Hi All,
There are few problems in HMC simulation that have not been fixed so far:
1. hmctest.py doesn't work right out of the box and gives error:
Command line:
/gem5$ build/X86/gem5.opt configs/example/hmctest.py
Simulation Error:
Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/home/muhammad/GEM5/gem5/src/python/m5/main.py", line 433, in main
exec filecode in scope
File "configs/example/hmctest.py", line 166, in <module>
m5.instantiate()
File "/home/muhammad/GEM5/gem5/src/python/m5/simulate.py", line 115, in
instantiate
for obj in root.descendants(): obj.createCCObject()
File "/home/muhammad/GEM5/gem5/src/python/m5/SimObject.py", line 1468, in
createCCObject
self.getCCParams()
File "/home/muhammad/GEM5/gem5/src/python/m5/SimObject.py", line 1423, in
getCCParams
setattr(cc_params, param, value)
AttributeError: '_m5.param_SubSystem.SubSystemParams' object has no
attribute 'enable_buff_div'
2. Running HMC simulaton from se.py or fs.py gives attribute error:
Command Line:
gem5$ build/X86/gem5.opt configs/example/se.py --mem-type=HMC_2500_1x32 -c
hello
Simulation Error:
Traceback (most recent call last):
File "<string>", line 1, in <module>
File "/home/muhammad/GEM5/gem5/src/python/m5/main.py", line 433, in main
exec filecode in scope
File "configs/example/se.py", line 284, in <module>
MemConfig.config_mem(options, system)
File "/home/muhammad/GEM5/gem5/configs/common/MemConfig.py", line 156, in
config_mem
HMChost = HMC.config_host_hmc(options, system)
File "/home/muhammad/GEM5/gem5/configs/common/HMC.py", line 261, in
config_host_hmc
for i in xrange(system.hmc_host.num_serial_links)]
AttributeError: Values instance has no attribute 'ser_ranges'
This problem is there from a while and Andreas pointed this out as well a
while ago:
"""
Hi all,
The HMC config script needs some attention. On first inspection:
1.hmctest.py creates a system.ser_ranges structure that is not created when
the HMC script is called from the “normal” config scripts
2. The HMC config functions in HMC.py make assumptions around how
system.mem_ranges is used to hold the various vault ranges, as opposed to
the system memory ranges.
It would be great if someone could tidy this up to match better with the
normal config path through e.g. fs.py or se.py. It would be best if the HMC
model simply took the first item in system.mem_ranges and used that to
derive both the ser_ranges and vault_ranges. I would suggest to keep it
simple and only support a single system.mem_range at this point.
Would anyone volunteer to sort this out?
Thanks,
Andreas
"""
Has anyone able to figure the problem out? HMC simulation model getting
more errors with time.
Please help fix this issue!!! (if any authors of the code are out there)
Thanks
Muzamil
_______________________________________________
gem5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
