Granted, the 1960s saw the rise of social program spending in America as well as the Vietnam War, both of which ballooned the federal budget and therefore makes the lunar program seem to cost less than it might have otherwise.Ĭhip makers are by their very nature less diverse than a national economy, and their moonshots are undoubtedly more costly relative to their size. Between 19, the US federal budget was a cumulative $2.06 trillion, and the lunar mission cost under 1.4 percent of that, and US gross domestic product from that same period adds up to $11.88 trillion, and the lunar program represented 24/10,000ths of GDP.
But perspective on what moonshots really cost is important. That number may seem like a lot, and if you adjust the total budget of the lunar missions between 19 to current dollars, it is more like $280 billion. NASA’s Apollo project landed people on the Moon, and was the culmination of a $28 billion effort by the United States government to be first to do so. What was true of the literal Ponte Vecchio bridge spanning the Arno River in Florence, Italy at the turn of the 10 th century or a cathedral built at the turn of the 13 th century was as true for the Apollo moon launch in 1969 and is also as true for the amazing processing devices that are being built during the AI revolution of the early 21 st century.Ĭhips are just another kind of stained glass, and projects that drive them are also moonshots. This thing is a Byzantine beast, but like every other complex device ever made, it is made of components that make sense in their own right, interconnected in clever ways to deliver what looks like a monolith. With Intel’s Architecture Day behind us and the Hot Chips 33 conference happening right now, this is a good time to take a hard look at the Ponte Vecchio device and see just what Intel is doing and what kinds of results the chiplet approach, with 2D EMIB stacking and 3D Foveros stacking combined in a single package, is yielding for a complex and powerful device. And it is pretty clear that the top brass at Intel are putting a lot of chips down on the felt table that the somewhat delayed “Ponte Vecchio” X e HPC GPU accelerator, which brings to bear all of the technologies and techniques the company can muster to create a powerful device, is the big bet for big iron. It is pretty obvious to everyone who watches the IT market that Intel needs an architectural win that leads to a product win in datacenter compute.
0 kommentar(er)
