Used to be you could buy a new computer every 3 years and get 2x the performance. Not anymore. Performance has hit a wall - or at least a steep hill. What’s this mean for the industry?
Moore’s law
Moore’s Law says that the number of transistors on a chip will double every 18 to 24 months. But Moore’s Law has been simplified to mean a doubling of performance every 18 to 24 months.
Not anymore.
Transistors ≠ performance. Yes, clock speeds have improved from the 1 MHz 6502 processor in the original Apple II to over 3 GHz today. But clock speeds have leveled out: in a third of a nanosecond light moves about 4 inches or 10 cm - and electricity is slower than light.
Another big piece of improved performance has come from wider data paths. Chips now move data in 64 and 128-bit chunks, rather than the 6502’s 8-bit bytes. Not much more growth there, either.
We’ve thrown transistors at performance issues: more and wider registers; bigger caches; deeper pipelines; intelligent branch prediction; smarter I/O management; and thousands of other enhancements.
More RAM? We’ve also been adding ever-larger on-chip caches that improve performance. SSDs further improve I/O through lower latency, an area we’re still learning about.
Multicore
We can’t make processors go faster. We can’t process more data per clock cycle. So how do we put twice as many transistors to work?
Stuffing more processors on a chip. And right now many of the brightest minds in computer science are struggling with the problem of getting usable work out of 8, 12 or 16 core CPUs.
Dual and quad core processors work pretty well because multitasking runs a lot of background threads. Those threads can use multiple cores and improve performance.
But outside video, image, voice and scientific apps, most personal apps - don’t need multicore architectures. Humans aren’t good multi-taskers.
The wall
We’ve hit a wall. We can still double the number of transistors. We can still double disk drive capacity. We can build faster interconnects, such as QuickPath, Light Peak and 10 Gb Ethernet. And SSDs also help performance.
But the easy wins are over. Going forward performance gains will be measured in single digit percents each year.
Implications
Information technology is driven by consumers, not the enterprise. What happens when a new PC is only 20% faster than your fully paid for three-year-old PC?
If it is a notebook, it can be smaller, lighter, more stylish and more rugged. But the desktop?
Future model differentiation will have to move on. Here’s where:
- Power. The server space is making greater power efficiency a differentiator. Mobile’s been pushing this for 15 years. You’ll see more.
- Integration. Open up in iPad or a MacBook Air and you see a tiny PC board, a few chips and a huge set of batteries. Battery life makes products convenient.
- Functionality. Integrating multiple applications, each with their own dedicated core, may enable consumer devices to collapse multistep workflows into a single devices. Capture, voice-edit, compress and upload video from a single candy bar sized device?
- Cost. The first low-res digital cameras cost hundreds of dollars and now they’re almost free. Huge market among the billions who live on less than $2500 a year.
The Storage Bits take
Moore’s Law driven market growth isn’t over. We can use our still growing technical capabilities to refine what we already do.
But the days of newer=faster are over. It’s newer=better: less power; smaller; cheaper; and - in cases like SSDs - overall system performance will improve too.
The good news for storage is that data production will continue to grow. Always on, always available consumer data systems will create ever more demand for storage.
In the enterprise this will affect storage architectures as well. When you can’t scale up, you have to scale out. Decomposable storage architectures will come to the fore.
Comments welcome, of course. The Apple ]['s motherboard style was the same as today's MacBook Air: a few chips on a PC board. Friends were always startled to see empty my Apple ]['s case was.
