arstechnica.com — Microsoft is working on 128-bit architecture compatibility with the Windows 8 and Windows 9 kernels. Consequently, the company is also forming relationships with major partners, including Intel, AMD, HP, and IBM.
Oct 7, 2009 View in Crawl 4
myztryOct 8, 2009
*in parallel - The legacy could be virtualised without having to build in system draining functions like capturing access to Program Files and implementing virtual 'shadow' copies of access.Those who get to the point of not needed the legacy can drop the virtualisation and with it the legacy overhead.
tdmethOct 9, 2009
@TsuruchiBrian: If you are doing register partitioning, you certainly do. The code needs to know which register (or register partition) to access the data. <a class="user" href="http://en.wikipedia.org/wiki/X86#x86_registers" rel="nofollow">http://en.wikipedia.org/wiki/X86#x86_registers</a>For example, you can access parts of the RAX via RAX, EAX, AX, AH, or AL. The os has to be aware of which registers it can access. The issue is a lot more complex than just slapping in a few extra registers. The OS has to be aware, The interconnects have to be in place (which exponentially complicates the chip design), etc.
tsuruchibrianOct 9, 2009
@tdmeth When I referred to "adding more registers", I was referring to creating a new architecture that has more registers than our current arches. This would include creating new compilers to create an OS that can use all these registers.You are right that this is complex, but so is creating a new 128-bit architecture. My question, is: What is the benefit to going to a new 128-bit arch with BIGGER(i.e. 128-bit) registers over going to a new 64-bit arch with MORE registers?
paulisleetOct 10, 2009
People like you are the reason that chose to register to vote as a Democrat
alandowneyOct 12, 2009
I'd be happy if Adobe could support it with a 128 bit Flash / Shockwave. They're not even up to the 64 bit, yet.
quantumstatejimOct 17, 2009
What about the plank length? Do you not understand the concept of a floating point number? I am not saying that you won't ever need that precision in scientific calculation, just that the significant figures si the bit that takes up the space mainly.
bobdigiOct 27, 2009
Christ, shut the f**k up you blabbering retard.