I'm using a x86/linux machine.
I'm trying to execute the programs of acstone benchmark.
Initially I used Altera's compiler and assembler.
The elf files generated by Altera's assembler are
little endian. When I try to execute the program in a
interpreted simulator I get the following output:
SystemC 2.1.v1 --- Apr 7 2006 16:29:02
Copyright (c) 1996-2005 by all Contributors
ALL RIGHTS RESERVED
ArchC: Reading ELF application file: 000.main.nios2
ArchC: -------------------- Starting Simulation
--------------------
Warning: (W562) sc_initialize is deprecated: use
sc_start(0) instead
In file: sc_simcontext.cpp:1199
Warning: (W540) sc_cycle is deprecated: use
sc_start(...) instead
In file: sc_simcontext.cpp:1213
ArchC Error: Unidentified instruction.
PC = 0
ArchC: -------------------- Simulation Finished
--------------------
Warning: (W540) sc_cycle is deprecated: use
sc_start(...) instead
In file: sc_simcontext.cpp:1213
ArchC: Simulation statistics
Times: 0.00 user, 0.00 system, 0.03 real
Number of instructions executed: 0
Simulation speed: (too fast to be precise)
When I use the assembler generated by ArchC i get this
output:
SystemC 2.1.v1 --- Apr 3 2006 19:21:55
Copyright (c) 1996-2005 by all Contributors
ALL RIGHTS RESERVED
ArchC: -------------------- Starting Simulation
--------------------
Warning: (W562) sc_initialize is deprecated: use
sc_start(0) instead
In file: sc_simcontext.cpp:1199
ArchC: Address out of bounds (pc=0x0).
ArchC: -------------------- Simulation Finished
--------------------
ArchC: Simulation statistics
Times: 0.01 user, 0.00 system, 0.00 real
Number of instructions executed: 0
Simulation speed: (too fast to be precise)
I checked this binary file and it's a big endian, but
the model is little endian. Why the assembler generates
a big endian file?
So I tryed to set the model to big endian e it worked
for that files generated by ArchC assembler, but it
should work also as little endian.
Guilherme Melo
Logged In: YES
user_id=1489337
Indeed... the assembler on version 1.6.0 of the ArchC
package had some issues to generate little endian objects
files. Those were fixed on the 2.x series.
Are you planning to migrate to the 2.x series?
-Alexandro.
Logged In: YES
user_id=1506481
Well, I was not planning to migrate to 2.x now.
I don't know how much effort I would need to migrate
and I don't have much time to finish this model.
Meanwhile I will use the model as big endian.
Guilherme