|
So there is probably a lot of information about the Pentium 4
that you really don't care about. I'm gonna try and wade
through all of it, to bring you specific statistics, and
technologies (with brief explanations) that make the Pentium 4
what it is. I'm gonna finish up with some basics about why
this chip will rock for us gamers.
First of all - the important stats:
|
400
MHz Internal Bus Speed (New)
800
MHz ALU (New)
Execution
Trace Cache (New)
SSE2
Instructions (New)
Advanced
Dynamic Execution (New)
Advanced
Transfer Cache (New)
Rapid
Execution Engine (New)
Hyper-Pipelined
Technology (New) |
Now, the explanation. For those of you who'd
prefer a very deep technical explanation, you'll need to check
out Intel's page (where I actually got all of this info).
For those of you - like me, who really only need the low-down,
here we go.
400
MHz Internal Bus Speed. There's really not that much to
explain here. To give you and idea of what's going on, I'll
speak about this in comparison with the 'older' Pentium 3's.
The older chips had internal bus speeds of either 100 or 133.
They would then have a clock multiplier of 5 or 7.5 or 10 -
whatever they needed to get to the rated speed of the chip. If
you had a P3 800MHZ then it had an internal bus speed of
100MHz, and a clock multiplier of 8 ( 8 * 100 = 800.... ).
With this newer P4, and it's internal bus speed of 400MHz, you
can see how fast chips will become. Imagine a 400MHz bus with
a clock multiplier of 10.... if you did your math right it'd
be 4000MHz - or a 4 GHz processor. Yeah... WOW! But don't get
your hopes up too high, I've read that they can only take the
existing architecture up to 3GHz. Only up to... as if that
weren't enough! In any case, the internal bus speed of the
P4's will allow for a much higher scale processor.
ALU. This stands for Arithmetic Logic
Unit. This is the part of the processor that will
calculate all of the math involving integers (whole math). The
point here is to be able to have a place to throw the 'simple'
math, and get it back quickly. As I mentioned before, it runs
at 800MHz - quick enough for ya?
Execution Trace Cache. This little puppy will help your
processor communicate with it's self. Part of the processor
works on one project, but has to 'pass' the results on to the
next step. This new execution trace cache will hold the data
that has already been worked on, until it needs to be used in
the next step. You could think of this part as something akin
to a bank. Part of your life is spent making money. The bank
will hold the money that you've made, until the other part of
you is ready to spend it. (Unless you're like me and the bank
holds no money 'cause it's spent as fast as you get it!). The
reason that this is so cool for processors is that
traditionally the individual parts have had problems with
'traffic jams'. Information was ready to be passed on, but the
part that was supposed to get it wasn't ready.
SSE2 Instructions. This stands for Streaming SIMD
Extensions 2. SIMD was a technology that came along after
MMX. We all know that MMX helped with the graphics. So did
SIMD... and now SSE2 will. Notice though, the first word and
last word. Streaming and extensions. Streaming refers to the
gamers all time favorite: streaming media - graphics, sound,
etc. Extensions refers to the fact that this new 'technology'
is an extension to the SIMD technology previously released.
Let me explain a bit here. The chip needs to know what to do
with certain 'projects' that it's given by the user. It has
'instructions' that tell it what to do. MMX was a set of
instructions, so was SIMD, and now SSE2 is. Again, SSE2 uses
SIMD but adds in special instructions for dealing with
multimedia applications.
Advanced Dynamic Execution. Before I can really explain
this function I need to mention another function of these
processors. It's called 'branch prediction'. This is a method
that attempts to predict what will need to be fed to the
processor next. Yeah there are brains inside the brains of
your computer! Well it gets better. You can think of ADE as
the brains inside the brains inside the brains of your
computer. Let's go over that again. The processor calculates.
Branch prediction attempts to help, by guessing what the
processor needs next. ADE attempts to help even further by
keeping track of what worked in the branch prediction and what
didn't work. In the end ADE helps reduce branch mis-prediction
by about 33%. Think I'm B.S.'n' you? Go read the
above mentioned article from Intel that speaks in more
technical terms what I just went over (you'll want to
scroll down to the part titled Advanced Dynamic Execution on
page 4)
Advanced Transfer Cache. This one is easy - not like
the last one! It simple is a statement about the fact that the
P4 can 'get and send' information inside its self faster. For
instance, the P3 exchanged information with the L2 cache at
16GB/s, whereas the P4 does it at 44.8GB/s. See the
improvement? Yeah - faster is better! : ) This also
means that your processor will spend less time just sitting
there waiting for information, and more time actually working
with the information.
Rapid Execution Engine. This is actually similar enough
to the ALU that I'm not going to go into it. All it is
referring to is the fact that the 'simple' stuff will go at
twice the speed (800MHz) of the normal internal clock
(400MHz), so you won't have to be waiting for any for the
'extras' to get done.
Hyper-Pipelined Technology. This is in reference to the
fact that the main 'guts' of the P4 chip have doubled. The P3
had a 10 stage pipeline for processing, and the P4 has a 20
stage pipeline. This will be a clear indication as well as the
internal bus speed, of the amount that the P4 will be able to
handle in the years to come. Having more 'piping' will allow
it to handle more traffic (think of the sewers - I'm not
going to expand on that!).
The bottom line for all of this is that the P4 was designed
with Intel's crystal ball vision of the future in mind. Intel
sees a computing world that will rely heavily on multimedia in
the next few years. I personally agree. The P4 has tested
lower than the P3 on certain applications. These
applications would be the type where the advanced SSE2, and
ADE have little to no hope of doing good. For instance.. how
can the processor 'guess' what I'm going to need next as I
type? But it can definitely guess that when you fire a rocket,
and you've only drawn the first frame of it's trajectory, it
still has to draw the rest.
In any case, the P4 will be very good for the future of
computing/gaming. It will be a lot like the P3 though.
Programmers had to write the code in to their games to take
advantage of the new technologies that became available with
the P3 - the same holds true for the P4. Right now I think the
only game that was written with the P4 in mind was
Sacrifice... So go get your P4 and wait for that next awesome
game to use these screamin' features. Me?... not enough money!
|