Multicore is unstoppable: the economics are driving it. Therefore we need software to exploit the huge opportunities that these new architectures offer. This is a permanent change that affects all computing.
Even if you are a decision maker in your organization, maybe you did not notice it, but someone in your payroll should be paying attention to it.
Or worse, your competitors are doing it…
Based on years of working in Multicore and Parallel Computing Open Parallel presents these recommendations for a basic Corporate Multicore Strategy
- Staff Survey
Start by doing an in-house survey about your capabilities. Maybe some of your developers are active Open Source Software (OSS) contributors in their spare time and have some of the skills that you will soon need.
Even better, they know more people that are keen to help
- Hardware replacement
Analyze which is the right Multicore hardware for you and how it will scale in the foreseeable future. Remember that vendors are ready to sell. Learn more about the opportunities offered through the Cloud.
Start to think about everything that can be optimized: floor space, energy consumption, maintenance, performance, and code. Multicore will be able to give you a positive answer to some if not to all of them.
- New Projects and Products
Multicore is not only about doing the same things faster. It is also about adding more options and features to the same problem with much more reliability. Think and explore different approaches to your markets, and don’t constrain your imagination. Dream about an option that was not possible before. You will be probably able to make it happen now
- Training in Parallel Programming
Start now to discuss with your staff how to train them in the new software techniques. If you need to create a Chair in parallel programming at the local university, start to do it now. It will be a good investment, and not just for marketing reasons
- Keep an eye on Universities
Some of them have been doing research on parallel programming for decades. They can have different models, experience in strange languages and maybe their recommendation is not what you are expecting. But their perspective and global network of talent is what you need to look forward 5-10 years in your strategy
- New Management Models
Be prepared to use other’s innovations as well as their source code
Be prepared that others will use yours.
Learn about Open Innovation
Look at new management models, either from fast growing companies or how OSS companies are moving ahead or how others are failing to deal with the OSS model
Do not hesitate to talk to multicore companies from New Zealand and abroad; confusion is everywhere and they will be happy to hear your experiences in exchange for theirs. There is room for everyone at this stage, but move fast: this window of opportunity won’t last forever
- Look outside your company
You will need to do joint ventures with IP providers. Be prepared to accept that independently of the country of origin, they can be better than anything else that exists in your own country. Innovation has no boundaries (and usually happens in the fringe)
- Piggyback on vendor’s ecosystems
They designed these chips and they need that software developers learn how to properly unveil their potential. They will do whatever they can to help sell your solutions, if this mean that they can keep their customers happy and sell more hardware.
They are learning to build solutions in a niche too, but it is not in their core DNA; it’s easier for them to partner with you.
Do not waste resources trying to create your own channels; look for the best synergy in a major vendor (or better, in a pool of them) and go for that ecosystem.
This is a paradigm shift in the market that could take 10-15 years to mature. Better to start now.
Good luck in your journey.