4 ms·

"The first generation of Itanium was hugely expensive, didn't run x86 code well, and couldn't easily masquerade as a plain Jane x86 machine." THAT was its bigg

"The first generation of Itanium was hugely expensive, didn't run x86 code well, and couldn't easily masquerade as a plain Jane x86 machine." THAT was its biggest blunder -- the x86 part. It was pointless, and a waste of silicon, and it also throttled the chip back with a lot of extra overhead. On top of everything else, it added NO value whatsoever, but it did make introducing the chip a lot harder, and designing it cost a lot more than it should have. Had they stuck to their original plan of pushing the architecture as far as they could and using a binary translator to provide x86 emulation, they would probably have had a barnstormer that made even IBM quake in its boots. Instead, they hobbled it with the biggest brick they could find. "Meanwhile, AMD looked at the same problem and came to a different conclusion." In addition to the list you provided, AMD started abandoning backward compatibility in Long mode. "Eventually they succeeded with a very substantial redesign of their hardware architecture (the core series) which paid huge dividends in the 2nd iteration as it proved to be far superior even to AMD's offerings." It took Intel less than 2 years to launch Core after AMD hit the market over the head with... er, Sledgehammer. (Pun not intended, but that WAS AMD's code name.) Core was a 5-year project -- obviously, they were already working on it, and turning the cranks almost to the point of melting motherboards was clearly just to extend the life of the Netburst architecture beyond where Intel thought they'd need to. I suspect that their original hope was that they be able to scale up the performance of Itanium far enough to exceed x86 performance even in emulation, and since Itanium was a simpler architecture (without hardware x86 compatibility, it's a FAR simpler architecture with a LOT more floating point throughput per clock cycle even now), it would be cheaper to manufacture. Intel blundered with the hardware x86 silliness and AMD caught Intel with its pants down. I would, however, argue that AMD's ccNUMA architecture was at least as big a deal as the fact that they introduced 64-bit computing to the mainstream -- which was huge.

InclinedPlane 16y ago
Fortunately Intel is not quite so dumb as to not have multiple irons in the fire at the same time. But they came closer to killing the company with Itanium & NetBurst than they ever had before. Itanium is a special case of the classic problem of trying to rebuild from scratch. If you do it right it can be a huge win, if you do it wrong you can kill the company. Generally speaking most times that such efforts fail it's due to ignoring the need to take the new architecture through enough iterations to gain maturity and parity with the old architecture, and ignoring the extensive costs of having 2 parallel architectures in play at the same time and the costs associated with transitioning from one to the other. The efforts that fail almost always assume that the new system will be 100% capable on day 1 and the switchover will be effortless and complete instantaneously (rather than costly, time consuming, and incremental).
- Tamerlin 16y ago
  Except that Intel wasn't in any danger of dying, I agree. Even with Itanium's steep costs and delays, NetBurst came pretty close to keeping pace with the Athlon and Athlon 64 -- close enough for Intel to compete with price. Admittedly, it was largely due to the fact that Intel had the edge in both fabrication technology and capacity, but either way AMD couldn't make enough processors to satisfy global demand. By way of example, recall that in the first month after Intel started shipping its first 64-bit x86 processor (a NetBurst Xeon whose code name I have forgotten), that processor outsold the combined annual production of G5, Athlon64, AND Opteron. AMD had a lead in many ways, but wasn't even in the ballpark as far as fabrication capacity. "The efforts that fail almost always assume that the new system will be 100% capable on day 1 and the switchover will be effortless and complete instantaneously (rather than costly, time consuming, and incremental)." Agreed -- and it's all the more reason that Intel should have relied exclusively on the binary translation technology that they inherited along with the Alpha development team, technology, and intellectual property.

InclinedPlane 16y ago

Fortunately Intel is not quite so dumb as to not have multiple irons in the fire at the same time. But they came closer to killing the company with Itanium & NetBurst than they ever had before. Itanium is a special case of the classic problem of trying to rebuild from scratch. If you do it right it can be a huge win, if you do it wrong you can kill the company. Generally speaking most times that such efforts fail it's due to ignoring the need to take the new architecture through enough iterations to gain maturity and parity with the old architecture, and ignoring the extensive costs of having 2 parallel architectures in play at the same time and the costs associated with transitioning from one to the other. The efforts that fail almost always assume that the new system will be 100% capable on day 1 and the switchover will be effortless and complete instantaneously (rather than costly, time consuming, and incremental).

Tamerlin 16y ago
Except that Intel wasn't in any danger of dying, I agree. Even with Itanium's steep costs and delays, NetBurst came pretty close to keeping pace with the Athlon and Athlon 64 -- close enough for Intel to compete with price. Admittedly, it was largely due to the fact that Intel had the edge in both fabrication technology and capacity, but either way AMD couldn't make enough processors to satisfy global demand. By way of example, recall that in the first month after Intel started shipping its first 64-bit x86 processor (a NetBurst Xeon whose code name I have forgotten), that processor outsold the combined annual production of G5, Athlon64, AND Opteron. AMD had a lead in many ways, but wasn't even in the ballpark as far as fabrication capacity. "The efforts that fail almost always assume that the new system will be 100% capable on day 1 and the switchover will be effortless and complete instantaneously (rather than costly, time consuming, and incremental)." Agreed -- and it's all the more reason that Intel should have relied exclusively on the binary translation technology that they inherited along with the Alpha development team, technology, and intellectual property.