Reserved for MF - Sequence 4788

[Dubslow]

So for the current C109 at i2958, how many curves at what bounds would get us up to t36 (or so)?

[Batalov]

It should be shot straight down. Chased with a beer.

Well, yes, say, t36 is simply 890 1e6 curves (ie a "t35") and optionally another 100 of 3e6 to make it to t36 and then 6 CPU-hrs on GNFS and on to the next iteration.

Next, for the c125, one would want the 2400 3e6s and some 11e6 curves

[Dubslow]

Well, I originally had no idea where you got the numbers from, but as I was reading YAFU's docfile, it said to see the GMP-ECM README for more details. After some fuss (gedit couldn't guess the character encoding for README, but it could for README.lib, stupid thing) I found a table. I'm guessing that's the "goto" reference for different t's and what curves/bounds to use?

[Batalov]

You can also do

echo 43144365336006471456390066354433525606788804425777803602784525134520707531477166398288695986225151664288960876332879535559137048007 | ecm -v 3e6

and it will report the needed curves for every level

The c131 is ready for gnfs?

[Dubslow]

I made a fuzzy

#!/bin/bash
# A script to print the bounds/curves of ECM necessary for a given size number.
# No argument prints the whole table.
# If arg > 65, then it assumes you entered the size of number, and prints the
# 	relevant lines of the table.
# Else if arg >= 20, then it assumes you entered the desired t value directly,
# 	and prints the relevant information.
# Else the arg is an error, print usage message. Arg == "-h" also produces
# 	usage message.

file=$HOME/yafu/ecm.conf
# Location of data table
# Line 2 has t20 data, line 3 has t25 data, etc.

function print_t_data {
	
	let n=$(( $((t-10)) / 5 ))
	# Convert from t-value to which line of ecm.conf needs to be printed
        # Line 2 has t20, line 3 has t25, etc., so t=5n+10
	# The /5 is integer division, so n is an int

	echo
	more +1 $file | head -n 1
	more +$n $file | head -n 2
	echo
	# Something of a hack. 'more +N file' starts printing at the Nth line,
	# and 'head -n N' prints N lines of the file, so the above prints
	# the first line of the file, followed by 2 lines of $file starting
	# at the $n'th line.
}

if [ -z "$1" ]; then
	echo; cat $file; echo;
	exit 0
elif [ "$1" = "-h" ]; then
	echo "Usage: $0 [size of number|desired t-level]"
	echo "If arg >65, assume size of number, else if arg >=20, assume"
	echo "desired t-level."
	exit 0
elif [ $1 -gt 65 ]; then

	sz=$1
	# Argument is digits in number to be factored

	let t=$((sz/3))
	# Apply the 1/3 rule

	echo; echo "You need to ECM up to approximately t$t."

	print_t_data
	
	exit 0

elif [ $1 -ge 20 ]; then

	t=$1
	# Assume argument is desired t-level

	print_t_data

	exit 0

else 	
# Arg is too small or NaN
	echo "Bad argument!"
	$0 -h
	exit 1
	# Calls itself with the -h switch to print usage
fi

bill@Gravemind:~/yafu∰∂ cat ecm.conf
t          B1   standard curves   Brent-Suyama curves
20       11e3           74                74
25        5e4          221               214
30       25e4          453               430
35        1e6          984               904
40        3e6         2541              2350
45       11e6         4949              4480
50       43e6         8266              7553
55       11e7        20158             17769
60       26e7        47173             42017
65       85e7        77666             69408
bill@Gravemind:~/yafu∰∂

In use:

bill@Gravemind:~∰∂ ecm 149

You need to ECM up to approximately t49.

t          B1   standard curves   Brent-Suyama curves
45       11e6         4949              4480
50       43e6         8266              7553

bill@Gravemind:~∰∂ ecm 131

You need to ECM up to approximately t43.

t          B1   standard curves   Brent-Suyama curves
40        3e6         2541              2350
45       11e6         4949              4480

bill@Gravemind:~∰∂ ecm 32

t          B1   standard curves   Brent-Suyama curves
30       25e4          453               430
35        1e6          984               904

bill@Gravemind:~∰∂ ecm 5
Bad argument!
Usage: /home/bill/bin/ecm [size of number|desired t-level]
If arg >65, assume size of number, else if arg >=20, assume
desired t-level.
bill@Gravemind:~∰∂ ecm -h
Usage: /home/bill/bin/ecm [size of number|desired t-level]
If arg >65, assume size of number, else if arg >=20, assume
desired t-level.
bill@Gravemind:~∰∂

(I didn't want to have to go through the bloat of installing gmp-ecm, since yafu runs the ecm just fine -- and besides, my scripting was out of practice )

Originally Posted by Batalov

The c131 is ready for gnfs?

Well nobody has reported ECM on it, I don't think.

[Batalov]

I'd run 1/4 of t45. If two more people do the same, it will be ready.

...Running gnfs now. Give it 10 hours.

[Dubslow]

That means you started the GNFS, but still want someone to run some ECM?

I'll start on 1200 curves at 11e6 then.

[schickel]

Originally Posted by Dubslow

That means you started the GNFS, but still want someone to run some ECM?

I'll start on 1200 curves at 11e6 then.

Actually, check the original post time and the "last edit" times and you'll see that he posted about the ECM, then came back an hour later and mentioned the GGNFS job.

You can still run the ECM, but it would be pretty much pointless.....

[akruppa]

Did 2k@11M.

[Batalov]

I went to the gym and didn't know if I'd come back to the comp before tomorrow, so it was easier to fire and forget (and the poly was decent - 8.4e-11). There is a chance of course that someone cracks it - I'll be up for another hour maybe; I'll kill the gnfs if so.

[Dubslow]

Originally Posted by schickel

Actually, check the original post time and the "last edit" times and you'll see that he posted about the ECM, then came back an hour later and mentioned the GGNFS job.

You can still run the ECM, but it would be pretty much pointless.....

I did check, and they were only 15 minutes apart. (22:41 OP time to 22:56 edit, local.) On the other hand, I'm only getting ~100 curves/hour, and I'll certainly be asleep before they finish (even if I put all cores to ECM) so I think I'll stop them. 108@11e6 total.