Transmeta again...

When I was on the communities of Loongson, I suggest them to seek another way to support Microsoft Windows. The way is that using some technologies similar with Transmeta processors, interpreting the x86 or x86-64 codes to the underlying MIPS64 instructions, but that is not the final target. The final target is to trap the Microsoft Windows to run the programmes written in the Loongson instructions sets with help of that smart interpreter. The benefits are that Microsoft does not need to develop another version of Windows for Loongson, and Chinese could find ways to spy the executing codes one by one, and secure the securities they need. I was laughed and denied there. Yeah, a little bit ridiculous. But...

 

I believe the future O/S should have a change for its essences, because it is more or less a machine, a software layer necessary but not always important. Because what we end users need are the applications, all kinds of applications, rather than to enjoy or be limited by the stupid framework of O/S provided by Microsoft or something else. So emphasising the Applications is much important than the passed more than 20 years since Windows NT started to dominated the computer world. From the aspect of the design of processor, there still leaves something overhead and not that important thing. That is the translation from the ISA towards its underlying core, designers tried to use many ways around them, and so many microarchitectures build up the stories for more than 3 decades. Why not just write applications with the codes for its core not some a certain ISA? Or in other words, a processor just needs to make sure the capabilities to compatible with current O/S, but let the O/S out of the charge of what kind of codes written for applications. If APPs for Android could not be run on iOS; and APPs written for iOS, they could not either work on Windows. So why should the programmers care about the things they should never care. Today people are using APPs in essence, and APPs should gain the 100% power from the processors, for many things to consider: power consumption is the first, because the resource is limited, they could never be reborn. And the good thing is that the future compatible PC would lost their dominance. The issue of compatibilities would be ignored eventually. So the processor manufacturers would not need to keep to be compatible with each other. They just need to support O/S, but not 100% controlled by that O/S.

 

If the future go that way, the good thing is that processors from both AMD and Intel could support Microsoft Windows, but the applications are not compatible with each other. This will help they both compete the market much fairly than ever before. It is just like PS4 and Xbox One. If you want to play Need for Speed on your PS4, when you have enjoyed it on the Xbox One in your friend's home, you just need buy the same game for PS4. You would also find the differences but you would eventually find the things worthy spending money on. If you love the Forsa Motorsport, then you have to get another Xbox One, then you own both of them. That is a good thing, you have chance to have them both, they are different!

 

So the future AMD processor should go some way like what Transmeta and HP Itanium did, just provides enough compatibilities, and eventually differ themselves from the Intel ones at levels. In other words, supporting the O/S should be split from supporting the applications in the very future...