CPU Quickie Definitions

Toggle sidebar Toggle sidebar
I

imported_Skorpio

Senior member
Aug 29, 2004
283
0
0
Basically I know that there are two types of cores....intel cores and amd cores......

Could any of you tell this newbie....what are cores? example: Prescott, Barton, Northword...etc etc

Is newcastle a good core? haha...good..lol

What exactly does the FSB do? Why is a larger number FSB better?

What is up with AMD's rating system? wh are their ratings like 2800+, 3500+?

What are micron's? Is a smaller number better because of less electron resistance?

Mucho thanks appriciated for helping this newbie out :)
 
I

imported_michaelpatrick33

Platinum Member
Jun 19, 2004
2,364
0
0
This would required a massive answer. I would go to some major websites and look at chipset/cpu reviews and gain an understanding.

www.anandtech.com
www.arstechnica.com
www.aceshardware.com to name a few out of hundreds

FSB is basically the speed at which the cpu and memory controller share information. AMD 64's have no FSB per se anymore because the memory controller is on the cpu itself and runs at full cpu frequency. However they still base their cpu multiplier (and HTT [the clock generator for the processor to northbridge communication][that is 200 x 5 =1000ddr] off the memory DDR frequency (which right now is 200DDR1)

The Intel memory controller runs at a 200mghrtz quad pumped (4 instructions per clock tick) for an effective 800mghrtz (the new Intel 3400EE with 925X board has a FSB of 266 quad pumped). The AMD64 memory controller has much lower latency and better performance but is not easily upgradeable due to having to do a major core revision. That is why it has advantages and weaknesses to have an integrated memory controller.

AMD rating system is simply a marketing tool created to show why a slower clocked processor runs as fast as a faster clocked processor. Frequency is not the only indicator of cpu performance. There is efficiency and design and instructions per clock etc.

Micron is the space between transistors on a core I think. The smaller micron the more transistors one can fit but the more problems (heat, gate leakage etc) can occur. Intel is now running into a wall trying to crank up frequency because of the heat and leakage in their latest cores.

There are many more cores than AMD and Intel Those are only two of the major X86 core companies. You have SUN, and IBM with their own RISC based cores along with a lot of other companies.
 
M

MiranoPoncho

Golden Member
Oct 7, 2004
1,441
0
0
Welcome!
Cores are the names for architechture that intel and amd give thier chips to diffirentiate from eachother.
F.S.B. is an acryonm for front side bus. Look at it like this. the 800 fsb, is actually a 200mhz, but puts four bits of data per clock cycle ie pulse. now thats what the data is shunted through. Prettymuch the width that data can travel, as the pontiac comerical said: wider is better. More fsb= more data able to be crammed. Now you ask of amd's rating system. It's has to do with the diffrences in the way the chip is manufactured. they amd chips are just more efficent per clock cycle than thier intel counterparts. look at it like this: the p4 is like the hoover dam. it has a lot of water(data) and it is really really big. the amd is more like venician aqua ducts, shoter, but more condensed. as for thier pwoesses, i can use this analogy: who would be better at the game(s) limbo? yao ming (intel, wide long pipeline) or gary coleman(short, small pipeline)? coleman of course! yao is big and tall, but dosent work well with 3 foot limbo polls. now take construction work(Compiling) coleman wins with his vertically challanged briggade, as yao and his fellow team mates are too tall to use a shovel. I ask you this for (media encoding) who can dismante and reassemble a bomb quicker? data, superman or the flash? basically, the amd rating system means, that because of it's shorter length, data travels a shorter distence, and it's more efficent. a 2800+ athlon 64 should intheory preform or out preform a pentium four, 2.8ghz, even though there is a 1ghz deficit. hope this wasnt confusing. happy computating.
 
G

gobucks

Golden Member
Oct 22, 2004
1,166
0
0
Hey drpootums, just out of curiosity, why no PCIe/socket 939? A 3500+/nForce4 Ultra/6600GT would perform the same, maybe for cheaper, with a much better upgrade path.
 
W

walla

Senior member
Jun 2, 2001
987
0
0
A micron is a unit of distance equal to 1x10-6 meter, or one one-millionth of a meter.

The micron process refers to spacing of design component for VLSI circuits. Without going into too much computer engineering detail, the small this number implies a greater number of transistors on a chip (quadratic relationship...halving the micron process will allow for 4 times more transistors based on area).

Of course, cramming more hardware onto a chip without a similar scale decrease in power consumption means power hungry and hot chips. There is also problems with line capacitance, too, because signals that are closer together will want to interfere with each other to a greater degree (capacitive/inductive crosstalk). So a lot of research is going into how to reduce these types of restrictive effects.

Thats probably more thant you wanted to know.
 
V

VIAN

Diamond Member
Aug 22, 2003
6,575
1
0
Intel is a company that produces CPUs
AMD is a company that produces CPUs.

A core is the CPU itself. When we refer to this core or that core, we mean which revision of the core of the current line of processors.

Intel's line of processors is Pentium4, which includes the following revisions:

Williamette: 180nm, 20 stage pipeline, 8KB L1, 256KB L2

Northwood: 130nm, 20 stage pipeline, 8KB L1, 512KB L2

Prescott: 90nm, 31 stage pipeline, 16KB L1, 1MB L2

When I use L1 or L2, I talking about cache levels 1 and 2. Level 1 cache is always smaller than level 2 because it needs to be much quicker. The bigger the cache the slower it becomes.


AMDs 3 lines of available processors are: Athlon XP, Athlon64, Sempron

Athlon XP:
-Palomino: 180nm, 256KB L2

-Thoroughbred: 130nm, 256KB L2

-Barton: 130nm, 512KB L2

-Sempron: 130nm, 256KB L2

Athlon64:
-ClawHammer: 130nm, 1MB L2, 64-bit support

-Newcastle: 130nm, 512KB L2, 64-bit support

-Sempron: 130nm, 256KB L2, no 64-bit support

Sempron is marketed as it's own processor line, but are based off of the other two, for this reason, I just noted it as a new core instead.


The FSB is just the pathway from the CPU to the Memory Controller. Increasing the FSB increases how fast the CPU commnunicates with the RAM. But if you increase just the FSB and not the RAM, then you are not gaining any performance. You must increase both. Athlon64 processors have the Memory Controller built into the CPU unlike other processor and therefore the FSB is the speed of the CPU itself.


AMDs rating system is there to fight against Intel's high clockspeeds. Clockspeeds aren't everything, but consumers think so. That's why Intel made the Pentium4 the way they did. These performance ratings are AMD's hopes of people seeing those numbers instead and think it's the clockspeed.


The press that's used to create the CPU is called a .xx micron press. Above I showed .15microns or .15um or .15micrometers or 150nm or 150nanometers. A nanometer is 1000 times smaller than a micrometer. We started to use 90.nm instead of .09um recently. This measurement, from what I know, is the size of one transistor.
 
I

isaacyang

Member
Nov 9, 2004
29
0
0
Williamette is 180nm
Palomino is also 180nm
Originally posted by: VIAN
Intel is a company that produces CPUs
AMD is a company that produces CPUs.

A core is the CPU itself. When we refer to this core or that core, we mean which revision of the core of the current line of processors.

Intel's line of processors is Pentium4, which includes the following revisions:

Williamette: 150nm, 20 stage pipeline, 8KB L1, 256KB L2

Northwood: 130nm, 20 stage pipeline, 8KB L1, 512KB L2

Prescott: 90nm, 31 stage pipeline, 16KB L1, 1MB L2

When I use L1 or L2, I talking about cache levels 1 and 2. Level 1 cache is always smaller than level 2 because it needs to be much quicker. The bigger the cache the slower it becomes.


AMDs 3 lines of available processors are: Athlon XP, Athlon64, Sempron

Athlon XP:
-Palomino: 150nm, 256KB L2

-Thoroughbred: 130nm, 256KB L2

-Barton: 130nm, 512KB L2

-Sempron: 130nm, 256KB L2

Athlon64:
-ClawHammer: 130nm, 1MB L2, 64-bit support

-Newcastle: 130nm, 512KB L2, 64-bit support

-Sempron: 130nm, 256KB L2, no 64-bit support

Sempron is marketed as it's own processor line, but are based off of the other two, for this reason, I just noted it as a new core instead.


The FSB is just the pathway from the CPU to the Memory Controller. Increasing the FSB increases how fast the CPU commnunicates with the RAM. But if you increase just the FSB and not the RAM, then you are not gaining any performance. You must increase both. Athlon64 processors have the Memory Controller built into the CPU unlike other processor and therefore the FSB is the speed of the CPU itself.


AMDs rating system is there to fight against Intel's high clockspeeds. Clockspeeds aren't everything, but consumers think so. That's why Intel made the Pentium4 the way they did. These performance ratings are AMD's hopes of people seeing those numbers instead and think it's the clockspeed.


The press that's used to create the CPU is called a .xx micron press. Above I showed .15microns or .15um or .15micrometers or 150nm or 150nanometers. A nanometer is 1000 times smaller than a micrometer. We started to use 90.nm instead of .09um recently. This measurement, from what I know, is the size of one transistor.
Click to expand...

 
You must log in or register to reply here.

TRENDING THREADS

COMPANY
RESOURCES
FOLLOW
Top Bottom