www-03.ibm.com — Since then, IBM scientists have continued to drive performance improvements to continue the path of Moore?s Law. Ten of IBM?s biggest chip breakthroughs chosen from the dozens of innovations coming from IBM?s labs in the past ten years include
May 3, 2007 View in Crawl 4
geminitojanusMay 4, 2007
"make it smaller"Intel had 65nm 19 months before AMD."faster"Clock rate fast? Intel Pentium D @ >7GHz. IPC fast? Conroe slightly nudges the Athlon 64 from its perch in most categories (and in others, completely blows it away)."cooler">120W Athlon 64 vs 95W Conroe"cheaper."You got me. But hey, if it's worth $20-50 to you to step down to AMD _and_ have to replace your motherboard the next chip generation, go for it.
geminitojanusMay 4, 2007
Fujitsu, Motorola (what's left of it anyways), NEC, Toshiba, Samsung, STMicro, Renesas Technology aren't American companies.
Closed AccountMay 4, 2007
Is cheating at chess considered an "innovation"?
refriaireMay 4, 2007
Fallout 2?
bigslackerMay 4, 2007
The engineering of all those companies do is done globally although the US is the engineering leader in pretty much all fields. A lot of foreign companies contract to US engineering companies or open development offices here.
geminitojanusMay 4, 2007
"The Penryn is essentially in the same generation as Core 2"Yeah, the Pentium II is essentially the same generation as the Pentium III too. Wait.."Just as Core 2 is a big architectural change from Pentium 4"Apples::Oranges. There is almost _nothing_ of Netburst in NGMA; they share a similar busing unit, they share some MicroOp Fusion techniques that came from their Israeli Microarchitecture development team, but that's _it_. "Phenom is a change from the Athlon 64 architecture"Phenom aka K10 aka K8L are all names for the same thing: K8+1, AMD's 9th generation architecture (silly to call it K10, but there's probably some trademark issues with K9). If we're comparing 9th generation architectures, we should compare Nehalem (Intel's 9th generation) to K10. Core 2 is an _eighth_ computing generation architecture (in fact, many people are incorrectly calling it P8; it is more accurate to call it P6-5 as it is the 5th generation of the P6 architecture [Pentium Pro, Pentium II, Pentium III, Pentium-M, Core]. P7 is "officially" Netburst).So, once again, if we are to compare _computing_ generations: 7th: Netburst vs K7 (Athlon XP). 8th: NGMA (Core 2) vs K8. 9th: NGMA+1 (Core 4?) vs K10."This is about comparing similar technology, not about what's currently available."If we want to compare similar technologies _today_, we have to talk about what's available _today_. If we want to compare this yesterday, then we have to compare it yesterday. It's not like I'm making up these terms; they're not only in the materials these companies use to sell these chips, but also in academics when studying this material. Intel definitely screwed with the timeline by using a completely new, never previously-used architecture (Netburst) instead of continuing on and building chips like the Pentium-M (also a 7th generation chip) from the get-go, but these are the facts. "We all know the Core 2 beats the Athlon 64, because it has 3 years advantage in technology."We all know the Athlon 64 (released circa 2003) beats the Pentium 4 (released circa 2000), because it has a 3 year advantage in technology, not counting the fact that Netburst was never tested in the market before it was released. See how specious your argument sounds?
geminitojanusMay 4, 2007
"Little kiddie self haters trying to deny facts"Don't bother insulting them. People will Digg you down on this site just if they don't agree with you, not based on any merit of the statement at all. To reply to your comment though: sure, many businesses set up camp here in the US because our IP laws are some of the most corporate-centric in the entire world; there's no better place in the entire world to file a patent than USA. However, even companies that are "completely" American do a huge portion of their development overseas (Intel in Israel, AMD in Germany, etc).
geminitojanusMay 4, 2007
"believe them."Do you work for IBM? Do you want to show me a working 45nm device they've developed? I'll sign an NDA in a heart beat, I just want proof they exist and they're not puffing smoke on something they haven't achieved yet (yes, I do believe they're working on High-K and 45nm, I just don't see any proof of the "breakthroughs" they've claimed). Meanwhile engineers at TSMC and Intel are already pumping out 45nm devices in sample quantities, Intel's cranking out High-K 45nm devices (which people have seen, and which have been publicly demonstrated and thusly known to work)."Last I heard, this company averages over 3000 patents per year and this is all stuff they develop."Well I certainly hope it's stuff they develop, else their patents aren't valid now are they? To put this in perspective: Intel published about 1,900 in the same period, HP pulled out 2100 patents, Microsoft published about 1500. So sure, they do a hell of a lot of research and a lot of land-grabbing over at IBM, but that's not necessarily a sign that very much work is being done (and there's no current estimate published on how many of those patents are actually /valid/ patents).