![]() |
|
|
#1 |
|
Nov 2003
22×5×373 Posts |
I have finished building a matrix for 2,1538M, but once again
I am running into memory limitations..... The matrix fits in 1.8G, but my biggest available machine only has 2G and the process is thrashing....... The matrix had 8.56M rows and 8.565M columns with a weight of 431.7M. I am going to do some more sieving over the weekend and hope I can bring it down just a bit smaller. This is a pain because I had already started sieving 3,517+. Now I have to switch back. I will also have to rebuild the matrix; another pain. |
|
|
|
|
|
#2 |
|
Sep 2004
22×5 Posts |
Hi Bob,
I'm a big fan of all the work that you and others that post on this forum do. How is 2,1538M going? I might be able to help out, but it largely depends on the task in hand along with my hardware suitablility. I have a Vista (32-bit) Core 2 Quad with 3GB DDR2 memory (dual channel mode), plenty of free HD space and broadband. May I ask a few questions: 1) What software do you use to solve the LA (do you have vista capable binaries)? 2) Is 3GB enough to accomplish the task? 3) Is this fire and forget, or does the software require parameter tweaking based on observed outputs (i.e. can a relative novice run it?) 4) Can the analysis be paused/continued, or is 1 complete run desirable? 5) And now the killer - approximately how long do you think this might/could take on a single core @2.4GHz - are we talking possibly 12-24 hours, or 2-3 weeks continuous (it's probably difficult to say - length of string and all, but I'm more interested if it's a night's processing rather than several days) Best wishes, Minty |
|
|
|
|
|
#3 | |
|
Nov 2003
746010 Posts |
Quote:
I generated another 4M relations over the weekend and am rebuilding the matrix now. I expect that it will fit easily within 2Gb and will take ~3 weeks to solve. (single threaded) |
|
|
|
|
|
|
#4 |
|
Sep 2004
22×5 Posts |
I was of course enquiring as to how the weekend work had gone to gauge whether you needed a PC with >2GB. Sounds like you've got everything under control - at 3 weeks this is certainly an undertaking if only in electricity costs.
I wish you all the best! |
|
|
|
|
|
#5 | |
|
Nov 2003
164448 Posts |
Quote:
but the weight increased to 446.8 million. I Also built a slightly larger matrix with similar weight. The code reads in the matrix without problem. It requires 1.78Gbytes of real memory. The code then attempts to malloc some auxiliary vectors and the malloc FAILS ---> the code will not run. |
|
|
|
|
|
|
#6 | |
|
Bamboozled!
"πΊππ·π·π"
May 2003
Down not across
1076310 Posts |
Quote:
Windows puts the system DLLs near the top of the address space. It's possible for malloc() to fail because of the fragmented virtual address space even though there's enough physical memory to hold everything. Paul |
|
|
|
|
|
|
#7 | |
|
Nov 2003
22×5×373 Posts |
Quote:
the OS to accomodate the /3GB switch. IT insists that it can't maintain systems unless everyone has the same configuration...... |
|
|
|
|
|
|
#8 | |
|
Bamboozled!
"πΊππ·π·π"
May 2003
Down not across
250138 Posts |
Quote:
My earlier offer still stands. Paul P.S. Having worked in IT support for many years, I have some sympathy with your IT people... |
|
|
|
|
|
|
#9 | |
|
Nov 2003
22×5×373 Posts |
Quote:
I will send you the final matrix and polynomial file via snail-mail... Remind me (via private email) of your snal mail address. Bob |
|
|
|
|
![]() |
| Thread Tools | |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Connectivity Matrix | Xyzzy | Lounge | 13 | 2017-02-21 18:29 |
| 12+256 matrix job | fivemack | Factoring | 11 | 2009-08-18 17:39 |
| GF(2) Matrix request | oslik | Factoring | 22 | 2008-11-02 12:53 |
| [Need help] about Matrix Polynomial | buan | Homework Help | 3 | 2007-07-17 15:07 |
| CWI Matrix Solver | R.D. Silverman | NFSNET Discussion | 6 | 2006-03-19 17:56 |