[FLASH-USERS] relocation truncated to fit

[Alexander Wagner]

Hi,

I've often had this problem often too. I added some compiler flags and I 
was successful after some trial and error changing MAXBLOCKS, N[XYZ]B, 
and the number of cpus. If you can't find a combination of these 
parameters that works, the size of your problem may be too big for the 
total amount of memory and cpus you have at your disposal.

Here are some relevant archived posts (which I had sent to Mikhail but 
not the list)

http://flash.uchicago.edu/pipermail/flash-users/2008-September/000307.html
http://flash.uchicago.edu/pipermail/flash-users/2010-August/000743.html
http://flash.uchicago.edu/pipermail/flash-users/2010-August/000744.html

Best wishes,
Alex


On 15/06/11 12:19 PM, Samuel Friedman wrote:
> Mikhail,
>
> I've had this same problem too!  Using the Intel compiler, I found 
> some linker flags to help (they deal with allocating more memeory or 
> something like that), but they didn't help too much.
>
> If anybody has any suggestions, I'd love to here them.
>
> Sam
>
>
> On Tue, 14 Jun 2011, Mikhail Klassen wrote:
>
>> When compiling a setup for a simulation of star formation inside a 
>> turbulent cloud, I receive the following error when
>> attempting to compile with maxblocks greater than about 900. After 
>> the linker stage in the compilation process, I
>> receive:
>>
>> dBaseDataByPointer.o(.text+0x184): In function 
>> `dbasedatabypointer_mp_dbasegetptrtozcoords_':
>> : relocation truncated to fit: R_X86_64_32S physicaldata_mp_unk4_
>> dBaseDataByPointer.o(.text+0x1f4): In function 
>> `dbasedatabypointer_mp_dbasegetptrtoycoords_':
>> : relocation truncated to fit: R_X86_64_32S physicaldata_mp_unk3_
>> dBaseBoundaryDataByName.o(.text+0x407): In function 
>> `dbaseboundarydatabyname_mp_dbaseputboundaryfluxes_':
>> : relocation truncated to fit: R_X86_64_32S physicaldata_mp_tflux_z_
>> dBaseBoundaryDataByName.o(.text+0x415): In function 
>> `dbaseboundarydatabyname_mp_dbaseputboundaryfluxes_':
>> : relocation truncated to fit: R_X86_64_32S physicaldata_mp_tflux_z_
>> dBaseBoundaryDataByName.o(.text+0x420): In function 
>> `dbaseboundarydatabyname_mp_dbaseputboundaryfluxes_':
>> : relocation truncated to fit: R_X86_64_32S physicaldata_mp_tflux_z_
>> dBaseBoundaryDataByName.o(.text+0x42b): In function 
>> `dbaseboundarydatabyname_mp_dbaseputboundaryfluxes_':
>> : relocation truncated to fit: R_X86_64_32S physicaldata_mp_tflux_z_
>> dBaseBoundaryDataByName.o(.text+0x436): In function 
>> `dbaseboundarydatabyname_mp_dbaseputboundaryfluxes_':
>> : relocation truncated to fit: R_X86_64_32S physicaldata_mp_tflux_z_
>> dBaseBoundaryDataByName.o(.text+0x441): In function 
>> `dbaseboundarydatabyname_mp_dbaseputboundaryfluxes_':
>> : relocation truncated to fit: R_X86_64_32S physicaldata_mp_tflux_z_
>> dBaseBoundaryDataByName.o(.text+0x44c): In function 
>> `dbaseboundarydatabyname_mp_dbaseputboundaryfluxes_':
>> : relocation truncated to fit: R_X86_64_32S physicaldata_mp_tflux_z_
>> dBaseBoundaryDataByName.o(.text+0x457): In function 
>> `dbaseboundarydatabyname_mp_dbaseputboundaryfluxes_':
>> : relocation truncated to fit: R_X86_64_32S physicaldata_mp_tflux_z_
>> dBaseBoundaryDataByName.o(.text+0x4bd): In function 
>> `dbaseboundarydatabyname_mp_dbaseputboundaryfluxes_':
>> : additional relocation overflows omitted from the output
>>
>> Has anyone else experienced this and knows a fix?
>>
>> Thanks!
>> Mikhail Klassen
>>
>>