Intel's new 48 core processor: Artificial intelligence gets a boost
Intel has demonstrated a fully programmable 48-core processor. It is called a Single-chip Cloud Computer (SCC) processor. Intel thinks it will pave the way for massive data computers powerful enough to do more of what humans can. The processor has 48 core (24 dual-core) with about 1.3 billion transistors and the processor consumes between 25 and 125 watts of power. Also, the SCC chips uses ordinary x86 processor cores, which means regular software such as Windows and Linux can run on them. Ultimately, Intel currently believes its assertive multicore approach will be the way computers get enough power for tasks such as vision and speech comparable to real humans and artificial intelligence.
No new breakthrough for Parallelism
With the new chip having many more cores than previous models, Intel expects programmers to eventually reckon with this trend. The issue now is that the computing industry is struggling to adjust to parallel programming because most PC software is geared to perform a single sequence of instructions, not many things in parallel. Parallelism is basically the concept of performing multiple tasks simultaneously. This effectively reduces the time needed to perform a given task. The Intel Tera-Scale research program is focused on the concept of utilizing many more cores than conventional to increase performance with parallelism. Unfortunately, the new chip hasn't addressed this age long problem. Intel CTO Justin Rattner said:
"This isn't a full solution"
A principal engineer with Intel's Microprocessor Technology Lab, Anwar Ghuloum said:
"Sometimes, the developers are trying to do the minimal amount of work they need to do to tap dual- and quad-core performance...I suppose this was the branch most discussions took a couple of years ago."
Key features
Some of the key features of the SCC processor include:
1 Cooling: there is a large copper heat sink and 2 fans for cooling
2 Router interface: A router interface links the two x86 cores with their own memory cache
3 Memory: the SCC can accommodate up to 64GB of memory
Compatibility? No problem
While the SCC chip may not have a good solution for the parallel programming challenge, it does have the advantage of some compatibility with existing computer designs. The SCC can run ordinary software for Intel chips, unlike the increasingly capable graphics chips produced by Intel rivals NVDIA and AMD.
"Our thrust is to maintain the compatibility and familiarity of the Intel architecture as we move to more and more performance,"
Conclusion
Intel is hoping to help solve the computing challenges that today require thousands of networked servers. These problems are considered ”here and now problems” compared to futuristic pursuits.
The SCC cores themselves are not as powerful as Intel's flagship Nehalem models, but collectively they pack a lot of power. In contrast, Intel's Larrabee processor, a many-core x86 chip under development for graphics acceleration, is a cache-coherent design that has a large amount of resources devoted to caching data. Ray tracing is a technique for rendering three-dimensional graphics using complex light interactions, allowing the creation of extremely detailed reflective surfaces with fantastic realistic results.
Thanks for reading
I can't wait to play League of Legends on one of this xD JK JK
Intel and Nvidia are 2 of my favorite "brands", after buying some cryptos i'm 100% going to buy stocks of those 2 companies, they are paving the way to the future, everything will sooner or later require a chip and who else makes chips and graphics cards besides nvidia and intel? (I know AMD does them do, but i never bought a AMD cpu so i can't speak abour the brand)
AMD is giving Intel a serious run. The Ryzen 7 1700 is on my wish list. Thanks for reading
Yeah ive heard about those!
Np, always glad to read science stuff
If we are ever going to have computers deliberately made for running AI then they will use ASIC chips, general chips have too many problems for them to be effective and efficient at such a niche task
Hmmm... can you please give me an example of this? I only heard about ASIC chips for bitcoin mining
ASIC is just Application Specific Integrated-Circuit
An example of an AI ASIC chip is the Intel Nervana NNP where it is specific to neural networks and is unable to process other tasks since it was built, to the atomic level, to run the mathematic processes of a neural network.