![]() |
|
|
#12 | |
|
May 2009
7 Posts |
Quote:
And - it works! Mlucas is running, using almost 100% CPU. I no doubt have some learning ahead about how to make this process reliable and about how to properly submit any results that may emerge - but I thank you for getting me going! What a great project! |
|
|
|
|
|
|
#13 |
|
∂2ω=0
Sep 2002
República de California
103·113 Posts |
Glad to hear it - if you would be so kind as to paste the first few checkpoint summary lines from the p[exponent].stat file and maybe provide the salient stats about your system, I'd appreciate it.
We are hard at work on the next rev of the code, which will include automated primenet-based assignment/results management, so hopefully the manual method will soon be a relic of the past. Have Fun, -Ernst |
|
|
|
|
|
#14 | |
|
May 2009
7 Posts |
Quote:
Here are the first few lines of p30000037.stat: M30000037: using FFT length 1664K = 1703936 8-byte floats. this gives an average 17.606316786545975 bits per digit INFO: Using real*16 for FFT sincos and DWT weights tables inits. Using complex FFT radices 26 32 32 32 M30000037 iteration = 2000 clocks = 00:06:33.000. Res64: 5EE121BB38C1F800 M30000037 iteration = 4000 clocks = 00:06:33.019. Res64: BC9436A6AE34472E M30000037 iteration = 6000 clocks = 00:06:33.009. Res64: B7A03B5D306A2112 M30000037 iteration = 8000 clocks = 00:06:32.879. Res64: 3A225247EE43DCAF M30000037 iteration = 10000 clocks = 00:06:32.870. Res64: C0638BB696014BB5 M30000037 iteration = 12000 clocks = 00:06:32.689. Res64: E6089592F7B13769 M30000037 iteration = 14000 clocks = 00:06:32.740. Res64: 48423623417BBCC6 M30000037 iteration = 16000 clocks = 00:06:32.810. Res64: 0467D80E8CD096D6 M30000037 iteration = 18000 clocks = 00:06:32.879. Res64: 1E2DFE8BEC60E3DC M30000037 iteration = 20000 clocks = 00:06:32.920. Res64: B219BE4359BB307D Restarting M30000037 at iteration 20000; Current Res64: B219BE4359BB307D Using complex FFT radices 26 32 32 32 M30000037 iteration = 22000 clocks = 00:06:27.360. Res64: E78420EB29DC7B0E M30000037 iteration = 24000 clocks = 00:06:27.290. Res64: 189BA6D6FC83B11D M30000037 iteration = 26000 clocks = 00:06:27.240. Res64: A8AD8CE2BDA24DC6 M30000037 iteration = 28000 clocks = 00:06:27.250. Res64: 471CB2681109CA8E M30000037 iteration = 30000 clocks = 00:06:27.120. Res64: 01BBB7E40F52C2D6 M30000037 iteration = 32000 clocks = 00:06:27.129. Res64: DE36CEBFE0652157 M30000037 iteration = 34000 clocks = 00:06:27.250. Res64: 2AE1F6D06A4F2B8B M30000037 iteration = 36000 clocks = 00:06:27.279. Res64: F8147582F3616674 M30000037 iteration = 38000 clocks = 00:06:27.519. Res64: 6EB3CF21D2397BBD M30000037 iteration = 40000 clocks = 00:06:27.149. Res64: 33622BB34F89E7D3 M30000037 iteration = 42000 clocks = 00:06:27.199. Res64: 77176A9F3790AC53 M30000037 iteration = 44000 clocks = 00:06:27.250. Res64: F5A339E79D4B6073 M30000037 iteration = 46000 clocks = 00:06:27.240. Res64: 301D5FD4980B1BF8 M30000037 iteration = 48000 clocks = 00:06:27.079. Res64: 56448A40031515FB M30000037 iteration = 50000 clocks = 00:06:27.189. Res64: 4FE910E37A7C3D9C M30000037 iteration = 52000 clocks = 00:06:27.120. Res64: D1774237645532CF M30000037 iteration = 54000 clocks = 00:06:27.019. Res64: 3557140ECE71C45F M30000037 iteration = 56000 clocks = 00:06:26.990. Res64: EB038003BC7EE052 M30000037 iteration = 58000 clocks = 00:06:26.930. Res64: 0BBFD2DB993DF5B0 And here is some information about the system: # uname -a HP-UX hp B.11.11 U 9000/785 2012702596 unlimited-user license # model 9000/785/C3750 # getconf KERNEL_BITS 64 It is a HP C3750, with PA-8700+ processor (875 MHz) with 2.25 MB on-chip L1 cache and 2GB of PC120 SDRAM installed. If there is any other information that would be helpful, just let me know. |
|
|
|
|
|
|
#15 |
|
∂2ω=0
Sep 2002
República de California
103×113 Posts |
Thanks, smoky -
0.19 sec/iter for a 1664K FFT @ 875 MHz translates to a similar per-cycle throughput as Prime95's optimized SSE2 assembly code gets on a Core 2 processor - not bad for high-level compiled code. Too bad the PA-RISCs aren't available n 3GHz+ models, eh? But - we do what we can, eh? Should be a fine platform for double-checking. cheers, -E |
|
|
|
![]() |
| Thread Tools | |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Quad CPU setup question | tcharron | Software | 19 | 2013-02-14 20:38 |
| GPU setup | bcp19 | GPU Computing | 4 | 2011-10-15 14:09 |
| Setup on Mac | RussJones | Information & Answers | 4 | 2011-01-04 14:08 |
| A Followup question on Mlucas... | DeadSpam | Mlucas | 41 | 2007-03-14 18:40 |
| A quick question regarding iterations in Mlucas... | DeadSpam | Mlucas | 4 | 2007-03-05 14:06 |