A newborn, nearby supernova with the potential to significantly improve our knowledge of the universe has been discovered by a supercomputer, two telescopes on opposite sides of the world, a sharp-eyed astronomer, and his helpful Oxford colleague. The Reg spoke with Peter Nugent, the California astronomer who first spotted the …
> "That's the nice thing about NERSC having all these cores available," he said. "We could just increase the load with just one change in one line in a piece of code and, boom, it just goes off and grabs more processors."
Uh, huh? You have to actually edit and recompile the code to change the number of CPUs? Sorry, that's hideously crude and primitive, cloud apps based on services like AWS or Appengine can do that without needing to be rebuilt, or even restarted. They should try not to boast about their weaknesses.
Thus spake the Master Programmer:
> "Cloud apps based on services like AWS or Appengine can do that without needing to be rebuilt. They should try not to boast about their weaknesses."
Who better follows the Tao? He who writes code of great elegance and power to fuel shopping carts and accounting packages, or he who writes ten thousand lines of dross that bears witness to a dying star?
...that 'line of code' may be part of the job's run control script.
Its not unusual for jobs run on systems bigger than a bog standard server to require the run control script to tell the OS what resources this run requires: not doing so is an excellent way of wasting computer power. The number of cores the job needs is an obvious candidate for such a request.
Re: Thus spake the Master Programmer
You sir, are fucking awesome.
Don't be daft. It's a simple configuration file change.
vi $filename.config; edit the string "UseCores=60" to "UseCores=120" ... and then run the file. No re-compilation required.
Why bother? I'll leave that as an exercise for the reader ...
Input for the mods ... and commentards.
One of the Grahams (Marsden or Bartlett) commented that they had hit "reply to post", but it had appeared as a new comment not a reply ... I pooh-poohed the comment, suspecting he was just hitting the "post your own message" button at the bottom of the comments. Which I *HAVE* done. I'm human, I make (mi)steaks.
But it just happened here in this thread, to me. I responded to the first reply to this article, using the "reply to post" button ... but my post shows up here as a new sub-thread. Only one post had been allowed by the mods when I posted.
I suspect an "off-by-one" error when only one post has been allowed ...
Why post here, instead of emailing the mods directly? It's probably an article where the computer savvy will read the comments ... thus giving the commentards a heads-up. The more eyes the better, when tracking down this kinda thang ... Over a third of a century of open source kinda gives a guy an opinion ;-)
Re: Input for the mods ... and commentards.
Ok we'll check it out.
Re AC: You've got to love smartasses. You're what, 15? Don't you know that supercomputers are not effing webservices?
But to be ontopic: way cool stuff.
34 teraflops per second? Wow!
It gets more powerful the longer it runs!
"34 teraflops per second"? "1.05 petaflops/sec"?
presumably you just mean "34 teraflops." and "1.05 petaflops"
FLOP = floating point operations per second.
come on reg!
So glad I'm not the only sad nerd who could't resist pointing this out. It just.... grates.
Floating Point Accellerators!
I, too, was pretty amused at the occurrences of 'flops/s'. From a lot of news sources I would mostly just roll my eyes a bit, but I do rather expect El Reg to not do such things, if only because it's the sort of thing they would mercilessly point out if done by others. :)
Picking nits, I know!
Why do you need the FLOP count anyway?
Isn't that, like, integer processing that we want here?
aren't 1a supernovae more useful for investigating dark energy than dark matter?
but if we were being very generous we'd point out that technically they're useful for both. let's assume we have cmb+sn1a data. then the sn1a data breaks degeneracies within the cmb data and lets us fix, for example, the energy density in dark energy. since the cmb is best fitted by a flat universe (energy density equal the critical energy density where in the presence of normal matter the universe would tend to stall and neither expand nor collapse -- that statement is obviously not true for a universe containing dark energy, but the geometry remains "flat" nonetheless) knowing the dark energy density gives us the density in clustering matter (baryons plus dark matter). we then also know, both from the cmb and from other sources such as the abundance of hydrogen and helium (and lithium etc) in the universe that the energy density in normal matter (baryons) is roughly 5%. putting it all together we get
70% dark energy (sn1a) + 5% normal matter (big bang nucleosynthesis/cmb) = 100% - x% dark matter. So sn1a ultimately tells us that we've got 20% dark matter.
in reality i suspect it was a typo (1a).
A closer look
A closer look at the picture and in all that exciting technological advances and there is a piece of wood holding the cable tray up on top of the nearest rack!!
I bet there is a dwarf turning a big wheel at the back of the room!!
Did Anybody notice...
What appears to be a Unix stereotypical geek in the top left of the picture?
No stinking titles used here
Are you sure thats not the spare dwarf, just in case they want to run more cores?
Probably the code includes the line
'Activate second dwarf with sharp stick'
Hiding in this picture
There are 23 ninjas hiding in this picture.
Where's the H1B worker?
Wonder if he also wears a belt with those suspenders.
I can only find 21... where are the other 2?
That is all.
Here's a title for you
Where are the big tape reels? Without them it's not a computer!
"discovered by a supercomputer, two telescopes on opposite sides of the world, a sharp-eyed astronomer, and his helpful Oxford colleague."
And a small dachshund, called Colin.
Rupert the parrot.
Jupiter's radius is only 70,000 km so object is not all that much denser than the gas giant. Still long live the Ringworld!
off topic warning
Wanted to post on other article about scientists finding diamond star core orbiting a pulsar but article doesn't have post section. " The object has more mass than Jupiter" and "its own radius has to be less than 60,000 km". Wow didn't Niven's ringworld contain about Jupiter's mass and it was 600 million miles circular? Granted its fiction and its a ring and not solid but thats compact for non exotic material. Yes pointless but in a posting mood.
Stirling Colgate, are you out there?
Looks like DigAs came to be, after all...and is finding just the sort of things you expected it to. Very Nice!
Thank you, El Reg
Thank you for this excellent article. Its good news about the rapid supernova discovery, but I particularly enjoyed the way your article laid out the details of how it was found.
type 1a supernova
One way of looking at it, is that a big enough diamond isn't around forever, or even for very long.