rivals
For now, NVIDIA's rivals are AMD and Intel, right?
I wonder if I've overlooked another strong rival 💧
I'm a little uneasy 💧
Disclaimer: Community is offered by Moomoo Technologies Inc. and is for educational purposes only.
Read more
Comment
Sign in to post a comment
みかんとリンゴ : Are there any rivals against Inter?
It's like they're not benefiting from AI.
Speaking of which, I saw the news that Intel dropped due to Dow's recruitment and NVIDIA was adopted
リック302 OP みかんとリンゴ : Intel is subtle if you look at it as an AI industry
NVIDIA is finally joining the Dow
I watch something like it's coming out as a rival horse for Broadcom
I was wondering if it was Intel or AMD when I looked at things like GPU design
The rivals were so subtle that I thought NVIDIA would be the best
What do you all think
I'm curious
洋介48 : From an IT engineer's point of view, it is a prerequisite for AI training that there are no NVDA rivals, that AI related source code downloaded from GitHub almost works with NVDA chips, and engineers around the world are developing AI software with NVDA's CUDA
exera : Certainly, the technologies such as GPUs and Kuda that have been built up until now are strengths$Microsoft (MSFT.US)$ Ya$Apple (AAPL.US)$ 、$Amazon (AMZN.US)$ Large companies that control products and terminals that have already penetrated to the general public and are essential to society and daily life$NVIDIA (NVDA.US)$ If we establish in-house technology that can ensure a certain level of quality even if not moderately, and our products only work because of that, the strength of the end product will be lacking$NVIDIA (NVDA.US)$ The battle might be scary, isn't it? Maybe it's a bit extreme![[undefined]](https://static.moomoo.com/nnq/emoji/static/image/default/default-black.png?imageMogr2/thumbnail/36x36 "undefined")
リック302 OP 洋介48 : The source code accumulated when performing AI deep learning was created by Kuda, and using it is technically (and economically) efficient
So there are no options other than NVDA
Is that OK to understand?
I looked it up but I'm an amateur
リック302 OP exera : I also feel like there are companies that can make a profit even if they spend a huge budget and time
Also, I wonder if the deep learning field is left to NVDA products to develop a strategy at that higher layer
sleepinglion : MPUs are more accurate than Intel GPUs for the best microfabrication technology, and the only thing that competes with microprocessing technology is that PCs work even without an Intel VSAMD GPU, but without an MPU, only Intel can make all of the box motherboards, MPUs, GPUs, chipsets, and communication modules (LAN chips), and Intel is the only company that doesn't have all compatibility issues, and other manufacturers eventually make products that match Intel products Sometimes AMD makes things better than Intel, but in the future, the military will eventually rise to Intel due to the reliability of compatibility, and in the future, what is thought to be Intel 1 strong is that when it becomes the quantum computer generation, AI should go there. Nvidia happened to specialize in image processing and was diverted to AI, but since designs without in-house production capacity are the main companies, there is a possibility that Arm may take that position...
At the end of the day, as long as distributed parallel processing is performed, the structure for connecting multiple servers will not disappear, and the connection will eventually become a network device, but the most reliable network chip is Intel or Broadcom, and Broadcom specializes in networks and doesn't make PC/server MPUs? Eventually, if Intel's standards are lost, it will be a hodgepodge of various companies, and if compatibility issues arise, responsibilities will probably be imposed on each other...
In short, it's the same as today's PCs, and what eventually becomes a bottleneck is I/O. The speed is only when information is extracted from the brain, so no matter how fast you can calculate in your head, if it takes time to answer, rabbits will be overtaken by turtles... the real game is until what you calculate at the data center reaches the place you use it
Until quantum computers become popular where they are used, the real strongest is Intel, or if AI can be completed with the GPU alone, it can't be said that it's Nvidia, but that's not true. Intel also makes Mazabo, and I wonder if most server mothers are Intel or SMC that are in place of this, such as Supermicro Computer and ASUS?
Nvidia is currently in the strongest position in AI technology, but I don't know what will happen when quantum computers become mainstream. I don't know how many years from now... If you think that 10 years from now, we may be in the quantum computer era... how many years ahead do long-term investments look ahead? It's going to be a story.
Wouldn't it be a story that would easily turn upside down when Intel and AMD started selling MPUs for servers at 100 times the price?
HONDA N-ONE : What about IOWN?
mya0 : GPUBased on the performance of,NVIDIA >= AMD >> IntelI think that's the order.NVIDIAwithAMDThe hardware is equal,CUDAI haveNVIDIAThe superiority of is rock-solid. On the other handIntelofGPUIt's not a blunder.
Now in the data centerAIProcessing technology is a hot topic, but due to privacy and security, on the computer sideAIThere is a possibility that they will want to use it. At that time,WindowsIt movesIntelDepartmentCPUsI'm makingIntelBetter thanGPUI'm makingAMDBut they combined itCPU/GPUI think it's going to turn into a big ghost by making it.
リック302 OP sleepinglion : Thanks for the detailed explanation and insight
It was a great learning experience
Intel has a wide range of technical capabilities that can set industry standards, and it means that its competitive advantage hasn't dropped yet
An MPU is always required to use a GPU, and as long as distributed parallel processing is performed, the speed of communication between servers and the capacity and speed of data exchange between components on the motherboard are also important.
Intel, which can set the standards necessary for hardware, is the strongest. For me, this is my understanding
The reason why Broadcom is raising stock prices is also understandable.
The scale of servers required for deep learning is getting bigger
NVIDIA means that the advantage of the Cuda is too overwhelming.
Quantum computers are also in the near future
View more comments...