Heh, this silly little rule does make quite a bit of sense. Is this why Andres Aitsen took a crap load of candidates up to 72 bits?

Since OBD just leveled up though, wouldn't the table look more like this?[code]Bits Target Actual
>= >=
77 11 [color=red]4[/color]
76 22 [color=red]10[/color]
75 44 [color=red]23 | 24[/color]
74 88 [color=red]25 | 26[/color]
73 all [color=red]53[/color]
72 uhoh 109[/code]The uncertainty being around where exactly 3321928417 has been TF'd to. :ermm:

[color=LemonChiffon]I really wish you'd fix the differences 'cause I just can't stop gazing at this table and willing it to shrink. There are only 35 rows and 26 are just due to the bit depth.[/color]

[QUOTE=lavalamp;173848]I really wish you'd fix the differences 'cause I just can't stop gazing at this table and willing it to shrink. There are only 35 rows and 26 are just due to the bit depth.[/QUOTE]

I've been meaning to ask Luigi if he is certain about the bit levels he shows.

I've been interpretting the silly little rule to apply to the presently achieved level, not the target level. When we get six at 77 bits or higher, that's when I'd start revising the levels. If you're going to target meeting the rule, my method has the practical frustration that one more at top level boosts the requirements all through the list - your list is a better planning guide because those levels will be good enough through level 11.05.

William

P.S. I just searched mersenneforum for 3321928417. It wasn't reported here, so either Luigi has a private update or an error.

[QUOTE=wblipp;173863]I've been meaning to ask Luigi if he is certain about the bit levels he shows.

I've been interpretting the silly little rule to apply to the presently achieved level, not the target level. When we get six at 77 bits or higher, that's when I'd start revising the levels. If you're going to target meeting the rule, my method has the practical frustration that one more at top level boosts the requirements all through the list - your list is a better planning guide because those levels will be good enough through level 11.05.

William

P.S. I just searched mersenneforum for 3321928417. It wasn't reported here, so either Luigi has a private update or an error.[/QUOTE]

I thought I had run it up to 75 bits, maybe I forgot to update it on the forum.
No problem, I will rerun it for a check with no CPU effort added, so

Taking 3321928417 from 73 to 75 bits,

Luigi :smile:

[QUOTE=wblipp;173863]When we get six at 77 bits or higher, that's when I'd start revising the levels.[/QUOTE]Can't be long now then, I've got 3321933541 61.3% of the way from 76 to 77 bits, and both em99010pepe and lfm have candidates reserved to 78 bits.

If you guys are reading this, how are you getting along with that?

[QUOTE=wblipp;173863]P.S. I just searched mersenneforum for 3321928417. It wasn't reported here, so either Luigi has a private update or an error.[/QUOTE]The last mention of 3321928417 I found was in this thread when Phinne said he'd take it to 72 bits, and he never posted back to say whether the range was completed.

Luigi,

What about all those factored exponents that you show at higher levels than I show - are you sure those are all correct?

William

[QUOTE=wblipp;173930]Luigi,

What about all those factored exponents that you show at higher levels than I show - are you sure those are all correct?

William[/QUOTE]

Yes, I am. I remember it because I had to modify my PHP script to accept ranges of already factored exponents. I don't remember well who submitted them all, they came in packs of 6 or more. :smile:

I guess it was PatrickSchmeer, look at post #[URL="http://www.mersenneforum.org/showpost.php?p=102728&postcount=172"]172[/URL] and following ones on Checkout thread, and post [URL="http://www.mersenneforum.org/showpost.php?p=102373&postcount=241"]241[/URL] on Report results.

In particular I found this one:
[url]http://www.mersenneforum.org/showpost.php?p=106553&postcount=274[/url]
But no detailed reporting about it was done. Instead, PatrickSchmeer vanished. :smile:
Those exponents were all already factored, and the previous ranges were all done, so I trusted his word.

Luigi

If you like, when I've done this exponent (in three hours), I can re-do all 159 factored candidates to 70 bits.

It'll only take slightly longer than running another candidate to 77 bits (I estimate 3 days 17 hours), so it's not really a big deal, and it would provide certainty of bitdepth.

[QUOTE=lavalamp;173950]If you like, when I've done this exponent (in three hours), I can re-do all 159 factored candidates to 70 bits.

It'll only take slightly longer than running another candidate to 77 bits (I estimate 3 days 17 hours), so it's not really a big deal, and it would provide certainty of bitdepth.[/QUOTE]

I don't think that's necessary. The post Luigi found explicitly mentions 26 already factored exponents, which is exactly the difference, and Patick said he finished them. I was apparently awaiting the promised report "soon." I'll just update it now.

William

Heh, regular expressions can be fragile little things at times, your last update to the code differed slightly from normal in two ways, neither of which would have mattered individually, but together they conspired to put two <BR>s in the middle of my application/xhtml+xml and wouldn't you know it, yellow screen of death in the 'fox. Fixed now though.

I see that 3321933397 has been taken to 97 bits now. :wink: I think I'll only take 3321933451 as far as 77.

Edit: Oh, and no factors for 3321933541 to 77 bits, unreserving.

Taking 3321930173 to 73 (4.5 freaking days on a 1.7GHz Willamette!).

Taking 3321933313 to 74.

3321933451 has passed 76, and is 7% of the way to 77.

Taking 3321933389 from 72 to 74 bits.

Luigi