Posted on Monday, October 26 2020 @ 11:11:21 CET by Thomas De Maesschalck
TechPowerUp noticed Intel chief architect Raja Koduri will speak at the Samsung Foundry SAFE Forum. At this virtual-only event, Koduri will host a presentation titled "1000X More Compute for AI by 2025."
We still don't know details about the deal closed between Intel and Samsung, but there have been rumors and speculation for quite some time now that Intel will tap Samsung's production capacity for Xe. In the past, Intel did everything in-house, but these days the chip giant sees itself forced to look outside as not only its 10nm process but also its 7nm node is in trouble.
Intel has a contract to develop an exascale supercomputer at Argonne National Laboratory, called Aurora. That supercomputer is using Intel's CPUs and the company's upcoming Xe GPUs. Since the company has problems with manufacturing and has to deliver the products (it is bound by several contracts) to its contractors and customers, it decided to look at external manufacturers for its products, specifically Xe graphics. Being that Mr. Koduri tweeted an image of him visiting Samsung Giheung Fab in Korea, and now presenting at the Samsung Foundry event, it is possible that Intel will tap Samsung's semiconductor manufacturing process for its Xe GPU efforts and that Samsung will be the contractor in charge.
Posted on Monday, October 26 2020 @ 11:05:05 CET by Thomas De Maesschalck
Besides process technology, foundries like TSMC are also innovating with packaging technology.
TSMC has reportedly started production with its 6th generation Chip-on-Wafer-on-Substrate (CoWoS) packaging technology. According to media reports, the new technology enables the placement of 12 stacks of HBM on a single package.
The new generation is said to enable a massive 12 stacks of HBM memory on a package. You are reading that right. Imagine if each stack would be an HBM2E variant with 16 GB capacity that would be 192 GB of memory on the die present. Of course, that would be a very expensive chip to manufacture, however, it is just a showcase of what the technology could achieve.
Posted on Monday, October 26 2020 @ 10:54:57 CET by Thomas De Maesschalck
Following the installation of Windows 10 May 2020 Update, various users have been unable to use the Windows 10 Reset feature. When attempting to use "Reset this PC", the feature fails to work and returns the following error message: "There was a problem when resetting your PC. No changes were made."
1. Open an elevated command prompt. To do this, click Start, type Command Prompt or cmd in the Search box, right-click Command Prompt, and then click Run as administrator. If you are prompted for an administrator password or for a confirmation, type the password, or click Allow.
2. Type the following command, and then press Enter:
dism /online /cleanup-image /restorehealth
3. Reboot your system, and attempt Reset this PC again.
A patch is expected to follow but Microsoft didn't provide a timeline.
Posted on Monday, October 26 2020 @ 10:48:45 CET by Thomas De Maesschalck
It's hard to believe The Elder Scrolls V: Skyrim is already nine years old! The game was extremely popular and many people are looking forward to The Elder Scrolls VI. The sequel to Skyrim was announced in 2018, with Bethesda saying it would be released after the launch of Starfield.
The developer said the next Elder Scrolls game will be the biggest engine overhaul in history. Photogrammetry will be used to deliver photorealistic game assets. As this is taking way to long, a fan decided to create a trailer of how "The Elder Scrolls VI: Hammerfell" should look like.
You can watch the concept below, it definitely looks impressive.
Posted on Monday, October 26 2020 @ 10:39:53 CET by Thomas De Maesschalck
Two more days to go until the reveal of AMD's Big Navi GPUs. In the meantime, there's a fresh rumor that the ASUS ROG STRIX Radeon RX 6800 XT will be able to offer a Boost frequency of over 2.5GHz. This card reportedly has a TGP of 290W. This is expected to be one of the fastest custom models.
The information is based on engineering board samples so final clocks and TGPs may vary but it is interesting just how far AIBs have managed to push AMD's Big Navi GPUs, even if internally. ASUS's ROG STRIX Radeon RX 6800 XT is the highlighted model today & there are three systems mentioned by Patrick that have been tested within 3DMark 11, all with variable clock speeds. The RX 6800 XT is reportedly running with a set TGP of 290W and a 1000 MHz memory clock which represents 16.00 Gbps memory dies.
Posted on Monday, October 26 2020 @ 10:34:32 CET by Thomas De Maesschalck
The first desktop versions of NVIDIA's Ampere GPU have been available for over a month now, albeit in limited quantity, but so far we haven't heard a lot about the GeForce RTX 30 mobile series. That changes thanks to a new leak spotted by VideoCardz.
It appears the GA102 GPU isn't going to make it to the mobile market, at least not for now:
The GA104 GPU would be used by another two Geforce SKUs: the RTX 3080 Mobile Max-Q and RTX 3070 Max-Q, a fresh leak from Gigglehd seems to suggest. There are surprisingly no Max-P variants for the higher-end GeForce RTX 30 SKUs.
The leak also reveals the first plans for GeForce RTX 3060 Mobile Max-P. Not much is known about this GPU, but if anything we heard about the desktop model is true, then this mobile SKU should also offer a cut-down GA104 GPU.
NVIDIA is expected to introduce its mobile Ampere GPUs in January 2021, around CES 2021. Interestingly, the leak suggests these GPUs will be paired with Comet Lake-H (10th Gen Core) CPUs and not Intel's new Tiger Lake-H CPUs.
Posted on Monday, October 26 2020 @ 10:27:59 CET by Thomas De Maesschalck
A fresh Geekbench 5 score of an system with the new Iris Xe Max GPU hit the web. The score reveals Intel's DG1 GPU is able to beat very low-end discrete graphics like the NVIDIA GeForce MX330. The latter scores 11,162 in Geekbench 5 OpenCL, while the system with the Iris Xe Max was able to post a score of 11,885.
The Iris Xe Max never aimed to compete in the performance bracket; therefore, it isn't a big shocker that the GPU barely beats Nvidia's GeForce MX330. The Geekbench 5 numbers showed that the Iris Xe Max was just 6.5% faster than the GeForce MX330. On the contrary, the GeForce MX350 delivered up to 16.4% higher performance than Intel's offering.
In comparison to the Radeon RX 550X, the two-year old AMD GPU outperformed the Iris Xe Max by up to 6%.
Intel(R) Iris(R) Xe MAX Graphics Compute Units 96 Maximum Frequency 1.50 GHz Device Memory 2.99 GB
Posted on Monday, October 26 2020 @ 10:15:33 CET by Thomas De Maesschalck
Two months ago, we wrote that Facebook is making a Facebook account mandatory for users of its Oculus headsets. Since the new policy went into effect, there have been some complaints from users who tried to merge their Oculus and Facebook account and ended up with an unusable VR headset.
Now news is spreading about another implication of the merger between Facebook and Oculus accounts. Facebook has decided that if you delete a Facebook profile that's linked to an Oculus headset, you will no longer have access to Oculus purchases:
This latest news comes from text Facebook displays when you try to permanently delete your account. "Deleting your Facebook account will also delete your Oculus information," the text warns. "This includes your app purchases and achievements. You will no longer be able to return any apps and will lose any existing store credits."
Important VR PSA: ???? Deactivating your Facebook profile disables your Oculus Profile. ???? Deleting your Facebook account takes away all your games, purchases, and progress. Source: Eli Schwartz pic.twitter.com/dSJIcIf0ki
Posted on Friday, October 23 2020 @ 16:54:21 CEST by Thomas De Maesschalck
The same Twitter leaker that posted a lot of accurate rumors about NVIDIA's Ampere GPUs earlier this year is now claiming that NVIDIA has a GA102-150 GPU with 7424 CUDA cores. The new variant has 1536 more CUDA cores than the GA104-based RTX 3070 but 1280 fewer than the RTX 3080.
The GA102-150 reportedly has a 320-bit memory bus and 10GB GDDR6X memory. Perhaps this model will be launched as the GeForce RTX 3070 Ti. The latter was originally envisioned as a model with a GA104 GPU with all 6144 CUDA cores enabled but it now seems NVIDIA plans to tap the GA102 instead.
As was mentioned yesterday, the RTX 3080 20GB and RTX 3070 16GB are no longer a thing. These higher-capacity cards were expected later this year but they recently got cancelled for unknown reasons.
Posted on Friday, October 23 2020 @ 14:07:55 CEST by Thomas De Maesschalck
During the company's third quarter earnings call, Intel CEO Bob Swan revealed that the Xe DG1 discrete GPU is shipping today and will be present in multiple OEM systems later this quarter. Furthermore, Swan announced that the chip giant now has working samples of the DG2 GPU. While the DG1 is expected to be a low-end solution that's not much more powerful than the integrated graphics found on the Tiger Lake laptop processors, the DG2 promises to take Intel into the enthusiast segment:
Our first discrete GPU DG1 is shipping now and will be in systems for multiple OEMs later in Q4. We also powered on our next-generation GPU for client DG2. Based on our Xe high-performance gaming architecture, this product will take our discrete graphics capability up the stack into the enthusiast segment. Beyond the CPU and the GPU, our customers tell us that they want a diverse range of AI solutions to fit every power level and performance needs from the intelligent edge to the data center.
The DG2 uses the Xe-HPG architeture and is manufactured by an external foundry.