• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Intel's gamble on high-speed computer chip [Itanium] not paying off

Page 2 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.
IMHO, people see Opteron's 64-bit backwards compatibility and forget that this also means that the AMD64 ISA is carrying all the baggage of the hideous monster that x86 has become. IA64 is a much cleaner architecture that should allow for better scaling as well as better cost/performance in the future.

As hideous as it may be, people are NOT ready to make the complete switch so a hybrid solution sounds like the best answer...for now.
 
Don't these doomsayers seem to show up every time Intel releases a new CPU? The Pentium Pro was supposedly a dud because it couldn't run 16bit software well. As it evolved into the PII and III, it did pretty well for itself. Then the P4 comes out, everyone rips it for being slower compared to the PIII and how AMD was taking over with the Athlon. That didn't last too long. Now Itanium is going to be a dud. I doubt it, if for no other reason than the company behind it. Intel may not have gotten it right with the first iteration, but like Microsoft, if they keep throwing enough money into it, they'll get it right eventually.
 
Originally posted by: NFS4
IMHO, people see Opteron's 64-bit backwards compatibility and forget that this also means that the AMD64 ISA is carrying all the baggage of the hideous monster that x86 has become. IA64 is a much cleaner architecture that should allow for better scaling as well as better cost/performance in the future.

As hideous as it may be, people are NOT ready to make the complete switch so a hybrid solution sounds like the best answer...for now.

Yes and no. Servers only need to have the apps they use ported over to IA64 to warrant a switch. AFAIK, many servers are dedicated to specific tasks. If I have a server that exclusively handles Oracle, then as long as Oracle runs faster/cheaper on IA64 than x86 I couldn't care less about backwards compatibility. If a few big apps provide better price/performace on IA64 than x86-64, then that will drive the rest of the transition on the server side.

Workstations will be trickier and will take more time to transition over (assuming that IA64 gains enough momentum to do this); the reason being that workstations are less task-specific than servers. Once workstations start transitioning over to IA64 common desktops will follow. That transition IMHO, will look a lot like the Win16-Win32 transition; most big apps will already have an IA64 version by the time consmers start transitioning and the less important ones that will run as legacy code (at 50% speed, if necessary).

I am a fan of IA64 because I believe it will lead to better performance all around in the long term, so I hope something like that happens eventually. Truth be told, whether or not IA64 will ever hit critical mass is very much uncertain right now. Let's just say I think my vision is a little idealistic, and I wouldn't put any money on it if I were a betting man. 😉
 
IMHO, people see Opteron's 64-bit backwards compatibility and forget that this also means that the AMD64 ISA is carrying all the baggage of the hideous monster that x86 has become. IA64 is a much cleaner architecture that should allow for better scaling as well as better cost/performance in the future.

true, but the only people who would really care are compiler writers. btw, x86 isn't the only architecture carrying loads of baggage. powerpc is also. and the winner for most historical baggage is ibm's current mainframe computers that also have to handle code written in the 1960s.
 
Yeah, The mainframe carries alot of baggage and yeah, you run can run 16 bit software on a 31bit plaform on 64 bit machine....... but looking the latest IBM zSeiries 990 server (t-rex) servers, they require 3.2 billion transistor MCM to deliver their current performance....and a 5184 pin LGA socket.... its makes IBM p690' MCM small.......and the Itanium 2 Tiny........ The big issue with Itanium currently isn't performance but compatablity...... In order to have existing code work on Itanium, you will have to recompile it.....which of course is a pain to do....... .....
Multi Core Chips aren't that new actually

IBM started it all with the POWER 4 which was a enhance dual core power 3 merged with a high speed cache..

HP has just released their PA-8800 dual core processor for their HP9000 series of servers..

SUN has just released their ultra sparc IV dual core CPU too.
 
Geez these guys are clueless:
In its biggest strategic mistake in a decade, Intel has spent an estimated $2 billion creating a high-speed computer chip, the Itanium, that most customers don't want and don't need.
Hmmmm don't need IA64 but do need x86-64 that makes sense.
rolleye.gif

Two weeks ago, Intel did an about-face, announcing that it would incorporate technology that mimics AMD's fast-selling Opteron chip into cheaper Intel chips being introduced later this year
Yup, they just pulled it outta their a$$es, they haven't been designing/engineering the solution for 2yrs + 😛

Thorin
 
Back
Top