Category: Hardware

Availability for the RTX 3000 cards is still up in the air, and AIB partners are jacking up prices whenever possible, yet Nvidia seems fine with releasing even more models that are designed to compete with AMD’s new RX 6000 cards. It is not like AMD is doing any better as far as availability is concerned, but Nvidia finally realized that the gap between RTX 3080 and RTX 3090 is a bit too wide, so an RTX 3080 Ti model sporting 20 GB VRAM is expected to launch early next year as a competitor for AMD’s RX 6900 XT. Previous rumors were suggesting that this model could be announced at CES 2021 in mid-January. According to new reports coming from Videocardz and Igor’s Lab, however, this does not seem to be the case anymore. Apparently, Nvidia is considering rescheduling the RTX 3080 Ti for a mid to late February launch, and instead rush the RTX 3060 (non-Ti) models for an early January launch.

These schedule changes are most likely caused by the production problems with the Samsung 8 nm nodes already acknowledged by Nvidia. The high-end SKUs seem to be more affected in this regard due to the die size, whereas the RTX 3070 and 3060 Ti models that integrate cut-down SKUs appear to get better yields as suggested by the slightly higher stocks.

When it comes to specs, the RTX 3080 Ti is rumored to feature the same number of CUDA cores as the RTX 3090 plus 20 GB of GDDRX6 memory for only $999 MSRP. It certainly looks like Nvidia could phase out the RTX 3090 with this move. Still, if the production problems still continue to be an issue well into 2021, those MSRPs could prove to be wishful thinking.

As for the non-TI versions of the RTX 3060 cards, Igor’s Labs and VIdeocardz are reporting that there will be 12 GB and 6 GB versions, each with different CUDA cores on top, to confuse the buyers even more. Nvidia did something similar a few years ago with the GTX 1060 cards that had 3 GB and 6 GB versions. We are not exactly sure why Nvidia is adding 12 GB of VRAM on the RTX 3060. Team green probably thinks that this could boost sales, but, of course, it all heavily relies on availability and pricing. The 12 GB version is supposed to launch first some time in mid-January, while the 6 GB version could possibly launch early February.

Igor’s Lab also reports that an RTX 3050 card based on an even more cut-down GA107 SKU with only 4 GB of RAM is to be expected at some point beyond February 2021. Additionally, word on the street is that Nvidia intended to launch an RTX 3050 Ti version, but rebranded it as the RTX 3060 6 GB meanwhile.

It has been a few weeks since Patrick Schur published information about the mobile variants of the RX 6700 series, but the software engineer has offered yet more insights into the upcoming RDNA 2 cards. According to Schur, Navi 22 GPUs will have 18.31 x 18.27 mm dies, which equates to a size of 334.52 mm². As Wccftech points out, a die size of 334.52 mm² would be over a third larger than the die of the RX 5700 XT.

Additionally, Navi 21 GPUs will offer up to a 192-bit memory interface split across six chips. Apparently, SKUs with a 192-bit bus width will also have a target TGP of 146 W, although Schur has previously claimed that the Navi 22 XT GPU has a TGP of 186-211 W. By contrast, Navi 22 SKUs with 90, 110 and 135 W TGPs will be limited to a 160-bit bus on five memory chips.

AMD supposedly plans to release Navi 23 and Navi 24 mobile GPUs too, which would make Navi 22 the top-end SKUs of the RX 6000M product stack. Navi 22 GPUs are also expected to offer up to 2,560 cores, 40 Compute Units (CUs) and up to 12 GB of VRAM. RX 6000M laptops should begin arriving in the first half of 2021, just in time to challenge NVIDIA’s RTX 30 series options.

So, NVIDIA announced the RTX 3060 Ti yesterday, but a blog post by the company has also confirmed that more RTX 3060 cards are on the way. There have been rumours that NVIDIA plans to release an RTX 3060 and RTX 3060 mobile early next year, although the company is yet to state that this is the case.

Nonetheless, NVIDIA describes the RTX 3060 Ti as being ‘the first member of the RTX 3060 family’, so it definitely has subsequent RTX 3060 releases planned. The same blog post also contains the benchmark chart that Videocardz published in mid-November, confirming the latter’s authenticity. NVIDIA leans heavily on the claim that the RTX 3060 Ti outperforms the RTX 2080 SUPER, although the chart makes a point of comparing the Ampere card with the RTX 2060 SUPER too, which one would hope is the case. We have included the chart published by Videocardz also as this includes data labels, which the official chart lacks.

Concrete details on the RTX 3060 are scarce for now, although Videocardz claims that NVIDIA will base the card on its GA106-400 GPU. Moreover, RTX 3060 is thought to feature 3,840 CUDA cores and will be offered with 6 GB and 12 GB of GDDR6 VRAM. Nothing is known about the mobile version, or versions of the RTX 3060, though. Wccftech reports that the graphics card will be available with AMD Ryzen 5000H and Intel Tiger Lake-H series processors, but no details about the RTX 3060 Max-P or Max-Q have been leaked so far.

A new Geekbench 5 entry shows that Vivo is working on a smartphone with the model number ‘V2056A’. According to the listing, the V2056A has a next-generation Qualcomm chipset. Based on previous leaks, it seems that the V2056A has a Snapdragon 888, which Qualcomm announced yesterday. There are now several listings on Geekbench with ‘lahaina’ motherboards, of which the V2056A is one. Additionally, the chipset can be identified as the ‘ARM implementer 65 architecture 8 variant 1 part 3396 revision 0’.

Geekbench also tells us that the V2056A has 12 GB of RAM and is running Android 11, which one should expect from a new flagship smartphone. The listing offers an insight into the performance of the Snapdragon 888 too, albeit only a limited one at this stage. The new chipset appears to have a 25% faster single-core performance than the Snapdragon 865, for example, along with an 11% advantage in multi-core tasks.

According to Geekbench, at least Vivo, OnePlus and Samsung have Snapdragon 888-powered smartphones in development. Realme and Xiaomi have confirmed that they do too, as has Motorola. Qualcomm has confirmed that 13 OEMs have already pledged to release smartphones containing the Snapdragon 888 though, as we have also reported.

Fans and enthusiasts have been waiting with bated breath for Qualcomm’s flagship SoC that will power most 2021 flagship smartphones. Standard naming conventions dictated that it would be called the Qualcomm Snapdragon 875. However, Qualcomm has somewhat of a surprise in store for us. The chipmaker has just announced that it will be called the Qualcomm Snapdragon 888 instead. While the name might seem a bit off, it makes sense, considering that the digits are considered lucky in some cultures.

Earlier leaks told us that the Qualcomm Snapdragon 888 would reuse its predecessor’s 1+3+4 core configuration. We should know more about it soon, as Qualcomm hasn’t divulged any information about the chip’s innards. However, we know for sure that the chipset will come with Qualcomm’s shiny new X60 5G modem that will support both sub-6GHz and mmWave networks. Other 5G-related features include support for both standalone and non-standalone (SA and NSA) networks, multiple SIM cards, and more.

The Qualcomm Snapdragon 888 also comes with a new AI engine that can perform up to 26 trillion operations per second (TOPS). It will go a long way towards improving a heap of camera-centric features like AI-driven portrait mode, colour correction, and more. The new NPU will undoubtedly help OEMs fine-tune on-device optimization features, such as optimizing battery usage, adjusting screen refresh rate, etc.

On the GPU side of things, the Qualcomm Snapdragon 888 is expected to run the Adreno 660 (tentative) GPU. Its specifics, much like the chip’s core configuration, will be revealed later. However, Qualcomm claims that the Snapdragon 888’s image signal processor (ISP) is up to 35% faster than its predecessor. The Qualcomm Snapdragon 888 will power a host of 2021 flagships from companies like OnePlus, ASUS, Sony, Realme, Oppo, Vivo and more. The chipset is expected to debut alongside the Xiaomi Mi 11, which may be launched soon.

The GeForce RTX 3060 Ti may be the cheapest RTX 30 series card, at least until it releases the RTX 3050 and RTX 3060, but custom cards will not reflect the price of the Founders Edition card. NVIDIA states that the Founders Edition will cost £369, €399 and US$399 when it launches tomorrow, making it significantly cheaper than the likes of the RTX 3070 and RTX 3080. However, leaked prices of custom RTX 3060 Ti cards are significantly higher than any of those prices.

According to listings on Dateks, NVIDIA’s board partners will charge anywhere from €484 to €664 for custom versions of the RTX 3060 Ti. These do not appear to be placeholders either, so most RTX 3060 Ti cards will cost between 20% and 66% more than the Founders Edition at launch. Prices will undoubtedly fall over time, but the scarce volumes of Founders Edition stock that NVIDIA releases makes its MSRPs somewhat meaningless when its partners ignore them.

Similarly, TechSpot spotted Amazon UK charging £490.18 for an MSI GeForce RTX 3060 Ti Ventus 2X OC. Like in the Eurozone, this price is over a third higher than NVIDIA will be charging for the Founders Edition card in the UK. It is a comparable story in the US, too. While the Founders Edition will be available for US$399, the cheapest custom card is US$449.99, or 12% more expensive. Based on the multiple retailer listings, custom RTX 3060 Ti cards should be orderable from tomorrow.

So, AMD has allowed its board partners to release Radeon RX 6900 XT cards, after all. There has been no official announcement from AMD on the matter, but the arrival of an RX 6900 XT by ASRock confirms that this is the case.

ASRock’s card is nothing more than a reference design, though. The company has not even put any of its branding on its RX 6900 XT, making it plain that this nothing more than the card that AMD has already announced. Unsurprisingly, ASRock has called its card the ‘Radeon RX 6900 XT 16G’.

The ASRock card also bears identical specifications to AMD’s reference card. So, that means that it has a 2,015 MHz game clock, a 2,250 MHz boost clock and a 300 W total board power.

ASRock has not stated how much its RX 6900 XT will cost. Presumably, it will cost approximately US$999 though, given how much AMD has marketed the price of the RX 6900 XT in comparison to NVIDIA’s more expensive GeForce RTX 3090.

It remains unclear whether we will see any custom RX 6900 XT designs, nor which other board partners will release cards. According to Videocardz, AMD’s partners should announce any custom cards after December 8, if there are any.

Rumours about the RX 6700 series have reared their head again. This time the rumour stems from Patrick Schur, a software engineer and active leaker on Twitter. Apparently, AMD has two SKUs planned, with the RX 6700 and RX 6700 XT set to join the RX 6800, RX 6800 XT and RX 6900 XT. Based on a recent EEC listing, AMD may be planning to release an RX 6900, too.

Schur states that the RX 6700 cards will use AMD’s Navi 22 GPUs. Both will be paired with 12 GB of GDDR6 VRAM, which is unsurprising considering that NVIDIA still has exclusivity over the newer GDDR6X variant. This correlates with a similar leak by @_rogame in September. By comparison, the RX 6800 series has 16 GB of GDDR6 VRAM, as does the RX 6900 XT.

Apparently, AMD will base the RX 6700 XT on the Navi 22 XT GPU, with the Navi 22 XTL used for the RX 6700. Additionally, AMD has apparently targeted a total graphics power (TGP) of between 186 W and 211 W for the RX 6700 XT, which would make it a lot less power-hungry than the RX 6800 series.

Meanwhile, the RX 6700 will have an upper target of 156 W. Schur states the card should have a minimum TGP of 146 W too, but adds that its minimum may be even lower. However, he is yet to see a Navi 22 XTL GPU with a TGP below 146 W. According to Videocardz, AMD plans to announce the RX 6700 series in January 2021, which is the same month as several NVIDIA RTX 3000 series launches. There is no word on pricing yet, though.

Tiger Lake-H has reared its head on UserBenchmark. The follow-up to Comet Lake-H is yet to be unveiled, but Tiger Lake-H is expected to mark the move from 14 nm to 10 nm for Intel’s 45 W processors. The laptop on UserBenchmark has benchmarked as ‘Insyde TigerLake’, denoting that it is probably a prototype of sorts. The machine has also been equipped with 20 GB of DDR4-3,200 RAM, an unusual combination for laptops. Additionally, UserBenchmark reports a base clock of 3.1 GHz and an average boost clock of 2.75 GHz, more indicators that this is a prototype or an engineering sample.

The entry indicates that the Tiger Lake-H processor supports Hyper-Threading, which is unsurprising. The processor also has eight cores, which Intel confirmed would be the maximum core count for the Tiger Lake-H series. Rumour has it that Intel will offer Tiger Lake-H processors with four and six cores, as it already does with the Comet Lake-H series. 

However, Tiger Lake-H processors will supposedly come with TDPs ranging from 35 W to 65 W, with a 45 W TDP offered for six-core SKUs. Similarly, Intel will utilise multiple iGPUs for the Tiger Lake-H series. All SKUs will have Iris Xe GPUs, but 35 W SKUs will be stuck with 32 Execution Units (EUs). By contrast, 45 W and 65 W processors will have 96 EUs.

Intel Tiger Lake-H-based notebooks are expected in the first half of 2021. However, it is unclear when Intel will unveil the Tiger Lake-H architecture.

Sapphire Tech has quite a few video cards based on the Radeon RX 6800 and Radeon RX 6800 XT, with most of them going above and beyond the specs and design of the reference models introduced by AMD. Now, three such video cards can be pre-ordered from Finnish retailer Multitronic, all of them being overclocked variants with custom cooling solutions.

The three models listed by the aforementioned retailer and their prices are the following:

Sapphire Radeon RX 6800 XT Nitro+ OC SE, 863.9 Euros (roughly US$1,020)

Sapphire Radeon RX 6800 XT Nitro+ OC Gaming, 838.90 Euros (around US$990)

Sapphire Radeon RX 6800 Nitro+ Gaming OC, 732.90 Euros (approximately US$870)

Before turning red because these prices seem outrageous, our non-EU readers should keep in mind that these are preliminary prices and there is also a rather large VAT included (Finland has a standard VAT rate of 24%, which applies to computer parts as well). Unfortunately, there is no shipment date mentioned in the listings, but it will probably take at least one or two weeks after the launch scheduled by AMD and some of its partners to take place on December 8.