News for hackers. Stuff that matters.
Re:Linux dominates supercomputers as never before by TallGuyShort on Tuesday June 24, 2014 @05:55PM (#7940913)
What I think is more amazing is that with the exception of the 2 Windows systems, EVERYTHING is either AIX UNIX or Linux, or a combination. UNIX started as a "microcomputer" OS, but it really has proven to be incredibly flexible. Personally, I think the ideas that have come out of the Solaris and BSD communities are invaluable and I always like to see them included in projects - I'm a little disappointed they have lost favor in the HPC community. They are similar enough to Linux and other UNIX-like OSs that targeting all of them is relatively practical in most cases, and users will find similar interfaces everywhere, but their ideas, communities and approaches are different enough that I think they would prevent it from being a monoculture.
Re:Linux dominates supercomputers as never before by _delirium on Tuesday June 24, 2014 @07:27PM (#7941354)
One of the things that imo has pushed some of the other Unixes out of HPC is the move to using cluster file systems. Any Unix supports a traditional NFS-mounted NAS solution for storage, as used to be typical in HPC clusters (and still is fairly common in smaller installations at universities). But only AIX and Linux really have solid cluster filesystems; there's nothing in the BSD world analogous to IBM GPFS, or to Lustre, and porting is nontrivial.
Re:Linux dominates supercomputers as never before by TallGuyShort on Wednesday June 25, 2014 @11:47AM (#7944676)
I think you're right. DragonflyBSD probably isn't widely used enough to be trusted by high-budget HPC projects anyway, but I was under the impression their HAMMER filesystem was a cluster file system. Looking at the website now I can't seem to find support for that, though...
Re:Linux dominates supercomputers as never before by _delirium on Wednesday June 25, 2014 @06:15PM (#7947069)
It's one of DragonFly's goals, but afaik the clustering support is scheduled for the HAMMER2 rewrite that's still in progress: http://leaf.dragonflybsd.org/mailarchive/users/2012-02/msg00...
Re:Linux dominates supercomputers as never before by kjs3 on Tuesday June 24, 2014 @10:01PM (#7941863)
> UNIX started as a "microcomputer" OS

One very minor nit...the PDP-7 where Unix started out cost US$72k+ in 1969 dollars. Nothing compared to the big iron of the day, but very far from a microcomputer (and, really, that market didn't exist until 1972 or so).

Re:Linux dominates supercomputers as never before by hga on Wednesday June 25, 2014 @11:46AM (#7944669)
And as far as capabilities go, the PDP-7 could address 64K of 18 bit words, more than double 1st generation microcomputers (those prior to the IBM PC and Apple Macintosh). Then again, I don't think it's generally known how much (very expensive, hand made) core memory was on that first UNIX(TM) PDP-7.

Maybe not that much, it next moved to the first model of the PDP-11, which had a max of 56 KiB memory of 8 bit bytes. Then to an 11/45, which could do 64 KiB of code, 56 KiB of data and 8 KiB of stack.

Re:Linux dominates supercomputers as never before by kjs3 on Friday June 27, 2014 @02:09AM (#7952902)
The 11/45 has a 16-bit address space optionally expanded to 18-bits with an MMU so could do 256KB of memory. With the MMU, you could take advantage of split instruction and data memory, which gave a pair of 64k address spaces mapped into the 18-bit space.

No PDP-11 AFAIK have a separate, dedicated stack space. You're thinking of the 8k dedicated I/O page.

Re:Linux dominates supercomputers as never before by hga on Friday June 27, 2014 @08:48AM (#7953712)
Ah, yes, for the 11/45 I was referring to what a single program could do.

However, as I recall (from the very early '80s when I was using both a 11/44 and 11/45), the MMU had a 8 KiB page size, so the natural thing to do was to divide the data as described, so a stack overflow would trigger a trap. Although after a bit of time with Google, I note the 11/40 manual says all PDP-11s have a location in which to put a stack boundary, so maybe that was a convention which matched what the MMU could do.