Safe Mode

JuanTutors · 2004-06-19, 00:25

Has anyone tried running Prime95 on their Windows PC on Safe Mode? Is this possible? Is this bad?

Clearly, if it were possible and weren't bad, it would only be a good idea to do when you are leaving your computer for an extended period of time like a few days. (You don't want to restart your computer that often!!)

I could think of a few logical reasons why this would or wouldn't matter, but I'm not looking for those. I'm actually asking if it's possible, and if so then if anyone has tried it.

Xyzzy · 2004-06-19, 02:30

.

Quote:

Date: Sun, 27 Apr 97 18:30:03 UT
Subject: Mersenne: Safe Mode?

I noticed an interesting thing yesterday when I was playing with Prime95. I
was trying to speed up the iteration time by terminating threads with Wintop.
After I had eliminated all non-essential threads, my iteration time dropped
from .203 to .202. At that time I had about 13 threads running. Just for fun
I rebooted into safe mode and the time dropped to .192. Obviously, I can't
get much work done in safe mode, but if I know that I will not be using the
computer for 12 or more hours, I guess it would be worth the time to reboot.

Any thoughts?

My current range is 1953800 or so... the reason my time is slow is because I
have no L2 cache at all! I was not aware of this when I bought my computer,
but I plan to add some on payday.

Anyways, it looks like safe mode is worth a 5% improvement. (At least on my
system...)

- -Mike

Uncwilly · 2004-06-19, 19:02

I found this out mysolf. I get a reasonable boost. Worth leaving it in safe mode if you are away for a few hours (like all day or more).

JuanTutors · 2004-06-21, 21:42

I tried this last night for the first time.

Working on M33796981
safe mode: avg iteration time 76.6 ms while idle
normal mode: avg iteration time 76.9 to 83.0 ms while idle, most often near 80 ms.

I only did this for the first time last night, so I don't have a plus or minus.

NOTE: The fluctuations in the iteration time in normal mode are from night to night, NOT from screen output to screen output. This is true even under the same conditions from night to night.

I mention this because it may well be that the safe mode iteration time may also fluctuate, and 76.6 ms may be one of the faster or slower times.

Does anyone else have any results?

Mike · 2004-06-22, 00:08

I tried it while factoring a 25,800,000 exponent:

Normal: 142-143 ms
Safe Mode: 148 ms

delta_t · 2004-06-22, 02:25

If you have a P4 with HT enabledd, running in Safe Mode in XP is actually slower than running normally. See here: http://www.mersenneforum.org/showthread.php?t=1643

lycorn · 2004-06-24, 10:15

Athlon 1200 (TBird core), Windows Millenium.
Exponent in the 11M range.
Safe Mode: ~93 msec/iter
Normal Mode: ~ 99 msec/iter
These are consistent results, measured over several months.
It pays to run in safe mode, if you are to leave the computer alone for a reasonable period of time (enough to make up for the time you loose shutting down and rebooting).
Assuming that shutting down and restarting (twice) makes you loose 4 minutes, @ 99msec/iter that amounts to ~2400 iterations. So in theory, any interval over 70 to 80 minutes would be enough to gain anything from running in SM, for the configuration and exponent described in this post. As it is a bit of a pain to shutdown and restart all the time, I only use SM when leaving the computer alone overnight or longer.

2004-06-19, 00:25	#1
JuanTutors "Juan Tutors" Mar 2004 23B₁₆ Posts	Safe Mode Has anyone tried running Prime95 on their Windows PC on Safe Mode? Is this possible? Is this bad? Clearly, if it were possible and weren't bad, it would only be a good idea to do when you are leaving your computer for an extended period of time like a few days. (You don't want to restart your computer that often!!) I could think of a few logical reasons why this would or wouldn't matter, but I'm not looking for those. I'm actually asking if it's possible, and if so then if anyone has tried it.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
SAFE MODE....	petrw1	Hardware	14	2009-03-25 13:17
Is libecm thread-safe?	frmky	GMP-ECM	5	2009-02-28 16:08
Prime95 in Safe Mode on Win98SE	Unregistered	Software	3	2006-06-07 04:51
Safe Mode/High Priority	A_Pryme	Software	12	2005-02-13 19:22
Safe Mode	Unregistered	Software	12	2004-07-16 01:06

2004-06-19, 19:02	#3
Uncwilly 6809 > 6502 """"""""""""""""""" Aug 2003 101×103 Posts 2A8F₁₆ Posts	I found this out mysolf. I get a reasonable boost. Worth leaving it in safe mode if you are away for a few hours (like all day or more).

2004-06-21, 21:42	#4
JuanTutors "Juan Tutors" Mar 2004 23B₁₆ Posts	I tried this last night for the first time. Working on M33796981 safe mode: avg iteration time 76.6 ms while idle normal mode: avg iteration time 76.9 to 83.0 ms while idle, most often near 80 ms. I only did this for the first time last night, so I don't have a plus or minus. NOTE: The fluctuations in the iteration time in normal mode are from night to night, NOT from screen output to screen output. This is true even under the same conditions from night to night. I mention this because it may well be that the safe mode iteration time may also fluctuate, and 76.6 ms may be one of the faster or slower times. Does anyone else have any results?

2004-06-22, 00:08	#5
Mike May 2004 Vancouver, Canada 100₁₀ Posts	I tried it while factoring a 25,800,000 exponent: Normal: 142-143 ms Safe Mode: 148 ms

2004-06-22, 02:25	#6
delta_t Nov 2002 Anchorage, AK 101100101₂ Posts	If you have a P4 with HT enabledd, running in Safe Mode in XP is actually slower than running normally. See here: http://www.mersenneforum.org/showthread.php?t=1643

2004-06-24, 10:15	#7
lycorn "GIMFS" Sep 2002 Oeiras, Portugal 3043₈ Posts	Athlon 1200 (TBird core), Windows Millenium. Exponent in the 11M range. Safe Mode: ~93 msec/iter Normal Mode: ~ 99 msec/iter These are consistent results, measured over several months. It pays to run in safe mode, if you are to leave the computer alone for a reasonable period of time (enough to make up for the time you loose shutting down and rebooting). Assuming that shutting down and restarting (twice) makes you loose 4 minutes, @ 99msec/iter that amounts to ~2400 iterations. So in theory, any interval over 70 to 80 minutes would be enough to gain anything from running in SM, for the configuration and exponent described in this post. As it is a bit of a pain to shutdown and restart all the time, I only use SM when leaving the computer alone overnight or longer.