Outline · [ Standard ] · Linear+ AMD & ATI, AMD a cumparat ATI ieri

Pana asteptam sa vedem ce face Renault-Nissan cu GM intre timp avem parte de o achizitie. AMD a imprumutat de la finantatorul sau 2.5 miliarde $ si a initiat cea mai mare achizitie din domeniul IT de-o bucata de vreme incoace (tranzactia finala e estimata la 5.6 miliarde $ urmand sa aiba loc un transfer de actiuni pentru a completa de la 2.5 la 5.6). Nu are rost sa va dau un link catre vreo stire. E suficient sa intrati pe: http://www.ati.com/ sau pe http://www.amd.com/ De mentionat ca nVidia face chipseturi pentru AMD iar ATI e concurentul de temut al lui nVidia si in acelasi timp ATI face chipseturi pentru INTEL pentru a completa oferta low-end a lui INTEL si reprezinta in acest moment 80% din toata afacerea cu chipseturi pentru ATI. De departe cea mai controversata achizitie. Motivul acestei achizitii ? Doar o speculatie cu privire la supravietuirea AMD dincolo de o limita de 5 ani. E vorba de viitorul platformei PC la care se pare ca INTEL e cu ceva inaintea AMD-ului. E vorba de "mini-cores". Toti ne-am obisnuit cu combinatia fericita CPU + GPU pentru grafica trasnet in jocuri. Analistii IT prevad deja moartea GPU-ului iar solutia e data de o posibila implementare la nivelul CPU-ului a tuturor functiilor grafice. Decat 1 - 2 nuclee complicate care mananca foarte multa putere se pare ca viitorul consta in o hoarda de nuclee mici si simple. Usor de conceput, usor de implementat, usor de extins. Vrei mai multa performanta ? Maresti numarul de nuclee deci mai adaugi putin silicon si gata, mult mai usor decat in trecut cand performanta crestea odata cu frecventa de tact a procesorului. INTEL se pare ca are o solutie de termen lung cat si una pentru termen scurt in viitorul imediat asa ca AMD care se prezenta regele performantei pentru consumatorii doritori de viteza maxima este putin pus la colt. Din strategia pe termen lung a lui AMD facea parte si achizitia unui grup major din industria IT si se pare ca ATI a fost cea mai potrivita alegere din cauza expertizei in domeniul grafic, mobil cat si a chipseturilor. ATI a fost interesat de accesul mult mai facil si ieftin la serviciile unor fabrici precum cele ale AMD-ului cu conditia ca aceasta tranzactie sa se materializeze. ATI va fi mai profitabil iar AMD isi va consolida mai puternic oferta cat si viitorul. Inca odata dupa o lunga perioada de plictiseala scena industriei IT ne arata ca este inca loc de schimbari majore. semnat, DigitalScribe This post has been edited by DigitalScribe: 25 Jul 2006, 16:24

Pana asteptam sa vedem ce face Renault-Nissan cu GM intre timp avem parte de o achizitie. AMD a imprumutat de la finantatorul sau 2.5 miliarde $ si a initiat cea mai mare achizitie din domeniul IT de-o bucata de vreme incoace (tranzactia finala e estimata la 5.6 miliarde $ urmand sa aiba loc un transfer de actiuni pentru a completa de la 2.5 la 5.6).

Nu are rost sa va dau un link catre vreo stire. E suficient sa intrati pe:

http://www.ati.com/

sau pe

http://www.amd.com/

De mentionat ca nVidia face chipseturi pentru AMD iar ATI e concurentul de temut al lui nVidia si in acelasi timp ATI face chipseturi pentru INTEL pentru a completa oferta low-end a lui INTEL si reprezinta in acest moment 80% din toata afacerea cu chipseturi pentru ATI.

De departe cea mai controversata achizitie.

Motivul acestei achizitii ?
Doar o speculatie cu privire la supravietuirea AMD dincolo de o limita de 5 ani. E vorba de viitorul platformei PC la care se pare ca INTEL e cu ceva inaintea AMD-ului. E vorba de "mini-cores". Toti ne-am obisnuit cu combinatia fericita CPU + GPU pentru grafica trasnet in jocuri. Analistii IT prevad deja moartea GPU-ului iar solutia e data de o posibila implementare la nivelul CPU-ului a tuturor functiilor grafice.

Decat 1 - 2 nuclee complicate care mananca foarte multa putere se pare ca viitorul consta in o hoarda de nuclee mici si simple. Usor de conceput, usor de implementat, usor de extins. Vrei mai multa performanta ? Maresti numarul de nuclee deci mai adaugi putin silicon si gata, mult mai usor decat in trecut cand performanta crestea odata cu frecventa de tact a procesorului.

INTEL se pare ca are o solutie de termen lung cat si una pentru termen scurt in viitorul imediat asa ca AMD care se prezenta regele performantei pentru consumatorii doritori de viteza maxima este putin pus la colt.

Din strategia pe termen lung a lui AMD facea parte si achizitia unui grup major din industria IT si se pare ca ATI a fost cea mai potrivita alegere din cauza expertizei in domeniul grafic, mobil cat si a chipseturilor. ATI a fost interesat de accesul mult mai facil si ieftin la serviciile unor fabrici precum cele ale AMD-ului cu conditia ca aceasta tranzactie sa se materializeze. ATI va fi mai profitabil iar AMD isi va consolida mai puternic oferta cat si viitorul.

Inca odata dupa o lunga perioada de plictiseala scena industriei IT ne arata ca este inca loc de schimbari majore.

semnat, DigitalScribe

This post has been edited by DigitalScribe: 25 Jul 2006, 16:24

[...]
In the past, a unique computer design firm was founded. Silicon Graphics was devoted to very specific specialized designs that did graphics faster than anyone else could - and for less.

The basic idea is simple. You first design one or more main processors that use a very large block of main memory to turn graphics commands into lists of drawing commands - in a new language you invent.

This main memory is designed in sections. Each section is totally dedicated to the main processors - unless a special crossbar switch reconnects one section of memory to the graphics subsystem. This way, either subsystem gets total control over some of main memory.

The video section uses advanced designs that feature RISC processors with specialized graphics engines. This subsystem runs independently of the main processor subsystem. It uses lists of OpenGL commands to produce the images.

This is about as far from a low cost general purpose hobbyist kludge as is possible. In the first microcompeters, the video chip literally drove the whole system. The video chip generated the system master clock and held total control of everything except for brief slices of time when it allowed the main processor to do things. That is the essence of the Commodore and Atari stuff. Earlier CP/M systems such as the S-100 bus and SS-50 designs were a bit more sophisticated.

Anyway, SGI was the first and main provider of computers for producing the video effects in movies and still has a lot of market share. But the SGI design is expensive and very difficult to upgrade. Its all about that crossbar switch that provides alternate access to main memory by two separate subsystems.

Now let me briefly mention AMD. The new AMD concept uses a crossbar switch inside the main processor - to separate the local main memory from the Hypertransport bus that connects to the main memory blocks of other processors. This is a simple and less efficient variation of the key design feature of Silicon Graphics systems. How about that?

Now if you replace one main processor with a special independent graphics subsystem designed by ATI - in an AMD multiple processor system - you get almost back to the original SGI design concept. You just have this delay inherent to the Hypertransport concept. Note that each general purpose processor still has the same old problem. Each main processor still shares its local memory with all other main processors - but the Hypertransport bus permits controlling of access to local main memory by the other nodes.

The Hypertransport concept has a lot more potential than anyone has realized - except maybe for this crazy old man who is a nothing in the industry. The best is yet to come.

Meanwhile, Intel is still locked into designing hardware systems that work well with Windows. Now I am not going to speculate on the usefulness of Plan 9, the recently rediscovered distributed operating system. Er, don't waste your time watching the movie Plan Nine From Outer Space either. Perhaps the world's worst movie - an Ed Wood specialty, making awful movies.

But nVidia has chipset expertise. ATI has some neat new ideas for independent graphics engines. Neither can touch a good RISC OpenGL card design but those ideas are still around, ready to be reused or reworked. The Cell by Sony/IBM/whatsis is a sort of cheap variation of the RISC powered OpenGL video subsystem idea. Honest. so if someone was smart enough to see the trees while standing in the middle of the forest....
[...]

sursa: http://www.theinquirer.net/default.aspx?article=33256

Long live INTEL. pacat de ATI

sursa: http://www.theinquirer.net/default.aspx?article=33287

Intel pulls ATI from Conroe launch

Own goal, own goal!

By Charlie Demerjian: Thursday 27 July 2006, 06:44

I WAS GOING to rant about the first real test of the AMD/ATI marriage happening later today at the Conroe launch, how will Intel react?

Will they take the high road, and do what is best for their users, or get cross and shoot themselves in the foot. Answer? High velocity lead podiatry.

[...]

AMD rulz



LE: naiba'n ea de eticheta...nu se vede partea esentiala a pozei. Ar trebui sa se faca ceva in acest sens!

This post has been edited by R19/16v: 27 Jul 2006, 12:13

Attached thumbnail(s)
27.jpg ( Size: 25.78k ) Number of downloads: 115