Rewriting the Technological Scenario

On the baseband chip, Intel is just a busy, but in another thing, Intel actually helped the wind group a busy, although Intel herself is estimated to be aware of.

That is in September 2007, Intel will acquire HAVOK, one of the three physical engines in the world, but in September of this world, Havok is acquired by the wind group with a price of 120 million meters. The reason why Intel helped the Great Breeze Group because I was in this world or asked to buy HAVOK's will, but I heard that the wind group had to buy, and Intel did not compete with the wind group.

Of course, why not fight, the key is because Intel did not see the value of this company. After the Intel acquired Havok because I didn't make a graphics card, I would like to be responsible by the CPU, resulting in the development of the graphics card to accelerate HAVOKFX directly .

A few months later, Ying Weida acquired another company of the world's three physical engine Ageia, British Weida's strong development, and opened the GPU physics acceleration era.

Therefore, Meng Qian wants to grab HAVOK from Intel's hands, and don't see how big the value of HAVOK. Some of his acquisition is even for patent fees, because the current ATI and Ying Weida have use HAVOK solutions. Microsoft's Xbox and Sony's PS also use the HAVOK solution.

However, there is no need to earn, this small money doesn't matter, it doesn't matter, just give a personal case, this, the company has arrived in the hands of the wind group.

After completing the acquisition, Meng Qian planted the company's original graphics department and the ATI-integrated GPU subsidiary Galang Technology with HAVOK's management, everyone gathered together to meet each other, and we will work together.

Then the first thing to announce is that HAVOK will become the independent subsidiary of the Great Breeze Group, which will not merge with Galaxy Science and Technology. The reason is also realistic, IBM has 6% of Galaxy Technology.

Then, everyone began to meet, the current CEO David, the current CEO of Galang, announced a good news. At the end of 2006, it was the AMD acquisition of ATI. After the world's wind group acquired ATi, ATI's R580 + listing, R580 + The listing brings a new series of product lines.

Among them, the high-end uses X1950XTX main, and the application of GDDR4 memory greatly improves the transmission bandwidth and performance. The RV570 hit of the high-end market, the RV570 has 36 PS units, 12TMU, 12ROPS, the same 3: 1 architecture, performance performance is very excellent, becoming the focus of medium and high-end markets. The low-end market RV560 performance exceeds G73, which has become one of the most cost-effective display chips in the middle and low-end graphics market in 2007.

Many people think it is true that God does not give ATI opportunities. If ATI has more anti-one year, maybe it is alive, this helplessness, I used to be AMD, this world is naturally a big wind group.

With this new series of product lines, the wind group acquired AIT after the graphics market, the market share took directly to the top 10 percentage points, and many users couldn't make a clear, I thought it was because the wind group acquired. So the ATI is done well, and a big wind group is promoted.

After reporting the sales of the product and the market situation, David began to talk about the company's development, and in Meng Qian's request, David took the person of the test of Havok, "you have a certain CUDA to Yingda. Learn about it? "

Everyone nodded, and CUDA has developed for a year. If you haven't learned CUDA, it is really OK.

"I don't know what kind of view is there on CUDA here?"

Havok's eyes are concentrated on a person, that is their chief technology officer, Claude, Claude very confidently. "Because everyone is discussing the potential of GPU, many people When talking about this problem, I will think of Ying Weida's CUDA, but at least from the current point of view, because of the purpose of designing CUDA, at least, in order to let more developers can use their products, but targeting, it is still a game. And video production areas.

True everyone began to pay attention to the potential of GPU, in fact, from this year, whether the original intention of the CUDA designed by Ying Weida is related to the potential of GPU, I said, but no matter whether it is related, Ying Weida can really take the ship in advance of CUDA, I created a little innate advantage, but Ying Weida could not use this innate advantage. Now it is still not good.

At least the development of CUDA is a bit slow. "

Meng Qiao nodded, Claude saw the nature of the problem, and is willing to speak these things when the first company meeting will speak.

David looked at Meng Qian's expression asked, "If you are responsible for the CUDA project in Ying Weida, what do you think?"

"There is no doubt thing is to strengthen the genericity of GPU, and the CUDA of Yingda does not help GPU real release potential. In other software applications other than games and video production, GPU is still idle, everyone should be clear, From the speed of using the floating point operation, the CPU is far lower than the GPU, and it should now be to dig the GPU universal possibilities. "

Meng Qian looked at Claude's eyes more appreciation, David continued to ask, "Can you talk about specific measures?"

"Simply, the things do what GPU on the game is to calculate the pixel color. In theory, in any application direction, we can completely prepare the corresponding array to determine the way GPU calculations, thus Implement the corresponding function, so this requires a library to complete the docking work, this is a point that I ignored after I saw the Yingda's CUDA, CUDA should not be just a programming environment, but should be an algorithm platform, Including the environment, including libraries, including hardware.

If you can create such an arithmetic platform, the value of GPU will greatly improve. "

"Then what do you think of GPU will shine in what areas?"

"The GPU should have its own use of martial arts for parallel computational demand." Claude thought, did not mention too specific direction.

However, Claude can have such an idea that Meng Qian has already thought this person is very good. At this time, Meng Qian gave David a look, David no longer continued, "I am very happy to hear the views of Mr. Claude. Because Mr. Claude's view is synchronized with us. "

This sentence comes out, HAVOK's high-level real thing, really, David's next thing to introduce, the wind group spent a few years has been in silent GPU computing platform, Xinghe computation platform, star river operation platform The huge function library, simple and convenient development environment, rich application tests, and compatibility with various hardware.

For now, the Star River Computing Platform is more like a GPU instruction manual for developers, which in how to use GPU to achieve more applications other than games and video production, and provide specials for developers. A convenient operation platform.

In the process of internal display, David highlights everyone to read the excellent performance of GPU in the depth study.

Meng Qian as a pioneer of deep learning theory, and then cooperates with the Shangxing River platform for GPUs in deep learning results, and has highly technically sensitive Craward has smell a fashion taste.

After introducing the Fin River platform, David summarizes, "Star River platform, GPU's physical acceleration development and technology is the three major development directions of our company. About the process, our goal is to achieve 55nm support within 2008, The Star River platform is developed with the physical acceleration of GPU, I hope that we can run together as soon as possible, create a perfect ecology. "

The conference room sounded a warm applause.

Meng Qian looked at the current team configuration and everyone's mental state, and it was still very practical.

...