Natural Science Forum / Physics / General Physics / February 2007

Kia ora,

I’ve recently done a crash course on microarchitecture, and below is a
brief synopsis of what I found.  I also formulated some algorithm base
blocks, that may need tweaking.  I’m still processing how to best
arrange them into the most dynamic methodology.  This require a little
more comparative research.

The essence of my idea is encapsulated in the acronym “O.L.D”
Translation T series.  The idea allows for a more dynamic and robust
system, that relies on some new algorithms, and refinement of old
standards, and where needed, reordering.

Though in earlier posts, I inferred the idea of time in a continuous
stream, the throughput over power ratio would be wasteful and create a
number of problems such as bottle necks, hotspots, etc.  The mention
of relating random-probability back to a continuous stream is a
natural throughput consideration, and inferred within the method
O.L.D  T -thus quantum computing.

Early estimations of current computer standards show areas for large
improvements and reforms to better align with the principles inherent
with O.L.D.

Within Pipelining:
These areas are, caching, bandwidth usage, prefetch, tree and branch
prediction, register naming, register order, compiling, sequencing,
clocking, floating points, parallel processing.

I am able to go into a detailed methodology, detailing the
implementation of improvements, though here I’m restricted to give a
general aims-list--synopsis.

The improvement results would lead to increase speed and throughput
efficiency, without increasing caching capacity, clocking speed, and
power usage.  In fact, one may and should expect decreases in these
areas, while optimising performances.  It is the new standard facing
microarchitecture, and better aligns with our growing consciousness of
Global Warming, and more efficient fuel consumption.  _Put simply,
consumers want more bang for their bucks.

Pipeline improvement according with the O.L.D principle:
·    Tree and branch prediction,
·    registry ordering & naming,
·    prefetch decode standing.

OLD T methodology directly affects these Core 3 components because
they best work completely synergistically, and improvements will allow
natural flow benefits to increase chances of parallel processing,
dynamic sequencing, dynamic clocking, better bus procedure, and
economical power usages.

Basically, O.L.D. is a multithreading algorithm and principle, with
the purpose of allowing dynamic sequencing, multithreading then
dynamic naming at unique clocking.  The best alignment of these Core 3
will ensure economy of scale, and coupled with new & reordered
measures, their control algorithms allow power scale to reflect direct
usages with minimal→0 waste.  Ideally Zero Waste.

I did a preliminary theory comparison to the simple scalar method.
Without going into great detail, OLD. T allows for savings in a number
of areas.  And  if I may be so bold to claim without tests and theory
regimes listed, the elimination of entire procedures.  If I didn’t
have a bug on my pc, I would have a proposal of microarchitecture,
logic handlers, and testing methods ready in a few days.  So for now,
all my notes remain in raw data form, with little explanatory
appendages.

Addressing the issue of a trust regime, I have a number of ideas that
will hopefully fit the appropriate billing.  In essence, a requirement
of storage space will serve an interface registry, without access
protocols to the core functions to ensure the integrity of the heart
registry while maintaining full functionality.  This is achieved
through the interface registry storage server.  The cost factors can
be minimised through a 4bit protocol… This has a limiting effect on
the desire to reach the most optimum resolve, but further cost
reduction can be created by allowing the protocol to serve only in
threatening circumstances, such as internet financial exchanges, and
downloading protocols.  The method allows for normal functionality and
exchanges, but ensures machine integrity is never breached.

The consideration of naming will be directed into trial confinement,
and installations will be granted write privileges in the interface
registry, and read privileges only with appropriate restriction in the
main registry.  Parallelism and multithreading will be used when
appropriate, and writes will follow OLD T logical tree building, the
cost factor thus is kept minimal.

As mentioned earlier, the K base will be established. One of its
primary jobs is storage of exchanges in global protocols for
evaluation, but especially monitoring the interface registry.  This
bases will allow assessments pinpointing of wasted and illegal
protocol, and establish whether any write protocols conflict with OLD
T logic.

OLD K logic’s full value is realised when its usage works
synergistically with OLD T and further divides all the branch prefetch
to its most efficient line.  OLD K will identify all protocols that
disable or create conflict, and order it into conformity.  If no
resolves are found, rouge code can be patched in 0ª state allowing
bypass, or prompt-deleted.

What does this mean?  With the advent of internet hacking and crime,
it will be impossible for anyone to access the computer core.  The
protocol will serve the most valuable need in today’s modern
computing,  Guaranteed Computer Security.