AMD and Intel processors, past and present generations Desktop PC?

[Zorim]

Any one who you can point me to website that shows early generation AMD and Intel desktop pc's that been released to the public until present. I hard time searching the net . i need that data for my homework. One more, to pair processors between amd and intel, is that the speed, by bit, or by date released.

Keep it UP!
Zorim

 

[Papa-Ming]

Something Like this:

http://www-106.ibm.com/developerworks/library/pa-microhist.html?ca=dgr-mw38MicroHistory

Or someone's little essay: ( sorry, don't know his name. )

AMD:
AMD first helped INTEL make their early chips like the 086, 186, 286, but they stopped making 386 because they were too similar to 486s. Then AMD just made “clone” chips. They earned the reputation of having cheaper, less reliable chips.

AMD later released the K5 core processors to compete with the original Pentiums and it ended up failing behind the Pentium 2.

In response, AMD released the K6 processor. It performed well when compared to its competition. Then came AMD's “big day” the K7 core.

It was made to kill the Pentium 3. It did so, forcing Intel to play catch up. AMD was able to release a 1GHz chip before Intel. They were also able to release chips afterward with higher clock rates. Intel was in a very bad spot, so they released the Williamette Pentium 4.

It was built using the 180nm process with 256K of level 2 cache. It sported a much lower IPC so it could gain high clock rates than the Pentium 3. However, when it was released with clock rates of 1.4GHz it was pummeled by AMD’s code-named Thunderbird Athlons.

Intel later released the Tualatin Pentium 3, which was built using the 130nm process and sported clock speeds as high as 1.4GHz and had 512K of level 2 cache. Meanwhile, AMD launched the first version of the Athlon XP, code named Palomino.

These chips again outperformed the Williamette Pentium 4s and the Pentium 3s of the day. Then Intel launched the Northwood, which was a large success for Intel. It was a Williamette with twice the level 2 cache, and it was built using the 130nm process, so it ran cooler and had a higher overclocking potential. AMD kept up with the Palomino core until the 2100+ (1.73GHz) model rating. Then they deputed with the Thoroughbred A core, with speeds up to 2200+ (1.8GHz).

With Intel launching models with higher clock rates than AMD’s, AMD was forced to come up with a solution. At first, they were going to make an early release of the Barton core. They thought that the extra cache would help them until they could release their Hammer cored CPUs.

Then a solution was found: add another layer of metal. This new core was called the Thoroughbred B. This was a great success for AMD. It was about the same in performance to the Pentium 4s up to 2.8GHz. Then they launched the Barton core processor.

It wasn't especially ground-breaking, but it was an improvement over the Thoroughbred B core. Barton was basically a Thoroughbred B core with twice the level 2 cache. It was just enough to get AMD through until the Hammer core was ready. It wasn't a huge success, mainly because the model numbers were off, but it still was a nice core with nice overclocks.

It ran on the 333MHz frontside bus and two (3000+ and 3200+) were used the 400 MHz FSB. The 3200+ was closer to a 2.8-3GHz Pentium 4 in performance than a 3.2GHz. The 3000+ was closer to a 2.8GHz Pentium 4 than a 3GHz Pentium 4.

After all this AMD decided to launch the Hammer core, in the form of the Opteron first, and later in the Athlon 64 form. Socket 754 was the home of single channel Athlon 64 processors. They went up to 3700+ in model ratings and 2.4GHz in clock speed. These chips were very powerful and put AMD back on the map. 3700+ chips are about the same as a 3.6 GHz Prescott or near 3.4 GHz Pentium 4 EE (Extreme Edition).

AMD also released the Athlon 64 FX series processors to compete with the P4EE. They did well, but they cost about $850 USD. Nice chips with large price tags, but the multiplier was left unlocked, as it was aimed at an enthusiast audience. The highest clocked FX chip (the Athlon 64 FX-55) at this point in time runs at 2.6GHz.

AMD's next step up in clock speed should be the FX-57 at (possibly) 2.8 GHz with 1MB of level 2 cache. It may be manufactured using the 90nm process, or it may use the older, and more proven, 130nm process. That won't happen until mid 2005 at the earliest. AMD's socket 939 3000+ and 3200+ 90nm chips run just as cool as the “old” Newcastle core 130nm processors, although AMD reports the 3500+ (2.2GHz) to have a higher TDP than the 3000+/3200+. It should be interesting to see how long it will take AMD to release higher clocked chips on the 90nm process (as in 2.4, 2.6 and 2.8GHz).
The 90nm die shrink has helped AMD in two main ways:


1.) It has reduced the core size by about 50%, reducing cost of production;
2.) It allows for higher clock rates.
There may be slight issues with these chips, but by now, BIOS updates have been released to correct the problems.

AMD has already demonstrated dual core Opterons and they seem to run fairly cool considering they are essentially two CPUs stuck onto one PCB. AMD said that they had single core Opterons at 45W. These chips will probably find their home in 1U and blade servers. Dual core Opterons are only around 95W, which is better than a single core Prescott Pentium 4. The Opteron was AMD's first x86-64 processor, but it was intended for use as a server processor. AMD didn't release the Athlon 64 until later.

AMD's model numbering:
AMD first launched the Athlon 64 on socket 754 with model of 3200+ (2GHz, 1MB L2 cache; later 2.2GHz, 512KB L2 cache). After, the 3000+ (2GHz, 512KB L2 cache) and 3400+ (2.2GHz, 1MB L2 cache; later 2.4GHz, 512KB L2 cache) were introduced.

Then came the 2800+ (1.8GHz, 512KB L2 cache), then the 3700+ (2.4GHz, 1MB L2 cache). The 3200+-3700+ were first made as Clawhammer chips, meaning they had 1mb of L2 cache. AMD knew how large and expensive these chips were, so they made the cheaper Newcastle core that has only 512K of L2 cache. To offset this, AMD increased their respective clock speeds by 200MHz so their model numbers wouldn't be affected. AMD only clocked these chips up to 2.4GHz so a 3700+ on the Newcastle core was never made. Now the main chips with 1MB L2 are the desktop replacement (DTR) chips, intended for use in notebook computers.


AMD also launched socket 939 chips for desktops. Their 3000+, 3200+, 3500+, 3800+, and 4000+ chips were clocked at 1.8GHz (512KB L2 cache), 2.0GHz (512 L2 cache), 2.2GHz (512 L2 cache), 2.4GHz (512 L2 cache) and 2.4ghz (1MB L2 cache), respectively. These model numbers are slightly askew compared to their Pentium 4 rivals. The 3500+ is closer to a 3400+ rating, the 3800+ is closer to a 3600+ rating and the 4000+ is closer to a 3800+ rating.
Later AMD launched the 3000+ and 3200+. These chips, though, are manufactured using the 90nm process. AMD also has since released the 3500+ on the 90nm process. The 90nm process has decreased the die size by a large amount, allowing AMD to make more chips on the same wafer. Yields are still unknown, but AMD says they are having good luck with it.

Whether they are telling the truth or not, the overclocks on these chips have been nice for a very early chip. 2.5GHz using simple air cooling is common with the 3200+ chips. New BIOSes need to be made since raising the voltage on these chips is a little different than the Newcastle and Clawhammer chips. When one sets the voltage in the BIOS to 1.5V, the chip actually gets 1.4V. This means that whatever you set the voltage to, the CPU receives 0.1V less.

Copied from here:
http://www.hardwareanalysis.com/

Also:
http://www.nationmaster.com/encyclopedia/Advanced-Micro-Devices

http://www.pcmech.com/show/processors/35/11

gl

 

[Zorim]

Thanks a lot man, its a history of the speed I should say.

One more thing, can I request if you can make a chart out it, I mean the the processors pairing each other w/ respect to time . Sorry for my english


Keep it UP!
zorim

 

[Erasmus354]

why do the guys homework for him lol?

I also thought that essay was quite biased, talking down to the AMD PR ratings through the entire piece. The AXP were rated higher than they should have been...but the A64 are rated lower than they really should/could be, not higher.

 

[Papa-Ming]

why do the guys homework for him lol?

I also thought that essay was quite biased, talking down to the AMD PR ratings through the entire piece. The AXP were rated higher than they should have been...but the A64 are rated lower than they really should/could be, not higher.

Well, I already had the links from when... so I didn't mind. And he'll read all of them. LOL!

That essay appears to be written by a poster from the link attached, and not by any formal review per say so......

But yeah, it's good you pointed out that the essay was not meant as a scientific history of the proceesor, rather how he might go about formulating his paper so forth.

gl