Sign in with
Sign up | Sign in
Your question
Closed

Intel releases new cheaper quad, AMD aims low-k 45nm

Tags:
  • CPUs
  • Quad
  • Intel
  • AMD
  • Product
Last response: in CPUs
Share
December 12, 2006 4:14:24 PM

I'm interested in some benchmarks.

-edit...wrong thing I'm replying to.

The partnership between AMD and IBM has proven to be quite the good deal over the past couple of years.
Related resources
December 12, 2006 4:40:15 PM

Im pretty sure by the time AMD has 45nm, Intel will be moveing on to 32nm, unless Intel has some sort of huge problem with 45 but they say they are are schedule.

Cheaper quad core ? hmmmm I think I will wait for Intel to get a different or faster bus before I would buy one (I know the aps dont point to much of any bottle neck)

Very interesting indeed !
December 12, 2006 5:23:34 PM

Quote:
I think the race for smaller nodes is not very efficient;the investment in the equipment is high,and getting to the next smallest node too fast decreases profits.

on the other hand,AMD fell asleep at the wheel getting to 65.Now they have got a workload that no one could envy.getting to 65,testing 45,incorporating ATI,producing integrated solutions,competeing with intel.and now alledged price fixing from ATI and nvidia,and records from as far back as 2000 bieng subpoenaed.

I dont envy AMD's workload,but they decided to get too self confident with k8.They are looking like one hit wonders right now.


this is all very true and it will be quite a difficult journey for AMD, but they do have one ace up their sleeve that intel does not. while Intel has burned just about every bridge with their fellow competitors in the industry, AMD on the other hand has gone to great lengths to reach out to its competitors. In the case of AMD/IBM partnership it looks like this may be the light at the end of the tunnel for the future of AMD.

eh, maybe :? :?:
December 12, 2006 6:10:15 PM

Quote:
whoever is planning technology over there should be demoted or fired.there is no reason why they let intel take the lead,other than laziness and arrogant overconfidence.

agreed
December 12, 2006 6:21:13 PM

It's not like AMD has been really sitting on its ass.. or it's run by people like Sharikou and MMM.
People here underestimate the effort required to design a new uArchitecture, or transition to smaller nodes, etc.
I'm pretty sure that AMD was perfectly able to see the avalanche-Core 2 coming at least one year ago (like i did, and i've nowhere near the knowledge of AMD engineers), when Intel announced that they were going back to the Pentium-M architecture (which was on par clock for clock with K8, except in FP code), integrate all the good stuff from Netburst, extend it to 4 issue and all the other enhancements.
So why K8L comes out (hopefully) only 2 years later?
I really think they couldn't do better.
December 12, 2006 6:43:19 PM

I think you're being a bit emotional here my friend ;) 
December 12, 2006 7:02:15 PM

Quote:
I think you're being a bit emotional here my friend ;) 
:D  :D  :D 

I wont deny how upsetting it is,but 3.5 years on the same die?now its 4yrs?
its pretty lame.

I do have to agree, they did stretch the 90nm process out for far too long especially considering they didn't introduce any major architectural changes to the CPU. If they had been upgrading the K8 uArch every 18 months or so then today i believe that it would perform very well against the Core 2 Duo.

AMD got complacent, and now they have to pay the price.
December 12, 2006 7:02:48 PM

Ok, i'll show it under another perspective.
After the 486 clone, AMD tries to design the K5.
It was ambitious, superscalar, with a pipelined FPU, and the first x86 CPU to decode the instructions internally into RISC and do OOO execution.
Cool eh?
But it runs hot as hell, it's extremely late to the market, and won't clock signifcantly higher than 100MHz (the 116MHz K5 was P-rated at 166) when the Pentium already is already at 200.
Then AMD acquires the K6 through Nexgen, then K7 is designed thanks to the massive acquisition of Digital engineers from the Alpha project.
Then AMD tries to design the K8, which was supposedly a 6-issue monster.. that project was aborted, and the new K8 is nothing but a K7 with slightly lengthened pipeline, 64bit extensions and it's main bottleneck removed: instead of the slow EV6 FSB which peaked at 400MHz, they introduced the memory controller and HT link.
Sure K8 was the most successful AMD CPU to date, but in terms of architecture, the big jump was done with the K7, and the K8 was just an evolution of it, not a real new design.
That's why i think that AMD probably didn't come out with something new earlier, because they couldn't.
But they've been pretty successful in refining their business model, achieving industry acceptance and recognition, and penetrating into new important markets such as Servers / Enterprise.
December 12, 2006 7:28:05 PM

Quote:
Jesus,man!Intel is invested in so many cutting edge technologies,at MIT and that,AMD has a mountain it needs to move,Intel is that mountain;and there is NO room for screwing around with aging technologies.


Totally agreed there! AMD can't even rely on their increased server sales to save them anymore as Intel has released their new Core 2 server chips. Once all contracts with Opterons are fulfilled i believe most people will turn back to the much better performing Intel Core 2.
December 12, 2006 8:09:59 PM

That's it, scream at the engineers. That will make them work faster. Did you have fun playing backseat CEO?

The only reason AMD was ahead for so long was that intel made a serious misstep with the P4. The P3 was below the K7 in per clock performance, but it was much closer than the P4. It took intel over 5 years to recover from that blunder, and everyone said it was a blunder back in 2001. intel is also in a position to have several development teams working on several different competing designs, while AMD typically only has one. That's what happens when your competitor earns $33,000,000,000.00 a year more than you.

I'm all for AMD coming up with a magical chip that is oodles faster than anything else on the market, but it's not like the Athlon was thought up overnight after a pep talk by the CEO. I remember the hype started almost two years before it was released, and AMD was getting pretty close to bankrupcy when it finally was released. intel had it easy recovering from Netburst because their Israeli design group was working on a P3/PM like design, and it still took 5 years for intel to roll back to a P3, add a few extensions, and release it.

Don't worry though, AMD historically has developed new architectures by buying EEs, from NexGen, from DEC, and now from ATI. Unless AMD just wanted to diversify its assets ATI's aquisition indicates the direction of AMD's next major enhancement. Perhaps dropping 48 or 96 or x PU's (aka pixel processors) into a 3 or 6 issue multicore chip, or something on a much lower level in the architecture. Something like that could be a Niagra-style design with excellent x86 abilities.
December 12, 2006 8:27:35 PM

AMD tried to innovate and the K5 was late, hot, and slow. Intel doesn't have some magical brain trust, they have an R&D budget bigger than AMD's revenue, plus almost twice AMD's revenue in cash. Intel can afford to sink $8 billion a year into 8 different design teams unlike AMD's $1 billion for 1 design team. If intel misses on 7 of 8 designs, they have a winner every generation. If AMD misses on a single design, they get buried. So, AMD buys design teams that have good designs.
December 12, 2006 8:46:20 PM

8 design teams VS 1 is an exaggeration, but i do agree on the principle.
Intel is a R&D powerhouse, which finds its peers only in the likes of IBM and Microsoft.
December 12, 2006 8:53:47 PM

Quote:
Thats not how it looks to the tech world;Amd has hurt themselves for not extending the lead,the 5% boost on am2 should have been a second year 939 revision.

Honestly it looks like AMD has nothing in its bag of tricks,we could debate the promise of barcelona all day and it wouldnt bring amd any closer.
they had 3.5 years to extend the performance lead ,that is alot of time.
They could have done better.

Totally agree! Barcelona could have been already released if they really were willing to do it; it was scheduled for early 2008 and they're bringing it in mid 2007 now that they're moving.
Blah! AMD fell asleep just like Intel did, they were really stupid; When the giant sleeps, you'd better have a healthy run rather than dance around him :wink:
December 12, 2006 9:36:55 PM

By this time they already know it; they thought they could squeeze K8 like intel did with netburst but forgot how netburst ended up. It's really true that it's hard to get the crown but much harder to keep it.
December 30, 2007 4:56:47 AM

what about samsung? i read somewhere that samsung is looking at helping amd out as well. and intel is first chip maker, sammy is number 2. if they partner up and start making stuff together, it could put amd back on its feet again, just like when k8 came out. either way, phenom is just another cpu. i mean it can game better than my x2 (or so i hope it would) but i hope they get back on their feet. maybe i wont have to go intel. but their next line of quads due out in just a month or two look fantabulous for my next build with crossfire (hope that isnt a mistake either) but two 3870 x2 boards, talk about heck of a score and gamability
December 30, 2007 6:25:54 AM

haha thats stuff from 06
December 30, 2007 9:31:09 AM

lol most of the ppl that posted in this thread i haven't seen posting here since that fellow left and took quite a few decent members with him.
December 30, 2007 9:48:53 AM

I don't understand.....It's telling me i've posted in this thread already but i know for a fact i haven't. I'm VERY confused.
December 30, 2007 5:22:09 PM

You have. . . look 9 posts up.
December 30, 2007 6:11:02 PM

I dont know who keeps digging up these ancient posts, but I do have my suspisions.

Cut it out. Dont make me talk to Brett or Mike.
Related resources
!