Saturday, August 16th 2008

AMD Shanghai Rushed in, Hits Market in Q4 2008

AMD seems to have decided that it won't face any more flack in the computing industry owing to the lack of success the Barcelona / Agena (K10) architecture had compared to a determined Intel that seems to be going for the kill with a string of products lined up, Intel in fact has looked forward four years in time based on its IDF slides. On the eve of IDF, AMD executives said the company's 45-nanometer Shanghai processor will enter the market by the fourth quarter of 2008. The AMD processor is designed to compete against the Intel Nehalem processor, especially in the volume two-socket server market.

AMD Shanghai is touted to be the server/enterprise version of the upcoming Deneb series 45nm quad-core processors. Unlike the Deneb that retains the 940-pin AM2/AM2+/possible AM3 socket, Shanghai would use the enterprise segment Socket 1207, there already are server boards with four sockets in the market, AMD plans to use this as something to flash before Intel, as for server applications, cost-effective Shanghai parts used in four-CPU configurations should provide high levels of computational power.
Source: eWeek
Add your own comment

68 Comments on AMD Shanghai Rushed in, Hits Market in Q4 2008

#51
blueskynis
suraswamiAMD should have patented lots of things, now with ATI with them, they should patent moe. Morons.
What is "moe"? :rolleyes:
Posted on Reply
#52
a_ump
yea i've read that and heard that from one of my friends friends, haha he was still talking about athlon x2 versus pentium D, hahaha:laugh: people where i live(elkins, westi virginia) are not computer savy at all, i'm only 16 and i'd go far enough to say i'm one of the most knowledgable about pc hardware in my school, and yet on here i'm miles behind sometimes haha :shadedshu whats west virginia coming to
Posted on Reply
#53
btarunr
Editor & Senior Moderator
blueskynisWhat is "moe"? :rolleyes:


The one on the right. He's probably dead by now. Too late for AMD to patent.
Posted on Reply
#54
KBD
blueskynisWhat is "moe"? :rolleyes:
more :)
Posted on Reply
#55
a_ump
lol btarunr:roll:
Posted on Reply
#56
WarEagleAU
Bird of Prey
I agree, I dont see why AMD didnt patent HTT and on die memory controllers. Now theyll have something better coming and wont patent that either...
Posted on Reply
#57
blueskynis
WarEagleAUI agree, I dont see why AMD didnt patent HTT and on die memory controllers. Now theyll have something better coming and wont patent that either...
HyperTransport (HT), formerly known as Lightning Data Transport (LDT), is a bidirectional serial/parallel high-bandwidth, low-latency point-to-point link that was introduced on April 2, 2001.[1] The HyperTransport Consortium is in charge of promoting and developing HyperTransport technology.
The HyperTransport Consortium is an industry consortium responsible for specifying and promoting the computer bus technology called HyperTransport. It is led by founding members Advanced Micro Devices, Alliance Semiconductor, Apple Computer, Broadcom Corporation, Cisco Systems, NVIDIA, PMC-Sierra, Sun Microsystems, and Transmeta.
They didn't invented it so they can't patent it...
Posted on Reply
#58
KBD
WarEagleAUI agree, I dont see why AMD didnt patent HTT and on die memory controllers. Now theyll have something better coming and wont patent that either...
It was not patented because i think it doesnt belong to them. Hyper Transport was developed by the Hyper Transport Consortium, that includes many companies including AMD. As far as IMC, i think that was developed by IBM and picked up by AMD.
Posted on Reply
#59
btarunr
Editor & Senior Moderator
The 'patent' is maintained by the SIG (those members). HyperTransport isn't something only AMD uses, NVIDIA has always been using it as a chipset interconnect (waaay back since the nForce 2 days). We all go "ZOMG QuickPath" when it comes to a chipset interconnect, but AMD and NVIDIA are 5 years ahead of Intel comparing HyperTransport to DMI.
Posted on Reply
#60
KBD
btarunrThe 'patent' is maintained by the SIG (those members). HyperTransport isn't something only AMD uses, NVIDIA has always been using it as a chipset interconnect (waaay back since the nForce 2 days). We all go "ZOMG QuickPath" when it comes to a chipset interconnect, but AMD and NVIDIA are 5 years ahead of Intel comparing HyperTransport to DMI.
So you are saying that HT technology is already patented by the HT Consortium?
Posted on Reply
#61
btarunr
Editor & Senior Moderator
KBDSo you are saying that HT technology is already patented by the HT Consortium?
Where's there's a technology, there's supposed to be IPR protection. Otherwise, why a consortium or special interests group?
Posted on Reply
#62
candle_86
by the SIG group yes, and the IMC has been used on server chips for mainframe systems for years. IBM's POWER5, and Sun Microsystems UltraSPARC T1 both have an IMC. The x86-64 are the only set of instructions AMD could have kept, but they like to use SSE and Intel and Nvidia trade instruction sets.
Posted on Reply
#63
X1REME
so why has amd given nvidia a licence to use Hyper Transport etc? if its not there's to decide
Posted on Reply
#64
btarunr
Editor & Senior Moderator
X1REMEso why has amd given nvidia a licence to use Hyper Transport etc? if its not there's to decide
AMD didn't make it, the HyperTransport Consortium did. In such tech consortia, members (mentioned in an earlier post) contribute funds and RnD staff for a common cause (like in the UN, countries send their armies/funds for common causes) . Both AMD and NVIDIA did (contributing research money, staff, etc.) and hence today AMD uses HyperTransport in its processors, NVIDIA uses it in its chipsets (for both Intel and AMD platforms). The AMD-NVIDIA affair took of when both needed something radical, and hence sought to brandish each-other. AMD used "the best performing chipset is made by NVIDIA, a leading GPU maker", while NVIDIA used "We not only make GPUs but are expanding into other spheres, now we have the highest performing chipset for AMD (nForce 1 didn't take off very well)" in their PR stunts.

The K7 (Athlon XP) failed badly (at one point) versus P4 (Willamette) which further evolved into Northwood and looked unbeatable at any price point. With the research that went into making Thoroughbred-B (at one point AMD did take the lead over Intel with the Athlon XP 2600+), on the parallel HyperTransport became a reality. The very first large-scale commercial implementation of HyperTransport was not by AMD in its processors (surprise!) but by NVIDIA in the nForce 2 series chipsets for AMD K7. NVIDIA used a 8-bit wide HT Link between the SPP and MCP that helped the feature-rich MCP-T southbridge.
Posted on Reply
#65
candle_86
It also helped the PCI bus bottleneck, the PCI lanes used the HT link to transmit data faster.
Posted on Reply
#66
$ReaPeR$
wiaki think ATi is still a company but owned by AMD
might be called a division after what i have read
like dreamworks is owned by disney


yes i agree with you
Athlon 64 beated the crap out of intel, while amd had a worth about 10% of intels total revenue?
you gotta love AMD, when their CPUs are slow, their GPUs arnt
and i dont think AMD will go away anytime soon, how?, even Intel kinda gives amd money, as in you might have guessed it, ATI Graphics, allmost all new Intel mbs have crossfire support :p
very good points!!!! i totally agree with you. its ironic but due to the laws concerning monopolies intel needs amd in order to stay in business like microsoft needs apple and linux:D:D:D
Posted on Reply
#67
X1REME
by $ReaPeR$ (August 19th - 2:01 AM) - Reply by: wiak;932635i think ATi is still a company but owned by AMD might be called a division after what i have read like dreamworks is owned by disney yes i agree with you Athlon 64 beated the crap out of intelvery good points!!!! i totally agree with you. its ironic but due to the laws concerning monopolies Intel needs amd in order to stay in business like microsoft needs apple and linux:D:D:D
Intel never paid amd money but instead paid other company to not use amd cpu`s.

Intel uses x-fire or sli to sell there useless motherboards, not as you suggest (Intel will get there own larabee very soon so that they can stop there support).

the only reason Intel might need amd to live is to copy there designs (opteron64 = nehalem) and that's it mate

what's monopoly got to do with competition and bankruptcy. so if amd dies from competition what's that got to do with Intel (DUH)
Posted on Reply
#68
PP Mguire
Because since there is no more CPU companies that would mean they have a monopoly and thats illegal in the United States. Meaning AMD could get some help to stay out of bankruptcy if that was really the case. But AMD isnt as "broke" as some people would think so that aint gonna happen anytime soon.
Posted on Reply
Add your own comment
May 7th, 2024 07:44 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts