Tac chien mang

style: Sến, tiểu nhân, ao làng

nguyễn đình đăng block nhưng vẫn dùng tài khoản khác để đi soi kẻ mà y đã block, đúng là hạng văn dốt võ dát, lèm bèm như một con đĩ già mồm!

Mao nói: trí thức là cục phân!
trí thức như nguyễn đình đăng: vô lại, phản quốc, mất gốc vong bản, một thằng ăn mày ở Nhật và bợ đít Tây tới mức ..... tởm lợm.

đăng định làm cách mạng màu cùng với thằng thắng đông A hay sao? đúng là hai thằng chúa màu mè đĩ bợm
không đủ tuổi đâu, đăng ạ!

Kaspersky Cloud Workload Security has been updated!

This release includes enhancements across:

Kaspersky Container Security 2.4
• AI-powered container image analysis via OpenAI API, allowing to get additional information about a scanned image
• SSO integration into your infrastructure
• AD multi-domain support within one installation
• Optimized scanning performance, enhanced process and size management
• Expanded security policy capabilities

Kaspersky Hybrid Cloud Security 6.4
• Now uses KES for Windows 12.12 and KES for Linux 12.4 as the Light Agent components

Explore in more detail >> https://kas.pr/p42a

Laser Hồng ngọc

Theodore Maiman invented the world's first laser, known as the "ruby laser" in 1960. Ruby crystal is composed of aluminum oxide, where some of the aluminum atoms have been replaced with chromium atoms. Chromium gives the ruby its vibrant red color. In a ruby laser, a ruby crystal is formed into a cylinder. A fully reflecting mirror is placed on one end and a partially reflecting mirror on the other. A high-intensity lamp is spiraled around the ruby cylinder to provide a flash of white light that triggers the laser action.

The green and blue wavelengths in the flash excite electrons in the chromium atoms to a higher energy level. Upon returning to their normal state, the electrons emit their characteristic ruby-red light. The mirrors reflect some of this light back and forth inside the ruby crystal, stimulating other excited chromium atoms to produce more red light, until the light pulse builds up to high power and drains the energy stored in the crystal—this produces what we typically think of as "laser light".

Although many different types of lasers have been invented since Maiman's device, the ruby laser is still used, mainly as a light source for medical and cosmetic procedures, and also in high speed photography and pulsed holography.

Atoms in Ruby Crystal Are Excited to Higher Energy Levels1. High-voltage electricity causes the quartz flash tube to emit an intense burst of light, exciting some of the atoms in the ruby crystal to higher energy levels.

Photons Are Emitted2.
At a specific energy level, some atoms emit particles of light called photons. At first the photons are emitted in all directions. Photons from one atom stimulate emission of photons from other atoms and the light intensity is rapidly amplified.

Mirrors Reflect the Photons3.
Mirrors at each end reflect the photons back and forth, continuing this process of stimulated emission and amplification.

Photons Are Emitted As Laser Light4.
The photons leave through the partially silvered mirror at one end. This is laser light.

GPU cho AI được bán giá 46.000 USD trên eBay

H100, chip mạnh nhất của Nvidia dùng để huấn luyện AI, đang được rao giá 46.000 USD (một tỷ đồng) trên eBay trong bối cảnh nhu cầu tăng vọt.

Theo John Carmack, cựu giám đốc tư vấn công nghệ của Meta, một số chip đồ họa Nvidia H100 đang được rao bán trên nền tảng eBay với giá 39.995-46.000 USD. Trong khi đó, mức giá được một số công ty bán lẻ đề xuất trước đó là 30.000-36.000 USD.

H100 là chip AI mới và mạnh nhất của Nvidia, cũng là bản kế nhiệm của A100 - có giá 10.000 USD và được mệnh danh là "ngựa thồ" của ngành trí tuệ nhân tạo, với 80 tỷ bóng bán dẫn bên trong!

Nvidia Reveals Hopper H100 GPU With 80 Billion Transistors

Today, at its GPU Technology Conference (GTC), Nvidia revealed details of its Hopper architecture and the Nvidia H100 GPU. We've known Nvidia has been working on next-generation GPUs for some time, but now we have some concrete specs. The Hopper architecture and H100 GPU are not to be confused with Ada, the consumer-focused architecture that will power future GeForce cards. Nvidia hasn't revealed any details on Ada yet, and Hopper H100 will supersede the Ampere A100, which itself replaced the Volta V100. These are all datacenter parts, and with steeper competition from the likes of AMD's Instinct MI250/250X and the newly announced Instinct MI210, Nvidia is looking to retake the lead in HPC.

As you'd expect given its legacy, H100 was designed for supercomputers with a focus on AI capabilities. It includes numerous updates and upgrades compared to the current A100, all designs to reach new levels of performance and efficiency. Hopper packs in 80 billion transistors, and it's built using a custom TSMC 4N process — that's for 4nm Nvidia, not to be confused with the generic N4 4nm process that TSMC also offers. For those keeping score, the A100 GPU 'only' had 54 billion transistors.

Nvidia didn't reveal core counts or clocks, but it did give some other details. H100 supports Nvidia's fourth generation NVLink interface, which can deliver up to 900 GB/s of bandwidth. It also supports PCIe 5.0 for systems that don't use NVLink, which tops out at 128 GB/s. The updated NVLink connection provides 1.5X more bandwidth than the A100, while PCIe 5.0 delivers double the bandwidth of PCIe 4.0.

The H100 will also support 80GB of HBM3 memory by default, with 3 TB/s of bandwidth — that's 1.5X faster than the A100's HBM2E. While the A100 was available in 40GB and 80GB models, with the latter coming later in the life cycle, both the H100 and A100 still use up to six HBM stacks, apparently with one stack disabled (i.e., using a dummy stack). Generally speaking, the H100 has 50% more memory and interface bandwidth compared to its predecessor.

That's a nice improvement, to be sure, but other aspects of Hopper involve even larger increases. H100 can deliver up to 2,000 TFLOPS of FP16 compute and 1,000 TFLOPS of TF32 compute, as well as 60 TFLOPS of general purpose FP64 compute — that's triple the performance of the A100 in all three cases. Hopper also adds improved FP8 support with up to 4,000 TFLOPS of compute, six times faster than the A100 (which had to rely on FP16 as it lacked native FP8 support). To help optimize performance, Nvidia also has a new transformer engine that will automatically switch between FP8 and FP16 formats, based on the workload.

Nvidia will also add new DPX instructions that are designed to accelerate dynamic programming. These can help with a broad range of algorithms, including route optimization and genomics, and Nvidia claims performance in these algorithms is up to 7X faster than its previous generation GPUs, and up to 40X faster than CPU-based algorithms. Hopper also includes changes to improve security, and the multi-instance GPU (MIG) now allows for seven secure tenants running on a single H100 GPU.

All of these changes are important for Nvidia's supercomputing and AI goals. However, the changes aren't all for the better. Despite the shift to a smaller manufacturing node, the H100 TDP for the SXM variant has been increased to 700W, compared to 400W for the A100 SXM modules. That's 75% more power, for improvements that seem to range between 50% and 500%, depending on the workload. In general, we expect performance will be two to three times faster than the Nvidia A100, so there should still be a net improvement in efficiency, but it's further evidence of the slowing down of Moore's Law.

Overall, Nvidia claims the H100 scales better than A100, and can deliver up to 9X more throughput in AI training. It also delivers 16X to 30X more inference performance using Megatron 530B throughput as a benchmark. Finally, in HPC apps like 3D FFT (fast Fourier transform) and genome sequencing, Nvidia says H100 is up to 7X faster than A100.

History of Graphics Processing Units

Back in 1999, NVIDIA popularized the term “GPU” as an acronym for graphics processing unit, although the term had been used for at least a decade prior to marketing the GeForce 256. However, the GPU was actually invented years before NVIDIA launched their proprietary NV1 and, later, the video card to rule them all.

1980s: Before there was the graphics card we know today, there was little more than a video display card. IBM made and introduced the Monochrome Display Adapter (MDA) in 1981. The MDA card had a single monochrome text mode to allow high-resolution text and symbol display at 80 x 25 characters, which was useful for drawing forms. However, the MDA did not support graphics of any kind. One year later, Hercules Computer Technology debuted the Hercules Graphics Card (HGC), which integrated IBM’s text-only MDA display standard with a bitmapped graphics mode. By 1983, Intel introduced the iSBX 275 Video Graphics Controller Multimodule Board, which was capable of displaying as many as eight unique colors at 256 x 256 resolution.

Just after the release of MDA video display cards, IBM created the first graphics card with full-color display. The Color Graphics Card (CGC) was designed with 16 kB of video memory, two text modes, and the ability to connect to either a direct-drive CRT monitor or a NTSC-compatible television. Shortly thereafter, IBM invented the Enhanced Graphics Adapter (EGA) that could produce a display of 16 simultaneous colors at a screen resolution of 640 x 350 pixels. Just three years later, the EGA standard was made obsolete by IBM’s Video Graphics Adapter (VGA). VGA supported all points addressable (APA) graphics and alphanumeric text modes. VGA is also known as Video Graphics Array as a result of its single-chip design. It didn’t take long for clone manufacturers to start producing their own VGA versions. In 1988, ATi Technologies developed the ATi Wonder as part of a series of add-on products for IBM computers.

1990s: Once IBM faded from the forefront of formative PC development, many companies began developing cards with more resolution and color depths. These video cards were advertised as Super VGA (SVGA) or even Ultra VGA (UVGA), but both terms were too ambiguous and simplistic. 3dfx Interactive introduced the Voodoo1 graphics chip in 1996, gaining initial fame in the arcade market and eschewing 2D graphics altogether. This hardcore hardware led to the 3D revolution. Within one year, the Voodoo2 was released as one of the first video cards to support parallel work of two cards within a single PC. NVIDIA entered the scene in 1993, but they didn’t earn a reputation until 1997 when they released the first GPU to combine 3D acceleration with traditional 2D and video acceleration. RIVA 128 did away with the quadratic texture mapping technology of the NV1 and featured upgraded drivers.

Finally, the term “GPU” was born. NVIDIA shaped the future of modern graphics processing by debuting the GeForce 256. According to the NVIDIA definition, the graphics processor is a “single chip processor with integrated transform, lighting, triangle setup/clipping, and rendering engines that is capable of processing a minimum of 10 million polygons per second.” The GeForce 256 improved on the technology offered by RIVA processors by taking a large leap in 3D gaming performance.

2000s: NVIDIA went on to release the GeForce 8800 GTX with a texture-fill rate of 36.8 billion per second. By 2009, ATI released the colossal Radeon HD 5970 dual-GPU card before being taken over by AMD. At the dawn of virtual reality in consumer technology, NVIDIA developed the GeForce Titan, which has become the forerunner of graphics technology since. NVIDIA sees multi-chip GPU architecture as the future of graphics processing, but the possibilities are endless.

History of Halogen Lamps
Invention of Halogen Light Bulbs

Since the beginning of the evolution of the incandescent lamp, engineers were trying to make a lamp that would last longer and give brighter light while consuming less energy. Life of the lamp is determined by lasting of a filament and if evaporation of the filament is lessened, life of the lamp will be longer. If you can raise the temperature of the filament lamp will shine brighter. For all this problems - halogen lamp is a solution. Not without flaws, of course.

Halogen lamp is a sort of incandescent lamp with a small amount of some halogen gas, most commonly iodine or bromine, as an atmosphere in a quartz or aluminosilicate glass bulb. In ordinary incandescent lamps, tungsten evaporates under the influence of the heat and deposits on the inner surface of the glass bulb darkening its surface. Major characteristic of the halogen lamp is relation between tungsten and halogen gas in the glass bulb (so-called halogen cycle). When tungsten evaporates, it reacts with halogen forming the halide, which does not deposit on the glass. When halide gets close to the tungsten, which has high temperature it dissolves into tungsten, which returns to filament, and to halogen, which returns to atmosphere in the bulb, to react again. That way glass stays clear and tungsten filament lasts longer. Because of the high temperature and need for bulb to be small so halogen can react with tungsten, bulb must be strong and resistant. That is why it is made from quartz or aluminosilicate glass. Because bulb is strong it is possible to rise pressure and concentration of halogen in the bulb which again gives better reaction between halogen and tungsten.

First lamp to use halogen gas (chlorine) was patented in 1882 but the first commercial halogen lamp that used iodine as a halogen gas was patented in 1959 by General Electric. It was developed by Elmer Fridrich and Emmet Wiley who worked at General Electric, in 1955. Since 1980, halogen lamp was improved and made lighter.