-core ARM chips clocked at 3 GHz being on cards. The 28 nm bulk process will also dish out AMD's next generation accelerated processing units (APUs).

Toshiba and GlobalFoundies in Talks for Chip Manufacture Deal

Prospects are looking brighter by the day for GlobalFoundries, as the firm is reportedly in talks with Toshiba to strike a chip foundry partner deal. The deal will give GlobalFoundries contract to manufacture any or many of the plethora of chips Toshiba uses in computer hardware and consumer electronics. On GlobalFoundries' side, Doug Grose, chief executive officer of the American foundry, told Nikkei that the two firms are in the final stages of negotiations.

"We had been considering outsourcing production to Samsung Electronics Co and GlobalFoundries since last year," said a Toshiba official. With this deal, Toshiba is looking for 28 nm and 40 nm class manufacturing, thereby cutting a chunk of its R&D, and making its engineers focus on chip design. Parallel to this, a deal with Samsung will allow the Korean electronics giant to manufacture low-cost image-processing chips, while GlobalFoundries manufactures products with the latest-technologies.

AMD Plans Successors to Evergreen Series GPUs within 2010

AMD is just about complete with its top-to-bottom lineup of DirectX 11 compliant GPUs, whose series is codenamed Evergreen. Under this, the company has a GPU and its derivative targeted at almost every price point, from the entry-level Radeon HD 5450, to the fastest graphics card - Radeon HD 5970. With the onset of NVIDIA's first current-generation GPU codenamed Fermi, NVIDIA might attempt to reclaim the performance crown, and infuse competition in the higher-end segment. Currently there's no news of lower-end derivatives of Fermi.

Probably in response to Fermi, AMD is readying a new GPU architecture slated for release within this calendar year, in the second half. "We are ramping the ATI Radeon HD 5000 series now and look forward to refreshing the entire lineup in the second half of next year," said Dirk Meyer, chief executive officer of AMD, during quarterly conference call with financial analysts. His reference to "next year" was of next fiscal year (FY 2010), in context of him speaking at a quarterly conference call with financial analysts for FY 2009. The new series of GPUs is codenamed "Northern Islands". At this point in time, nothing much is known about the new GPU series. TSMC, AMD's main foundry partner for GPUs and motherboard chipsets, earlier claimed that it will be ready with a 28 nm production node by the end of this year. Whether the new GPU series makes use of the new fab technology is anybody's guess.
Return to Keyword Browsing