« first day (1427 days earlier)      last day (3358 days later) » 

kevin holzer
kevin holzer
09:09
Hi. I get hlzr.net/measureproblem.png out when I think I should get something more like 8 m. Any advice? I was also having problems running QuantityUnit (to check the inputs of a function), I think the problems I was experiencing are related...
 
2 hours later…
Szabolcs
Szabolcs
10:39
Can someone please test if the following crashes the front end?

1. Create a new notebook
2. Enter this code and evaluate it: `Image3D@DiskMatrix@Reverse[{528, 320, 456}/2]`. If you have enough memory, it should work.
3. Save the notebook and close it.
4. Open the notebook. **At this point my front end crashes.**
I wonder if this is because the front end is only 32-bit on OS X and thus can't use a lot of memory.
I'm curious if it also crashes on Windows.
@kevinholzer Does the problem persist after restarting the kernel?
@IstvánZachar On OS X it's fine. Maybe this is a Windows-specific problem.
István Zachar
István Zachar
11:51
@Szabolcs thanks for checking!
@Szabolcs No, it saved and opened normally, and can rotate the image correctly (though slowly). Win7 64-bit, Mathematica 10.3.0.0
Jacob Akkerboom
Jacob Akkerboom
12:33
@Szabolcs Interesting question, a quick look at some answers 1 2 does not support that 32 bit programs are more restricted in their memory use (correct me if I understood you wrong). If anything, it looks like 64 bit programs usually use more memory, which lets them fill up all the RAM faster. With the same stack size, the stack should also overflow sooner I think. But maybe you know more
Szabolcs
Szabolcs
12:57
@JacobAkkerboom Well, with 32 bit pointers we can only address 2^32 = 4 GB of memory. For some technical reason that I don't understand that is often just 2 GB in practice on some operating systems. I thought that maybe the decoding process for the Image3D was memory-inefficient and this limit got exceeded, and that's what caused the crash. The FE is trailing behind on OS X in still being a 32-bit executable (on Win/Linux it's 64-bit).
The kernel is of course 64-bit on OS X.
rcollyer
rcollyer
13:30
@Szabolcs I don't see a crash, what version are you running?
 
1 hour later…
Szabolcs
Szabolcs
14:43
@rcollyer 10.3 on OS X. Did you also test on OS X?
@rcollyer Actually all of 9.0.1, 10.0.2, 10.2 and 10.3 crash.
I don't keep 10.1 anymore.
Szabolcs
Szabolcs
15:12
Has anyone tried to measure a shortest path on a smooth surface which is given as a mesh?
rcollyer
rcollyer
15:42
@Szabolcs yes, with 10.3 on OSX, I don't get a crash. I'll check the others in a bit.
blochwave
blochwave
15:54
OOH 10.3.1 is on its way!
2
"This update resolves key regressions identified in recent
Mathematica releases. Version 10.3.1 addresses performance
slowdowns related to sending strings over the Wolfram Symbolic
Transfer Protocol, as well as to numerical interpolation. It also
delivers more robust import of image file formats, refined
support for international language and translation, improved
travel and Wikipedia data manipulation, and other cloud
deployment, image processing, and user interface fixes."
Szabolcs
Szabolcs
@rcollyer Could it be the graphics hardware then? I use a 2014 MacBook Pro with NVIDIA GeForce GT 750M. I also tried to do it after restricting the OS to use the integrated graphics card (Iris Pro), and it still crashes.
@rcollyer I did report it (CASE:3486912), but I'm not really expecting anything. Just wanted to let WRI know.
P. Fonseca
P. Fonseca
16:19
@user21 Hi Olivier, I have one question concerning an error I kept receiving on NDSolve, that I don't know if it should be reported or not as an improvement request...
I have one single ODE (pressure along height), with a derivative that has a sharp change, managed with a If (two fluids, one on top of the other, which densities depend on pressure, but also, resulting into an abrupt transition of density, when one fluid ends and the other starts, somewhere in the middle of the independent variable height).
For convenience, I placed the density also as a function (like and auxiliary variable, and so, my system has three equations: pressure[0], pressure'[lheight], density[height]), so to extract its interpolation function directly from NDSolve, instead of "rebuilding" it later.
I kept receiving the error NDSolveValue::nderr at the transition point... By listing both pressure and density on option DependentVariables, the error message just went away... which tells me that the problem was kind of fictitious. Is this to be expected? Can this be improved (should I report it?) Thank you.
(on the above text: lheight is actually a typo, and should be read as just height...)
halirutan
halirutan
@P.Fonseca It is unlikely that user21 sees your ping because he was absent for too long from chat. Basically, only the usernames that are expanded when you start writing @... are really pinged.
P. Fonseca
P. Fonseca
@halirutan thank you! didn't know that... I'll "ping" him on his e-mail... :-)
halirutan
halirutan
@P.Fonseca Yes, that's better.
Jacob Akkerboom
Jacob Akkerboom
16:42
@Szabolcs Ah, you are right of course (this answer supports you). Thanks for reminding me how this works.
I will do a test with ReadList, note that ReadString["!cat"] did work for me
I also have a NVIDIA GeForce GT 750M (2048 MB) and I have a mac, so maybe I'll do that test after
Ok, I suppose ReadList does indeed not work. I get this in my terminal: 
Jacob Desktop $ echo "1 \n 2 \n 3"| ./streamTestScript.sh
ReadList::readt: Invalid input found when reading 1 \n 2 \n 3 from !cat.
{$Failed}
Wait, I should have used echo -e it seems, then it does work
The script that works with ReadList, is 
#!/Applications/Mathematica.app/Contents/MacOS/WolframScript -script
Print@ReadList["!cat"];
Then I get 
Jacob Desktop $ echo -e " 1 \n 2 \n 3"| ./streamTestScript.sh
{1, 2, 3}
Wolfram Blog
Wolfram Blog
Wolfram Blog: Wolfram Programming Lab Now Available in Beta
posted on December 15, 2015 by Rob Morris

The Wolfram Language provides a unique opportunity to revolutionize programming education, and we’ve been working on ways to deliver the language for students and educators. Today we’re making available a beta version of Wolfram Programming Lab. You can access it free on the web in our Wolfram Open Cloud. There are subscription versions that provide [...]

rcollyer
rcollyer
@Szabolcs thanks. Hopefully, they can reproduce it.
Jacob Akkerboom
Jacob Akkerboom
16:58
@Szabolcs My front end also crashes
user21
user21
@P.Fonseca, hi! That error tag seems unfamiliar, I'd need to see the code to comment if this an issue or no@r. If you could post it here or send it to me that would be great. And @halirutan is right I seldom hang out here - so email or a question is better for me to see it. Seasons greetings to everyone here.
 
2 hours later…
kevin holzer
kevin holzer
18:31
@Szabolcs Thanks! That fixed it for me :)
 
3 hours later…
Sjoerd C. de Vries
Sjoerd C. de Vries
21:45
@belisarius No more "learning to fish"?
@blochwave Interestingly, the newby of this Community post of a week ago seems to already have it. I wonder how...
blochwave
blochwave
22:34
@SjoerdC.deVries As always I get the email "your site is scheduled to be the first to receive the upgrade"...then many more get it before me.
Sjoerd C. de Vries
Sjoerd C. de Vries
@blochwave Well, I got nothing so far and I have a Premier License (TM). The guy in the refence is the only one I heard it from
blochwave
blochwave
Also worth mentioning for people on here:
"Updating from Mathematica or MathLM 10.3 to 10.3.1 will not
require a new activation key. Users will only be prompted to
enter a new key if installing Mathematica or MathLM for the first
time or upgrading from a version that predates Version 10.3.

Because this update contains important fixes, anyone who has
already been assigned an activation key for Mathematica or MathLM
10.3 will automatically be provided with access to the installers
for Version 10.3.1 through the Wolfram User Portal, regardless if
Is the quote from my email
Sjoerd C. de Vries
Sjoerd C. de Vries
22:51
@blochwave Hmm, if I go to the wolfram portal and click on the upgrade button at the end of the 10.3 line I am offered to upgrade to 10.1. I guess, I'll wait a while

« first day (1427 days earlier)      last day (3358 days later) »