[Gems-users] ATMTP: Problem with strange Transaction-Aborts

[Philipp Tölke <toelke+gems@xxxxxxxxx>]

Hi Everyone,

I experience strange Transaction-Aborts with ATMTP:

|6686637022   1 [1,0] TID -1 XACT ABORT 1 caused by 0 [ 0, 0 ] xid: 0
|address: [0x0, line 0x0] delay: -1903297570  PC [0x0, line 0x0]  *PC 0x0
|'illtrap 0'
|failed assertion 'value >= 0' at fn void Histogram::add(int64) in
|common/Histogram.C:100
|failed assertion 'value >= 0' at fn void Histogram::add(int64) in
|common/Histogram.C:100
|At this point you might want to attach a debug to the running and get to
|the
|crash site; otherwise press enter to continue
|PID: 30696
|
|Abort (SIGABRT) in main thread
|The simulation state has been corrupted. Simulation cannot continue.
|Please restart Simics.
|Starting command line. (May have skipped commands in script files.)
|[cpu1] v:0x0000000000012f88 p:0x0003e672f88  mov %g0, %o0

The code 'mov %g0, %o0' is from the compiled begin_transaction-inline,
which I copied from the microbenchmarks.

I think this means, that a transaction fails the moment it is started. I
do not understand, why the delay is negative.

What can I do about this? How is it caused?

If it matters: I'm trying to run 332_ammp from the SPECOMP2001-Suite for
my Bachelors-Thesis.

Thanks,
-- 
Philipp Tölke

Attachment: signature.asc
Description: Digital signature