someone could clarify this for me.
Thanks,
Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
n (if ruby) src/mem/ruby/system/Sequencer.cc in
Sequencer::handleLlsc(). So if an LL/SC from another thread overlaps,
this guarantees that a thread's SC does not complete the store if the last
load linked was issued by the same thread.
On Mon, Nov 19, 2012 at 10:45 PM, Anusha wrote:
>
ble to find logic that will
prevent thread 3 from executing LL and SC.
~ Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
You hit the nail on its head! Thank you so much that detailed answer which
clarified things a lot.
The confusion I have is during LL we never check if lock is already taken
for that address or am I missing that part? As you correctly mentioned, the
lock must be free to begin with.
~ Anusha
)). Am I missing
something here?
Any help would be really appreciated.
~ Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
Hi,
I tried adding a map(std::map) to L1 cache(MOESI_CMP_directory-L1cache.sm).
I added the datatype to RubySlicc_Exports.sm. (external_type(std::map,
primitive="yes");)
When I tried to compile, it gave me the following error. Syntax error at
/home/anusha/gem5_9370_4th/src/me
I wanted to add a note on the subject that it is related to slicc.
On Sun, Feb 17, 2013 at 7:33 PM, Anusha wrote:
> Hi,
>
> I tried adding a map(std::map) to L1
> cache(MOESI_CMP_directory-L1cache.sm). I added the datatype to
> RubySlicc_Exports.sm. (external_type(std::map,
Update on this.
I synced to 9370 and I haven't had problems creating checkpoints with ruby
and restoring:
*Command to create checkpoint:*
./build/ALPHA_MOESI_hammer/gem5.opt -d *outputdir* configs/example/ruby_fs.py
--cpu-type=timing --ruby -n 4 -b *Benchmarkname*
*
*
*Command to restore checkp
No, it did not create any checkpoint.
-Anusha
On Mon, Jun 25, 2012 at 10:46 AM, biswabandan panda wrote:
> I guess, There is nothing wrong with it. It's fine. Are u able to create
> checkpoints?
>
> On Mon, Jun 25, 2012 at 9:11 PM, Anusha wrote:
>
>> When I tr
id you passed to create checkpoint. I think
> total number of canneal is less than x264.
>
> On 6/25/12, Anusha wrote:
> > No, it did not create any checkpoint.
> >
> > -Anusha
> >
> > On Mon, Jun 25, 2012 at 10:46 AM, biswabandan panda
> > wrote:
> >
Checkpointing works for streamcluster. I am working on canneal. I think it
would work.
Thank you so much for the help.
Thanks,
Anusha
On Mon, Jun 25, 2012 at 11:03 AM, biswabandan panda wrote:
> It seems u were using simsmall input set and your rcS script is Canneal
> but u have forgot
/streamcluster/output.txt 4
echo "Done :D"
/sbin/m5 exit
/sbin/m5 exit
Thanks,
Anusha
On Mon, Jun 25, 2012 at 5:48 PM, Anthony Gutierrez wrote:
> Nowhere in the information you've provided do I see a checkpoint being
> created. Maybe I'm missing something.
>
> -Tony
I use the precompiled PARSEC binaries. Instead of using
parsec/install/bin.ckpts/ binaries, I had used parsec/install/bin/binaries.
Hence such an error message I guess.
-Anusha
On Mon, Jun 25, 2012 at 6:04 PM, Anthony Gutierrez wrote:
> I see, so you've compiled into the parsec binary
stable version, I had no trouble creating a ALPHA
FS ruby checkpoint and now I am having issues.
I receive the following message.
Runtime Error at MOESI_CMP_directory-L1cache.sm:233, Ruby Time: 6061396337:
Invalid RubyRequestType, PID: 21009
Is anyone aware of the cause?
Thanks,
Anusha
and the significance of it
Thanks,
Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
le
0: system.tsunami.io.rtc: Real-time clock set to Thu Jan 1 00:00:00 2009
What am I missing here?
Thanks,
Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
-flags=ProtocolTrace
--trace-file body_mem.out.gz -d results/new_bodytrack
./configs/example/ruby_fs.py -n 4 -b Bodyno --cpu-type=detailed --caches
--l2cache -F 50
Global frequency set at 1 ticks per second
info: kernel located at: /home/anusha/m5_system_2.0b3/binaries/vmlinux
P_directory --> to obtain
ProtocolTrace, do you checkpoint with fs.py and restore using ruby_fs.py?
Would that work
4. The reason I wish to checkpoint is, I am interested only in trace of
ROI.
Thank you Nilay for all the answers you provided so far.
Thanks,
Anusha
On Tue, Jul 17, 2012 at
stats
./blackscholes 4 /parsec/install/inputs/blackscholes/in_16K.txt
/parsec/install/inputs/blackscholes/prices.txt
echo "Done :D"
/sbin/m5 exit
/sbin/m5 exit
Please let me know if this problem was fixed or if I am missing something?
Thanks,
Anusha
___
Hi,
Does anyone know the version of previous gem5-stable (current one is 9073) ?
Thanks,
Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
Thanks Steve. I would use the hash ids in the future
~ Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
ruby_fs.py.
- Does that mean I cannot obtain protocoltrace in ROI with current version (
(f75ee4849c40)
Thanks,
Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
Hi,
What does the --ruby flag in fs.py options signify?
Thanks,
Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
So, it looks like you can create a checkpoint using fs.py and restore using
ruby_fs.py.
To reduce the simulation time of PARSEC
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
taken in timing mode. Am I correct in saying this?
Based on the following thread, I made the above statement.
*http://www.mail-archive.com/gem5-users@gem5.org/msg03973.html*
*
*
*I received a very similar message as the above link.*
*
*
*~ Anusha
Can't unserialize 'system.cpu0:locked'
@ cycle 8115519521000
[paramIn:build/ALPHA_MOESI_CMP_directory/sim/serialize.cc, line 228]
Memory Usage: 739476 KBytes
Thanks,
Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim
I have added the following line in Benchmarks.py
'Blackscholes':[SysConfig('blackscholes_sm.rcS', '512MB')],
Here is blackscholes_sm.rcS script
#!/bin/sh
# File to run the blackscholes benchmark
cd /parsec/install/bin.ckpts
/sbin/m5 dumpstats
/sbin/m5 resetstats
./blackscholes 4 /parsec/in
The checkpoint is part of the application. I did obtain the patch and I
receive the following message. I am trying to root cause the reason.
Switch at curTick count:1
info: Entering event queue @ 8115519521000. Starting simulation...
8115519521000 2Seq Begin >
restore-with-cpu was timing, it was
working. Now, with the new patch, even this combination fails with above
assertion
Thanks,
Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
After you updated the patch, the old combination of cpu-type=timing and
restore-with-cpu =timing works.
But cpu-type=detailed and restore-with-cpu=timing fails with the same
assertion
command line:
/home/anusha/gem5_9123/build/ALPHA_MOESI_CMP_directory/9123moesi.opt
--debug-flags=ProtocolTrace
Has anyone been successful in creating checkpoint in timing mode and
restoring in detailed mode with current version of gem5? (9123-
281b3ac0e0a1)
Thanks,
Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo
Can't checkpoints be created in timing mode too?
~Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
ssage. (
http://www.mail-archive.com/gem5-users@gem5.org/msg04621.html )
7. Previously, I was using version 8642 and did not face these issues
and hence the question.
Thanks,
Anusha
___
gem5-users mailing list
gem5-users@gem5.org
http://m5sim.org/cg
33 matches
Mail list logo
