Message Boards Message Boards

Avoid instability of Mathematica 11.3?

Following a recent upgrade from 11.2 to 11.3 I have found that Mathematica repeatedly crashes (black rather than coloured code) when using some custom functions that I have developed. It is not my Mac as similar problems occur on Macs belonging to my colleagues. Reverting to 11.2 or earlier solves the problem completely. I am still trying to isolate some trigger code examples that I can send to Wolfram support. Are others experiencing such crashes?

POSTED BY: Rudolph Kalveks
20 Replies

The entire X server crashing is most likely to be a graphics driver problem.

Things worth trying would be to use a different driver or start Mathematica with the -mesa command line flag.

POSTED BY: Ilian Gachevski
Posted 5 years ago

Thanks for the resolving this issue. As you said, now it works perfectly by adding command line flag -mesa to the Mathematica.

POSTED BY: HL Hariyanto
Posted 5 years ago

Today, I moved Mathematica to another computer. The command line interface works perfectly, but the GUI doesn't work. When I open Mathematica, after loading the kernel in the splash screen, my laptop is suddenly reboot. I post a video for that. I have tried all the possible answers on the internet, but I found nothing that resolve the issue. I hope this forum will help me.

I use Mathematica Student 11.2.0 for Linux 64-bit. The operating system is Ubuntu 16.04 LTS.

This is some useful information about the machine.

$ math
Mathematica 11.2.0 Kernel for Linux x86 (64-bit)
Copyright 1988-2017 Wolfram Research, Inc.

In[1]:= Needs["Benchmarking`"]                                                  

In[2]:= Benchmark[]                                                             
Test 1 of 15: Data Fitting ...
Test 2 of 15: Digits of Pi ...
Test 3 of 15: Discrete Fourier Transform ...
Test 4 of 15: Eigenvalues of a Matrix ...
Test 5 of 15: Elementary Functions ...
Test 6 of 15: Gamma Function ...
Test 7 of 15: Large Integer Multiplication ...
Test 8 of 15: Matrix Arithmetic ...
Test 9 of 15: Matrix Multiplication ...
Test 10 of 15: Matrix Transpose ...
Test 11 of 15: Numerical Integration ...
Test 12 of 15: Polynomial Expansion ...
Test 13 of 15: Random Number Sort ...
Test 14 of 15: Singular Value Decomposition ...
Test 15 of 15: Solving a Linear System ...

Out[3]//InputForm= 
{"MachineName" -> "user-lenovo-g41-35", "System" -> "Linux x86 (64-bit)", 
 "BenchmarkName" -> "WolframMark", "FullVersionNumber" -> "11.2.0", 
 "Date" -> "July 30, 2019", "BenchmarkResult" -> 0.417, "TotalTime" -> 33.2, 
 "Results" -> {{"Data Fitting", 2.207}, {"Digits of Pi", 0.886}, 
   {"Discrete Fourier Transform", 1.488}, {"Eigenvalues of a Matrix", 3.789}, 
   {"Elementary Functions", 1.918}, {"Gamma Function", 1.119}, 
   {"Large Integer Multiplication", 1.192}, {"Matrix Arithmetic", 0.814}, 
   {"Matrix Multiplication", 4.328}, {"Matrix Transpose", 2.034}, 
   {"Numerical Integration", 2.939}, {"Polynomial Expansion", 0.277}, 
   {"Random Number Sort", 2.173}, {"Singular Value Decomposition", 4.447}, 
   {"Solving a Linear System", 3.589}}}

$ cat /proc/version
Linux version 4.4.0-157-generic (buildd@lgw01-amd64-037) (gcc version 5.4.0 20160609 (Ubuntu 5.4.0-6ubuntu1~16.04.10) ) #185-Ubuntu SMP Tue Jul 23 09:17:01 UTC 2019

$ uname -a
Linux user-Lenovo-G41-35 4.4.0-157-generic #185-Ubuntu SMP Tue Jul 23 09:17:01 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux

$ dpkg-query -l | grep -i wolfram
ii  wolframscript                                 1.2.0-22                                                 amd64        WolframScript (for the Command Line)

$ cat /proc/cpuinfo 
processor	: 0
vendor_id	: AuthenticAMD
cpu family	: 22
model		: 48
model name	: AMD A8-7410 APU with AMD Radeon R5 Graphics
stepping	: 1
microcode	: 0x7030105
cpu MHz		: 2200.000
cache size	: 2048 KB
physical id	: 0
siblings	: 4
core id		: 0
cpu cores	: 4
apicid		: 0
initial apicid	: 0
fpu		: yes
fpu_exception	: yes
cpuid level	: 13
wp		: yes
flags		: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc extd_apicid aperfmperf pni pclmulqdq monitor ssse3 cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt topoext perfctr_nb bpext perfctr_l2 cpb hw_pstate ssbd vmmcall bmi1 xsaveopt arat npt lbrv svm_lock nrip_save tsc_scale flushbyasid decodeassists pausefilter pfthreshold overflow_recov
bugs		: fxsave_leak sysret_ss_attrs spectre_v1 spectre_v2 spec_store_bypass
bogomips	: 4392.01
TLB size	: 1024 4K pages
clflush size	: 64
cache_alignment	: 64
address sizes	: 40 bits physical, 48 bits virtual
power management: ts ttp tm 100mhzsteps hwpstate cpb [12] [13]
(... and other 4 cores ...)

$ cat /proc/meminfo 
MemTotal:        7047424 kB
MemFree:          416284 kB
MemAvailable:    5734552 kB
Buffers:          251908 kB
Cached:          5078552 kB
SwapCached:          100 kB
Active:          4359208 kB
Inactive:        1796672 kB
Active(anon):     718612 kB
Inactive(anon):   125160 kB
Active(file):    3640596 kB
Inactive(file):  1671512 kB
Unevictable:         168 kB
Mlocked:             168 kB
SwapTotal:       7813116 kB
SwapFree:        7807336 kB
Dirty:                 4 kB
Writeback:             0 kB
AnonPages:        825572 kB
Mapped:           321716 kB
Shmem:             18352 kB
Slab:             348384 kB
SReclaimable:     310036 kB
SUnreclaim:        38348 kB
KernelStack:       11504 kB
PageTables:        41828 kB
NFS_Unstable:          0 kB
Bounce:                0 kB
WritebackTmp:          0 kB
CommitLimit:    11336828 kB
Committed_AS:    5966752 kB
VmallocTotal:   34359738367 kB
VmallocUsed:           0 kB
VmallocChunk:          0 kB
HardwareCorrupted:     0 kB
AnonHugePages:         0 kB
CmaTotal:              0 kB
CmaFree:               0 kB
HugePages_Total:       0
HugePages_Free:        0
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:       2048 kB
DirectMap4k:      170996 kB
DirectMap2M:     3928064 kB
DirectMap1G:     4194304 kB
(with Mathematica in command line running)
POSTED BY: HL Hariyanto

Hello. I know it's an old thread that I'm going to revive - sorry - but I just landed here while searching info about the crashes I saw myself:

I've upgraded on a RasPi 3B+ from Mathematica 11.2 to 12.0. Now, every pass of the built in mathematica benchmark crashes the system directly into a hard reboot in Test 9 (I think). 11.2 completed all benchmarks without any problem on the same machine. Did an apt-get update; apt-get upgrade, apt-get dist-upgrade afterwards, but this did not change a thing. The original raspian image was set up in about Feb 2018.

Steps to reproduce:

Needs["Benchmarking`"]
Benchmark[]

About the system (same happened with older kernel before upgrade):

$ cat /proc/version 
Linux version 4.19.57-v7+ (dom@buildbot) (gcc version 4.9.3 (crosstool-NG crosstool-ng-1.22.0-88-g8460611)) #1244 SMP Thu Jul 4 18:45:25 BST 2019

$ uname -a
Linux raspberrypi 4.19.57-v7+ #1244 SMP Thu Jul 4 18:45:25 BST 2019 armv7l GNU/Linux

$ dpkg-query -l | grep -i wolfram
ii  wolfram-engine                                12.0.1+2019062401                          armhf        Mathematica® and the Wolfram Language
ii  wolframscript                                 1.3.0+2019062801                           armhf        WolframScript (for the Command Line)

$ cat /proc/cpuinfo 
processor       : 0
model name      : ARMv7 Processor rev 4 (v7l)
BogoMIPS        : 38.40
Features        : half thumb fastmult vfp edsp neon vfpv3 tls vfpv4 idiva idivt vfpd32 lpae evtstrm crc32 
CPU implementer : 0x41
CPU architecture: 7
CPU variant     : 0x0
CPU part        : 0xd03
CPU revision    : 4
(...and so on, 4 cores...)

$ cat /proc/meminfo 
MemTotal:         943748 kB
MemFree:          270556 kB
MemAvailable:     516936 kB
Buffers:           41588 kB
Cached:           256476 kB
[...]
(This is with mathematica already running.)

I'm happy to provide further details if needed or open a new thread for this if you think I should.

Thanks a lot for any response or the Wolfram team looking into this and kind regards, js

The entire RPi rebooting is likely due to hardware instability.

Possible things to look at would be the power supply voltage and/or CPU temperature.

POSTED BY: Ilian Gachevski
Posted 5 years ago

I recently upgraded from mma 7 to 11.3. I'm running Windows 7 Professional 64 bit. I keep having issues with 11.3. For example, after evaluating a simple expression (e.g. 1+1) in an old notebook, a popup message appears "One or more dynamic objects are taking excessively long to finish....", and the notebook freezes. There are no dynamic commands (e.g. Manipulate[ ]) in the notebook. Wolfram Support's suggestion was to do a "clean start", but the problems did not go away. Numerous other stability issues have meant that I have gone back to v7, which is comparatively rock steady. I wonder if my problems are because I have both 7 and 11.3 installed, although that was suggested as being viable in the installation procedure for Windows.

POSTED BY: chris slinger

Got some crashes too, Mac OSX, 16GB of memory

POSTED BY: Samuel Jones

Please can you specify your Mathematica version and any other relevant info.

POSTED BY: Moderation Team
Posted 5 years ago

Long-time user, big-time fan, and first-time replier:

I just upgraded to 11.3 (Mac OS X x86 64-bit) and the following crashed the kernel:

Map[Normalize, 
   RandomVariate[
    MultinormalDistribution[ConstantArray[0, 6], IdentityMatrix[6]], 
    10^6]];

I was able to isolate it to Map and to the size of the list, where 10^4 works, but 10^6 does not. Therefore, I had to go back to 11.2, where there is no problem.

POSTED BY: Sean Ford

@Nicholas, @Rudolph: With help from our Technical Serviices group I found some of your examples.

I verified that a factorization crashed in 11.3 (I do not recall who sent which example but this was from a notebook that used a large expression named coulomb). There was a prior simplification example that I ran to completion. I did not try to test it with multiple runs because it involved some steps that have interactions between front end interface and kernel (via Monitor).

I also verified that a Simplify example from a notebook named with "HarmonicAnalysys" in the name was crashing in 11.3.

I suspect both crashes have a common cause involving PolynomialGCD internals. Both appear to have been addressed for version 12, in that I could do several runs of each with no crash (also they seem a bit faster). I apologize for any inconvenience this problem has caused.

POSTED BY: Daniel Lichtblau

So the problem is not just with the Macs in Imperial College, London…(where many of us have given up with 11.3 - so are sticking doggedly to 11.2).

Interesting that you just need to call Simplify on a big piece of code to create the crash. Our problems similarly involve transformations of complicated functions that require a lot of memory.

I suspect a memory management problem in 11.3. It may be specific to Mac OSX as the programmer at Wolfram that I exchanged some of my problem files with was running another system and could not replicate the crashes.

It would be nice if Wolfram would wake up, take interest and fix such a fundamental error.

POSTED BY: Rudolph Kalveks

I had this problem for two months using Simplify within some complex code. I stripped down examples so that they would crash nearly 100% of the time and sent them to Wolfram. They were still too big for the Wolfram people to analyze. The bottom line is that 11.3 was so unstable that it is completely useless to me. I reverted to 11.2 and have been happy ever since ... (except for the fact that Wolfram still hasn't fixed 11.3 ... I check from time to time).

I am using the latest Mac laptop with 32 GB of memory ... but the crashes also happened on my older laptop with 16GB of memory. The program was nowhere near the memory limits and was not grinding for hours ... it could happen within minutes of executing the code.

POSTED BY: Nicholas Warner
Anonymous User
Anonymous User
Posted 6 years ago

Ditto, I use a mac (love it!), I had had issues with earlier 11, but 11.2 is good and 11.3 seem fine so far.

I suggest two things:

A typical mistake is to not see WR's installation instruction page (for mac) which says to (uninstall) the old version before installing the new (it's ok, you can still correct). go read that off the wolfram help pages and go step by step, but read #2 first.

You may need to rename these two folders: /Users/whoami/Library/{Wolfram,Mathematica} (when mathematica is not running). Restart your re-installed mathematica. You'll LOOSE settings and DATA stored in those two places (remediating that i don't mention), possibly paid download content (many people don't need to worry about it at all). The FOLDERS will be remade with defaults that are "surely compatible" with 11.3. (if you keep your 11.2 Library folder, and you may need to, you are bound to fix problems that might cause. as far as i know there are none currently but there has been in the past, issues re-using the Library/app folder's old settings)

Apple has great forums (the above advice about ~/Library can be found there and at the following places).

A good online resource for fixing minor issues on Apple devices, to avoid Support calls or trip to Apple store, is:

https://www.lifewire.com/fix-4102586

and

http://osxdaily.com

POSTED BY: Anonymous User

Hi, after upgrade to 11.3 it is a horror ( linux Mint 18.3). Mathematica leaves the kernel without any message. This happens during Simplify. I have 32 GB of memory, during these operations Mathematica uses less than 4GB. When I separate critical calculations into a new notebook, then it passes through.

When you get these crashes, do you know offhand whether anything was happening in the user interface, such as scrolling of the mouse hovering over code? I've seen problems involving those sorts of situations (though on Linux), one of which I did manage to reproduce and report.

POSTED BY: Daniel Lichtblau

I have not yet experienced any problems with the latest version:

In[3]:= $Version    
Out[3]= "11.3.0 for Linux x86 (64-bit) (March 7, 2018)"

... and I definitely like it - congrats to WR !

POSTED BY: Henrik Schachner

The advantage of LTS releases is that they keep getting bug fixes, but they don't (often) get new bugs introduced, as no new features are added. Thus in the LTS the number of bugs goes down over time, while in the Current there's a balance between fixes and breakages.

I think that before experimenting with an LTS release, Wolfram should address a much bigger problem: bugs in certain functionality areas just don't get fixed. That's why I keep complaining about the Graph stuff—there are many serious bugs that haven't been touched at all through several releases. Actually, this functionality area isn't getting any new features either, so the bug number at least isn't going up ...

[To be fair, some Graph-bugs did get fixed, but most serious ones I found (silent wrong results, random results, etc.) still persist after multiple reports.]

POSTED BY: Szabolcs Horvát

I wish we had two Mathematica branches: one very stable and another one with the latest features. For example, people from NodeJS have the LTS (Long Term Support) and Current releases for that. In my case I would be using the LTS version for all the tools that I release for my team and Current only for exceptional exploratory work.

And yes, I have found and reported some crashes, but exclusive to 11.3.

POSTED BY: Gustavo Delfino

Running 11.3 under Windows 10, I have the general impression that the software is less stable particularly when accessing the documentation center for the first time. Things seem to momentarily hang, and sometimes crash.

POSTED BY: Todd Allen

I did not find 11.3 to be any more crashy than 11.2. It must be something specific that is present in your code that causes the crash.

POSTED BY: Szabolcs Horvát
Reply to this discussion
Community posts can be styled and formatted using the Markdown syntax.
Reply Preview
Attachments
Remove
or Discard

Group Abstract Group Abstract