Perkembangan AMD dari Awal sampai Akhir

[gabol]

Tweet

Salam kenal untuk para punggawa prosesor disini...

Gini, saya pengen tahu tentang perkembangan prosesor AMD dari sejak

pertama dibuat (seperti seri 8080-nya intel) sampai yang terakhir

dibuat (FX series ya kalau gak salah). Kalau ada yang punya artikelnya,

tolong di share disini donk... soalnya saya pengen banget tahu

perkembangan step demi stepnya. Kalau bisa yang sekalian ada keterangan

fitur, konstruksi dan fungsi yang diusung prosesor seri tersebut. Saya

mohon kalau ada yang selengkap mungkin ya, soalnya masih newbie nih..

Thnx sebelumnya.

[Magician]

Tweet

panjang pak.. cek aja di arstechnica.com kalau mau tau masing masing detilnya, tapi disana nggak semua ada, hanya yang arsitekturnya terkenal aja.
Kemudian kalo mau tau sejarahnya bisa dilihat di http://www.amd.com/us-en/Weblets/0,,7832_10554,00.html

[gabol]

Tweet

panjang pak.. cek aja di arstechnica.com kalau mau tau masing masing detilnya, tapi disana nggak semua ada, hanya yang arsitekturnya terkenal aja.
Kemudian kalo mau tau sejarahnya bisa dilihat di http://www.amd.com/us-en/Weblets/0,,7832_10554,00.html

Kalau seperti yang ada di
http://www.klik-kanan.com/fokus/intel.shtml
(penjabaran yang semacam itu) punya referensinya gak (tapi buat AMD) ?

[Magician]

Tweet

Coba ke sini.
http://www.localcolorart.com/encyclopedia/AMD/

Yang ini sesuai kronologi. Intel/AMD
http://www.pcmech.com/show/processors/35/2

Ini lebih ringkas, tapi dimulai dari K5
http://translate.google.com/translat...language_tools

Anyway, google aja lah.. banyak kok

[gabol]

Tweet

Wah makasih atas masukkannya
(Yang di localcolorart lumayan lengkap tuh )
Bung Magician, kalo gugling kata kuncinya pake apaan ya? Koq aku nyasar mulu ya ?
Btw yang lain ada yang bisa kasih masukan (referensi) lain ? Biar semakin lengkap
Anyway makasih banyak buat bung magician.

[Magician]

Tweet

ketik aja AMD CPU history

[gabol]

Tweet

ketik aja AMD CPU history

Makasih..
Aku coba dulu.

[Scorpion14]

Tweet

Intinya mulai berkembang pesat cpu amd mulai dari k6 yaitu teknologi NexGen yang telah dibeli oleh AMD pada saat NexGen ingin meluncurkan seri....ehm lupa gw.....
NexGen mempunyai teknologi RISC yang membuat processor bekerja efektif dan cepat.....dan menggunakan juga teknologi yang dipakai oleh Apple MII.......kemudian setelah NexGen dibeli oleh AMD teknologi itu kemudian di tambahkan beberapa fitur yang lebih canggih yaitu 3DNow dan beberapa fitur lainnya.......ini membuat intel kwalahan pada saat intel mengeluarkan P2 yang baru.....sedangkan AMD masih menggunakan motherboard yang lama (pentium) tapi bisa bersaing dengan P2 bahkan mengalahkannya............
naaah inilah teknologi AMD sampai jaman Soket A.....bahkan sampai sekarang.....jadi Teknologi AMD adalah perpaduan 3 Teknologi yaitu AMD, NexGen, Apple......
NexGen dulunya berdiri dari Orang2 Intel itu sendiri..........

[gabol]

Tweet

Wah baru tau nih.....
Thanx atas infonya.

Ada yang mau kasih masukan lagi???
Mumpung gratis lho

[andhee]

Tweet

History of The AMD Microprocessor: from 8- to 64-Bits.

By Bit-Depth

When Jerry Sanders founded AMD in 1969, he made commitment to product quality one of the corporation’s top goals, as well as a means of corporate differentiation. Even before the company had launched or manufactured a product, it was decided that AMD would test all of its equipment to ensure it conformed to the stringent MIL-STD-883 standard, without passing on the end cost to the end consumer. This standard, developed and maintained by the US military, still exists in updated form and is in use today. Simultaneously offering a high-quality product at a cost-effective price is a puzzle that has tied companies in knots before, but it’s a goal AMD has succeeded at admirably throughout the last thirty-five years.
Before the Microchip: AMD from 1969 – 1978
Although today AMD is known for CPUs and flash memory, the company once had a much larger portfolio of devices it designed and manufactured. From the start, AMD was built as a company that not only developed its own proprietary solutions and designs, but also licensed and built chips based on the technology of other companies. Often, in these cases, AMD made it a priority to offer a higher-performing variant, a lower-cost variant, or both. This commitment to quality and price paid off: AMD launched its first RAM chip (the Am9102) in 1975, released a clone of the 8080A Intel processor, and released the first iteration of the Am2901 family all within 12 months. The Am2901 was a “bit-slice” processor based on a modular design that allowed 4-bit ALUs to be arranged in parallel for simultaneous processing. Using 8 Am2901 chips, for example, a system could perform 32-bit calculations efficiently, even though each chip was only 4 bits wide.

AMD began building its Austin facilities, expanded manufacturing in the Philippines, and grew rapidly, while improving its various processor and memory designs.
A 16-bit Party with an 8-bit Detour: 1978 - 1990
It’s generally known that IBM’s decision to pick the Intel-designed 8088 CPU to power the IBM PC was instrumental in catapulting that company to dominance. What isn’t quite as well known is that it had a fundamental impact on AMD as well. In the early 1980s, IBM maintained a policy known as “second sourcing”, which required at least two sources for any product it was using. This meant that even if the first manufacturer failed to deliver volumes IBM wanted, they could still pick up product from a secondary (and ready-to-produce) source.

The fact that IBM required these types of agreements says something about the amount of influence they wielded over the semiconductor industry and the sheer volume of product they were capable of needing; it also meant that Intel needed a manufacturing partner. AMD was initially brought on board to handle IBM’s volume overflow. This, in turn, required that AMD be licensed to produce hardware built to the x86 specification. The terms of AMD’s cross-license negotiation with Intel aren’t generally known, but they began while negotiating CPU production for IBM’s product lines. AMD secured the rights to produce 286 and 286-derived hardware as a result of these agreements, and also their position as a second-source manufacturer for 8088, 8086, and any other CPUs IBM might require.

AMD made ground quickly in this market, selling processors to lower-cost OEMs and producing higher-clocked versions of older designs that undercut competing solutions in the market. Because they had full access to the core microcode and CPU design structure, there was never an issue of compatibility. Compared to some CPU designs from third party vendors (which were plagued with issues), AMD was in good shape. While this situation did, unfortunately, create some problems for them (consumers not generally knowing which third-party builders were good, and which had problems), AMD landed important design wins in this period. Am286 chips power the Space Shuttle to this day, while A 3113 timer chips are used in the Tomahawk line of cruise missiles.

Computing, however, was still essentially a 16-bit game. Even though the 80386 processor design (first released in 1985, with AMD’s own Am386DX following in 1991) arrived not long after the release of IBM’s 8-bit PC, most operating systems of the day were written for 16-bit systems. IBM’s OS/2 project attempted to move the industry to a native 32-bit solution years ahead of Windows® 95, but failed to establish a significant foothold commercially.

As computer usage increased and consumers began demanding more powerful systems with multimedia capabilities, it was obvious that a more advanced operating system environment would have to be developed. This, in turn, requires the use of processors with the ability to provide those functions.

The value of a processor with a wider data path (8-, 16-, 32-, 64-bits) is that it increases the amount of data that can be handled and processed inside the CPU during a single cycle. While this is not a magical fix (a program must be re-compiled in order to take full advantage of the additional data path width), it offers a meaningful advance in CPU capabilities over a longer period of time.

Each quantum leap in desktop computing power has lead to the system software evolving and taking advantage of the hardware’s capability. DOS was a 16-bit textual OS with only a very primitive GUI sub system available (and provided mainly by third-party vendors). Windows 3 introduced some 32-bit capabilities but was mainly a 16-bit operating system. Windows 95 and NT truly took advantage of 32-bit capabilities (NT from the ground up, 95 in a hybridized structure), and once again, it was here that we saw real advances in usage-level computing.

The 32-bit Era: From Am386DX to K5 (1991 - 1995)
The launch of Windows® 3.0 in 1990 began a new era in desktop computing. Windows 3.0 and 3.11 shone on a 386, and AMD's Am386DX was a huge success, selling over a million units months after being introduced. The Am486DX and Am486® microprocessor designs followed in the next few years, extending AMD's reputation as a provider of high-quality product at a good price. Up to 1995, however, none of AMD's designs differed fundamentally from the established x86 processor they were based on. Some of them ran at faster clock speeds, or offered different cache amounts, but at heart they were only minor deviations from already established processors.

By 1993 it was clear that AMD's role as a second-source/alternate producer of cloned competitor products was over. Competitive realities, court cases, and the expiration of their initial agreement to do so all dictated that AMD either find new business opportunities to explore or develop their own microprocessors. Although their first independent effort (the 5x86 chip) was fairly well-received and offered competitive performance, it was still essentially an extension of the Am486 microprocessor design taken to the highest level that architecture could reach. It was an impressive achievement—but it wasn't a design AMD could truly call “their own.”

The K5 (released in 1995) was AMD's first attempt at their own competitive x86 architecture. Although internally quite advanced, it was also over a year late and ran at lower clock speeds than expected. Instead of being a high-end competitor, it was positioned as a lower-cost alternative. Once it became apparent that the K5 design simply wasn't good enough to carry them forward, AMD went looking for an alternative design they could bring to market quickly. In 1996, AMD acquired NexGen and their CPU designs. AMD quickly rebuilt Nexgen's NX686 CPU for Socket 7 and released the AMD-K6® processor.

The K6 Family: 1997 - 1999
Unlike the short-lived K5, the AMD-K6 processor was a major product for AMD. First released in 1997, the AMD-K6 processor was the first alternative CPU design that gave competitive performance in business and desktop applications without choking on floating-point math-a critical component of gaming and some multimedia tasks. While the AMD-K6 didn't offer the strongest x87 FPU available; the CPU became tremendously popular in midrange and budget systems.

Again, this combination of high performance and low cost made AMD a market leader in the emerging sub-$1000 segment. AMD followed the successful AMD-K6 with the AMD-K6® -2 processor, which added support for SIMD (Single Instruction Multiple Data) instructions and moved on to an advanced form of the original Socket 7, now called Super Socket 7. This new form factor added support for a 100 MHz FSB, and kept the aging platform standard competitive with other designs. Because AMD offered a variety of AMD-K6-2 models, the CPU also became known as a drop-in replacement for users who had CPUs that only ran up to 200 MHz. The AMD-K6-2 400 reused an obsolete multiplier setting allowing it to run at 400 MHz even on older motherboards.

The AMD-K6-III processor followed the AMD-K6-2 in early 1999. The AMD-K6-III was an AMD-K6-2 with one major difference-this CPU incorporated 256K of L2 cache. The performance increase was startling; in a single bound the AMD-K6-III became one of the fastest chips on the market. Competition in the budget segments (where AMD was strongest at this point) had stiffened considerably since 1997, however, and AMD was definitely in need of something stronger, more powerful, and more effective than the aging AMD-K6 core. As impressive as the AMD-K6-III was for its time, it also fundamentally lacked floating-point power, an issue no amount of L2 cache could improve.

The Value of the AMD-K6® Family and the Road to AMD-K7™:
The AMD-K6 processor family was an integral part of AMD's success for several reasons. First, despite its lack of pure FPU performance, it was the first x86 CPU design that wasn't derived from an Intel product and was capable of challenging the mainstream leader in all market segments. AMD's decision to offer models of the chip that could be used to upgrade older platforms (as well as to continue extending the older platform in the first place) at excellent prices brought the company tremendous business and enabled it to continue developing its next-generation technology, code named AMD-K7™.

In developing the AMD-K7 architecture (later known as the AMD Athlon™), AMD had to realistically evaluate both the strengths and weaknesses of its earlier AMD-K6. While the CPU's largest single drawback had been its FPU performance there were other issues to be considered as well. The AMD-K6's largest single weakness as a platform base wasn't the processor, but the motherboard/chipset. Despite being based on the trusted Socket 7 platform, Socket 7 chipsets and boards varied widely in quality and capability. Because end users often have a hard time determining what system component creates an instability or problem, AMD found themselves saddled with a reputation for being “less stable” than the competition.
At this point, AMD had two options. They could've ignored the problem and swept it under the rug of “not our fault,” or they could deal with it. In typical fashion for a quality-conscious company, they chose to deal with it, and began working harder with their Taiwanese partners to increase chipset and motherboard stability. While there was no magical “turning point” in this process and there were still some definite bumps in the road, AMD platform stability and reliability has increased dramatically since the days of Socket 7.

The AMD-K7™ Microprocessor Cometh: 1999 - 2003
Immediately upon its unveiling in August 1999, it was clear that the AMD Athlon™ (code name: AMD-K7) processor was something different from any processor AMD had ever built. In place of the AMD-K6's single non-pipelined FPU unit, AMD built a multiple-pipelined FPU capable of executing multiple floating-point instructions in parallel. This massive FPU has proven to be one of the AMD-K7's most enduring strengths; it guarantees that even in legacy code the Athlon isn't left behind.

The entire market has shifted and evolved since the launch of the AMD Athlon processor. AMD was first-to-market with a 1 GHz CPU, and the first desktop manufacturer to ship desktop CPUs in volume at that speed. Later generations of the AMD Athlon architecture introduced on-die L2 cache at full processor clock, SSE instructions were brought on board with the AMD Athlon XP processor, and AMD became the first mainstream CPU manufacturer to support DDR memory in the fall of 2000. 2001 brought the introduction of the 760MP/760MPX chipsets, and AMD again offered a highly competitive/attractively priced multiprocessor server solution in the form of the AMD Athlon MP processor. It was not, however, a bed of roses. From 1999 to 2003, AMD dealt with the dot-com bubble burst, a heavy increase in competitive pressure, and several delays to its next-generation K8 CPU that forced the further extension of the AMD Athlon XP processor's design.

By 2003 it was clear that the AMD-K7 architecture was in need of a refresh. While it still offered competitive and price-effective performance, it was increasingly overshadowed by newer solutions capable of taking advantage of advances in CPU and programming technology.

Two 64-bit Roads Diverge: AMD64 vs. IA-64
Before I move on to 64-bit computing in its current form, I want to talk for a moment about IA-64 and AMD64 technology. AMD64 is sometimes also referred to as x86-64 as a generic term, and is marketed by Intel as EM64T. All three of these terms refer to the same core set of technology and the same AMD-driven architecture.

AMD64 is the name given to AMD's 64-bit extension of the x86 architecture. If you recall, the original 8086 processor was a 16 bit chip, with an 8-bit knockoff (the 8088), and the x86 architecture specifies 8 GPRs (General Purpose Registers) 8 FPRs (Floating-Point Registers), and 8 SIMDs (Single Instruction, Multiple Data) registers. In the 8086, each of these registers was 16-bits wide. When Intel developed the 80386 design, they extended the register width to 32 bits (the resulting update to the x86 ISA is referred to as IA-32), but they left the number of registers alone.

AMD64 not only extends the IA-32 instruction set, it doubles the available number of GPRs and SIMD registers to 16, up from 8 each. This offers a tremendous theoretical advantage to x86 CPUs switching to 64-bit mode, even in programs that might not inherently benefit from 64-bit memory addressing. Even though the AMD64 ISA still inherently lacks some of the register combinations available to other CPU designs, AMD has proven it's quite capable of squeezing tremendous performance out of the twenty-five year old x86 architecture.

As you undoubtedly know, 64-bit support is a major feature of AMD's eighth generation processor family. That support is currently available in all shipping AMD Opteron™ and AMD Athlon 64 processors, while AMD Sempron™ is 32 bit only, but offers many of the AMD Athlon 64 processor's other speed improvements. AMD's goal with the AMD64 instruction set is to extend 64-bit computing on the desktop, workstation, and server platforms, at prices consumers, businesses, and IT corporations can afford in each market segment. AMD64 is not an entirely new standard (it builds on the existing IA-32 architectural platform), but it extends a proven standard in a method similar (and possibly superior) to what Intel did back in 1985 with the 80386.

Intel's in-house 64-bit solution uses the IA-64 ISA, is called Itanium, and is a complete departure from the x86 architectural standard. Without going into too much detail here, it's accurate to say that these are two products designed according to two fundamentally different paradigms. AMD designed the AMD Opteron processor for backwards compatibility without sacrificing performance in the future, and chose to extend existing standards rather than break from them completely and force the market chaos such a maneuver creates.
As of this writing, Itanium and AMD Opteron are targeted in different markets and have different focuses. Given the infancy of the 64-bit industry in general, it's not clear which architectural approach will triumph in the long run. Intel's decision to license and manufacture processors that are compatible with AMD's 64-bit standard says volumes, however, about what the future of the market might be.

More Than Just a 64-bit Pretty Face: The K8 Impresses All Segments
AMD's current eighth generation technology is more than just a 64 bit extension, however. When compared to either seventh generation AMD Athlon™ XP or other competitive solutions, the
AMD Athlon 64 and AMD Opteron™ processors offer strong performance, thanks to the addition of SSE2 support and an onboard integrated memory controller. The AMD Athlon 64 processor also uses HyperTransport™ technology, a point-to-point bus architecture AMD developed and licenses through the HyperTransport Technology Consortium.

The AMD Opteron and AMD Athlon 64 processors have reinvigorated the company's market share and its bottom line, while an attractive combination of performance and thermals allow AMD to offer solutions in mobile, desktop, server, and enterprise performance segments. The release of Windows® XP 64-bit edition (it's already available in beta) should help jump-start 64-bit adoption.

In short, K8/Hammer is sitting pretty.

Why We Need 64-bit CPUs:
It's a fair question to ask why we need 64-bit chips at all. After all, 32-bit chips have advanced tremendously since 1991 and Windows 3.11. A year from now (over two years after the
AMD Athlon 64 processor launched), the vast majority of software in the world will still be running…32-bit. Does this mean the capability to use 64-bit software is irrelevant? Absolutely not.

Stepping back and looking at the computer industry, it's easy to see how a quantum leap in software often follows after CPUs go through a major bit-change-but this leap does not appear immediately. The full potential of a 64-bit system won't be recognized in Windows XP 64-bit; it may take a few years. It may take until Longhorn. The last time the industry shifted from 16 to 32 bits, it took nearly a decade. So no, buying a 64-bit chip today doesn't magically make 64-bit software appear.

If you want a good example, however, of something a 32-bit chip can do that a 16-bit chip can't, take a look at Windows 95 in comparison to DOS or even Windows 3.11. Wolfenstein 3D required a 286, Quake ran on a Pentium, or possibly a fast 5x86. The difference between the two is more than just an increase in clock speed-it's a difference in how much data the CPU was capable of handling per cycle.

This type of evolution, however, takes time; only now are we beginning to see software that takes advantage of 64-bit computing. Undoubtedly 2005 will bring more titles or applications that can do so, but it will still be a small market, used for enhancements or additional features.

In the not-too-distant future, 64-bit CPUs will be ubiquitous enough that a game developer, application author, or OS manufacturer will design a killer product that doesn't just use a 64-bit system for enhancements, but utilizes its capabilities as a fundamental part of the end product. AMD's x86-64 standard does that, and it's a standard based on the same fundamental precepts Jerry Sanders put in place in 1969 when he required certification to the MIL-STD-883. Thirty-five years after their founding, AMD is still committed to providing quality products at a good price that consumers can rely on. Though they've definitely had rough patches and hit rough times (some caused by macroeconomic events far outside anyone's control), the past twenty years has seen AMD move from being a second-source provider to a designer and innovator of technology in their own right.