#computer_components

And so far only in one test Yesterday, the Chinese company Loongson finally released its 3A6000 processor to the market , which in a number of tests is actually not inferior to the Ryzen 3 3100 and Core i3-10100F . It turned out that it can even compete with the Core i5-14600K! True, only at the same frequency. [caption id="attachment_84905" align="aligncenter" width="780"] Core i5-14600K[/caption] Recent tests have shown that in the UnixBench test the new Chinese CPU performs at the level of the Core i3-10100, if the frequency of the latter is lowered to 2.5 GHz, like the 3A6000. And in the Spec CPU 2006 test, the same thing happens when comparing the Chinese new product with the Core i5-14600K! Moreover, in one of the modes the latter is even a little slower. The Chinese quad-core Loongson 3A6000 processor is capable of competing with the Core i5-14600K. This indicates the great potential of the Loongson architecture, but, of course, in practice the frequencies still differ by more than twice. In addition, these are separate tests and sometimes separate modes in these tests, so we still don’t have the full picture. But we can say for sure that buying a Loongson 3A6000 for a home PC in order to overclock it well and reach the performance of modern AMD and Intel CPUs is a bad idea. Apparently, the overclocking potential is either in general or is still quite weak. Regular overclocking allows you to increase the frequency from 2.5 to only 2.63 GHz, and liquid nitrogen allows you to achieve 3 GHz, but here the motherboard BIOS limitation comes into force. So perhaps extreme overclocking is possible to higher frequencies.

Significant changes in the top five The list of the most productive supercomputers Top500 has been updated once again. Let us remind you that this happens once every six months.  [caption id="attachment_81909" align="aligncenter" width="740"] AMD[/caption] While several supercomputers with performance of about 2 exaFLOPS or higher are on the way, the leader is still the Frontier system, which was the first to overcome the 1 exaFLOPS mark and remains the only one of its kind, far ahead of its closest competitor.   The Frontier supercomputer based on AMD components remains the most powerful in the world. [caption id="attachment_81910" align="aligncenter" width="714"] AMD[/caption] This competitor, by the way, has changed. If six months ago the Supercomputer Fugaku system with a performance of 442 PFLOPS was in second place, now it is Aurora with a performance of 585 PFLOPS. But it is worth saying that now this is only part of the system that has already been put into operation. As a result, Aurora will have performance in excess of 2 exaFLOPS. This system is built on Intel Xeon CPU Max 9470 processors and Intel Data Center Max accelerators. Frontier, remember, relies on AMD components.  In third place is also a newcomer: the Eagle system, owned by Microsoft, with a performance of 561 PFLOPS. But Supercomputer Fugaku dropped to fourth place. 

The company may release a special version of the RTX 4090 for China with performance limitations Nvidia is ramping up shipments of GeForce RTX 4090 GPUs to China ahead of the ban, which comes into force on November 17th. [caption id="attachment_78169" align="aligncenter" width="780"] RTX 4090 GPU[/caption] Nvidia ramps up RTX 4090 GPU shipments to China ahead of supply ban According to MyDrivers, several video card manufacturers have confirmed that the RTX 4090 series will be completely banned from sales in China after November 17. It is advisable for every supplier to purchase and preserve them while they are still available. It is reported that in order to sell more adapters before the ban, various graphics card brands are increasing the supply of RTX 4090. Benchlife resource claims the same. Interestingly, since the new US sanctions rely on performance metrics (4800 TOPS for AI), Nvidia believes it can circumvent the ban by limiting the corresponding performance of the RTX 4090 at the BIOS level. Much the same as it was done during the mining boom.

RX 7800 XT looks better The Radeon RX 7900 GRE video card has become available for free sale in European retail. [caption id="attachment_76366" align="aligncenter" width="780"] Radeon RX 7900 GRE[/caption] Let us remind you that the adapter initially appeared in retail in China, and was available on the global market as part of ready-made PCs. However, now the situation has changed. Radeon RX 7900 GRE can now be purchased at retail in Europe The RX 7900 GRE is sold at least in Spain, where they ask for 720 euros. For comparison, the RX 7900 XT costs 890 euros there, and the RX 7800 XT costs 580 euros. That is, the new product is located approximately in the middle between the two indicated models. True, if you look at European aggregators, the RX 7900 XT can be bought for 835 euros, and the RX 7800 XT can be found for 540 euros. Be that as it may, if we turn to the tests, it turns out that, despite the fact that technically the RX 7900 GRE is close to the RX 7900 XT, in fact, in terms of performance it is not particularly faster than the RX 7800 XT, and the latter turns out to be much more profitable. Let us remember that the RX 7900 GRE has a recommended price of $650, and in this case, it would be more profitable.

GTX 1080 Ti probably won't cope either We've already looked at tests of modern video cards in Alan Wake 2, and the demands of the game with ray tracing turned out to be incredibly high. But TechPowerUp didn't check one feature. Let us remind you that information recently appeared that due to the use of mesh shaders in the game, video cards older than the Radeon RX 6600 and GeForce RTX 20 will either not work at all or will work poorly. And we have at least GeForce GTX 1070 tests.  [caption id="attachment_76167" align="aligncenter" width="600"] GeForce GTX 1070[/caption] The GeForce GTX 1070 is definitely for retirement. The game works, but the performance is incredibly slow. In Full HD, with minimal settings, the video card can only produce 15–20 fps. Moreover, before launching the game even warns that the video card is not supported. This is the first time that a game in non-trace mode explicitly states that it does not support a GTX 10 generation graphics card.  https://youtu.be/kMHxkrABRNs Activating FSR 2 in Quality mode does not help the situation, since performance increases slightly. The same can be said about the Performance mode. That is, the GTX 1070 is not suitable for Alan Wake 2 at all. If you look at benchmarks from the past, the GTX 1080 Ti was faster than the GTX 1070 by about 30%. But even if the increase in Alan Wake 2 is the same, you still won’t be able to play normally due to severe drawdowns. 

From a trusted source Well-known insider kopite7kimi shared fresh data about GeForce GTX 40 Super video cards.  [caption id="attachment_76144" align="aligncenter" width="780"] RTX 4070 Ti Super[/caption] According to his information, the RTX 4080 Super will remain on the AD103 GPU but will receive it in a full configuration with all 10,240 active CUDA cores. For comparison, the regular RTX 4080 has 9,728 of them active.  There will not only be an RTX 4080 Super but also an RTX 4070 Ti Super. However, this option raises questions regarding the memory configuration. The adapter was previously credited with 20 GB of memory, and to implement it, you need a bus with a width of more than 256 bits, which the AD103 has. Of course, Nvidia can reduce the bus to 160-bit, but on a video card with a price of at least $1200, it will look strange. However, Nvidia has had 3D cards in the past with non-standard memory-to-bus width ratios.  The RTX 4070 Super will receive an AD104 GPU with 7168 active CUDA cores, that is, it will be closer to the RTX 4070 Ti with its 7680 cores.  There will also be an RTX 4070 Ti Super. The name is very strange, but this is not the first time it has appeared in rumors and leaks. Such an adapter should receive an AD103 GPU or even an AD103 with 8448 CUDA cores. Information about the possible use of AD102 looks strange compared to the AD103 of the RTX 4080 Super, but the latter could still also be based on an older GPU.  Insider kopite7kimi was the first to accurately describe the parameters of GeForce RTX 30 cards, and he also described RTX 40 cards more than once. Overall, it is one of the most reliable sources in this segment. 

In this case, the video card will probably be transferred to a new GPU The Benchlife resource shared details about the future GeForce RTX 4080 Super - they are reported to have been received from Nvidia partners. [caption id="attachment_72862" align="aligncenter" width="780"] GeForce RTX 4080[/caption] GeForce RTX 4080 Super will get 20 GB of memory, but there will be other improvements According to these same unnamed partners, the GeForce RTX 4080 will have 20 GB of GDDR6 memory - 4 GB more than the current version. This automatically means that the memory bus width will also increase - from 256 to 320 bits. If so, then the new model should switch to another GPU - and it could be the flagship AD102 (but cut down for the GeForce RTX 4080 Super). Let us remind you that details about the GeForce RTX 4070 Super appeared earlier, and it may undergo similar changes: it will receive more memory and a memory bus of a larger capacity 

6-seater cars will begin transporting passengers in central Tokyo in 2026 Honda Motor Company has announced that it plans to launch a self-driving taxi service in Japan along with Cruise and General Motors (GM). The joint venture will be created by the partners in 2024, and the driverless taxi service itself should start operating in 2026. First, driverless taxis will transport passengers in the center of Tokyo, and then beyond. [caption id="attachment_71932" align="aligncenter" width="612"] Self-driving taxis[/caption] To transport passengers, a Cruise Origin car will be used without a driver's seat and steering wheel; the interior of this vehicle can accommodate up to 6 people. At first, the service will be launched with several dozen cars, and then their number will increase to 500. Passengers will be able to order an unmanned taxi using a mobile application and pay for the service there. Honda and General Motors will launch self-driving taxis in Japan [caption id="attachment_71933" align="aligncenter" width="780"] Self-driving taxis[/caption] It is known that Cruise Origin cars have level 4 autonomy, that is, they can move without a driver if a number of conditions are met. According to international standards, there are 5 levels of car autonomy: the first involves only a little automatic assistance to the driver - this could be, for example, an autonomous emergency braking system; Level 5 is a fully self-driving car. In April of this year, a law came into force in Japan allowing the movement of driverless cars with Level 4 autonomy in some cases.