Message boards :
Number crunching :
Today's average Core 2 Duo PC versus Cray-2 (1985) supercomputer
Message board moderation
Author | Message |
---|---|
![]() ![]() Send message Joined: 21 Jul 07 Posts: 147 Credit: 2,204,402 RAC: 0 ![]() |
Can today's average Core 2 Duo PC beat the above supercomputer in terms of GFLOPS & MIPS? |
Fred W Send message Joined: 13 Jun 99 Posts: 2524 Credit: 11,954,210 RAC: 0 ![]() |
Can today's average Core 2 Duo PC beat the above supercomputer in terms of GFLOPS & MIPS? Well, according to what I can find, the US Army reckon the Cray 2 had a peak of 1 GFlop and other (www) sources indicate about 825 MIPS. My O/C'd E6400 C2D is reported by Boinc as performing at 3.15 GFlops and 6971 MIPS unless I entirely misunderstand the figures (which is entirely possible :) F. |
![]() ![]() Send message Joined: 18 May 99 Posts: 534 Credit: 5,475,482 RAC: 0 |
Can today's average Core 2 Duo PC beat the above supercomputer in terms of GFLOPS & MIPS? I remember the Cray machine at Los Alamos Labs in the late 70 and early 80 time frame, with instruction times in the low pico second range. It used a couple of CDC 7600 as file server systems. Long time ago now. When we finally figure it all out, all the rules will change and we can start all over again. |
![]() ![]() Send message Joined: 21 Jul 07 Posts: 147 Credit: 2,204,402 RAC: 0 ![]() |
Can today's average Core 2 Duo PC beat the above supercomputer in terms of GFLOPS & MIPS? Where did you get the program to measure your PC's MIPS & GFlops? I'm looking for it too. And can it be run on Windows 98? From the GFlops & MIPS figures you mentioned, it took about 20 years for the processing power of a 1985 supercomputer to make it to the desktop. Similarly, I read that a classic Pentium (no MMX) system of the mid-90s had the processing power of the 1976 Cray-1 supercomputer. Computers have come a long way. If Moore's Law applies over the next two decades, we would have today's supercomputer on our desktops in 2027. |
![]() Send message Joined: 9 Jun 99 Posts: 30977 Credit: 57,275,487 RAC: 157 ![]() ![]() |
Can today's average Core 2 Duo PC beat the above supercomputer in terms of GFLOPS & MIPS? "Who" knows, mabey sooner. ![]() |
Fred W Send message Joined: 13 Jun 99 Posts: 2524 Credit: 11,954,210 RAC: 0 ![]() |
Where did you get the program to measure your PC's MIPS & GFlops? I'm looking for it too. And can it be run on Windows 98? Didn't measure them specially. Just looked at the Boinc "Computer Summary": Measured floating point speed 3149.69 million ops/sec Measured integer speed 6971.64 million ops/sec May not be totally accurate but probably close enough for current purposes. F. |
Aaron Send message Joined: 14 Jan 07 Posts: 1 Credit: 3,027 RAC: 0 ![]() |
If Moore's Law applies over the next two decades, we would have today's supercomputer on our desktops in 2027. We already do... Take a look at the latest Cray Machinery, the X3 and X4. They are built around AMD Opteron Processors... only Cray has made the systems scalable to 120,000+ CPU's. Spec'd out, what are we talking? 240,000+ Cores??? Holy Bat Flaps Bat Man, thats alot of BOINC'ing. My point is this though, if the standard "supercomputer" is being built on consumer level technology... well gee, welcome to the 22nd Century Buck Rodgers. All they are doing is scaling things up. Basically its distributed computing on the xxx,xxx scale, inside ONE box. Screw GHz ratings these days... ITS THE CORE WARS!!! Im waiting for my 128 core AMD Phenom rite now. |
archae86 Send message Joined: 31 Aug 99 Posts: 909 Credit: 1,582,816 RAC: 0 ![]() |
While this is nearly true regarding the CPUs, it is not true at all regarding the interconnect communication. Issuing results on a time scale of days from servers in Berkeley, and only gathering up any conclusions requiring consideration of what went on in more than one result weeks to years later is many orders of magnitude longer latency than are obtained in "one box distributed computing". And it needs to be--not so many problems can readily be broken down into independent chunks which hardly ever need to talk to each other. The lead of single-processor execution speed in supercomputers vs. consumer desk-top machines has shrunk mightily since 1975, which is bad news for the sort of problem that does not parallelize nicely. |
siliconflux ![]() Send message Joined: 31 Dec 02 Posts: 18 Credit: 140,154,946 RAC: 0 ![]() |
Can today's average Core 2 Duo PC beat the above supercomputer in terms of GFLOPS & MIPS? Moore's Law will eventually reach a plateau with regards to single-core electron-based technologies. At the very least, the exponential increase will be reduced. Rumor has it that we have already reached a point where the effective heat dissipation cannot keep up with near future processing demands. Then again, we can simply keep combining cores until we run out of case space or heatsinks. Hopefully by then, quantum computers will be out. Unless the CPU manufacturers continue with Core-type technologies, we might b |
![]() ![]() Send message Joined: 14 Mar 04 Posts: 357 Credit: 650,069 RAC: 0 ![]() |
Can today's average Core 2 Duo PC beat the above supercomputer in terms of GFLOPS & MIPS? There's odd stuff out there that could change things. One of the more interesting was an idea/HW that IBM patented about 10 years ago. Trinary logic- instead of 1/0, you have 1/0/-1. Basically a lot more horsepower with the same amount of HW, but the logic is a mess, not to mention problems connecting it to the real world. ![]() ![]() Boinc Button Abuser In Training >My Shrubbers< |
©2025 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.