Processors - Page 74

AMD is currently preparing its next-gen Van Gogh line of APUs for ultra-low power notebooks, with new leaks teasing that we are to expect a quad-core APU powered by the Zen 2 architecture as well as an RDNA 2-based GPU.

The news is coming from Moore's Law is Dead, with Tom reporting that AMD's next-gen Van Gogh APUs will be made on the 7nm node by TSMC and will feature a monolithic design packing a CPU, GPU, and I/O IPs -- all ready for ultra-low power notebooks and ready for 2021.

AMD's new Van Gogh APUs will not only have Zen 2 CPU cores and RDNA 2 GPU cores, but it'll also pack LPDDR6 memory which offers more bandwidth -- great for gaming and content creation on-the-cheap (and inside of ultra-low power notebooks of the future). AMD is using the FF3 socket for its new Van Gogh APUs which is an compact low power socket, something that AMD will use to fight against Intel's new Tiger Lake-U processors.

AMD has just launched its new Ryzen 5000G series processors with the introduction of the new Zen 3-based Ryzen 7 5700G, Ryzen 5 5600G, and Ryzen 3 5300G chips.

It starts with the AMD Ryzen 7 5700G which is a new 8-core, 16-thread CPU which clocks at up to 4.6GHz and has 20MB of cache in total. There are 8 compute units at up to 2GHz here, with the CPU using up to 65W of power. Next up is the Ryzen 5 5600G with 6 cores and 12 threads at up to 4.4GHz with 19MB of cache in total, 7 compute units with the GPU at up to 1.9GHz and the same up to 65W power consumption.

AMD's lowest member of the new Ryzen 5000G processors is the new Ryzen 3 5300G which is a new 4-core, 8-thread CPU with clocks of up to 4.2GHz, 10MB of cache and 6 compute units at up to 1.7GHz and up to 65W power consumption.

NVIDIA has been firing on all cylinders at its annual GPU Technology Conference (GTC 2021) where it has unveiled its first-ever CPU, its next-next-gen GPU architectures through to 2025, an insane new AI workstation and also a next-gen Atlan SoC.

The new NVIDIA Atlan SoC is a new AI-enabled chip for autonomous cars of the future, with NVIDIA pushing a huge 1000 trillion operations per second (TOPS). NVIDIA is using its next-gen in-house ARM-based Grace CPU, as well as its next-gen Ampere Next GPU technology for its new Atlan SoC for future autonomous vehicles.

NVIDIA is using the new LPDDR5X memory technology that offers double the bandwidth, and 10x the energy efficiency over the previous-gen LPDDR4X standard. We are told to expect Atlan SoC sampling to begin in 2023, while the chips will come off the production line and into the new waves of autonomous vehicles in 2025 and beyond.

NVIDIA has just announced its first-ever CPU with the introduction of Grace, a new ARM-based processor designed for AI and servers -- and not for your gaming PC.

The new NVIDIA Grace processor has 500GB/sec of memory bandwidth thanks to the use of next-gen LPDDR5X memory, which is attached to the GPU through Cache Coherent NVLink to the GPUs at 900GB/sec, with CPU to CPU communications having up to 600GB/sec of bandwidth.

Huang explained: "Leading-edge AI and data science are pushing today's computer architecture beyond its limits-processing unthinkable amounts of data. Using licensed Arm IP, NVIDIA has designed Grace as a CPU specifically for giant-scale AI and HPC. Coupled with the GPU and DPU, Grace gives us the third foundational technology for computing, and the ability to re-architect the data center to advance AI. NVIDIA is now a three-chip company".

We've been hearing about AMD's next-gen Zen 4 architecture and Ryzen 7000 series or codename Raphael CPUs for over 6 months now, but now we have some more information thanks to new leaks.

A new leaked AMD roadmap has been teased by @sepeuwmjh which teases the Zen 4-based codename Raphael CPUs, and has been tweaked into a detailed roadmap by @Olrak29_ that gives us a great look at what to expect from AMD's upcoming slate of Zen processors.

The new codename Raphael CPUs will reportedly use the Zen 4 CPU architecture, pack on-board RDNA 2-powered graphics, support DDR5 and PCIe 5.0 technologies, and arrive on TSMC's fresh new 5nm node. AMD will be shifting to the new AM5 socket for the Ryzen 7000 series CPUs according to these roadmaps. Before that we have codename Warhol which will reportedly use the Zen 3+ architecture, DDR4 and PCIe 4.0 technology and be made on the 6nm node.

We are finding out some more information on Intel's next-gen Sapphire Rapids architecture, with the next-gen Xeon platform set to pack some juicy specs and technologies when it drops.

The new tease is from VideoCardz with an anonymous tipster sending a slide that shows the specifications of Intel's upcoming Sapphire Rapids platform. We're looking at up to 56 cores per CPU, DDR5 memory support, 64GB of HBM2e memory per socket (!!!), PCIe 5.0 support with up to 80 lanes of PCIe 5.0 goodness, and so much more.

Sapphire Rapids will scale between 1 and 8 sockets, so you will be able to cram in many of those 56-core CPUs into future supercomputers. But the most surprising things here are the 80 PCIe 5.0 lanes and 64GB of HBM2e memory on-package with up to 1TB/sec of memory bandwidth per socket. Delicious.

If you are in the market for a kick ass new monitor + CPU + gaming bundle then you might want to check out this Lenovo 27-inch QHD 165Hz gaming monitor + AMD Ryzen 9 5950X processor + Xbox Game Pass for PC bundle.

The bundle itself runs until April 9 which will match together the Lenovo G27Q 27" QHD (2560 x 1440) IPS 165Hz gaming monitor, a AMD Ryzen 9 5950X processor which offers 16 cores and 32 threads of Zen 3 goodness, as well as a 3-month Xbox Game Pass for PC subscription -- for a total of $1178.98 -- not bad at all.

Considering AMD's own Ryzen 9 5950X isn't the easiest processor to find right now, getting a bundle like this could be your way into Zen 3 quick. Lenovo's own monitor isn't bad at all, a 27-inch 2560 x 1440 monitor with a super-smooth 165Hz refresh rate and 1ms response time.

Intel has finally launched its Rocket Lake-S family of processors with the flagship Core i9-11900K chip being put through its paces, and cooled with LN2 cooling to break some OC world records.

The world's best overclockers have got their hands-on the new Rocket Lake-S processors and pushed the flagship Core i9-11900K past 7GHz and even cranked DDR4 speeds through the roof, pushing DDR4-7156MHz which is simply insane. A bunch of different motherboards were used by different overclockers including the GIGABYTE Z590 AORUS Tachyon, MSI MEG Z590I Unify, ASUS ROG Maximus XIII APEX, and the ASRock Z590 OC Formula motherboards.

ASUS took to their official Twitter to shout from the roof tops that their ROG Maximus XIII Apex motherboard was used for some extreme overclocking sessions with overclocker "Raud" breaking the long-standing 4 minute barrier in the SuperPi 32M test using the ASUS ROG Maximus XIII Apex motherboard with an Intel Core i9-11900K processor and LN2 cooling. He reached an insane 7304MHz on the CPU and secured 10 world records and 7 global first places for ASUS.

Intel is launching its new Rocket Lake-S family of 11th Gen Core processors led by the flagship Core i9-11900K, which when installed into GIGABYTE's mid-range B560 AORUS motherboard can be clocked up to 5.1GHz across all CPU cores.

The new Core i9-11900K was installed into a GIGABYTE B560 AORUS motherboard, with GIGABYTE releasing a new BIOS that enables support and overclocking for Intel's new Rocket Lake-S chips. The new BIOS also enables higher memory overclocking, with GIGABYTE now supporting up to DDR4-5300MHz memory.

This is quite the achievement as Intel's mid-range B-series motherboards used to suck for overclocking, but now GIGABYTE's unleashing of the new BIOS for their B560 series motherboards means that mid-range users can grab a high-end CPU and overclock the crap out of it. Not only that, but unlocked memory XMP overclocking is also here which is another gold star for GIGABYTE and their mid-range motherboard offerings.

We are so damn close to the new Intel 11th Gen Core CPUs being released, with Rocket Lake-S nearly upon us it is time to tease the 13th Gen Core CPUs -- yeah, not the 12th Gen Core CPUs, the new 13th Gen Core CPUs with Intel itself confirming Raptor Lake-S is coming.

VideoCardz leaked out a slide that shows Intel's next-gen Raptor Lake-S is a CPU coming in 2022 and that it would be an upgrade to hybrid technology, pack better CPU cache for gaming and new LPDDR6X memory support. Alder Lake will offer the first 'Performance Hybrid Architecture' with new CPU cores as well as PCIe 5.0 technology.

Intel's 12th Gen Core range with the Alder Lake CPU architecture offers the new Performance Hybrid Architecture with new CPU cores and DDR5 support with up to 4800MHz RAM support, as well as next-gen PCIe 5.0 technology with up to 48 PCIe 5.0 lanes for some seriously insane bandwidth for SSDs, GPUs, and everything in between.

Intel's new Core i9-11900K has been pushed to an insane 7GHz with exotic LN2 cooling and a huge 1.873V running through the new flagship Rocket Lake-S processor.

The new Intel Core i9-11900K processor was used in an ASUS ROG Maximus XIII Apex motherboard, and pushed to its absolute limits with LN2 cooling. The CPU was pushed to 7048MHz with a mind boggling 1.873 volts, with 'ROG Fisher' taking all of the 3 top overclocking records on the CPU-Z validation website.

That's not the only overclocked Core i9-11900K processor we have to talk about today, with Indian YouTube channel PC Wale pushing his Core i9-11900K chip to 6.5GHz on the ASUS ROG Maximus XIII Hero motherboard. PC Wale also used LN2 cooling on the overclocked 11900K processor.

Qualcomm has just unveiled its new Snapdragon 860, which is just the Snapdragon 855 with some makeup on -- it's an overclocked Snapdragon 855 pretty much.

The 'new' Snapdragon 860 from Qualcomm has its Kryo 485 cores at the same speeds as the overclocked Snapdragon 855+ chip but the company has used the 7nm node for its new Snapdragon 860 -- over the use of the new 5nm node for its Snapdragon 888 in order to save costs and make the Snapdragon 860 cheaper.

XDA Developers says the new Qualcomm Snapdragon 860 doesn't just have the same Kryo 485 cores at the same clocks as the Snapdragon 855+ but also the same Adreno 640 GPU. But we do have some new goodies: up to 16GB of RAM is now supported under the new Snapdragon 860, and there's no 5G modem integrated into the Snapdragon 860.

Intel is unleashing its next-gen 11th Gen Core CPUs and while that is all happening, the company decided to unleash its new -- and first public demo of their new 11th Gen Core H-Series processors codenamed Tiger Lake-H.

We should be looking at Intel's new flagship Core i9-11980HK processor, with Intel teasing the 8-core, 16-thread CPU with GPU clocks of up to 5.0GHz. The 5GHz claim was confirmed by VP and GM of Intel, Roger Chandler, during his GDC Showcase 2021 event.

Inside, Intel is using its new Willow Cove architecture and its new 10nm fabrication process -- it is a worthy note here is that these new 11th Gen Core H-Series CPUs are the first non-14nm high-end CPUs from the company. Inside of the laptop on show was the new flagship -- yet named -- Core i9 Tiger Lake-H processor.

Intel has finally unleashed its next-gen 11th Gen Core CPUs, with the company promising unmatched overclocking and game performance out of the new Rocket Lake-S family of processors.

One of the most exciting things about the new 11th Gen Core CPUs is that Intel is debuting its new Cypress Cove architecture, with up to 19% more IPC performance over the current-gen Comet Lake-S offerings. Intel is also introducing PCIe 4.0 technology, faster memory, and more. Rocket Lake-S helps Intel catch up to Zen 2 and Zen 3 from AMD... for now.

Intel's new flagship Core i9-11900K processor is an 8-core, 16-thread CPU at up to 5.3GHz -- down from the 10-core, 20-thread CPU that is the Core i9-10900K. But the IPC uplift from the new Cypress Cove architecture alongside other improvements Intel has made under the hood of the 11th Gen Core CPUs is enough to overpower the lesser core and thread count.

Intel's new Core i9-11700K had its first review posted nearly a month before the CPU officially launches, with our friends at AnandTech posting their review just days ago -- but now it's HardwareLuxx's turn with an early review on the same Core i7-11700K processor.

The new Core i7-11700K processor is an 8-core 16-thread CPU that has a base CPU clock of 3.6GHz and Max Turbo 3.0 boost CPU clock of 5.0GHz. HardwareLuxx used an ASUS ROG Maximus XIII Hero motherboard, where in The Division 2 at 1080p it beats the Core i7-9700K, but loses to the Ryzen 5 5600X, Core i9-10850K, and the rest of the high-end flagship CPUs on both sides of the fence.

It does better in Battlefield V and Shadow of the Tomb Raider at 1080p, while power consumption while running Cinebench R23 saw the new Intel Core i7-11700K processor hitting 195W before it dropped to 125W and its default PL1 value. HardwareLuxx says that there is around a 15% performance increase with the new Rocket Lake-S offering, but we have to wait for the final BIOS and reviews to see how that goes.

We now have some much more definite answers on gaming performance from Intel's new flagship Rocket Lake-S processor -- with the Core i9-11900K being around 11% faster in 1080p gaming over the current-gen flagship Core i9-10900K processor.

In new performance slides the company shows the new Core i9-11900K benchmarked in 4 different games: Total War: Three Kingdoms, Gears 5, Grid 2019 and Microsoft Flight Simulator -- all at 1080p and High graphics settings. The new Core i9-11900K is 13%, 9%, 8%, and 14% faster than the Core i9-10900K in those games, respectively -- 11% average across the games.

Intel's new Core i9-11900K processor offers 8 cores and 16 threads at up to 5.2GHz with Turbo Boost 3.0 technology, and all-core Turbo of up to 4.7GHz. This is done with a 125W TDP and the Z590 chipset also now offers PCIe 4.0 tech which is a first for Intel and now catching them up to the last couple of years of their competitor in AMD with PCIe 4.0 technology.

AMD has announced it will be officially launching its next-gen Zen 3-based EPYC processors on March 15, with an all-new all-digital reveal event.

The new 3rd Gen EPYC processors are based on the latest Zen 3 architecture, and will succeed the current 2nd Gen EPYC processors that are powered by the Zen 2 architecture. AMD will have a heap of models, with 19 new EPYC processors to be rolled out over the coming months.

The flagship AMD EPYC 7763 is a beast, packing 64 cores and 128 threads with a base CPU clock of 2.45GHz and boost CPU clock of 3.5GHz with 256MB of L3 cache and a 280W TDP. There are plenty of other models, with CPU and thread counts ranging down from there.

Intel's new 11th Gen Core processors are still a few weeks away, but AnandTech has performed some magic trick and has a review of the Core i7-11700K close to a month before its official launch.

The thing is that Intel hasn't even announced the new series of CPUs and have been light on details of its Rocket Lake-S family of chips even though there's been a mountain of leaks over the last few months. AnandTech purchased a retail version of the Intel Core i7-11700K and even though they've signed an NDA, there's no new information shared that breaches that contract.

Dr. Ian Cutress of AnandTech explains that the information shared in his Core i7-11700K review: "Units obtained by that method are not under NDA by definition, and we obtained the Core i7-11700K for this review at retail, and as such we are not under NDA for any information we have obtained through using this processor".

Intel has been slapped with a massive $2.2 billion fine by a federal jury in Texas over the chip giant violating patents held by Dutch semiconductor company VSLI Technology LLC.

The patent violation is over two patents held by VLSI along with NXP Semiconductor getting a slice of the pie, as it owns a part of the patents, with the split of $2.2 billion going into $1.5 billion and $675 million. VSLI acquired NXP Freescale Semiconductor back in 2015, with the patents transferred to VSLI in 2019.

The two patents are first (patent 7,523,373) allows a processor to switch to a low-power state called "sleeping" when it's not in use, which is done through two voltage regulators where one of them will lower power consumption while the other provides enough power to keep the chip's memory in an active state. Patent 7,725,759 allows the chip to switch between programs super-fast, boosting processing speed.

AMD is really aiming for the stars with its upcoming Zen 4 architecture, with the latest rumor on the next-gen EPYC Genoa CPU family is that it will feature up to an insane 96 cores and 192 threads.

The 96-core, 192-threads of CPU power are done through 12 chiplets and on the SP5 (LGA 6096 socket), with 12-channel DDR5 memory support and PCIe 5.0 support. The new rumors are coming from ExecutableFix, which tweeted out that the new Zen 4 architecture would pack some serious horsepower.

The new Zen 4 EPYC family of processors would pack 12 chiplets each with 8 compute cores, with a separate I/O chipset. Instead of increasing the core count per chiplet, AMD is increasing the number of chiplets in the packaging -- with Zen 4 EPYC packing 12 chiplets that usher in up to 96 cores and 192 threads of CPU domination.