MATLAB: Do I receive a segmentation violation when I perform calculations using sparse matrices in MATLAB 6.5.1 (R13SP1)

_spmultopMATLABnumericsnumerics.dllspmultop

I use the following code to define a sparse matrix and then multiply with a scalar:
B(1) = sparse(1,1);
B*0
I receive the following segmentation violation:
 ------------------------------------------------------------------------
        Segmentation violation detected at Mon Apr 25 11:00:29 2005
 ------------------------------------------------------------------------
 Configuration:
   MATLAB Version:   6.5.1.199709 (R13) Service Pack 1
   Operating System: Microsoft Windows XP
   Window System:    Version 5.1 (Build 2600: Service Pack 2)
   Processor ID:     x86 Family 15 Model 2 Stepping 9, GenuineIntel
   Virtual Machine:  Java 1.3.1_01 with Sun Microsystems Inc. Java HotSpot(TM) Client VM
     (mixed mode)
 Register State:
   EAX = 00000000  EBX = 24f81da0
   ECX = 00000000  EDX = 00000001
   ESI = 24f82100  EDI = 00000000
   EBP = 00dfd304  ESP = 00dfd270
   EIP = 7a624d8a  FLG = 00010246
 Stack Trace:
   [0] numerics.dll:_spmultop(0x24f81da0, 0x24f82100, 2, 1) + 410 bytes
   [1] numerics.dll:_spBinaryArrayOps(0x24f81d10, 0x24f81650, 2, 1) + 231 bytes
   [2] numerics.dll:_spBinaryMatrixOps(0x24f81d10, 0x24f81650, 0, 1) + 169 bytes
   [3] numerics.dll:_spFullBinMatrixTimesFcn(1, 0x00dfd414, 2, 0x00dfd4dc) + 39 bytes
   [4] m_interpreter.dll:_inExecuteInternalFcn(206, 2, 1, 0) + 1183 bytes
   [5] m_interpreter.dll:enum opcodes __cdecl inIntFcn(enum opcodes,int,int,int)(77, 0x010000ce, 1, 2) + 90 bytes
   [6] m_interpreter.dll:int __cdecl inInterp(enum inDebugCheck,int,int,struct inPcodeNest_tag volatile *)(2, 0, 0, 0x012de5e8) + 2554 bytes
   [7] m_interpreter.dll:_inInterPcode(2, 0x00dfdc54, 0, 0) + 193 bytes
 <snip>

Best Answer

This bug has been fixed in Release 14 Service Pack 3 (R14SP3). For previous product releases, read below for any possible workarounds:
There is a bug in MATLAB 6.5.1 (R13SP1) that affects the way MATLAB handles sparse matrix initialization.
When a sparse matrix is created using this syntax, the amount of storage allocated for nonzero matrix elements is initialized to zero. This can be verified using the following command:
nzmax(B)
This causes an error when performing calculations using sparse matrices with zero storage allocation.
To work around this issue, use the following alternative syntax to define a sparse matrix:
B = sparse(1,1);
Related Question