Archived sieve reservations and discussions - Page 3

Ken_g6 · 2009-09-06, 18:57

Reserving 40T-45T [Range 1] Ken_g6

cipher · 2009-09-06, 19:07

Quote:

Originally Posted by gribozavr

cipher:
In range 1 we have:
21282782 falctors total
16429092 factored numbers
39248305 numbers total in 480000-484999.txt

~77.2% of factors are not duplicate.
~41.9% of numbers in 480000-484999.txt have been factored.

Thanks Gribozavr, I send a PM to MooMOOO it looks like he is nowhere to be found also. Can any one else take over stats and Range Reserve/completion duties? Can you Gribozavr?

thanks
cipher

Joshua2 · 2009-09-06, 19:26

If it is possible I would like to be added a moderator to help out with this thread. All ranges complete, reserving 45-65T
http://www.sendspace.com/file/def1b1 for both previous ranges

cipher · 2009-09-07, 05:28

Sieve reservations for 1<k<10M, 480000<n<485000:

Code:

Range        Reserved By        Status
6T-10T        cipher           complete       
10T-12T       joshua2          complete
12T-15T       joshua2          complete       
15T-17T       cipher           complete       
17T-22T       ken_g6           complete       
22T-25T       amphoria         complete       
25T-35T       joshua2          complete       
35T-40T       cipher           in progress       
40T-45T       ken_g6           in progress       
45T-65T       joshua2          in progress       
65T+                           available

Sieve reservations for 1<k<10M, 485000<n<490000:

Code:

Range		Reserved By		Status
1-10G		various (see below)	complete
10G-11G		cipher			complete
11G-50G		cipher			complete
50G-55G		MooooMoo		complete
55G-100G	Mini-Geek		complete
100G-150G	Mini-Geek		complete
150G-200G	mdettweiler		complete
200G-400G	Mini-Geek		complete
400G-500G	mdettweiler		complete
500G-600G	MooooMoo		complete
600G-1000G	pschoefer		complete
1000G-1200G	amphoria		complete
1200G-1500G	amphoria		complete
1500G-2000G	amphoria		complete
2000G+                    		available


1500G+		available

Since moomooo is not updating Range/Reserved Here is the continuation and updated one.

I have not updated the stats, but if any one wants to update it and post it feel free. If there is any error in range reserved etc please let me know.
Thanks cipher

Ken_g6 · 2009-09-07, 16:57

40T - 45T complete (Range 1)

Joshua2 · 2009-09-08, 17:27

45-65T complete 1.45 factors / second. Reserving 65-90T
http://www.sendspace.com/file/s2cjj6

Joshua2 · 2009-09-09, 08:30

are all 3.6.4 versions the same speed? I had downoaded something slightly different I think. I'm going to try a X2 Neo and see how it does and it seemed the file was different when I downloaded it here. I run x64 on my C2Q and Neo. Any optimized versions? I have visual studio if that helps.
reserve 2t-10t range 2

Ken_g6 · 2009-09-09, 15:09

0.3.4/0.3.4a versions should be the same. I made a couple of minor tweaks that shouldn't affect the binaries at all a day or two ago. The _x64 versions are the fastest; -sse2 versions are 80% as fast, but work on 32-bit; and the plain 32-bit version is slow.

You might be seeing the difference that P size makes. Speed is proportional to log2(P/kmax), for P > 2*kmax.

Edit: Optimized versions for X2? No; but I benchmarked the 32-bit versions on an Athlon64. I expect it's about half as fast clock-for-clock as Core 2; but I would be curious whether the SSE2 version is faster than the 64-bit version (which I didn't try.)

Joshua2 · 2009-09-09, 19:14

I'm getting 6.7M p/sec ~2T on range 2 on the Neo X2 64 bit. It is a ultraportable chip competitor to the Atom running on like 10W-15W. 256kb cache was a tad faster even though it has 512kb. I doubled the block size and made the checkpointing less often as well, cause I figured that would make it a little faster.

Ken_g6 · 2009-09-09, 19:47

Did you try the 32-bit SSE2 version? It actually looks to be faster on my Athlon64!

Joshua2 · 2009-09-10, 04:14

That raised it to 10.15M p/sec!

2009-09-06, 19:26	#25
Joshua2 Sep 2004 13·41 Posts	If it is possible I would like to be added a moderator to help out with this thread. All ranges complete, reserving 45-65T http://www.sendspace.com/file/def1b1 for both previous ranges Last fiddled with by Joshua2 on 2009-09-06 at 19:28

2009-09-08, 17:27	#28
Joshua2 Sep 2004 13·41 Posts	45-65T complete 1.45 factors / second. Reserving 65-90T http://www.sendspace.com/file/s2cjj6 Last fiddled with by Joshua2 on 2009-09-08 at 17:31

2009-09-09, 08:30	#29
Joshua2 Sep 2004 13·41 Posts	are all 3.6.4 versions the same speed? I had downoaded something slightly different I think. I'm going to try a X2 Neo and see how it does and it seemed the file was different when I downloaded it here. I run x64 on my C2Q and Neo. Any optimized versions? I have visual studio if that helps. reserve 2t-10t range 2 Last fiddled with by Joshua2 on 2009-09-09 at 08:32

2009-09-09, 15:09	#30
Ken_g6 Jan 2005 Caught in a sieve 395₁₀ Posts	0.3.4/0.3.4a versions should be the same. I made a couple of minor tweaks that shouldn't affect the binaries at all a day or two ago. The _x64 versions are the fastest; -sse2 versions are 80% as fast, but work on 32-bit; and the plain 32-bit version is slow. You might be seeing the difference that P size makes. Speed is proportional to log2(P/kmax), for P > 2kmax. Edit: Optimized versions for X2? No; but I benchmarked the 32-bit versions on an Athlon64. I expect it's about half as fast clock-for-clock as Core 2; but I would be curious whether the SSE2 version is faster than the 64-bit version (which I didn't try.) Last fiddled with by Ken_g6 on 2009-09-09 at 15:27*

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Soapbox Discussions	only_human	Soap Box	41	2019-11-16 15:46
n=390000 discussions and old reservations	TimSorbet	Twin Prime Search	31	2010-05-22 23:13
Sieve reservation discussions	jmblazek	Twin Prime Search	27	2007-01-31 21:41
Primegrid discussions	pacionet	Twin Prime Search	17	2007-01-20 11:22
Automated PRP discussions	ltd	Sierpinski/Riesel Base 5	20	2006-09-02 22:19

2009-09-06, 18:57	#23
Ken_g6 Jan 2005 Caught in a sieve 5·79 Posts	Reserving 40T-45T [Range 1] Ken_g6

2009-09-07, 16:57	#27
Ken_g6 Jan 2005 Caught in a sieve 395₁₀ Posts	40T - 45T complete (Range 1)

2009-09-09, 19:14	#31
Joshua2 Sep 2004 13×41 Posts	I'm getting 6.7M p/sec ~2T on range 2 on the Neo X2 64 bit. It is a ultraportable chip competitor to the Atom running on like 10W-15W. 256kb cache was a tad faster even though it has 512kb. I doubled the block size and made the checkpointing less often as well, cause I figured that would make it a little faster.

2009-09-09, 19:47	#32
Ken_g6 Jan 2005 Caught in a sieve 5·79 Posts	Did you try the 32-bit SSE2 version? It actually looks to be faster on my Athlon64!

2009-09-10, 04:14	#33
Joshua2 Sep 2004 1000010101₂ Posts	That raised it to 10.15M p/sec!