Cannot Read the Star Data in Stellarium 32bit
Carte ▾ ▴
Loading large star data files
2008-05-16
2012-ten-09
-
-
Hi Garry
This is the same trouble as with the scripts namel SP3. I remember the same problem may occur with Vista. I am near to try it only information technology takes a while to re-create the catalogues to my Vista installation.
No Vista with 512 MB of memory volition run all the star catalogues. It takes quite some time to load stellarium however.
Barry
If you would like to refer to this annotate somewhere else in this projection, copy and paste the following link:
-
Hi Garry,
Wow over 1 Gb of bandwith...
Downloaded all catalogs on Vista Bones SP1 with two Gb of retention,
All run correctly with no errors.
Every bit Barry says, it could exist a sp3 effect,If you would similar to refer to this comment somewhere else in this project, copy and paste the post-obit link:
-
Barry, you seem to be online right at present and then I'd thought u would attempt to catch you here.
Did yous always go the SVN version 3151 to work? I'one thousand withal stuck with the enable-automobile-import error.If you would like to refer to this annotate somewhere else in this project, copy and paste the post-obit link:
-
Hi Varkhan1
No. I got 3144 to piece of work with Qt-4.four.0 but 3151 hangs in the linking phase with the error I put in the mail service. I have tried to compile with Vista, XP and Linux, all with the same problem. Information technology says to read about enable-auto-import but I tin can't detect out where. I volition have another look tomorrow.
Barry
If you would like to refer to this comment somewhere else in this projection, copy and paste the post-obit link:
-
Further
No I did not need to download them once again. I have them on a data DVD but it takes 4 minutes to copy from the DVD to the HDD.
Barry
If you would like to refer to this comment somewhere else in this project, copy and paste the post-obit link:
-
Many thanks for all your comments.
In fact I have the aforementioned problem loading Star Itemize viii on two different machines, one with 512MB RAM and the other with 2GB of RAM.
The same problem occurred with Windows XP and SP2 (earlier I installed SP3), and also when using Stellarium 0.9.0. So I don't retrieve it's an SP3 problem.
On the 512MB car, star catalog 8 won't load no matter what I practise. On the 2GB auto, if I remove itemize 7 catalog 8 does load and works ok.
It seems unlikely that true cat viii is corrupted, but I suppose it's a possibility. (I don't want to download it once more -- 600 MB is a bit big!!!)
Thank you,
Garry
If yous would like to refer to this comment somewhere else in this project, copy and paste the following link:
-
Your detailed error study makes it easy to diagnose the trouble:
"MapViewOfFile failed: eight" ways co-ordinate to
http://msdn.microsoft.com/en-us/library/ms681382(VS.85).aspxERROR_NOT_ENOUGH_MEMORY, Not plenty storage is bachelor to procedure this control.
I should say you lack virtual memory. That is, your operating system propably uses a 2G/2G split of the 32 bit accost space. And 2G virtual accost space in the userspace is not enough.
Solution: use a different operating system (or Servicepack or any). Or build a 64bit windows binary. If you can do this, please tell us how you have done information technology, because you will be the first one how tries building for 64bit Windows.
Yours,
JohannesIf you would similar to refer to this comment somewhere else in this project, copy and paste the following link:
-
Thanks very much for all the suggestions.
Johannes, I have tried increasing the virtual memory, but on two unlike machines under XP SP3 it will not permit me increase it beyond 4GB, even though at that place is plenty of hard disk space. Am I right in saying that if ane could increase virtual retentivity plenty then the large file is likely to load, or is it more than complicated? Can ane modify the split of the accost space or do y'all know how to override the apparent system limitation of 4GB.
Thanks,
Garry
If y'all would similar to refer to this comment somewhere else in this project, copy and paste the following link:
-
Hello Garry,
on a 32bit auto an address is merely 32 bits long. This means, that there can exist only pow(2,32) = 4294967296 different addresses. And since one accost points to one byte, the maximum accost infinite of a process is 4294967296 bytes = 4GB. Even if you have 64bit windows running on a 64 fleck machine, stellarium.exe is a 32 bit binary and therefore can accost just 4GB.
Now not all of the 4GB vest to stellarium.exe. Simply the operating organisation - which is windows in your instance - reserves some of this address infinite to itself. 32bit Linux usually makes a 3GB/1GB split: 1GB to the operating arrangement, 3GB to the application (userspace). Windows on the other hand seems to be fatter, and often makes a 2GB/2GB split up.
The stellarium star catalogues are eigther loaded into memory, or mmaped into the address infinite of stellarium.exe. So in either case true cat 0-eight consume at to the lowest degree 1287783926 bytes of address space. Plus the administrative data - some precomputed values, whose initialization accounts for the loading time of the large star files in mmap mode - 104857600 bytes, totally 1392641526 bytes. This leaves 754842122 bytes only for the rest of stellarium.exe, and all the dlls that are also loaded into userspace. And in your instance this is not enough.
Then what is different between your case and the other windows users, whose operating systems can accommodate the residuum of stellarium.exe plus windows dlls in 750MB? I exercise not know. Probably the windows dlls. Merely I take quite successfully avoided windows for nearly 10 years, and lack windows operating arrangement knowhow. Therefore I cannot help yous.
If you lot really want to load stars8.true cat, I see only these ii practical options for you lot:
ane) make your windows installation similar to the installation of Barry or another user, for whom the loading works.
2) abandon windows and apply Linux.I cannot recommend the third solution to yous, which would exist building a 64bit windows stellarium past yourself.
Sad if this did not really help,
Yours
JohannesIf you would like to refer to this comment somewhere else in this project, re-create and paste the following link:
-
Hullo Garry
My son gave me a hint to improving speeds on your computer. Make a special partition of four GB on the hard deejay and allocate information technology exclusively for the virtual memory. This tin avoid the fragmentation that happens afterwards time on when it is in the normal operating position. (Simply like Linux does)
Barry
If you would like to refer to this comment somewhere else in this project, copy and paste the post-obit link:
-
Hi Johannes,
Many thank you for the beautiful explanation of what seems to be happening -- it makes a lot of sense now.
I will encounter if any of my contacts knows how to change the 2GB/2GB split or if information technology is possible and / or risks messing things up completely. Maybe another Stellarium user may take some ideas on this?
At whatever charge per unit, I'll report back if I make any progress,
Many thank you,
Cheers,
Garry
If you lot would like to refer to this comment somewhere else in this projection, re-create and paste the following link:
-
The culling Barry is talking most requires the cosmos of a new sectionalization. There are ii known problems with this method.
1. Information technology requires you to create another partition, which in nearly cases tin can't exist undone without a reinstallation of windows.
2. The second sectionalisation of a hard bulldoze is considerably slower than the start, causing paging performance issues.Another, perhaps better culling, is described in the post-obit steps.
1. First you volition need a defragger that tin do this. I recomend the free JKDefrag. http://www.kessels.com/Jkdefrag/
two. Delete your page file (Yep, I mean it. Completely delete information technology.)
3. A reboot volition be required. When yous reboot, boot into safety fashion by repeatedly pressing F8 earlier windows boots.
4. Run your defragmenter. In JKDefrag you might want to run it from the control prompt with the option -a 7
5. Wait for the defragger to cease. This will probably take a few hours depending on the amount of information you take on your disk.
half dozen. Reboot into windows, and increase the page file to 4 gb.The difference here is that windows now has continuous gratis space to put the page file in, rather than splitting it in chunks, which equally Barry says is very bad. And when your page file is not continuously being adjusted by the system (make sure you turn this off) information technology stays defragmented because none of pagefile.sys is beingness moved.
If yous would like to refer to this comment somewhere else in this project, copy and paste the following link:
-
Farther note:
To control pagefile options in XP, Rightclick computer, select backdrop, become to advanced, select settings under performance, goto the advanced tab, and click modify.
Yous tin trust JKDefrag, if you make up one's mind to use information technology. I use it as my only defragger, except when I need to defrag the kicking files using windows.
The post-obit is an entry in the JKDefrag log:
Time Fragments Bytes Clusters Location
09:31:19 i 2097152000 512000 C:\pagefile.sysThis shows that my pagefile has no fragment(i fragment means it's all in 1 piece). My entire pagefile is on the same partition as my windows installation, and I don't endure the performance problems of adding another segmentation.
Varkhan1
If y'all would similar to refer to this comment somewhere else in this project, copy and paste the post-obit link:
-
Many thanks for the suggestions Varkhan1.
Maybe I'k missing something but my pagefile is already 4 GB and the system won't let information technology go across this for the reason given by Johannes. Every bit outlined by Johannes it seems to me that it is a memory/pagefile size trouble equally a result of unfavorable splitting (2GB/2GB) or whatsoever rather than a speed problem caused by defragmentation. Speed-wise there is no problem. Won't what you are suggesting just increase the speed and not the memory/pagefile size? It would be overnice to make the split say 3GB/1GB or two.5GB/1.5GB and run into what happens.
Perchance I've misunderstood?
Thanks,
Garry
If you would like to refer to this annotate somewhere else in this project, copy and paste the following link:
-
Garry, you are correct.
The procedure I was discussing will only increment the performance of your page file, non make information technology larger.
The style to make the split 3GB/1GB is discussed here:
http://support.microsoft.com/kb/833721
This involves modifying your boot file; so if y'all cull to do information technology, delight be very careful!
What the /3GB option does is limit the system to 1 GB of virtual space, and gives the other three to programs.
I don't know if this will help you at all, simply this is the way to change the split, if you desire to attempt it.Ane affair I've noticed...
I ran stellarium with all star files. The program used upward 216 MB of memory and 279 MB of pagefile. This didn't come close to making me run out of retentiveness/pagefile. You say y'all have a 2gb organisation like myself. I don't encounter why you are running out of room.From your original postal service:
Loading mmap:stars_8_2v0_0.cat: 8_2v0_0; Fault: SpecialZoneArray(8)::SpecialZone
Array: MapViewOfFile failed: 8, page_size: 65536
initialization failed
Why does it say page_size 65536?
There is something special about this number, though I'm non certain what it is. Information technology doesn't seem similar it could be luck that that'south the last row in older versions of MS Excel (such as Excel 03)!
Read this too: http://www.dq.winsila.com/tech-news/excel-2007-the-65535-65536-trouble.html
Oh, and this number also happens to be 2^sixteen.
The x86 platform has a word size of sixteen bits, or 65536.
I don't know what this might have to do with your trouble, though. Merely information technology seems odd that your page_size would be 65536.If you would like to refer to this comment somewhere else in this projection, copy and paste the following link:
-
Thank you for the link, which seems to be really helpful!
> I ran stellarium with all star files. The plan used up 216 MB of memory and 279 MB of pagefile. This didn't come close to making me run out of memory/pagefile. You say you have a 2gb organization similar myself. I don't see why you lot are running out of room.
I think Garry does non run out of memory, simply out of address space. I also wonder why this is the case.
Perchance the "Performance Monitor" equally described in http://support.microsoft.com/kb/823440/ can give some insight in this problem.Johannes
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
-
Never mind. I constitute out what it means by page_size. :)
If you would like to refer to this comment somewhere else in this projection, re-create and paste the following link:
-
There has been some other develpment which I don't understand at all.
On the automobile with 2GB of existent memory, it previously prevented me from increasing the virtual retention beyond 4GB (4095MB to be precise). As far as I know I oasis't made any changes but it now allows me to increase the virtual retentiveness far beyond 4GB. I increased it to 16GB, and the machine reported that in that location was in fact 16GB of virtual memory after rebooting (I suppose it wasn't telling lies!), but even so the same error from stellarium concerning the #8 star file.
As y'all guys seem to be saying, it'due south almost as if the page file size is actually but 65536 (=2^16) or that it tin only address that much or something.
I oasis't been game enough yet to try the 3GB/1GB split, but the links you've pointed out seem to be very helpful in that regard.
Thanks once again for your continued help.
Garry
If you lot would similar to refer to this comment somewhere else in this project, copy and paste the following link:
-
In the DOS surface area, before the introduction of the paged memory management in the 80368, all processes - and fifty-fifty the operating system - shared the same address infinite. Everybody is happy that this disaster id over now. Different processes practice not share the same address space any more. And then it makes good sense to have 16GB swap sectionalization, or pagefile, however you telephone call it. With a 16GB swap file you will be able to run many memory consuming processes in parallel on a 32 bit machine. But still each of the processes is limited to 4GB address space. However massive paging in this situation might boring down your machine downwards to 0.
Johannes
If you would similar to refer to this comment somewhere else in this project, copy and paste the following link:
-
Now I'm dislocated. I didn't know windows limited you to 4gb. Mine will become at least to sixteen Gb. I think you lot are confusing it with memory, which in 32 fleck is capped at 4GB. Notwithstanding, equally Johannes mentioned higher up, your trouble is likely not memory or page file, just address space.
If you would like to refer to this comment somewhere else in this project, copy and paste the post-obit link:
-
PS: forgotten to explain: virtual memory comes in pages, the size of each folio is the page_size. On windows I found information technology to be more often than not 64kB (65536 byte), while on Linux it is 4096 byte. Yous need not worry about this, but the programmer must take this into account for some reason (whose caption is beyond the scope of this thread).
Johannes
If y'all would like to refer to this comment somewhere else in this projection, copy and paste the following link:
-
I think your final message Johannes, regarding the 65536 byte folio size, is the central to it and I'll pursue this, and let you know if I make any progress.
For some reason or other on two separate machines, each with 2GB of real retentivity and plenty of virtual memory (4GB), both using v 0.9.1 of stellarium and running XP SP3, it can't cope with this 65536 limit on the page file.
Incidently this also happened with SP2 and version 0.ix.1 on both machines and v 0.9.0 under XP SP2. [The versions of stellarium I utilise are executables (I don't accept the knowledge to build them myself), downloaded from the Sydney mirror site a think.]
Besides, although its probably non relevant whatsoever more than, 1 of the machines really volition allow me to increase the virtual retentivity to at least 16GB while the other stops at 4095MB, and both have enough of costless disk space (similar 50GB).
Thanks again.
Garry
If you would similar to refer to this comment somewhere else in this project, copy and paste the following link:
-
Hi all,
the aforementioned error occured to me as well and all suggestions didn't change the behaviour of the program.
Neither increasing pagefile.sys nor editing boot.ini with the /3G parameter.But i found out, the problem what occurrs is related to the page_size:
(Loading mmap:stars_8_2v0_0.cat: 8_2v0_0; ERROR: SpecialZoneArray(eight)::SpecialZone
Array: MapViewOfFile failed: eight, page_size: 65536)##
Later on googling for a while i found the following link:
http://technet2.microsoft.com/windowsserver/en/library/94d21089-411b-4bce-a823-49a77a46e7661033.mspx?mfr=true
Hither is of import the department:
LargeBufferSize
Value Type: REG_DWORD
Default: PAGE_SIZE (4096 bytes on i386, 8192 bytes on Alpha)
Description: The size, in bytes, of large buffers used by AFD. Smaller values use less memory and larger values can meliorate performance.I accept no idea why Stellarium needs this parameter, merely when you create that entry using "regedit" (start --> execute) Stellarium is loading (after a reboot of the estimator, otherwise the modify of the registry won't take any result) the stars_8_2v0_0.cat.
Go to
HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Afd\Parametersand create a new (correct click in the main expanse) DWOR-value.
Name it "LargeBufferSize" (without quotes),
the value (checkbox dezimal) is 65536
Go out regedit & reboot.
Thats it.
Please study if it works on your car.But actually it should, cause it solved the trouble on my machine.
Cheers,
Rico
If you would like to refer to this comment somewhere else in this projection, copy and paste the following link:
-
Howdy all,
unfortunately i have to say my answer before doesn't piece of work.
I don't know why, just in one case Stellarium could read the stars_8_2v0_0.true cat and and then i thought its cause of the things i did inside the registry.
Just after another start of the application it doesn't work again.
So i was bit to quick to mail service already a supposed solution.Any way... if somebody will observe a solution someday, i would be happy too :)
Thanks,
Rico
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
pollardprionsprould1946.blogspot.com
Source: https://sourceforge.net/p/stellarium/discussion/278769/thread/0f598794/
I am using Windows XP SP3, with 2GB of memory. I am having problem getting the largest star file (stars_8_2v0_0.cat) to load in Stellarium version 0.nine.one.
The only fashion I can get information technology to work is to remove some of the smaller star files (say 5, half dozen and 7). Information technology doesn't seem to be an mmap problem as eight will load if 5, vi and 7 are removed even if the mmap option is Not used. The error message I get is as below.
Loading star data...
Loading stars_0_0v0_1.cat: 0_0v0_1; stars: 5013
Loading stars_1_0v0_1.cat: 1_0v0_1; stars: 21999
Loading stars_2_0v0_1.cat: 2_0v0_1; stars: 151416
Loading stars_3_1v0_0.true cat: 3_1v0_0; stars: 434064
Loading stars_4_1v0_0.true cat: 4_1v0_0; stars: 1725497
Loading mmap:stars_5_2v0_0.cat: 5_2v0_0; stars: 7669011
Loading mmap:stars_6_2v0_0.true cat: 6_2v0_0; stars: 26615233
Loading mmap:stars_7_2v0_0.cat: 7_2v0_0; stars: 57826266
Loading mmap:stars_8_2v0_0.true cat: 8_2v0_0; Mistake: SpecialZoneArray(viii)::SpecialZone
Array: MapViewOfFile failed: 8, page_size: 65536
initialization failed
finished, max_geodesic_level: seven
If anyone has any suggestions, I'd capeesh information technology very much.
Cheers,
Garry
If you would similar to refer to this comment somewhere else in this project, copy and paste the following link: