#computing_hardware

[phonemantra-blog]

Almost all parameters are unknown The Chinese company Loongson produces not only some of the most modern Chinese processors, but also GPUs. And its new development is designed to compete with Nvidia accelerators for AI, although they are far from the most productive and modern. [caption id="attachment_85288" align="aligncenter" width="600"] Loongson introduced LG200 AI accelerator[/caption] The accelerator (or its GPU) is called LG200. The characteristics of this solution, unfortunately, are unknown. The block diagram shows that the GPU consists of 16 small ALUs, four large ALUs, and one huge ALU or special purpose unit. Loongson introduced LG200 AI accelerator But the performance is known, albeit for the whole node: from 256 GFLOPS to 1 TFLOPS. Here, unfortunately, the details are again unknown, so it is unclear for which mode the performance is indicated, but even if it is FP64, the figure is quite modest, since modern monsters Nvidia and AMD offer 50-60 TFLOPS or more. At the same time, Loongson’s solution is a GPGPU, that is, it supports general-purpose computing. Unfortunately, there are no details here yet. Separately, we can recall that Loongson promised next year to release a video card that can compete with the Radeon RX 550 , whose performance (FP32) is just over 1.1 TFLOPS. It is possible that the LG200 will be a direct relative of this adapter.

[phonemantra-blog]

It supports Android 12 and Debian 10 The shelf of single board computers has arrived: the Banana Pi BPI-M7 based on the Rockchip RK3588 SoC is presented. The SoC features a quad-core 2.4 GHz Arm Cortex-A76 and four 1.8 GHz Arm Cortex-A55 cores, an Arm Mali-G610 MP4 GPU, and a 6 TOPS Neural Processing Unit (NPU) for acceleration. AI tasks. [caption id="attachment_82595" align="aligncenter" width="780"] Banana Pi BPI-M7[/caption] Banana Pi BPI-M7 will go on sale in versions with 8, 16 and even 32 GB of LPDDR4X RAM, and the total amount of eMMC flash memory can be 64 or 128 GB. Also on the board measuring 92 × 62 mm there is an M.2 slot connected to the PCIe 3.0 x4 bus. Banana Pi BPI-M7 introduced, one of the most powerful single board computers [caption id="attachment_82596" align="aligncenter" width="496"] Banana Pi BPI-M7[/caption] Banana Pi BPI-M7 is equipped with Bluetooth 5.2 and Wi-Fi 6 adapters, as well as a pair of 2.5 Gbps Ethernet ports. HDMI and USB-C are used to connect monitors, and both support image output at 8K resolution, but the frame rate is different - 60 and 30 fps, respectively. The computer is compatible with Android 12 and Debian 10 OS. Unfortunately, the cost has not yet been announced.

[phonemantra-blog]

The division into Pro and regular models will return The Videocardz resource has published the first images of AMD Ryzen Threadripper 7000 processors. It's finally confirmed that AMD is going back to two lines for its monster CPUs. We are waiting for Threadripper Pro 7000 processors for workstations and regular Threadripper 7000 processors for enthusiasts. The latter should just be sold at retail. In addition, the latter will receive packaging with orange accents, while the former will have silver accents. [caption id="attachment_71237" align="aligncenter" width="780"] Ryzen Threadripper 7000 processors[/caption] The new processors will use the sTR5 socket and include from 8 to 12 chipsets, like the Epyc Genoa server ones. Technically these are very similar processors. The first images of Ryzen Threadripper 7000 processors have been published It is known that motherboards for the new CPUs will support DDR5 memory and PCIe 5.0 cards. Probably, the difference in the number of channels in the RAM controller will also return. Let us remind you that the top models of the new line will be 96-core CPUs, which will allow for a giant leap in performance compared to the already very powerful Ryzen Threadripper 5000. The announcement of new processors will take place tomorrow.

[phonemantra-blog]

And not only in Counter Strike 2 AMD had to disable its latest Anti-Lag+ feature in all games. The latest Radeon Software Adrenalin 23.10.2 driver comes out just a week after the previous one, where Anti-Lag+ received support for Counter-Strike 2.  [caption id="attachment_70272" align="aligncenter" width="780"] AMD[/caption] Let us remind you that it turned out that Valve Anti-Cheat security software automatically blocks players who have the Anti-Lag+ function enabled. The problem affects not only Counter-Strike 2 but also several other competitive online shooters.   AMD has disabled the new Anti-Lag+ feature, which caused gamers to be blocked in Counter-Strike 2 The thing is that the method by which Anti-Lag+ works in games through the driver is identified by anti-cheat programs as one of the ways to bypass the game’s protection. Anti-Lag+ does not require developers to integrate it into a particular game but instead modifies game DLLs.  It is unclear what AMD will do next with its development. Perhaps it will simply block its activation in those games where this may be a problem, leaving the function active for everyone else. Let us remind you that the Anti-Lag+ function itself is designed to reduce delays.  

[phonemantra-blog]

These are almost complete copies of the RX 6700 and RX 6700 XT AMD today introduced the Radeon RX 6750 GRE 10GB and Radeon RX 6750 GRE 12GB video cards. Unlike the RX 7900 GRE, which is an independent model, these two new products are almost complete copies of the RX 6700 and RX 6700 XT. But at the same time, the new products are much cheaper, and because of this, they create competition for the same RX 7600. You can verify this by looking at the tests that have already been published by Chinese resources.  [caption id="attachment_70259" align="aligncenter" width="780"] AMD[/caption] As you can see, the youngest of the new products is slightly ahead of the RX 7600, while the older one is much faster and almost identical to the RX 6750 XT. At the same time, let us remind you that AMD’s new products are very cheap. Initially, only prices for China were known, but there are also retail prices for the US, which is strange considering that in the US and Europe, these adapters will reportedly only be available as part of pre-built PCs.  One way or another, for the US we are talking about $270 and $290 for the younger and older versions, respectively. The difference is surprisingly negligible, and the older card looks much more advantageous: it has more memory, a wider bus, and more stream processors.  AMD is selling for $290 while the GeForce RTX 4060 Ti sells for $400-$500. [caption id="attachment_70260" align="aligncenter" width="556"] AMD[/caption] [caption id="attachment_70261" align="aligncenter" width="563"] AMD[/caption] At the same time, the price of the RX 7600 is just $270, that is, we can say that both new products compete with this adapter and the older of the new products looks more profitable.   As for competing with Nvidia, the RX 6750 GRE 12GB is almost identical to the RTX 4060 Ti. Moreover, the latter costs $400.

[phonemantra-blog]

On average, there is almost no difference between the new and old generations Yesterday, Intel introduced the Core i9-14900K/KF, Core i7-14700K/KF, and Core i5-14600K/KF processors of the Raptor Lake Refresh generation. Today their reviews appear on the Internet. [caption id="attachment_70245" align="aligncenter" width="780"] Intel[/caption] Let us remember that only Core i7 models received more cores than their predecessors, while all other new products are content with only increasing frequencies.  Because of this, which is logical, the performance increase for all new models, except for the Core i7, is not particularly large.   If we talk about the Core i5-14600K, on ​​average, according to TechPowerUp tests, in various applications this CPU is only 2-3% faster than the Core i5-13600K, and this difference can be neglected. In games, as expected, the situation is the same. Interestingly, the consumption of the new product is slightly lower than that of its predecessor: 93 W versus 105 W. This is an average based on data from 47 applications.  The Core i9-14900K is the same amount faster than the Core i9-13900K, which makes the new product identical to the Core i9-13900KS model. The latter was simply renamed Core i9-14900K since even the frequencies of these CPUs are the same.   The Core i7-14700K can be called the most interesting among the new products, especially considering the unchanged price. However, this is only on paper. In reality, the increase is also very modest. The new CPU is on average faster than the Core i7-13700K by about 5%.  Of course, as always, when it comes to modern multi-core processors, you need to look at those tests that are important to a particular user. Because, for example, in Blender Core i7-14700K is no longer 5%, but 14% faster than its predecessor. And in several other applications that actively use all cores, the situation is approximately the same, since after all, the new CPU has four more small cores.   Overall, however, the new generation is, as expected, only a very slight acceleration of the previous generation. Intel needs new CPUs to fill the gap before the release of Arrow Lake, which is not expected until the end of 2024.

[phonemantra-blog]

Because of this, GPD is unable to supply its set-top boxes to customers. GPD, known for its portable laptop-game console hybrids, is experiencing supply problems with its new Win Max 2 model. And AMD is to blame for this, according to GPD.  [caption id="attachment_70240" align="aligncenter" width="780"] AMD[/caption] GPD accuses AMD of missing delivery dates for Ryzen 7000 hybrid processors GPD says AMD is unable to supply the pre-agreed number of Ryzen 7000 APUs on which the new console is based. More precisely, we are talking about the Ryzen 7 7840U. GPD received the first batch of these APUs, but they still haven’t received the second, although they should have received it a long time ago. As a result, GPD accused AMD of violating the terms of the contract, but it is not yet clear whether this is all formalized in legal form.  It is also unknown what volumes of Ryzen APUs we are talking about. However, you need to understand that AMD does not produce its own CPUs and GPUs. The company's modern solutions are produced by TSMC.  

[phonemantra-blog]

The basic version with 10 GB of memory costs this much, but there is also an option with 12 GB of memory. Today, AMD officially unveiled the Radeon RX 6750 GRE 3D card. It has two options that don't differ much in price ($305 and $325) but are very different in terms of features. [caption id="attachment_70227" align="aligncenter" width="780"] AMD[/caption] The base Radeon RX 6750 GRE with 10GB of GDDR6 memory has 36 compute units, a 160-bit memory bus, and 80MB of fast Infinity Cache memory. GPU frequency - 2189–2450 MHz, consumption - 170 W. This is how AMD describes the latest Radeon RX 6750 GRE graphics card for $305 [caption id="attachment_70228" align="aligncenter" width="780"] AMD[/caption] The Radeon RX 6750 GRE version with 12 GB of GDDR6 memory received a GPU with 40 computing units and a frequency of 2439–2581 MHz. The memory bus width was 192 bits, the Infinity Cache volume was 96 MB, and the consumption was much higher - 230 W. Both versions will be offered in China by AMD partners and major PC builders as pre-configured systems.

[phonemantra-blog]

Next will be Core and Core Ultra Intel has unveiled its new 14th-generation Core processors, and now the company has confirmed that these will be the last Core I CPUs.   We already knew that the upcoming Meteor Lake would switch to the Core Ultra brand. In addition, there will be just Core models. But there will be no more Core i models.  It is worth understanding that the 14th generation Core will also include mobile processors that will be released later. They will still be named Core i.   Intel says goodbye to the Core i brand. [caption id="attachment_70101" align="aligncenter" width="780"] Intel[/caption] Let us remember that the first Core processors appeared in 2006, but these were the Core Duo and Core Solo models. The first Core I came out in 2009. 

[phonemantra-blog]

So far tested on a CPU with a consumption of more than 800 W Intel seems to be hinting to us that in the future processors may well be able to consume more than 1 kW of power. Together with Submer, the processor giant announced what participants say is a revolutionary development in the field of single-phase immersion cooling technology, providing heat dissipation from processors with power up to 1000 W.  [caption id="attachment_69912" align="aligncenter" width="780"] Intel[/caption] The system is designed to “revolutionize data center cooling by reducing the number and cost of components required to comprehensively capture and dissipate heat in chips with a TDP of over 1,000 watts.”  Intel is preparing the market for processors that consume more than 1 kW of power. For testing, Intel used an unnamed processor that consumed more than 800 W. It is quite possible that this was a modern representative of the Sapphire Rapids line.  [caption id="attachment_69913" align="aligncenter" width="780"] Intel[/caption] The companies note that the new development will reduce customer costs, it is universal and can compete with systems based on direct liquid cooling. The companies also note the simplicity of the design and the ability to use 3D printing to create elements of this design. The development will be officially presented at the OCP Global Summit event, which will be held from October 17 to 19.