Welcome Stranger to OCC!Login | Register

Video Cards Article (148)

NVIDIA GeForce RTX 2080 Ti & RTX 2080 Founders Edition Review

NVIDIA GeForce RTX 2080 Ti & RTX 2080 Founders Edition Review

» September 18, 2018 05:00PM

XFX RX 570 RS 8GB XXX Edition Review

XFX RX 570 RS 8GB XXX Edition Review

» July 28, 2018 05:00PM

A Look at RX Vega 64 Efficiency

A Look at RX Vega 64 Efficiency

» June 28, 2018 05:00PM

GTX 1070 Ti Overclocking Guide

GTX 1070 Ti Overclocking Guide

» April 15, 2018 05:00PM

SAPPHIRE NITRO+ Radeon RX VEGA 56 Limited Edition 8G HBM2  Review

SAPPHIRE NITRO+ Radeon RX VEGA 56 Limited Edition 8G HBM2 Review

» January 15, 2018 05:00PM


Video Cards News (715)

AMD Potentially Readying Another Polaris GPU

Category: Video Cards
Posted: September 20, 2018 08:59AM
Author: Guest_Jim_*

Once again the patches to Linux from AMD have revealed something new, but this one is a bit more unusual than some we have seen before as the new PCI ID is for a Polaris GPU. Polaris was first available in the Radeon RX 400 series of graphics cards two years ago, and was then refreshed for the RX 500 series. This new ID is 0x6FDF, which the Phoronix writer states he has not seen referenced anywhere previously. It is listed as part of the "POLARIS10" family, and for those of you thinking that might mean this is based on an older version of Polaris, as the RX 500 series was Polaris 20, the Linux driver lists Polaris 20 in this same family, so this could be another version of Polaris 20 or something new. Whatever it is, its ultimate destination is also unknown, as it could easily be a specialized product for a specific market.

Something that makes this discovery particularly interesting is that typically when a company is going to launch a new GPU it reserves multiple PCI IDs, but here it was just one added. These reserved IDs are not always used and sometimes the difference between the devices they are assigned to is as simple as being an OEM product or not. Still, this is curious as it might mean we could see a new Polaris-based product in the future, along with Vega 12, Vega 20, Raven2, and Picasso. Of those GPUs, it is only Vega 20 we have much information about as it is almost certainly the 7 nm Vega GPU destined for compute applications in the Vega Instinct card. Vega 12 is still a mystery and while we know Raven2 and Picasso would be the GPUs in future APUs, we do not know the relative placement of these two products.

Source: Phoronix



NVIDIA Shares General Performance Graphs for Turing GPUs

Category: Video Cards
Posted: September 13, 2018 01:38PM
Author: Guest_Jim_*

Next week marks the launch of NVIDIA's Turing-based GeForce RTX 20 series graphics cards, but while it is only a week away, there is still a lot of speculation about how these cards will perform. NVIDIA was keen to show off the architecture's ray tracing performance when they were announced, but there was little information about performance for traditional rendering methods, or the information was clouded by the use of DLSS, Deep Learning Super Sampling. From what I have seen, DLSS works by using a pre-trained neural network to better upscale images for a game, so it can actually be rendered at a lower resolution, significantly improving performance. Anyway, now NVIDIA has shown off another graph with some generalized performance information on it at GTC Japan 2018.

According to one graph, the RTX 2080 is capable of 4K at 60 FPS, without DLSS, and is faster than the GTX 1080 Ti. With DLSS, the performance is even greater. Just how much greater though is unknown because, like many marketing graphs, there is no y-scale, or really even much of a y-axis (the only mark is for 4K 60 FPS). While WCCFtech does have the graphs, there is no information on what games or benchmarks were used to arrive at these performance values, or what settings were used.

According to VideoCardz last week, the embargo on RTX 2080 and RTX 2080 Ti reviews ends on September 19, so we should learn more then.

Source: WCCFtech



Embargoes for GeForce RTX 2080 and RTX 2080 Ti Reviews May Have Changed to September 19

Category: Video Cards
Posted: September 11, 2018 09:29AM
Author: Guest_Jim_*

This is likely not the news those interested in the upcoming GeForce RTX 20 Series want to see, assuming it is accurate. According to VideoCardz, reviewers have been having a hard time getting samples for testing, and this has become enough of a problem that NVIDIA has pushed the embargo date back on the RTX 2080 to September 19. The pre-orders for the cards are supposed to start shipping on September 20. This date also matches the embargo date for RTX 2080 Ti reviews, where previously the dates would have had the RTX 2080 reviews out two days prior (September 17). The embargo on articles covering the Turing architecture the RTX 20 series uses is supposed to lift on September 14.

Complicating this situation all the more is that apparently drivers have not been provided yet. This means even if you have a review unit, testing performance is potentially pointless for the time being, assuming you can get it running without proper drivers. The situation with how drivers are being delivered is actually something HardOCP wrote about a couple weeks ago, as its sources indicated drivers are being delivered exclusively by NVIDIA and not AIB partners, and only after signing NVIDIA's multi-year NDA.

Source: VideoCardz and HardOCP



Vega 20 Supports XGMI Based on Linux Patches

Category: Video Cards
Posted: September 5, 2018 11:30AM
Author: Guest_Jim_*

While it is true that currently our best knowledge on the Vega 20 GPU from AMD will be destined for compute applications, this first 7 nm GPU is still a source of some interesting news. Today it was spotted in Linux patches explicit support for XGMI for Vega 20. This technology is a high-speed, peer-to-peer interconnect, so it would allow multiple GPUs to work together and it is based on Infinity Fabric. Global Memory Interconnect, GMI, is already used for connecting the dies on Threadripper and EPYC CPUs from AMD with XGMI, inter-chip GMI, powering the connections between EPYC processors on multi-socket boards.

For a GPU destined for compute applications, such an interconnect makes sense, as NVIDIA has already demonstrated with its NVLink technology. Unfortunately we do not know if XGMI will need a special hardware connector between the cards or if it can be run over PCIe. We do know Vega 20 is supposed to support PCIe 4.0. Still, there is an intriguing possibility with XGMI on the GPU, and that is the potential for it to be leveraged to connect GPU and CPU, for the appropriate workload. That possibility is purely speculative at this point, but is still an interesting concept.

The expectation is for Vega 20-based products to launch later this year, but the only one known of is the Vega Instinct compute card, which is not meant for gaming.

Source: Phoronix



Strange Brigade is First Game with Vulkan Multi-GPU Implementation

Category: Video Cards
Posted: September 5, 2018 10:57AM
Author: Guest_Jim_*

It appears this news is actually from a week ago but it went unnoticed by for a while. Strange Brigade is the latest game from Rebellion Developments, the same company behind the Sniper Elite series, and it worked with AMD to implement support for both Vulkan and DirectX 12. Both of these are low-level APIs that can reduce CPU overhead by reducing the need for the CPU to translate commands and improve access to hardware features. One example of something possible with these APIs is for the game to directly allow for multi-GPU configurations, instead of relying on a separate, driver-side technology like Crossfire and SLI. While DirectX 12 had this capability from the beginning, Vulkan more recently received it and now Strange Brigade is the first game to actually use it. According to AMD, so take it with the appropriate amount of skepticism, using this mGPU capability with Vulkan achieves nearly perfect scaling with the RX Vega 56. The graph shown at the source has the RX Vega 56 with mGPU disabled reaching 53.8 FPS with Vulkan at the Ultra Preset running at 4K. That is actually fairly impressive on its own, but with a second RX Vega 56 it reached 102.3 FPS, which is a factor of 1.9x faster.

While systems with multiple discrete GPUs are not common, this capability still has a great deal of potential because, unlike Crossfire and SLI, it is possible to implement it to work with mismatched GPUs. This applies for both mismatched GPUs for performance and for manufacturer, so you could potentially see a benefit when an integrated GPU is used with a discrete GPU, and even when using both an AMD and NVIDIA graphics card.

Source: AMD



More Vega 12 Code Appears in Open Source Drivers

Category: Video Cards
Posted: August 24, 2018 07:46AM
Author: Guest_Jim_*

AMD's decision to create and support open source drivers for its GPUs on Linux has led to some interesting reveals, as unreleased and even unannounced products will appear in the code. Some months ago a Vega 12 GPU appeared, and it is still not known what exactly it is, but now the driver code gives us a look at the hardware. From the code shown by Phoronix, it appears Vega 12 is a cut down design compared to the full size Vega 10 used in the RX Vega 56, RX Vega 64, and other discrete Vega-based cards. While Vega 10 has 16 compute units in each of the four compute engines, resulting in 64 in total this Vega 12 GPU is listed as only having five per engine, so a total of 20. That does make this a significantly cut-down chip. It is also marked as stepping 4, while Vega 10 is stepping 0, and that it supports stereo rendering.

Some are speculating this might be the discrete mobile Vega GPU AMD showed off at CES but has yet to launch. Others think this could be used to replace the Polaris-based RX 500 series, and if this is the case it will almost certainly use a different memory controller as well, to work with GDDR5 or GDDR6 instead of HBM2. For now, only AMD knows but as more work is done on this software we can expect the product to come closer to an announcement.

Source: Phoronix and Github (Commit)



MSI Reveals Custom NVIDIA GeForce RTX GPU Series

Category: Video Cards
Posted: August 20, 2018 10:36AM
Author: Guest_Jim_*


The release of NVIDIA's newly announced GeForce RTX 20 series powered by the company's Turing architecture has been hotly anticipated by many. Of course it is not just the release of the reference version of the graphics card, but those from AIB partners. Now we have information from MSI on what its versions of the GeForce RTX 20 series of graphics cards.

Delivering the top thermal performance for their class are the GAMING TRIO and DUKE series, featuring new premium designs and triple fans. The VENTUS features a minimalist dual-fan design and the AERO has a blower-type design that will exhaust the hot air out the back of your case. For those of you looking for some more advanced cooling than the GAMING TRIO or DUKE though, there will be a SEA HAWK version as well, with hybrid air and liquid cooling. This will make liquid cooling accessible for many with ease of installation and its maintenance-free design.

By the way, the NVIDA Store page for the series is up currently, so you can visit there if you want to check out reference specifications or pre-order a Founders Edition version of the RTX 2080 or RTX 2080 Ti. The RTX 2080 will cost $799 and the RTX 2080 Ti comes in at $1199. Both will ship on or around September 20, so in one month. The RTX 2070 is also listed at $599, but is only marked to notify you when it becomes available.

Source: MSI



Intel Teases Future Discrete GPU on Twitter with 2020 Target

Category: Video Cards
Posted: August 16, 2018 06:05AM
Author: Guest_Jim_*

For several months Intel has been making moves to bring its own discrete GPU to a reality, such as the hiring of Raja Koduri, the former head of Radeon Technologies Group. Such a move by the company to enter this market is of great interest to many, if only to see a disruption to the currently duopoly between AMD and NVIDIA. Now we have a little more information on Intel's efforts via a 'We will set our graphics free' tweet from @IntelGraphics, including a teaser video. The video makes several claims about the successes of the Intel GPUs built into many of their CPUs, and then ends with a silhouette of something resembling a graphics card that is stated as coming in 2020.

Previously there were rumors putting the reveal of an Intel dGPU at CES 2019, and perhaps this will still be true, but as only a paper reveal and/or only professional parts, not consumer parts. Undoubtedly Intel's interest in entering the GPU market is also to produce and sell them for use in data centers and for machine learning applications. Hopefully whatever they produce will also be able to game well.

Source: Twitter (@IntelGraphics)



NVIDIA Shows Off Quadro RTX Cards at SIGGRAPH 2018

Category: Video Cards
Posted: August 14, 2018 07:06AM
Author: Guest_Jim_*

Technically this is only half of the announcement so many have been waiting for, but it is still an important half. NVIDIA announced three new graphics cards using its Turing GPU architecture, but these are all targeting the professional workspace. It is still the first official glimpse at the capabilities of the Turing architecture, and so we can get an idea of what NVIDIA is working with for its next generation of consumer-aimed GeForce GPUs.

The three cards announced are the Quadro RTX 8000, Quadro RTX 6000, and Quadro RTX 5000. One of the features of Turing these cards all share are RT Cores, as NVIDIA calls them, which are responsible for the accelerated ray tracing the manufacturer is pushing very heavily. According to NVIDIA, both the RTX 8000 and RTX 6000 are capable of 10 GigaRays/sec while the RTX 5000 reaches 6 GigaRays/sec. This lines up with the CUDA and Tensor core counts also stated, with both the RTX 8000 and RTX 6000 having 4608 CUDA core and 576 Tensor cores, while the RTX 5000 has 3072 CUDA and 384 Tensor cores. NVIDIA is claiming the 4608 CUDA cores will allow the Turing GPU to reach 16 TFLOPS and 16 IOPS (integer operations per second). All three cards will use Samsung-produced GDDR6 memory, with the RTX 8000 sporting 48 GB, RTX 6000 halving it at 24 GB, and the RTX 5000 coming in at 16 GB. Utilizing NVLink to connect two RTX 8000 cards though gets you up to 96 GB with up-to 100 GB/s data transfer.

The estimated price of these cards is $10,000 for the Quadro RTX 8000, $6300 for the Quadro RTX 6000, and $2300 for the Quadro RTX 5000. They are expected to be available in Q4 of this year. Unfortunately, we do not have any specific information on the next round of GeForce cards, although NVIDIA has shared a teaser video (below) that supposedly contains hints, and the belief is an announcement is coming at Gamescom 2018, which starts next week.

 

 

Source: NVIDIA



Radeon Pro WX 8200 Announced by AMD for $999

Category: Video Cards
Posted: August 13, 2018 09:36AM
Author: Guest_Jim_*

This week marks SIGGRAPH 2018, a conference focused just on computer graphics, so naturally it is a good time to announce new graphics products and technologies, for both consumers and professionals. Actually announced yesterday, AMD has revealed its new Radeon Pro WX 8200 workstation graphics card, which is based on the Vega 10 GPU and costs $999. The Vega 10 GPU is what you will find in basically all discrete graphics cards using Vega, including the RX Vega 64 and RX Vega 56. As far as specs go, this card is similar to the RX Vega 56, so a slightly cut down version of the Vega 10 GPU, but being a workstation card has special features such as ECC memory and dedicated AMD Secure Processor for IP-sensitive tasks.

Something interesting about the memory for the WX 8200 is that while it is still HBM2, like we see used in all discrete Vega cards, it is clocked higher, reaching 2.0 Gbps, as opposed to the 1.89 Gbps seen on the more expensive ($2199) WX 9100. This boost in speed could prove helpful for its intended use of professional workloads, including real-time visualization, physically-based rendering, and VR.

Announced at the same time, two new plug-ins are coming for AMD's Radeon ProRender, the company's high-performance physically-based rendering engine that CAD designers and 3D artists can take advantage of. These two new plug-ins are PTC Creo for allowing designers and engineers to quickly created rendered visualizations of their products, and Pixar USD viewport, for those building a USD Hydra viewport as this adds path-traced rendering for accurate previews.

AMD will also be showing off AI rendering with ROCm and Radeon Pro WX graphics card, real-time viewport raytracing for DCC and CAD applications, Cloud ProRender for expanding rendering capacity with the power of the cloud, and PIX on Windows for performance tuning and debugging DirectX 12 games.

Source: AMD and Anandtech



NVIDIA Releases New Driver for World of Warcraft: Battle for Azeroth

Category: Video Cards
Posted: August 1, 2018 08:11PM
Author: Grilka8050


Today NVIDIA launched a new driver for World of Warcraft: Battle for Azeroth. This driver also includes optimization for Monster Hunter: World. These drivers provide the best experience for GeForce gamers because the engineers work up to the last minute to optimize and perfect the driver for the gamer’s experience. This (and all Game-Ready Drivers) are WHQL-certified by Microsoft. This driver update also includes fixes to previously-reported issues. Even if you don’t play WoW or MH, I’d still download this driver. Happy Hunting!

You can download the new driver from the GeForce website.

Source: NVIDIA



NVIDIA GeForce 11 Series Launch Schedule Leaked

Category: Video Cards
Posted: July 24, 2018 10:53PM
Author: Grilka8050


Apparently, NVIDIA released some details to its AIB partners about their next generation of video cards. There have been a lot of rumors surrounding their release, with this adding even more fuel to the ever-expanding fire. This email that was sent to a YouTuber, Gamer Meld, is the source of our leaked information. It includes a launch schedule and some other interesting industry information.

The Schedule:

August 30th, 2018 - GeForce GTX 1180 

September 30th, 2018 - GeForce GTX 1170 and 1180+ 

October 30th - GeForce GTX 1160 

What an exciting autumn! That GTX 1180+ is an intriguing one, since is it like a Ti variant or just a GTX 1180 that's factory overclocked or with more memory bandwidth (or both)? There were some more juicy details in the email, too, including a “confirmation” of rumors about NVIDIA having a surplus of 10xx GPUs. This is the most likely reason that the launch dates were pushed back so far. As such, prices will probably fall between now and the August 30th launch date. 

Source: WCCFTech




NVIDIA GTX 1170 Benchmark Potentially Leaked

Category: Video Cards
Posted: July 22, 2018 01:30AM
Author: Grilka8050


The GTX 1170 has already been labeled a “1080 Ti killer,” based on previous leaks. This newest leak, if true, will confirm it. It allegedly has a 256-bit memory interface and 16GB of GDDR6 memory, with about a 2.5GHz clock speed. If these rumors prove true, this video card will be 38% faster than the GTX 1080.  

There are a few problems with this leak, however, so take this information with a grain of salt. Firstly, it’s based off a picture of a screen (rather than a screenshot), which can obviously be faked quite easily. Second, 3DMark’s online database doesn’t have a locatable entry on this photo. Third, this is displayed as “GTX 1170,” instead of “generic VGA”, which is what usually happens with early graphics card leaks. 3DMark usually has specs but no name with a graphics card leak this early on. The last red flag is that the clock speed is so high. A 2.5GHz speed out of this would be extremely unlikely. It looks like this new card will be one to watch out for, but as for these specs being true, I’m not holding my breath.

Source: Wccftech




ASRock GPU Roadmap Shown and Lacks New GPU Series

Category: Video Cards
Posted: July 9, 2018 06:53PM
Author: Guest_Jim_*

Earlier this year ASRock entered the video card market by launching its Phantom Gaming series. All of the original cards are based on AMD's RX 500 series of GPUs, but now there are Vega-based Phantom Gaming cards too. A roadmap for future ASRock graphics cards has appeared and it demonstrates a continued commitment to staying in the market with a MK2 series launching apparently in August. Unfortunately what are not shown on the roadmap are MK2 versions of the Vega cards or any mention of a new GPU series from AMD, such as a potential RX 600 series.

The roadmap does stretch to around February or March, so this might be an indication that AMD is not launching anything new until then, or it might be the roadmap is not complete. There is also no mention of any MK2 Vega cards, but then these cards only just launched, so it may be premature to tease second versions. It may also be premature for a newcomer to the graphics card market to reveal a GPU manufacturer's unannounced products, so it is a little hard to interpret the significance of this information.

Source: Videocardz



Potential Prototype NVIDIA GeForce PCB Appears in Pictures

Category: Video Cards
Posted: June 26, 2018 06:16PM
Author: Guest_Jim_*

In the ongoing series of rumors and speculation concerning the next series of GeForce GPUs from NVIDIA, we now have what looks like a prototype PCB for an NVIDIA graphics card. This prototype has a number of interesting features on it that might be present because it is a prototype or are indicative of what we can expect in a release product. On the list of likely to be indicative is the 12 GB of GDDR6 memory, and according to WCCFtech the Micron spec sheet indicates these will enable 14 GB/s over a 384-bit memory bus. Making the list as potentially just a prototype feature is the three 8-pin PCIe power connectors. Each 8-pin connector is specified to provide up to 150 W, so that is a potential 450 W for the GPU, memory, and board itself before going out of the spec (which is quite possible). It is possible this much power is only available to open up overclocking headroom, to see what the card will be capable of and is not how a final product will ship.

A curious feature present on the prototype board is the inclusion of NVLINK 2.0 connector. Thus far this connector has not been used on a consumer part, though it does offer more bandwidth than SLI connectors. Also, with both DirectX 12 and Vulkan supporting explicit multi-adapter for allowing multiple GPUs to work together without relying on SLI or AMD Crossfire (and even allowing both NVIDIA and AMD GPUs to work together) it would seem the future of SLI could be in question. It is possible the connector is there just because it is present on NVIDIA prototype PCBs and means nothing more than that. The image also shows four fans covering the VRMs so at least this design expects them to be handling a lot of power, but again, that might tell us more about the prototype than any future product.

Unfortunately the GPU itself is not installed, so we do not get a look at that. It will probably be in the next few months before the next GeForce series is released and by then many of our questions will likely be answered.

Source: WCCFtech



Industry Watchers Suggest NVIDIA May Launch New GeForce Series Late Q3

Category: Video Cards
Posted: June 22, 2018 09:10AM
Author: Guest_Jim_*

In a Digitimes article covering the gaming notebook sector, the news outlet cites industry watches expecting NVIDIA to launch its next GPU series late in the third quarter. There have been a lot of rumors and speculation concerning when these highly expected GPUs will finally launch, with a number of suggested launch and announcement dates already behind us. This new expectation is based on the belief there will be significant increase in demand for gaming notebooks in H2 from gamers wanting to upgrade their systems following the release of the new GeForce cards.

Technically this is still a rumor and WCCFtech does point out one scenario that is probably not what any gamer would want to see; a staggered launch. It is possible, if highly unlikely, for NVIDIA to launch these expected mobile GPUs first with the desktop parts following later. Also something to consider is the reporting that NVIDIA has an oversupply issue, with one Taiwan OEM actually sending back 300,000 GPUs. NVIDIA would likely not want to diminish the interest in these unsold parts with a new series launching, but with demand too low to sell them and significant interest in the next series, that might be exactly what it needs to do.

Source: Digitimes and WCCFtech



NVIDIA Apparently Having Over-Supply Issue

Category: Video Cards
Posted: June 20, 2018 08:54AM
Author: Guest_Jim_*

With the launch of the next series of GeForce GPUs still not on any public calendar, a lot of people have been asking questions as to why. According to some sources, it looks like a reason has to do with the current supply of graphics cards from NVIDIA, specifically the company made too many. With demand for GPUs being so high of late, from both gamers and cryptocurrency miners, NVIDIA apparently decided to increase manufacturing to satisfy the market. The problem is that demand from miners has been dropping recently, so fewer units are moving. While this has allowed pricing for a number of graphics cards to come closer to MSRP lately, it means there are storage facilities having to hold these unsold GPUs, and NVIDIA does not want to launch something new while sitting on product.

To give you an idea of just how severe this issue is, Seeking Alpha points out NVIDIA typically has significant influence over its partners, but one of the top three Taiwan OEMs still sent back 300,000 GPUs to the company. Additionally, this is the result of demand decreasing, and not the used-parts market being flooded with parts. NVIDIA has been buying up GDDR5 lately, too, which could be used to assemble boards with the unsold GPUs.

While there are still reasons to believe NVIDIA will launch its next GeForce series at least before the end of the year, if not in the coming months, this news does demonstrate how complicated launching a product can be. Something more to consider is how this might impact NVIDIA in the future, with how it deals with supply at times of potentially unstable demand.

Source: Seeking Alpha



Cooling Fan Manufacturer Expects Shipments to Increase in H2 2018

Category: Video Cards
Posted: June 19, 2018 01:46PM
Author: Guest_Jim_*

If you are looking to upgrade your GPU but want to wait for the next series from AMD or NVIDIA before making a purchase, Power Logic just shared some information you might be interested in. Even if you are not familiar with Power Logic's name, you might know their products as they make cooling fans for computers, including those for graphics cards. The company is expecting a significant increase in orders in the second half of this year, which would be fed by people purchasing newly released GPUs.

Unfortunately H2 2018 is not very specific so we still do not know when we can expect new graphics cards, except by the end of the year. Still, this indicates something is coming in the months ahead.

Source: DigiTimes



AMD RTG Senior Vice President of Engineering Pledges New Graphics Products Yearly

Category: Video Cards
Posted: June 9, 2018 08:34AM
Author: Guest_Jim_*

Computex 2018 has ended now and neither AMD nor NVIDIA provided any specific information on whatever their next consumer-targeting GPU will be. NVIDIA even said it will be a "long time" before its next series and while AMD did show off a new GPU, it was the 7 nm Vega 20 going to Radeon Instinct compute cards. AMD President and CEO Lisa Su did state though that 7 nm gaming GPUs are coming, but did not say when or if she was referring to Vega or potentially Navi, the following architecture.

None of that is exactly good news if you are looking forward to an upgrade over what is currently available, but David Wang, the Senior Vice President of Engineering at AMD's Radeon Technology Group made a pledge that is encouraging. The promise is that AMD will be releasing new graphics products yearly, with these products representing a new architecture, process change, or "maybe incremental architecture changes." This plan is to provide some kind of consistency and continuing to stimulate the excitement for something new coming.

Wang also stated that he feels "AMD had lost momentum" in the consumer markets as effort was put into chasing after AI. Perhaps ironically, the Vega 20 Radeon Instinct GPU shown off at Computex 2018 is actually targeting AI workloads. Still, this statement shows an interest in returning focus to the gaming-side of GPUs we all enjoy.

Source: PCWorld



Many NVIDIA GPUs May Need Firmware to Support Latest DisplayPort 1.3/1.4 Displays

Category: Video Cards
Posted: June 8, 2018 06:35PM
Author: Guest_Jim_*

If you are looking at a new monitor that uses either DisplayPort 1.3 or 1.4 and are using an NVIDIA GPU, you might need to update its firmware. Apparently many NVIDIA GPUs need the update to address issues of blank screens on boot, or a hang on booting up the system. The company is providing a firmware update tool that will detect if the update is needed and allow the user to install it, which you can find at the source link below. If you are experiencing the issues mentioned, NVIDIA also has some possible workarounds to try until you get the update installed. First, try using a DVI or HDMI connection, then try a different monitor if you have one. After that NVIDIA recommends changing the boot mode from UEFI to Legacy, or the reverse, and finally trying a different graphics device, like a secondary or integrated GPU so you can get Windows and install the firmware.

The GPUs supported by this firmware update include the Titan X (Maxwell), Titan X (Pascal), and Titan XP, the GeForce GT 1030 up through the GTX 1080 Ti, the GTX 950 through the GTX 980 Ti, and the GTX 745, GTX 750, and GTX 750 Ti.

Source: NVIDIA



NVIDIA CEO States Next GeForce Series Not Coming Soon

Category: Video Cards
Posted: June 4, 2018 02:46PM
Author: Guest_Jim_*

This is probably not the news some of you were hoping to hear coming out of Computex, but Jensen Huang, NVIDIA's CEO, stated the next series of GeForce GPUs will not be coming soon. In fact he says it's a long time from now, but made no definite statement. It looks like we will be with the GeForce 1000 series for a good bit longer, but at least it appears we will be hearing about what comes next in August. These statements were made at an invite-only event.

Source: VideoCardz



NVIDIA Will Discuss Its Next Mainstream GPU at HotChips

Category: Video Cards
Posted: June 1, 2018 10:03AM
Author: Guest_Jim_*

This year's HotChips symposium on high-performance chips will be held from August 19 to August 21, and now we know on August 20 NVIDIA will be discussing its next generation mainstream GPU. While this does not necessarily tell us anything about when the highly anticipated GeForce GTX 1100 series will launch, at least we know we will learn more about its architecture and capabilities on that day. Currently, the dominant speculation is it will be based on the Turing architecture and be manufactured on the same 12 nm process as Volta. Beyond that everything is rumor and thus not necessarily reliable. On the bright side, it is entirely possible the GTX 1100 series will launch before HotChips 2018, but only time will tell.

Source: Phoronix



PowerColor to Launch RX Vega 56 Nano Edition at Computex

Category: Video Cards
Posted: May 15, 2018 07:34AM
Author: Guest_Jim_*

For anyone looking forward to a Vega Nano graphics card, it looks like PowerColor has heard you. Last month an image appeared showing what appeared to be such a graphics cards card from PowerColor, and now PowerColor has apparently confirmed it will be bringing out an RX Vega 56 Nano Edition graphics card for those wanting a powerful GPU in a short package.

As the name of the product suggests, this is a special version of the RX Vega 56 graphics cards already available, featuring a shorter PCB and single-fan cooling system. It is not the RX Vega Nano AMD had shown off at SIGGRAPH 2017, sadly, but then this will apparently be launching at Computex 2018 next month. There has been no official word on an AMD RX Vega Nano product since it was first shown off last year, but perhaps if PowerColor is successful with this product, other companies will start making similar units.

Source: VideoCardz



GTX 1170 Specs Possibly Leaked

Category: Video Cards
Posted: May 14, 2018 09:26AM
Author: Guest_Jim_*

Last week an entry for the GTX 1180 appeared in the TechPowerUp GPU database and now we have some rumored specs for the GTX 1170. According to WCCFtech, it will feature 2688 CUDA cores with a 256-bit GDDR6 bus connected to 8 GB or 16 GB of the memory. This memory might come in slower than the GTX 1180, at only 12 Gbps. The core clock is supposed to be around 1.5 GHz but boosts up to 1.8 GHz with what is expected to be a TDP of 140 W to 160 W. Potentially this combination of specifications would have it come in just behind the GTX 1080 Ti, though with a much lower TDP. The MSRP might be around $499, which would put it above the price of the GTX 1070.

WCCFtech also states the GPU in the GTX 1170 is a GT104, which suggested it is the 104 class GPU on the Turing architecture. (The current GTX 1080 and GTX 1070 use the GP104 GPU.) While this is in line with previous rumors concerning the Turing and Ampere architectures, it does differ from the TechPowerUp entry for the GTX 1180, which claims it will be using a GV104, meaning Volta. Supposedly we can expect to see these two GTX 1100 series cards launching in July, so should have all of that cleared up in the coming months, if at least that part of the rumor is accurate.

Source: WCCFtech



GeForce GTX 1180 Appears in TechPowerUp GPU Database

Category: Video Cards
Posted: May 7, 2018 12:48PM
Author: Guest_Jim_*

There has been much speculation about the next series of graphics cards from NVIDIA for quite some time, and now one of them seems to have appeared in the TechPowerUp GPU database, along with some specifications. It is the GeForce GTX 1180 now listed, and according to TPU it uses a GV104 GPU, which means it is derived from the Volta architecture, and was manufactured on a 12 nm process. While the step from the 16 nm process of the current Pascal-based GPUs found in the GTX 1000 series to 12 nm is not as sizeable as going from 28 nm to 16 nm (the leap from Maxwell, the GTX 900 series to Pascal), this will likely still result in improved performance and/or efficiency. The GPU has 3584 shading units, 224 TMUs, 64 ROPs, and an SM count of 28. The GPU is matched with 16 GB of GDDR6 memory through a 256 bit memory bus, so we have a bandwidth of 384 GB/s.

It is worth noting this would be an engineering sample, so some of the specifications may change, including the clock speeds. According to the database the base clock is 1405 MHz with a boost clock of 1582 MHz. The graphics card has 1 HDMI and 3 DisplayPort outputs and will use one 6 pin and one 8 pin PCIe power connector. While that would add up to a potential 300 W of power (the PCIe slot specification allows it to provide 75 W, as does the 6 pin connector, and the 8 pin specification is for 150 W) the TDP is only 200 W. This should provide plenty of room for overclocking, and for comparison the GTX 1080 has a TDP of 180 W, which is 20 W lower with only a single 8 pin connector is limited to 225 W of power, without going past the slot and power connector specifications (which many modern, quality components can safely, as they are overdesigned).

Source: TechPowerUp



ASRock Decided to Restrict European Graphics Card Sales, Not Banned by AMD

Category: Video Cards
Posted: May 7, 2018 09:34AM
Author: Guest_Jim_*

It looks like the news from last week about ASRock's Phantom Gaming graphics cards not being for sale in Europe was not entirely accurate. That news came from the German Tom's Hardware review of one of the cards, when a company representative inquired as to how the review unit was gotten, and then saying it was AMD that told ASRock to not sell the cards in Europe. Since then there has been a lot of speculation as to why, but Jason Evangelho, a contributor at Forbes, decided to reach out to ASRock and got a different statement.

According to this other source, which only discussed intentions for Europe, so this does not necessarily have any bearing on the North America or Asian markets, ASRock decided to not sell its new Phantom Gaming graphics cards in Europe in a way that would get them to normal consumers. They will not appear in retail shops, online or physical. The reason is because ASRock originally wanted to enter the graphics card market to take advantage of the cryptocurrency mining boom, and so in Europe the company is intending to only sell them to miners and for industrial use. The minimum order quantity is 500 units.

This new statement contradicts those from the Tom's Hardware Germany site, but the Forbes contributor says he confirmed it twice over the phone and once over email. Perhaps if interest in cryptocurrency declines enough, or if the supply of cards is enough we will see Phantom Gaming cards appear in European retailers, but for now it appears that is not going to be the case, and not because of AMD's influence.

Source: Forbes



ASRock Blocked from Selling Phantom Gaming Graphics Cards in Europe By AMD

Category: Video Cards
Posted: May 4, 2018 05:30PM
Author: Guest_Jim_*

At the end of March, ASRock entered the graphics card market with its Phantom Gaming cards, which use AMD's Polaris GPU. Unfortunately if you live in the European market, the company is not allowed to sell them there because AMD has not agreed to allow them to, according to an ASRock Sales Manager. This was discovered after ASRock reached out to Tom's Hardware to learn how it acquired its review sample for the review posted on the German Tom's Hardware site. If the product cannot be sold in Europe, the company has no interest in marketing it there either.

It appears what is going on is an example of regional sales bans, which are not uncommon. As VideoCardz points out, Colorful or Colorfire graphics cards are also not found in Europe because of such bans that are to avoid an oversaturation of board partners in a region. Of course someone in Europe could still choose to import ASRock graphics cards, but would need to pay the additional taxes and shipping costs associated with doing so.

It is a potentially important bit of context here that ASRock is a relative late comer to the graphics card market. Perhaps with the next series of AMD GPUs ASRock will be able to negotiate and enter Europe, assuming it wishes to continue a presence in the graphics card market and is involved when that series launches.

Source: VideoCardz



AMD Vega 20 Results Appear for 3DMark11

Category: Video Cards
Posted: April 30, 2018 08:06AM
Author: Guest_Jim_*

Last week AMD shared it had a 7 nm Radeon Instinct card in its lab, and apparently someone there decided to run 3DMark11 with it. Compared to the 14 nm Vega Frontier Edition, it does lose with a graphics score of 26,323, compared to 28,938, but it is also an engineering sample with early drivers, so if/when we see new results, the scores will most likely increase. The clock speed of the GPU is likely incorrect in the 3DMark11 results, but the 32 GB of memory aligns with rumors and expectations for the GPU.

The Radeon Instinct GPU is targeting machine learning applications with its Vega 20 core, so it is currently unlikely we will see it deployed in a gaming graphics card. Being a 7 nm part as well, it will only be manufactured in limited quantities this year, and by the time next year it could be ready for mass production, AMD might move on to the Navi GPU architecture.

Source: VideoCardz



PowerColor Prototype RX Vega Nano Spotted

Category: Video Cards
Posted: April 23, 2018 09:16AM
Author: Guest_Jim_*

At SIGGRAPH 2017 AMD showed off many things, but one that has not materialized yet has been holding the interest of at least some people. The RX Vega Nano is the apparent successor of the R9 Fury Nano and would feature a Vega GPU with HBM2 in a small graphics card, but after being held up for the world to see, there has been nothing. Recently someone was able to snap pictures of prototype RX Vega Nano made by PowerColor at a launch event for the Ryzen 2000 series of CPUs in Munich. There is no word on if this will ever become a consumer product, but it is still interesting to see the short shroud handing just past the length of the PCB with a single fan for cooling. One of the changes from what was seen at SIGGRAPH was the addition of a 6 pin PCIe power connector alongside an 8 pin.

Source: Tech Power Up



NVIDIA Also Ending 32-Bit OS Drivers

Category: Video Cards
Posted: April 9, 2018 10:47AM
Author: Guest_Jim_*

Almost forgot this was announced along with the ending of Fermi-GPU driver support; NVIDIA will cease releasing drivers for 32-bit operating system this month, though like with Fermi-GPUs, critical security updates will be delivered through January 2019. This will impact Windows 7, 8/8.1, 10, Linux, and FreeBSD, so if you are running any of these systems and are still using the 32-bit version, your NVIDIA driver support will be ending soon. Many modern games already require 64-bit operating systems though, and you should be installing the 64-bit drivers if that is what you are running, so the number of impacted gamers should hopefully be small. Still, if you are running a 32-bit system with a GTX 400 or 500 series GPU, you might need to look into upgrading if you want to continue receiving driver upgrades.

Source: NVIDIA



Random Pic
© 2001-2018 Overclockers Club ® Privacy Policy
Elapsed: 1.2377350330   (xlweb1)