Re: Mersenne: Error message from prime95 on an old Win95 box
Try http://www.sysinternals.com/ntw2k/source/regmon.shtml to get the name and value of the registry entry. (or maybe anyone knows what's going wrong) Greetings, Mohk Am Mittwoch, 17. Juli 2002 03:34 schrieb A T Schrum: Hi Folks, I didn't find a reference to this problem. My old PentiumMMX 200 Mhz box running Win95 OSR2 (with tons of patches) now has Prime95 2.26.1 on it and it runs reasonably faster (about 20ms faster at 768K FFT size). But upon startup, Prime95 reports Can't write registry value and continues on. Should I be concerned? Thanks, -Allan _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Error message from prime95 on an old Win95 box
At 09:34 PM 7/16/2002 -0400, A T Schrum wrote: I didn't find a reference to this problem. My old PentiumMMX 200 Mhz box running Win95 OSR2 (with tons of patches) now has Prime95 2.26.1 on it and it runs reasonably faster (about 20ms faster at 768K FFT size). But upon startup, Prime95 reports Can't write registry value and continues on. Should I be concerned? I doubt it. Prime95 should be trying to create a registry entry to run the program at bootup. If you uncheck the Options/Start at Bootup menu item the problem should go away. I'm curious though. Do other Win95 users have the same trouble? _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Re: P-1 limits
Recently I've been translating Prime95's P-1 limit-choosing algorithm into another language for a P-1 utility of mine. Let me offer some further explanation. (George, please point out any errors!) Inputs to the algorithm: exponent of the Mersenne number to be factored, how far (as a power of 2) trial factoring of that Mnumber has reached, whether the P-1 effort precedes a first-time LL test or a double-check, how much memory is to be available during P-1 factoring, plus empirically-derived constants and fudge factors in various formulas. The algorithm is independent of specific systems or CPUs. That is, it does not consider or estimate actual time needed to perform an operation, or consider a given CPU's relative performance on integer vs. floating-point instructions. Its basic unit of measure for the cost of a procedure is the FFT squaring (= a transform, a convolution, then another transform). For GCDs, which do FFT multiplications that are not squarings, the algorithm estimates the number of FFT transforms needed by the GCD, then divides by two to get the equivalent number of squarings. If a P-1 factoring run with particular B1 and B2 bounds will require 175,000 FFT squarings in stage 1, a GCD that will take as much time as 3,000 FFT squarings, and 122,000 FFT squarings in stage 2, the algorithm considers that P-1 run's total cost to be 300,000 -- it is assumed that the time required for other parts of the run can be neglected in comparison. (George -- Doesn't a two-stage P-1 require two GCDs, not just one? And it looks like some of the cost comparisons don't include even one GCD ... was that a deliberate omission, or am I overlooking something?) Dickman's function (see Knuth Vol. 2, or previous discussion on this list) is used to calculate the probability of the P-1 procedure's finding a factor. Perhaps you've wondered what role the how-far-trial-factored input plays. Well, if a Mnumber has been trial-factored to 2^63 (with no success, or else we wouldn't be considering P-1 now), then P-1 will not find any factors of 63 bits or less. But if the Mnumber had been trial-factored to only 2^57, then there would still be some chance of finding a factor between 2^57 and 2^63 during a P-1 run, in addition to the chance of finding a factor above 2^63. The algorithm takes this into account. Continuing the example from above -- If that 300,000-squaring P-1 run has a 1.25% probability of finding a factor, then it would balance an L-L test costing 300,000/.0125 = 24,000,000. L-L tests are considered to cost 1.015 times their number of FFT squarings, to balance the historically observed 1.5% rate of erroneous L-L results, so an L-L test on an exponent of, say, 22,000,000 would be considered to cost 22,000,000 * 1.015 = 22,330,000 squarings. In this case, the P-1 limit-choosing algorithm would decide that a P-1 cost of 300,000 with probability of 0.0125 would not be worthwhile on an exponent of 22,000,000, but would be worth a try on an exponent of, say, 25,000,000 because 300,000/.0125 is less than 25,000,000 * 1.015. The above example is somewhat backwards, though, because the algorithm actually considers only one given exponent's L-L cost versus the costs and probabilities for several combinations of B1 and B2 bounds, and picks out the best combination of P-1 cost and probability for that exponent. Prime95 considers B1 values starting at 10,000 then going up in increments of 5,000. For each B1, Prime95 considers B2 values starting with B2 = B1, then going up in increments of B1*0.25. So, for B1 = 10,000, Prime95 calculates costs and probabilities for B2 = 10,000, 12,500, 15,000, 17,500, and so on. Then for B1 = 15,000, it considers B2 = 15,000, 18,750, 22,500, 26,250, 30,000, and so on. When does the algorithm stop incrementing B1 and B2 values it examines? Either (1) when the B2 stage would require too much memory (one of the algorithm inputs was how much memory is to be available during P-1 factoring), or (2) when the relative worth of the P-1 run falls below 90% of the most worthy P-1 bounds combination so far considered. By relative worth and worthy, I mean the amount by which [L-L cost multiplied by probability of P-1 success] exceeds [P-1 cost]. The shape of Dickman's function guarantees that criterion (2) is eventually satisfied at some increment of both B1 and B2. Where the first-time/double-check matters: For a double-check, the L-L cost is as illustrated in the example above. For a first-time L-L, the L-L cost against which P-1 cost is compared is twice that value -- i.e., the algorithm multiplies the exponent by 1.015 * 2 = 2.03 to get the number of L-L squarings. That's because a successful P-1 factoring before a first-time test will eliminate the need for both the first-time and double-check L-L tests rather than only the double-check L-L test. Formulas for calculating the numbers of FFT squarings required
Re: Mersenne: Error message from prime95 on an old Win95 box
No go. The box was unchecked. I checked it, restarted Prime95, and the error message was not there. So I unchecked it, restarted Prime95, and the error message came back. George Woltman wrote: At 09:34 PM 7/16/2002 -0400, A T Schrum wrote: I didn't find a reference to this problem. My old PentiumMMX 200 Mhz box running Win95 OSR2 (with tons of patches) now has Prime95 2.26.1 on it and it runs reasonably faster (about 20ms faster at 768K FFT size). But upon startup, Prime95 reports Can't write registry value and continues on. Should I be concerned? I doubt it. Prime95 should be trying to create a registry entry to run the program at bootup. If you uncheck the Options/Start at Bootup menu item the problem should go away. I'm curious though. Do other Win95 users have the same trouble? _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Error message from prime95 on an old Win95 box
Sounds like the opposite problem: Prime95 is trying to delete a registry entry that doesn't exist. I had one do that to me recently. Rather than uncheck the box, manually edit ( in prime.ini ) the line windows service=1 (or whatever line it has to that effect) to ...=0 and it will no longer see a need to try to delete the registry entry. And the box will now show as unchecked. Hope that helps, Steve Harris -Original Message- From: A T Schrum [EMAIL PROTECTED] Date: Wednesday, July 17, 2002 5:26 PM No go. The box was unchecked. I checked it, restarted Prime95, and the error message was not there. So I unchecked it, restarted Prime95, and the error message came back. George Woltman wrote: At 09:34 PM 7/16/2002 -0400, A T Schrum wrote: I didn't find a reference to this problem. My old PentiumMMX 200 Mhz box running Win95 OSR2 (with tons of patches) now has Prime95 2.26.1 on it and it runs reasonably faster (about 20ms faster at 768K FFT size). But upon startup, Prime95 reports Can't write registry value and continues on. Should I be concerned? I doubt it. Prime95 should be trying to create a registry entry to run the program at bootup. If you uncheck the Options/Start at Bootup menu item the problem should go away. I'm curious though. Do other Win95 users have the same trouble? _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Re: P-1 limits
Recently I wrote: Well, if a Mnumber has been trial-factored to 2^63 (with no success, or else we wouldn't be considering P-1 now), then P-1 will not find any factors of 63 bits or less. Tom Cage has reminded me that he and others have found factors with P-1 that are smaller than the posted bit limit in the current nofactor.cmp at http://www.mersenne.org/gimps. AFAIK, the number of such findings is too small to make any significant difference in the probabilities calculated by the P-1 limit-choosing algorithm, but it would be well to remember that it can still happen. Richard Woods _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Error message from prime95 on an old Win95 box
At 06:01 PM 7/17/2002 -0400, A T Schrum wrote: No go. The box was unchecked. I checked it, restarted Prime95, and the error message was not there. So I unchecked it, restarted Prime95, and the error message came back. I reactivated an old Win98 box for debugging. You are correct. The error message is harmless. I've got a fix ready for download at ftp://mersenne.org/p95v227.zip The only new feature a v22.7 is SSE2 based trial factoring for factors above 2^64. It is about 4 times faster than v22.6. Since most P4 users are getting assigned prefactored exponents, the speedup is of no value. However, if you are working on 10 million digit numbers, you may get assigned a number that needs factoring and the new code will certainly help. _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
Re: Mersenne: Error message from prime95 on an old Win95 box
Hi Steve, My current prime.ini has Windows95Service=0, and the Win95 Service box is not checked. However, whenever the program starts, it gives me the Can't write registry value error. I looked into the latest source (source22.zip) and I see that for Win95 systems, it seems to always try to delete the key: HKEY_LOCAL_MACHINE\ Software\ Microsoft\ CurrentVersion\ RunServices\ Prime95 but cannot because it is not there (as you said). This is in the routine Service95(). But this is not supposed to be called unless Prime95 is setup as a Windows service. For some reason, on my Win95 box, Service95() is always called. I tried deleting the Windows95Service option from prime.ini, then set Prime95 to be a service and it properly changed the prime.ini file. I do not use the -A option with Prime95, so the prime.ini in the Prime95.exe directory is used. But whether the Windows95Service option is deleted or just set to zero, I still get the error. The code in Prime95.cpp looks fine, but I must be missing something. But wait, there's more. Just for grins, I tried it on my Win98 SE box and it shows the same behavior! If I uncheck the Start at bootup box, restart Prime95, I get the exact same error message. So this is not a Win95-specific problem. I always had my Win98 box set with Start at bootup so I never noticed this before. And just for more fun, I found another problem with Prime95 in how it operates. I have my Prime95 setup with Tray Icon. If I start Prime95, double-click on the icon to open Prime95, do a File- Exit real quick, the program hangs and stops responding to all mouse clicks. At this point, you must kill it. This is reproducible on my Win98 and Win95 box. If you wait a few more seconds before doing the File- Exit, then there is no problem. Note to everyone else: None of these problems are critical to the operation of Prime95. They amount to minor annoyances, so please don't blow this out of proportion. The details I supplied are to help George figure this out sometime when he is looking for something to do :-) Thanks for the air time. Regards, -Allan Steve Harris wrote: Sounds like the opposite problem: Prime95 is trying to delete a registry entry that doesn't exist. I had one do that to me recently. Rather than uncheck the box, manually edit ( in prime.ini ) the line windows service=1 (or whatever line it has to that effect) to ...=0 and it will no longer see a need to try to delete the registry entry. And the box will now show as unchecked. Hope that helps, Steve Harris _ Unsubscribe list info -- http://www.ndatech.com/mersenne/signup.htm Mersenne Prime FAQ -- http://www.tasam.com/~lrwiman/FAQ-mers
