Incoming links: Critical path
This is the mantra of the semiconductor industry:
- power and area are the main limiting factors of chips, i.e., your budget:
- chip area is ultra expensive because there are sporadic errors in the fabrication process, and each error in any part of the chip can potentially break the entire chip. Although there areThe percentage of working chips is called the yield.In some cases however, e.g. if the error only affects single CPU of a multi-core CPU, then they actually deactivate the broken CPU after testing, and sell the worse CPU cheaper with a clear branding of that: this is called binning www.tomshardware.com/uk/reviews/glossary-binning-definition,5892.html
- power is a major semiconductor limit as of 2010's and onwards. If everything turns on at once, the chip would burn. Designs have to account for that.
- performance is the goal.Conceptually, this is basically a set of algorithms that you want your hardware to solve, each one with a respective weight of importance.Serial performance is fundamentally limited by the longest path that electrons have to travel in a given clock cycle.The way to work around it is to create pipelines, splitting up single operations into multiple smaller operations, and storing intermediate results in memories.
The Supermen: The Story of Seymour Cray by Charles J. Murray (1997) Updated 2025-01-10 +Created 1970-01-01
Borrow from the Internet Archive for free: archive.org/details/supermenstory00murr
Initial chapters put good clarity on the formation of the military-industrial complex. Being backed by the military, especially just after World War II, was in itself enough credibility to start and foster a company.
It is funny to see how the first computers were very artisanal, made on a one-off basis.
Amazing how Control Data Corporation raised capital IPO style as a startup without a product. The dude was selling shares at dinner parties in his home.
Very interesting mention on page 70 of how Israel bought CDC's UNIVAC 1103 which Cray contributed greatly to design, and everyone knew that it was to make thermonuclear weapons, since that was what the big American labs like this mention should be added to: en.wikipedia.org/wiki/Nuclear_weapons_and_Israel but that's Extended Protected... the horrors of Wikipedia.
Another interesting insight is how "unintegrated" computers were back then. They were literally building computers out of individual vacuum tubes, then individual semiconducting transistors, a gate at a time. Then things got more and more integrated as time went. That is why the now outdated word "microprocessor" existed. When processors start to fit into a single integrated circuit, they were truly micro compared to the monstrosities that existed previously.
Also, because integration was so weak initially, it was important to more manually consider the length of wire signals had to travel, and try to put components closer together to reduce the critical path to be able to increase clock speeds. These constraints are also of course present in modern computer design, but they were just so much more visible in those days.
The book does unfortunately not give much detail in Crays personal life as mentioned on this book review: www.goodreads.com/review/show/1277733185?book_show_action=true. His childhood section is brief, and his wedding is described in one paragraph, and divorce in one sentence. Part of this is because he was very private about his family most likely note how Wikipedia had missed his first wedding, and likely misattribute children to the second wedding; en.wikipedia.org/wiki/Talk:Seymour_Cray section "Weddings and Children".
Crays work philosophy is is highlighted many times in the book, and it is something worthy to have in mind:
- if a design is not working, start from scratch
- don't be the very first pioneer of a technology, let others work out the problems for you first, and then come second and win
Cray's final downfall was when he opted to try to use a promising but hard to work with material gallium arsenide instead of silicon as his way to try and speed up computers, see also: gallium arsenide vs silicon. Also, he went against the extremely current of the late 80's early 90's pointing rather towards using massively parallel systems based on silicon off-the-shelf Intel processors, a current that had DARPA support, and which by far the path that won very dramatically as of 2020, see: Intel supercomputer market share.