I believe the old PTLStats are gone. The new YAML-based stats can be turned
on by adding

-yamlstats ptlstats.yaml

to your simconfig parameters. These stats are in a human readable file
format or you can use the mstats.py script from the utils/ folder to do
some extra stuff like filtering.



On Mon, Nov 14, 2011 at 1:59 PM, Addison Mayberry <[email protected]>wrote:

>  Thank you very much for your help. I am now using the latest release as
> you recommended and I am able to run the code succesfully. However, I can't
> figure out how to read PTLsim's statistics files in 0.2 as I can't find
> PTLstats anywhere in the directory. I know that it's mentioned in a couple
> places on the website that you were beginning to transition to a new
> statistics system, has PTLstats been completely removed in favor of that?
> And if so, can you point me to some information (or even names of source
> files) for the new system?
> -Addison
>
>
> On 11/09/2011 11:46 AM, avadh patel wrote:
>
> This bug is fixed in last release.  I strongly recommend to use the 0.2
> release.  After that release we have some minor bug fixes that you can grab
> via github's master branch.  If you'r new to git, please read Chapter 3 of
> Git Community Book (http://book.git-scm.com/) which explains the basic
> branching and merging.  As we and other users discover bugs we push the
> fixes on the master branch on github.  My workflow with Marss is like for
> every project I create a new branch and all the bug fixes are done onto
> 'master' branch. Then I do a 'git merge master' to patch my project code.
>  With this model you dont have to publish anything related to your project
> and still you can use git for your code management.
>
>  - Avadh
>
> On Wed, Nov 9, 2011 at 9:26 AM, Addison Mayberry <[email protected]>wrote:
>
>> Hello, I am a new MARSS user picking up some earlier work involving MARSS
>> done by my group. First off, I'd like to say thank you for building such a
>> great system, it's been extremely helpful for our work thus far and I'm
>> looking forward to getting more involved with it in the future. We are
>> using MARSS 0.12, which I know is outdated at this point.
>>
>> My problem is that any time I drop into PTLSim, either using PTLcalls or
>> simconfig -run from the terminal, MARSS prints out some start information
>> and then quickly segfaults. I have tried this on two different images, one
>> derived from the parsec benchmark image on the MARSS site (converted to
>> QCOW2 to allow for checkpointing) and one that I made myself using qemu-img
>> with the latest version of Ubuntu. I have also tried both images on two
>> separate machines running different distributions, all had the exact same
>> behavior.
>>
>> The error message produced by running in gdb is:
>>
>> Program received signal SIGSEGV, Segmentation fault.
>> 0x0000000000590778 in x86_sse_stvwu (ra=<optimized out>, m=0x2108638) at
>> ptlsim/lib/logic.h:22
>> 22    inline void x86_sse_stvwu(vec8w* m, const vec8w ra) { asm("movdqu
>> %[ra],%[m]" : [m] "=xm" (*m) : [ra] "x" (ra) : "memory"); }
>>
>> And I can provide a backtrace if that would be helpful. As I said, I'm
>> using 0.12, but because the problem is so persistent between images and
>> systems I have the feeling it's something very simple that I'm doing wrong.
>> Any support would be greatly appreciated.
>> -Addison
>>
>> _______________________________________________
>> http://www.marss86.org
>> Marss86-Devel mailing list
>> [email protected]
>> https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
>>
>
>
>
>
> _______________________________________________
> http://www.marss86.org
> Marss86-Devel mailing list
> [email protected]
> https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel
>
>
_______________________________________________
http://www.marss86.org
Marss86-Devel mailing list
[email protected]
https://www.cs.binghamton.edu/mailman/listinfo/marss86-devel

Reply via email to