[Gems-users] Having trouble getting LogTM running

["Jim Leek" <jrleek@xxxxxxxxx>]

My goal is to duplicate the results from a LogTM paper, but I am having a lot of trouble installing the OS to get the first checkpoint. Could someone give me a hand?

I'm using Gems-2.0 with Simics 2.2.19, on a 3.4 Ghz Xeon with Ubuntu installed.
I have built Ruby, but not Opal.  I don't think it's required.
I downloaded the Solaris 10 CDs. (8/07 release)
I edited system.conf to set the correct values to run naked-check-create.sh. 
I then ran the naked-check-create.sh script: ./naked-check-create.sh 32 65536  (I understand the LogTM papers used 32 processors)

The script got through the install setup fairly quickly, but it now it appears to be frozen, spinning on some kind of loop.  The installation xterm says:

Preparing system for Solaris install

Configuring disk (c0t0d0)
        - Creating Solaris disk label (VTOC)

Creating and checking UFS file systems
        - Creating / (c0t0d0s0)

Beginning Solaris software installation

But the terminal I'm running simics on says:

[qlc24B_2 info] Note: Abort Command Received - Something probably went wrong
[qlc24B_2 info] Note: Abort Command Received - Something probably went wrong
[sc0 info] OBP 0x6006
[sd0 info] Unimplemented SCSI Command: <Unknown SCSI Command> (0x35)

Top is showing:

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
2798 jrleek    25   0  466m 417m 3688 R  100 20.6 819:12.57 simics-common       

In other words, it's been stuck at this point for about 13 hours.  That doesn't seem good.  


-----

Of course, maybe I'm using the wrong 'version' of simics?  The Quick start guide says I should use the executable in gems-2.0/simics/home/MOSI_SMP_bcast, but the script uses the one in: ${SIMICS_DIR}/home/sarek.

Thanks,
Jim