Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Prospects for 128 Bit Processors ( John Mashey SGI 1995)

yarchive.net

Post reply on HN
Print view View original

21–25 of 25 posts

Re: The Prospects for 128 Bit Processors ( John Mashey SGI 1995)

#21

Post by jburgess777 »

torginus wrote: post #19
bArray wrote: post #16

That exchange is quite a good prediction all things considered. > Note that "minor" implementation issues like die space, routing, and gate delays, especially of 128-bit adders & shifters are non-trivial, so people aren't going to rush out and build 128-bitters for fun, just as people matched timing dates of their 64-bitters to their expected markets. I think we're stuck with 64 bit for quite a while. The circuit siz…

Dunno, a 128bit integer add is essentially 2x as expensive as a 64 bit one, a multiply is 4x. We've already had SSE in the early 2000s that could do computations like this in a single cycle (tho on multiple 32 bit numbers, not 128-bit ones). I think on x86, we're still limited to 48 bits of address space on 64 bit systems, not sure if this has changed, but even 64 bits is so vast (16exabytes) , that you'd need a supe…

X86 CPUs with support for 57 bit of virtual address space have been produced by both Intel and AMD: https://en.wikipedia.org/wiki/Intel_5-level_paging.

Re: The Prospects for 128 Bit Processors ( John Mashey SGI 1995)

#22

Post by AlotOfReading »

anttihaapala wrote: post #13
Aardwolf wrote: post #8

Personally I'm disappointed 128-bit floating point (quadruple precision) never properly made it into CPU's (sure it appeared in some niche ones here and there, but not in what we actually use today). After all, in the 1980's they had 80-bit ones, it's not even that far off, and they had millions times less transistors then. While probably niche and applications that need higher precision using their own custom types…

The problem is for most practical uses where you need fast calculations the 64 bit precision is enough. For example there is hardly any physical calculation that would need more precision. 64 bit float can be used to measure the Earth-Sun distance to 30 micrometre precision. 128 bits does not just add anything generally useful. For monetary calculations you should be using decimals instead of binary anyway.

Quads make it feasible to write correctly rounded double functions. You can almost do the same things with double-double, but the result is slower and implementing most of the standard functions remains a bit of an open problem.

Re: The Prospects for 128 Bit Processors ( John Mashey SGI 1995)

#23

Post by jdashg »

bArray wrote: post #16

That exchange is quite a good prediction all things considered. > Note that "minor" implementation issues like die space, routing, and gate delays, especially of 128-bit adders & shifters are non-trivial, so people aren't going to rush out and build 128-bitters for fun, just as people matched timing dates of their 64-bitters to their expected markets. I think we're stuck with 64 bit for quite a while. The circuit siz…

I am compelled to point out how much "For uint64_t and UINT64_MAX you will be waiting a long time" really undersells it. If we assume that the uint32_t version runs in one second, the u64 version will take 136 years. In reality, on the machine in my lap, u32 takes 0.6s, so we can expect the u64 version to take at least 81 years! Statistically, "you" will not be waiting a long time, because you will die first.

"128 bit values? Even longer." Yes, 2^64 times longer than u64, and 2^96 times longer than for u32! 2^96 just about 80e27. Age of the universe is ~13.8 billion (13.8e9) years, or ~435e15 seconds.

If we consider a more moderate 96-bit size, and I started at the very beginning of the universe, I will only have been able to run that program just over 25 times by now. That's all.

Post reply on HN