There were a couple of patches sent to the list that should have
resolved this problem.
http://article.gmane.org/gmane.comp.emulators.m5.users/2608
http://article.gmane.org/gmane.comp.emulators.m5.users/2598
Ali
On Apr 13, 2008, at 12:49 AM, Clint Smullen wrote:
I have seen previously discussion from the middle of March ([m5-
users] error running splash2 in beta 5, with followups on [m5-dev])
discussing the following error message:
Assertion `tags->findBlock(mshr->addr) == __null' failed.
Running full-system Alpha, I am able to get the error to occur with
the addition of two configuration likes the the configs/example/
fs.py as follows:
test_sys.physmem.latency = '10ns'
test_sys.l2.latency = '100ns'
Running the simulator with '-n 4 --caches --l2cache', no need to
specify a script, as it occurs even before reaching the prompt. This
is booting the kernel off the website along with tsb_osfpal compiled
from the alpha-system package. I am unable to get the error to occur
with only one core, though it does occur with two cores, though it
did not occur with eight.
I attempted commenting out the assertion, as previously mentioned.
The system is able to boot with four cores, but multithreaded
workloads have a tendency to get fatal errors (fatal within the
simulated machine, crashing only the program, not the simulated
system).
I am also getting this error in a less asinine custom configuration,
where the speed difference between the cache and memory is small
(40ns and 50ns, respectively), though these settings did not cause
it in the example configuration.
I could not find any further responses that had indicated this issue
had been resolved, and was hoping someone would know something more.
Thanks,
- Clint Smullen
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
