Mobile Phones vs. Supercomputers of the Past 247
An anonymous reader writes "The recently published Top 500 list of the world's fastest supercomputers is based on the Linpack benchmark developed decades ago by Jack Dongarra. This same test has been ported to Android mobile phones, which means that we can compare the performance of our phones against that of the supercomputers of the past. For example, a tweaked Motorola Droid can hit 52 Mflop/s, which is more than 15 times faster than the CPUs used in the 1979 Cray-1." But even today's most powerful cellphones don't come with an integrated bench.
Integrated bench (Score:3, Interesting)
It's sad. I was at the Computer Museum in Mountain View a few years ago, where they had a Cray-I in a corner of the lobby, just sitting there used as a bench. It's not even labeled; some visitors think it's just furniture.
Re:Things like this... (Score:3, Interesting)
At the risk of appearing pedantic it's worth pointing out that not as many people thought that the world was flat as is commonly believed -
http://en.wikipedia.org/wiki/Myth_of_the_Flat_Earth [wikipedia.org]
Re:Things like this... (Score:5, Interesting)
If you were born in 2004 you would have missed out on everything. All you'd know is multi-core processors, terabytes and petabytes, touchscreen everything, wireless internet everywhere, 24/7 access to everyone you don't really know and directions to anywhere from anywhere available in your pocket. You'd have no appreciation for any of it and probably know nothing about computers because modern operating systems are far better than offerings in the 90s.
Trust me when I say you were born at the right time.
Re:1979 tech still wins (Score:5, Interesting)
Ah but all work on the Cray-1 was programed to be as parallel as possible, so cpu to cpu isn't an appropriate comparison. Much more useful is device output, in which case the 136 MFLOPS is significantly better performance than the 52 MFLOPS.
That is of course not considering that the designers of the Java applet that runs the benchmark admit that you're moreso benchmarking the java effeciency of a given device with their app than the full performance of the device.
Well, also not considering the $6m to $8m price tag of the Cray-1 vs the $200 (after rebate and 2 year plan) price tag of the Droid. Even factoring in inflation, I think my droid wins the performance-per-dollar crown by a little bit.
It does mean though that the intial statement "15 times faster than the cpus in the cray-1" is not quite reality. more like 5 times faster.
Re:Crays did proper work (Score:5, Interesting)
Not rendering bitty little colour screens or scanning for viruses. Plus the code was written to extract every last drop of power out of the architecture. So when you compare the amount of WORK a machine from the 70s or 80s did (my university's mainframe had a FORTRAN complier that needed less that 131kWord of memory - today the GRUB bootloader is bigger than that) with a more modern box, with all its overheads and inefficiencies, the balance isn't as great as the scoffers might think.
Does that make it any less impressive that a cell phone is putting up these kinds of numbers? Does it make it less impressive that you can code up an Linpack in Java, throw it at a JVM and rely on JIT compiler to optimize the DAXPY for the hardware on the fly? I think it both of those things are pretty damn impressive.
Cray did Last Starfighter, iPhone/Android better (Score:5, Interesting)
Today, the games you can play on your iPhone/Android or even the aging Nintendo DS have better graphics!! Resolution is a lot lower (not 3000x5000!) but at the screen size it certainly looks much better - and rendered in real time!
Re:Ridiculous Comparison (Score:5, Interesting)
Actually, for mobile devices, the most important metric is performance per unit of power instead of just performance per unit time. After a certain speed/throughput has been reached, nobody cares how fast the CPU is, only how long the battery lasts.
For scientific purposes, back when Cray was building systems, you got charged by the second you had access to the computer. So you carefully composed the solution to your problem to make darned sure every whizz-bang aspect of the computer was doing something useful all the time. Today, you just want to play a game for a while, then make a voice call, and don't want the battery to fizzle out before you get home (and maybe have some juice left for watching a show during your train ride home.)
Mobile devices don't try to match the throughput of all parts of the system, because it's not in anybody's interest to keep the I/O subsystem saturated close to capacity 100% of the time you're using your Droid/iPhone... in fact, they turn them off (go into a low power state) and do aggressive power management that is coordinated system-wide.
Just choose your date of Supercomputer ... (Score:2, Interesting)
When I started work as a computer programmer the Supercomputer of the time was the CDC6600 which had just taken the crown from the Ferranti Atlas.
When I took early retirement about 7 years ago, I often carried four devices which each needed about the power of the 6600 to function effectively:
A mobile phone
An MP3 player
A PDA (mainly used as an ebook reader)
A GPS (OK, I didn't carry this all that often)
A composer/researcher was using our University Mainframe (not quite that powerful) to produce music - his jobs typically ran for a whole 8 hour nightshift with an output of some 30 seconds of "music".
Re:Things like this... (Score:3, Interesting)
In fact articles from that time where talking about how to use More's law together with an estimation of how long a calculation would need to run, to decide when to buy the computer to finish said calculation quickest (provided that you couldn't or wouldn't upgrade the computer while the calculation was running).
This included economic calculations about the price of hardware, inflation and interest rates.