Sweet 16?

I just saw the article "AMD calls end to core growth on server chips" at Techworld.com. The gist of the article is that AMD has decided to produce server chips with no more than 16 cores. There were some interesting future directions outlined and hinted at by the end of the article, too.

What seemed most disturbing to me was the limit on the number of cores being self-inflicted. Surely we can't have reached the maximum number of cores that are possible to squeeze onto a chip? The whole "right turn" idea to add cores rather than try to cool processors reaching rocket engine temperatures was less than 10 years ago. I'm not sure where the physics starts to overshadow Moore's Law, but I thought I'd  heard that a few more generations of smaller wire sizes in processor dies were still possible. So why not push more and more cores into the same package?

It might be that the average server application (and, perhaps even more so, consumer applications) can't scale well beyond some fixed number of cores. How many cores does it take to type and post a tweet or update your Facebook status or to watch a streaming video? Would any of those tasks be faster or somehow enhanced if there were twice the number of cores available?

If we stop increasing the core counts in the next 5 years, how will new chips keep fulfilling the ever-growing hunger for more performance by consumers? Maybe it won't be about faster and faster application exeuction, but more about less energy consumption while maintaining a level of performance. I guess at some point we'll stop being concerned about Gigahertz or core counts because all processors will be able to do many of the same tasks in about the same amount of time.

I do know that power consumption is going to be a major driving design force as HPC moves closer toward Exascale platforms.  Thus, if the THX-1138 processor draws power twice as fast as the CFM602 processor, I would be more likely to build my system equipped with the former.
For more complete information about compiler optimizations, see our Optimization Notice.

Comments


Thanks, Cyrus.

The point of the blog is to make others aware of AMD's plans and to specualte about what that might mean for other semiconductor companies that make processors (like Intel). Everyone expects more performance from chips every 18-24 months. If we cap the number of cores that are going to be put into future processors, how will we achieve better performance over the previous chips? Maybe execution speed won't be the measure, but something like energy efficiency will be a more important performance metric.

One thing that I didn't mention in this blog was the possiblity of heterogeneous chips. Put some number of general purpose cores into the processor and use the extra space for cores that can do specialized computation faster than a general CPU. Things like graphics or signal processing or streaming video decoding. This is an idea I've seen kicked around for the last 5-6 years. Maybe we'll see something more than speculation in the next 5-6 years.


Raf - "CFM602" is the license plate of the old Rambler station wagon that my parents got to haul the family (2 adults, 3 kids) around when I was much younger.

I didn't want to make any veiled claims on actual products, so I used "nonsense" strings for model numbers.