On Apr 16, 2008, at 1:38 PM, Meng-Ju Wu wrote:
Ali, thank you very much.

(1) I modified the src/arch/alpha/linux/process.cc.

      SyscallDesc("ftruncate", unimplementedFunc) =>
SyscallDesc("ftruncate", ftruncateFunc)

    Now, the ftruncate() works.

I want to implement other system calls. What the files should I look at?
Most of the syscalls are implemented in src/sim/syscall_emul*. Taking a loot at how they're implemented is probably the best way to get started. Then function pointers are simply inserted into the syscall tables. Some architecture and os specific syscalls are implemented in kern/*/*_syscalls.* and arch/alpha/linux/*.


(2) The memory system configuration is

      system = System(cpu = [CPUClass(cpu_id=i) for i in xrange(np)],
               physmem = PhysicalMemory(range=AddrRange("2048MB")),
               membus = Bus(), mem_mode = 'timing')

    This is the largest number I can set.

    There are 9 spec2006 benchmarks has the problem

" terminate called after throwing an instance of 'std::bad_alloc' "

    I tried to use a smaller dataset in 410.bwaves. It  finished
correctly without any problem.

    What should I  do to avoid this problem or correct it without
changing the dataset?
So it did work with a smaller dataset? Two possibilities arise:
1) There is a bug. Either with the way we allocate memory or some how we're emulating mmap wrong. It's not impossible, but I think it's pretty unlikely.
2) The benchmark needs that much memory. I found the following:
Memory SPEC CPU2006 requirements to memory have grown significantly: the typical memory size is 1 GB for 32-bit systems, exclusive of OS/ overhead, but more may be required. Typically, 64-bit environments will require 2GB for some of the benchmarks in the suite. More memory will be needed if you run multi-copy SPECrates: generally 1GB for 32- bit, or 2GB for 64-bit, for each copy you plan to run.

So it very well may require > 2GB of memory for that dataset.

Ali

_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users

Reply via email to