Why have computers been stuck at around 3 gigahertz for years now? And where's Pentium 5,6,7 etc?

Question

Back in the 1990's we went from 100Mhz to about 1.1Ghz. I could easily see the progress computers were making back then. Now, every time I check the latest computers they all top out at around 3.06Ghz. Surely, I MUST be missing something?

Answer 1

Well there are physical limitations on how fast a clock signal can be. I think both AMD and Intel realized that you can improve the speed of a CPU without adding more speed to the clock. They are improving on things like true threading of execution and shortening of execution pipes. To prove the point, a 1.5 Ghz Intel Pentium M processor was "faster" than a 3.0 Ghz Pentium 4. I don't think you'll see another clock speed jump anytime soon but you will see faster CPUs.

Answer 2

Newest processors that I have seen being sold are running around 3.7MHz (Intel Pentium Extreme Edition 965 Presler 1066MHz FSB LGA 775 Processor Model BX80553965). The main reason why in my opinion as why they have not jumped up higher like you were describing is that there has been a lot of development into 64 bit processors verses the standard 32 bit that most systems run on today. The technological jump from moving from 32 to 64 bit processors is light years higher then continuing to make 32 bit processors running at 4-8 Ghz by increasing the allowed RAM for a computer from 4GB today to "17,179,869,184 gigabytes or 16 exabytes of RAM"!!!!!!!!
Also there are larger cache sizes of today's 3.0 verses 3.0 two years ago (512KB vs 2MB today)

Answer 3

You can't always rate a processor on MHz. It's about the number of units of work it can complete for each clock tick multiplied by the actual clock speed in MHz.

You also have to consider than AMD spent time working on a 64 bit processor instead of simply making their existing processor faster. Granted 64 bit technology isn't catching on quickly, but it will get there.

AMDs chips for a given clock speed do more units of work per clock tick than a Pentium at the same clock speed. That's why AMD labels their processors like 3500+ Actual clock speed is more like 2.5GHz or something, but it compares to a Pentium 3500MHz.

Answer 4

wow, you are seriously missing something
You can't determine things from their outside ghz speeds
Well for starters
the single core 3.77ghz Intel Extreme has been out for 2 years now
3ghz is pretty much old technology now

AMD is probably the leading in gaming cpus and cpu performance
FX-57 2.8ghz stock, thats faster than the 3.77ghz dual core (not single)
yes, that is true from tests, and guess what... if you have a phase changer, say hello to a 3.6 ghz AMD, equal to a 5ghz Intel
study the architecture of both (don't feel like it for me) and u'lll understand. or u can just suck it up and learn

Answer 5

pentium 5=core duo
pentium 6=core2 duo

they rebranded the pentium line as Core...


they stopped making jumps in processor speed beecause the computers running the really fast chips (there was a 4.0 Ghz chip for awhile) where overheating...

really overheating...

and they used a lot of electricity...

plus, now the chip manufacturers are making more powerful chips that run cooler and use less electricity by running at lower clock speeds..

Answer 6

Yes, AMD made the jump for 64 bit processors, and Intel is still behind somehow. In the meanwhile, Microsoft came out with XP 64bit, so the trend will go in that direction. Belive me, 3 GHz is too much power for too many applications today...

Answer 7

there is much more to the speed of a processor. the speed of clock cycles hasn't moved very fast, the the power difference from the first 3 gig pc's to the latest dual core pc's is night and day

Answer 8

the Pentium 5 and 6 if you will are the dual core and Pentium extreme processors check out Intel.com for all the info

Answer 9

they are not exactly stuck at 3GHz. now it's up to 3.6GHz. BUT do not forget the new tech called (Dual-Core) which i have on my PC that i bought last week. the speed of mind is 2.8GHz. But since i have Dual-Core, means that my cumilitive speed is (2.8 * 2) which mean that the speed 5.6 GHZ. And that's really high....

Answer 10

Nope, you are not missing anything. the manufacturers hit a brick wall in the voltages and SIZE of the chips. If you will remember, the
first Pentium chips were about 5.00 volts, then they started dropping to 3.52, into the 2 volt range, and now into the 1 volt
range. It takes TIME, even at the speed of light to get a wire or transistor from Zero ( Ground) ( which can be thought of as a ZERO
bit ), up to the working voltage, say 1.8 Volts. This lag is just to
fill the wire ( In a CPU it would P channel or N channel doped
silicon material ) full of electrons, or to drain a conductor of all electrons. By lowering the PEAK voltage, one can speed up the
process. Think of it as climbing a 45 degree hill. If you climb a 5
Unit hill, it takes twice as long as it would to climb a 2.5 unit hill.

The second factor is the size of the wiring and traces and transistors in the CPU. Technology has come a LONG way from the first integrated circuits, which would only have a few thousand transistors. The first CPU was generally called the
2002, a 2-bit processor for calculators from Intel. It had roughly 2000 transistors. I have a photograph of a celebration at Intel the day when they hit the "profit" ratio for manufactured chips that passed inspection - 10%. Then came the 4004 ( four bit with roughly 4000 transistors. ) then the 8008, the 8080, the 8086, then the
infamous 8088 used on the first IBM PC. The transistor count has gone up, as you pointed out from 8 bits, to 16, to 32, to 64, BIT, CPUs with many Millions of transistors. If you remember the
SLOT 1, or 'chocolate bar chip" that was common a few years ago, this is an example of manufacturers trying to get millions of
transistors and CACHE as close together as possible without
making a 6 inch square CPU. The Slot 1 assembly had typically 6 chips on it, using the standard chip "size" and voltage of the day, and had very powerful use of the L1 and L2 cache.
But, -- the SIZE of the traces has reached a manufacturing limit using "standard" techniques. Now you get "Dual Core" CPUs,
and the CACHE ( L1 and L2 --level 1 and Level 2 ) is reaching 2
Megabytes, in order to speed up the CPU's output, since the
speed, and size has maxed. I have read of new experimental techiniques with single ATOM channels for CPUs, and other
techniques to make the size of the parts smaller. The other practical limit, that you should easily guess, is the HUGE power requirements of the new CPU's, since the average first PC might have a 100 Watt power supply, and now 600 Watt supplies are common. This power as to "go" somewhere, and it generates tremendous heat in CPU and must be cooled, or the traces and
layers will break down. Finding ways to get rid of this heat is a major concern. One guy in Antarctica was able to overclock his 800 MegaHertz CPU into the 2000 MHz range, but, not eveyone can keep their computer at Minus 72 degrees.
As someone pointed out in an answer above, CPU "clock frequency" is not the entire factor in "speed", and better written Cache programming, more cache memory, better design,
and even more efficient software will make a computer faster as well.
A while ago, computer designers discovered Russian computer experts, who, lacking the resources of Silicon Valley, had designs and computer algorithms which did twice the work of the rather bloated "standard" programming of North America. By using
better designs, and clever programming, with LESS code, the
CPU's are now doing much more work than before.
Something simple like CACHE RAM, with better usage can radiacally improve a CPU's output. My explanation of cache,
in an earlier answer is as follows:
__________________________________________________

cache is memory, just like any other memory on your computer.
The difference between RAM ( random access memory) and cache
memory, is the location and speed, and how it is used.

It was discovered that by PREDICTING what memory locations
" Were GOING to be used ", the computer could speed up the
the process of reading the next set of memory locations,
by having a "predictor" program load the memory into VERY
fast memory, close to, or inside the CPU ( Central Processing UNIT ie the Pentium chip ). There are different levels of Cache RAM, usually called L1, L2, L3 etc. with
L1 Cache being built right on the chip itself, or on chips
built onto the CPU chip module at the factory. L2 Cache ram
is usually added in slots very close to where the CPU ( Pentium), is located on the motherboard. Believe it or not,
the computer chips run so fast, that electricity, moving at the speed of light, takes TIME, to travel across the circuitboard, so that the L1 cache is in the heart of the CPU, and the L2 add on cache is as close as possible to the
core to speed up the transfere.
In early Pentium ( and AMD, Cyrix, IBM, etc. clones,) the
L1 cache was controlled by what can be considered an entire computer WITHIN the CPU. This "computer" had a program, with
its own ram, and rom, and it was pre-programmed at the factory. Its job, was to watch what the (big) computer was doing, and predict, what the computer was going to do NEXT. For example, if you were running a program that used program lines 345, 346, 347, 348, then it would predict that you
would use lines 349, 350, 351, 352, etc, and load them from the " SLOW" ram cards in the ram slots, into the blazing fast Cache ram, right inside the CPU itself, where the
CPU could access it immediately. There are different kinds of cache ram within the "Cache" " Predictor" program, such
as TAG ram, Dirty ram, etc. which are used to keep track of what ram is loaded, used, and changed, by the running software that the CPU is currently using. Tag ram would keep a list of the ram locations that were loaded into the L1
cache, and the Dirty ram, would keep track of which L1 cache
locations had CHANGES made - when the "predictor" program
decided that NEW locations of slow RAM were going to be used, it had to put back the old RAM locations' memory contents. It Could ignore memory values that WERE NOT CHANGED ( saving time otherwise spent writing the information BACK to the slow RAM ), but would have to write
back any CHANGED ram back to the slow ram (marked dirty).
AMD invented very clever and powerful "predictor" programs in some of its early chips, and this made such a dramatic difference that it out-performed the much more expensive Pentium chips. The "predictor" computer program component is now extremely important in achiving faster speeds. CPUs typically have 64k ( K is one thousand (( 1024 bits) ), 128 K, 256K, or 512K of cache built in, and many motherboards can have slots near the CPU that the user can add more
L2 cache ram, for use by the predictor CPU program section.
Cache ram is typically much faster, but much more expensive to construct in the factory, so that chips with more cache are usually a lot more expensive. Since CPU speeds are reaching a limit for the moment, the manufacturers are increasing the L1 CACHE to 1 or more MEGA BYTES, on the high end chips.
SOOooooo... cache is not "just" fast memory. It is a program, inside the CPU, running constantly to predict what memory is "going" to be used. This program goes out to the slow ram, and grabs the memory contents, and loads it into
L1 cache BEFORE the computer asks for it, to speed up memory transfere. If the predictor is correct in GUESSING the next memory locations, the entire computer speeds up greatly....



_________________________________________________


There are 4 -CPU, Dual -core, motherboards available, since just putting "more" of the same old CPU will increase the speed.
As you pointed out, there is an almost halt in the given clock speed at the moment, but researchers around the world are
VERY busy finding new breakthroughs. I would hazard a guess, since I have been watchng the computer industry since 1978,
that you will see blazing speeds once again, in while, when so-called
experimental chip fabrications technologies are brought to the factory floors.
You can easily DOUBLE the speed of any computer you want, if you can find a way to keep it at about -72 degrees.... ie. getting rid of the heat is not a trivial factor.

There dozens of articles in all the Computer and Scientific magazines, and you can search the web for articles on experimental techniques, but everyone has accepted that at the moment, speeds are increasing VERY slowly. you might notice in the COMPUTER magazines, everyone is selling LIQUID cooling,
and going to extreme lengths to cool cases and CPUs and Video cards ( which really are CPU's ! ).
The as-yet-undiscovered techniques of the next generation of CPUs will have lower voltages, lower POWER consumption,
smaller fabrications, and ingeniuos ways to get rid of the heat.
Better use of caching, better algoithms in CPU process programming, and Dual/ Quad/ Multiple Processor techniques
will also be used.

Until then, you just have to wait!

hope this answers your question

robin