They do ship internationally but the shipping cost to Canada is almost twice the price of the book.
I'd love a signed copy of the shattering peace but I'll have to settle for an e-book version.
Any plan for a tour in Ontario?

Didn't take a screenshot but the machine I worked on this weekend has 132 cores, 1056 threads and 48TB of RAM

As400 was created in 1988.
Linux 1991
Windows 1985
Unix 1971

Latest release was less than a month ago and the roadmap is still there and strong.

It's not really possible to begin with since it's impossible to fire a public sector employee in France.

AFAIK Intel announced phasing out of 32-bit CPU, not of 32-bit mode in 64-bit CPUs.
It all depends on how you define the most widely used architecture, currently ARM is more commonly used than x86.
Remember, Intel wanted to kill x86, they didn't want x86_64.

Motorola, knew that CISC was not the way to go forward and went into PowerPC... things didn't go that well with IBM and Apple and they didn't have the huge PC market to allow them to survive the design mistakes they made.

x86 absolutely works but it's like the IE6 of CPUs, all the others do better or much better on almost everything (Quick Sync being a very notable exception) but everybody uses it because it's the de facto standard.
I wouldn't hold my breath for 32 bit removal!

True but even the core x86 instructions are broken down into micro operations.

At least they tried to innovate with itanium (even though I think it would have been terrible long term).
The problem (beside the price and the performances less than stellar) was that AMD created amd64 (now x86_64) and the industry went the easy way and said no to itanium

It would be slower. They switched to an underlying RISC architecture because it's way easier to optimize with current technology than cisc.
But this "emulation" isn't without overhead, that's partly why x86 is so bad compared to almost everything else despite the massive r&d done by Intel and AMD.

Part 2 was something else.
First of all, it was extraordinarily vague, second, it needed a brute force method, all in all not a great part.

I still used some regular expression to read the input data, not that it was absolutely necessary but it works pretty well.

That was really a fun puzzle (after a very frustrating few minutes I have to admin!).

RPG is very expressive with array index out of bound and number overflow, this helped me a lot avoiding silent bugs!

After (way too much) thinking, the data structure can be greatly simplified, the program runs in a few minutes and uses a few KB of memory, victory!

OK I can slice everything to keep memory under control, compile, run and wait and wait and wait... After several minutes we're still at 30ish iterations. Those puzzle are supposed to be solved in under 15s by any current hardware... something must be wrong!

Part 1 was pretty straightforward with a simple nested loop and dynamic table.

Part 2 description is short and sweet: do 75 iterations instead of 25: easy, just change the outer loop to do 75 iterations, compile, run and BAM! memory goes through the roof!

Not that RPG isn't suited to solve it (on the contrary) but the logic isn't trivial. Also part 2, while not looking very different from part 1 at first, needed a completely different solution. A very fun challenge!