Nvidia Reportedly Stopped Maxwell GM204 and GM200 Production

Winding down production of a soon to be discontinued product is industry standard practice. When you see production of current models cease, that’s when you know the next generation is just around to corner. Two weeks ago, we brought you news that Nvidia may have stopped GTX 980Ti production. Now it seems that Nvidia has also ceased production of 2 other Maxwell-based GPUs in preparation of Pascal.

According to HWBattle, both of the GM204 based GPUs are no longer being supplied to AiB partners. This means the GTX 970 and 980 will disappear off store shelves sometime between the next 2 months given a normal logistics situation. At the same time, this means the replacements for GM200 and GM204 are well on their way and may arrive within 2 months as well. This is perfect for the late May launch at Computex.

As we’ve reported before, the replacements will be the 3 GP104 chips that will be called GTX 1080, 1070 and 1060Ti. These are the GP104-400, GP104-200 and GP104-150 respectively. Furthermore, the usual practice of launching reference cards first followed by custom ones will not be followed by the GTX 1070. Instead, the custom cards will launch at the same time as the reference models. With Computex just around the corner, we’ll all find out soon what GeForce Pascal truly is like.

Gainward Phoenix GTX 970 4GB Graphics Card Review

Introduction


The GTX 970 has been with us for around a year now and in that time it has cemented itself as possibly one of the best Maxwell-based cards available. Obviously the Titan X is the most powerful and the GTX 980Ti is the obvious option if you have £600 to spend, but the GTX 970 is the right amount of power to sustain consumers into the next generation of cards without taking a massive hit come resale when Pascal and HBM v2 is released.

Today in the test bench is the Gainward Phoenix GTX 970. It’s nice to see manufacturers still pushing this lower line considering how popular the GTX 980Ti has been and how saturated the market already is with competing cards.

I can’t really go into a GTX 970 review without at least touching on the issues that were present with the VRAM and miss advertised specifications. When it was first released, the GTX 970 seemed like the perfect card, 980 performance at a reduced price, then reviewers and consumers started to notice the drop in performance at high VRAM loads even though it was well within the VRAM limit of the card. NVIDIA decided to utilise an altered DDR5 memory architecture on this card which increased the speed of the first 3.5GB, but severely hindered the last 512MB. Along with that, the cores, ROP and TMU’s were all advertised higher than they really were. All that being said, the GTX 970 is still a cracking card and one of my all time favourites.

Packaging and accessories

When I opened the shipping box, I was surprised to how large the actual retail box was; the bright colours are certainly enticing.

The rear of the box is simple with key information listed. Some of the more important features regarding NVIDIA and the power of the GTX 970 are outlined with graphics.

The accessories are the usual lot, molex to PCIe 6-pin connector, DVI to VGA adapter, driver disc and installation manual.

Nvidia Prepping Full GM204 990M for Gaming Laptops

With the inevitable march towards efficiency due to improved architectures and process technology, laptop GPUs have never been closer to their full-fledged desktop cousins. Right now Nvidia’s top mobile GPU, the GTX 980M, only has 1536 shaders, less than the number found in the GTX 970. That’s all set to change with the arrival of the 990M, based off the same GM204 as the once flagship GTX 980.

With 20148 shaders, the 990M will have 33% more shaders than the 980M. While clocks will likely be below that of the desktop 980, performance should still be on par with the 970. This would put performance below previous rumours but should still be respectable for 1440p gaming though 4K is still a ways away. Even with the lower clocks and Maxwell’s inherent efficiency, laptops boasting the 990M will likely still require a substantial power delivery system and an equally beefy cooling system.

With 28nm long in the tooth and 16nm just a few months away, it begs the question why Nvidia is introducing a new mobile flagship. One possibility is that Nvidia wants to target gamers who want both performance and mobility. Another reason might also be AMD’s Nano GPU. With the Nano, AMD has substantially improved their efficiency, meaning a cut-down Fiji die may very well compete in the same segment as the 990M. Nvidia may simply be introducing this card to retain their mobile crown. Whatever the reason, the card is set to launch later this year.

Thank you TechRival for providing us with this information

Nvidia Announces Quadro M4000 and M5000 Maxwell-Based GPUs

NVIDIA announced the successors to the Quadro K5200 and K4200 graphics cards and the two new cards have been dubbed the Quadro M5000 and the Quadro M4000. These two new cards are based on the Maxwell 2 architecture just as the Quadro M6000 announced back in April was, and these aren’t just some wash up rebrands of previous cards but offer a performance that should be double that of the predecessors.

NVIDIA isn’t always quick to share all the details on these new cards, but we can predict the missing information based on the information we got and the general knowledge of the systems and chips. The first thing we don’t fully know is the actual GPU, but based on the CUDA cores it is safe to assume that it is a fully enabled GM204 with 2048 CUDA cores and the full 256-bit memory bus. The M4000 is using the same GPU, but it only has 1664 active CUDA cores. Both cards are almost equal when it comes to memory as they both feature 8GB GDDR5 memory. The M5000’s memory is clocked slightly higher and it also features a software based ECC support.

A thing that didn’t change much over the predecessor Quadro cards K4200 and K5200 is the power consumption. They all draw power from a single 6-pin PCIe power connector and the M5000 comes with a TDP of 150W, same as the K5200. The Quadro M4000 got a slight bump up to a 120W TDP over the previous 105″.

The new generation Maxwell GPU has a lot of benefits over the older Kepler and one of them is being able to native support up to four 4k monitors and both of these cards can do that with four DisplayPort connectors. The Quadro M4000 is a single slot card that doesn’t have room for more connectors, but the Quadro M5000 being a dual-slot card also comes with a DVI connector.

NVIDIA never discloses the prices of these cards, they leave that up to the card partners, but it’s safe to assume that it will be around the same as the predecessor cards, $2000 and $1000 for the Quadro M5000 and Quadro M4000 respectively

Official Titan X Specifications Leaked, Launch Set in 24 Hours

VideoCards has just leaked the official specifications of NVIDIA’s GTX Titan X, having them match everything that has been leaked before.

As we all know, the Titan X houses the GM200 core, which is the successor of the GM204. One of the problems found in the GM204 was the lack of FP64 performance and if the GM200 is able to address that, we are looking at a worthy successor to the GK110-based original Titan chip.

There are two inconsistencies however, one linked to the other. The memory clock was thought to be 2000 MHz (8000 in total), however the table shows the memory is 1753 MHz. This in turn affects the memory bandwidth due to the fact that a 2000 MHz memory clock would have given the card a 384 GB/s output, but instead only 336 GB/s are outputted if the memory frequency turns out to be the one displayed in the table.

Thank you WCCF and VideoCards for providing us with this information

GTX 980 Ti or GTX Titan Incoming?

The Nvidia GeForce GTX 980 and 970 have barely arrived and the dust hasn’t even settled yet. That doesn’t hold anyone back though and we’re getting new leaks on the next generation of cards, or rather just the GPU so far.

A Czech IT publication got their hands on images and information on the new GM200 chips. If the normal schedule holds up, we can expect cards with this GPU in about 2 months time. These are rumours and unconfirmed information, but the site has been accurate in their predictions and leaks about the GTX 980 and 970, as well as many other products.

The future graphics card equipped with the Gm204 GM200 chip might be named GTX 980 Ti, or it could be a possible new GTX Titan. The GM200 has a die size of 551 mm² and is still built on the 28 nm technology. It is said to have 20-22 SMM’s and 2560-2816 CUDA cores. The memory bus runs on a 384-bit interface and the amount should be increased to 6 GB.

Allegedly the new GPU can perform 50% higher than the current GTX Titan Black and double that of the 780 Ti. All of this while consuming less power. The two mentioned cards use about 250 watts each, where the new cards are speculated to use just 200 watts, or possibly even less. That isn’t much for that kind of power.

According to the rumours, the core is done and test samples have been sent to customers. The source also revealed a dual-core flagship GTX 990 using GM204 cores, something that I’m sure will get consumers excited in many. The possible launch time is mentioned as December, which would fit well into the normal schedule from tape-out to finished product. Merry Gaming Christmas indeed. Or could this indeed be scheduled for CES 2015 in January?

Thank you PCtuning for providing us with this information

Images courtesy of PCtuning

Final GeForce GTX 980 Specifications Revealed

The new flagship GPU from Nvidia; the GeForce GTX 980, has been rumoured for some time now. When Nvidia announced their Game24 event, it didn’t take much to put two and two together and figure out that it would be presented during that show. Actually, they will be presenting two new cards, but it will be the GTX 980 that steals the show while the GTX 970 is more of a gap filler.

Since then we’ve seen a steady stream of leaks, photos and specifications. We’ve seen both the stock-card itself and some of the after market coolers. Now thanks to videocardz.com we have the final specifications to share with you.

The new Nvidia flagship is a card that deservers to carry this title, at least when you look at the pure specifications. The card uses the new GM204 processor and it is the first GPU based upon the Second Generation Maxwell architecture. The new processor still uses the 28 nm fabrication node, something many people had hoped to be smaller.

It features 16 Streaming Multiprocessors and has 64 Raster Operating Units (ROP), that’s twice that of the Kepler GK110. With 5.2 billion transistors packed into the GM204 the GPU has a die size of 398 mm². The L2 cache has also grown quite a bit, it has actually quadrupled into 2 MB.

Some of the new features mentioned are Dynamic Super Resolution, which is basically a new up-scaling technology advertised as “4k Quality on a 1080p display”. It also has Third Generation Delta Colour Compression; a technology that will help increase memory efficiency on the limited 256-bit interface. The last new technology we get teased with is the Multi-Pixel Programmable Sampling; a technology that improves sample randomization and reduces quantization artefacts.

The above was the processor specific numbers and while relevant, they’re not the most important thing for us now, so lets summarize what the GTX 980 really packs. It has 2048 CUDA cores, 128 TMU’s and 64 ROP’s. The card is equipped with 4 GB GDDR5 memory on a 256-bit interface. The memory will run at 7012 MHz effectively (1750 MHz) while the GPU itself has a base clock of 1126 MHz and a boost of 1216 MHz. 

These are some really nice figures for a card that only draws 165W TDP. This is only two-thirds of the power consumption the GK110 has, with its 250W. We also get confirmation that the GTX 980 will support HDMI 2.0 among its five display outputs, making them a DVI, a HDMI 2.0 and three DisplayPort 1.2 connectors.

The retail price, while still only a rumour, seems to be set at a MSRP of $599; about the same price as the current GTX 780Ti.

Thank you Videocardzfor providing us with this information

Images courtesy of Videocardz

Nvidia GeForce GTX 980 & 970 Specifications Revealed

The new Nvidia GeForce GTX 980 and GTX 970 graphics cards are set to launch very soon, bringing with them a wave of features that will make them the new flagship Nvidia GPU’s. There has been a lot of speculation on the cards specifications in recent weeks, but now it looks like most of the final specs have been revealed.

The GeForce GTX 980 is set to feature 2048 CUDA cores via 16 Streaming Multiprocessors Maxwell (SMM), while the GTX 970 features 1664 CUDA cores via 13 SMMs, leaving a 384 CUDA core difference between the two cards.

The GeForce GTX 980 (GM204) is being sold as the replacement for the current GK104 hardware and since the new architecture is set to offer improved power efficiency and performance, that shouldn’t be a difficult task.

We don’t have long to wait until these cards are fully revealed, but early indications are that they’re going to be a worthy step up from the current 7xx models.

Thank you VideoCardz for providing us with this information.

Image courtesy of VideoCardz.

Are Nvidia Preparing to Launch GeForce GTX 880 at Gamescom?

It’s not secret that Nvidia have a strong trade show presence and with Gamescom just around the corner we’re expecting to see a few big things from the company at this years event. Not only have we finally confirmed that the Nvidia Shield is very much real, but now further rumours point towards their new Maxwell based GPU being near completion.

The Maxwell architecture, the same architecture which was used in the new GeForce GTX 750 and 750 Ti GPU’s is the main focus for Nvidias GTX 8xx series. Leaked images of 3DMark suggest that the card is made of high-end parts and it is believed at this time that the screenshot it of the GTX 880 (GM204) performance figures.

It’s only a month until Gamescom and with the exception of the Titan-Z, it’s been a while since we had a new range of cards from the green team, even longer since we had something that was even worthy of the attention of your average consumer, given the bonkers price tag of the Titan series cards. It’s likely that Nvidia will at least tease their new cards along side their new tablet and Tegra K1 hardware, but one thing is for certain, GTX 8xx series cards are very close.

Thank you TechPowerUp for providing us with this information.

Image courtesy of TechPowerUp.

Rumour: Nvidia GTX 880 Pictured, 8GB and 375W TDP

Over at the Chinese technology website MyDrivers they have managed to get pictures of an engineering sample of what is believed to be a GTX 880 graphics card. The card’s GPU has the marking 1421A1 which apparently means it was manufactured within the first 21 weeks of this year according to VideoCardz. The prototype card has 8GB of GDDR5 with 16 double-backed SKHynix H5GQ4H24MFR 0.5GB modules, the standard version might therefore have 4GB but 8GB models will probably be made available too. 

Despite the power efficiency of Maxwell, which we have seen with the GTX 750 Ti, this card comes with three PCIe power connectors: two 6 pins and one 8 pin meaning a maximum power draw of 375W. Of course it is unlikely a Maxwell card will need this much power since it is more efficient than Kepler and no single-GPU Kepler card required more than a 6 pin and 8 pin. The extra power connectors are probably for engineering sample testing purposes.

The main part of the graphics card aside from the GPU die area is hidden with pixellation: whatever is behind it is clearly important. Based on earlier rumours it might be ARM cores or there could just be unique identifiers in these parts of the board that might lead to the person who leaked these photos being reprimanded by Nvidia. The GPU die is rumoured to be GM210 by the source by other websites have suggested GM204 is more likely since this GPU die is larger than GK104 but smaller than GK110.

Source: MyDrivers

Images courtesy of MyDrivers

NVIDIA New GeForce GTX 880 & GTX 870 Coming in Q4

The GPU war is still going strong and it looks like Nvidia will be launching their next high performance single GPU graphics card no later than Q4-2014. The new card is set to be the GTX 880 and will launch along side the new GTX 870.

The new cards are to be based on the new GM204 hardware, which is belied to still use the 28nm silicon fabrication process, due to delays at TSMC in implementing the new 20nm process, which is why we switched to the new Maxwell GPU’s, but didn’t switch to a new fabrication process. Of course, as we saw with the GTX 750, Maxwell is a great bit of engineering, and the extra time this gives Nvidia to fine tune 20nm could be a good thing.

With a rumoured 3200 CUDA cores, 200 TMUs, 32 ROPs, 256-bit wide GDDR5 memory interact and more, the new GM204 hardware sounds pretty awesome, and should offer steady improvements over the current GK104 architecture that powers their current line of high-end GPUs, but I guess we’ll have to wait until October / November to find out.

Thank you TechPowerUp for providing us with this information.

Image courtesy of TechPowerUp.

Nvidia Maxwell GM204 GPU Gets Detailed

More details have emerged on Nvidia’s upcoming 20nm Maxwell GPUs which will form the new GTX 800 series, however, as always these are very premature and somewhat vague details. As a result we encourage readers to remember they could be untrue, inaccurate or subject to change. The GM204 is the part in question and GM204 is believed to be the direct successor to the GK104 Kepler GPU that currently makes up the GTX 770, GTX 760, GTX 680, GTX 670 and GTX 660 Ti. Like GK104 the GM204 makes use of a 256 bit wide memory bus which also implies the card will come as standard with 2GB of GDDR5 VRAM with an option for 4GB models if Nvidia vendors choose to double up. We’ve also heard about the existence of the flagship GM210 GPU which will succeed the GK110 GPU, GM210. As a result we expect the GM210 to lead the top of the GTX 800 series (for example the GTX 890, GTX 880 Ti, GTX 880), while the GM204 GPU will likely make up the mid-range (GTX 870, GTX 860 Ti, GTX 860). Speculation suggests that it is possible the GM204 GPU may use Maxwell architecture but make use of the 28nm process to keep the costs down, this leads to further logical reinforcement that GM204 will be a mid-range Maxwell GPU.

Source: Expreview

Image credit to HQWallPapers.Org and Expreview

Nvidia GTX 800 Flagship “GM200” GPU Spotted In Transit

Recently we brought you some news about the upcoming Nvidia GTX 800 series and now some more details have come to light. While these are rumours, and as such should be treated with all the relevant precautions, the information still seems interesting. Nvidia’s GM200 Maxwell-based flagship GPU has been spotted in transit in India. The GM200 GPU will be the top of the line Maxwell based GPU die that will probably form the top next-generation Nvidia GPUs. The GM200 is going to be replacing the current flagship GK110 GPU (which forms the GTX Titan Z, Titan Black, Titan, 780 Ti and 780). The value of the transit is so significantly large that it can be inferred that this is one of the first GM200, if not the first, units being shipped from TSMC since they begun 20nm production of next-gen GPUs. Another potential explanation is that it could be a Loadboard interface for testing GM200 based GPU samples, hence the high price because it is not just a GPU but machinery for testing them. Early GM204 samples shipped with a 10,000 to 30,000 INR value per unit for some perspective, meaning this is 200 times the value. The part in transit is rumoured to be a 20nm node sample produced by TSMC on Nvidia’s request. Given the early stages of development a final launch of the GTX 800 series based on the GM200 GPU is at least 8 months away if not longer. The expected release date for the GTX 800 series based on this new development is the second quarter of 2015.

Source: WCCFTech

Image credit to HQWallPapers.Org and WCCFTech

GM204 Graphics Cards Already Being Tested By Nvidia

Nvidia is pretty good at keeping secrets, but it’s impossible to close all the holes and stop every leak from escaping. We have little idea about what GM204 actually is, but going on the current history of Nvidia hardware naming, we can safely guess that it’s a set of next-gen graphics hardware that will follow GK104, the current hardware in the newer GeFoce 760, or even the GK110 which is in cards like the Titan.

The product codes were spotted on a shipping declaration from April which was sent by Nvidia to its test labs in India, before being published on the Chinese forum over at Chiphell. According to the packaging the GM204 graphics unit features 8GB of GDDR5 memory which does suggest its a high-end gaming class graphics card, but we really no idea if this is a standalone GPU, mobile GPU, or something else altogether.

The rumour mill is working hard on this one, as it’s rumoured that GM204 could be the successor to the current GPU range, bringing around 10% performance improvements over the current generation. It’s not rumour that next-gen hardware is in the works at Nvidia, that much is a certainty, but for now we’ll have to wait a little longer to see if this leaked document is real or not.

Thank you Chiphell for providing us with this information.

Image courtesy of Chiphell.

Nvidia GM107 GPU Detailed, Will Form GTX 750 and GTX 750 Ti

VideoCardz have managed to get a hold of the new Nvidia GM107 GPU die layout. The GM107 is based on the next-generation Maxwell GPU architecture and according to rumours it contains a very different layout to its predecessor “Kepler”. In the GM107 die, which is expected to form the GTX 750 and GTX 750Ti, there are five SMM (streaming multiprocessors) blocks which each contain 128 CUDA cores. This makes it different to Kepler where one SMX unit contained 192 CUDA cores. If these rumours and the above image is true it means that the GM107 features 640 CUDA cores. Apparently all five SMM units are enabled on the GTX 750 Ti meaning it gets all 640 CUDA cores while the GTX 750 is rumoured to have one of these disabled meaning 512 CUDA cores in total.

Apparently the GM107 and another smaller (read: slower performing) chip, the GM108, are both to be built on the 28nm process with the Maxwell architecture. In terms of the higher end the GM206, GM204 and GM200 are the three GPU dies that will succeed the GK106, GK104 and GK110 respectively. These GPUs are expected to be built from Maxwell with a 20nm, not 28nm, process.

Image courtesy of VideoCardz.com, Via TechPowerUp