> On Jan. 25, 2013, 4:13 p.m., Nathan Binkert wrote:
> > Seems like overkill to me.  If you do this, then you can't do any math 
> > using SQL and you have to suck out values to do anything.  If that's the 
> > attitude, why even bother using sqlite at all?

You can't do math in sql, but that probably wasn't what you wanted to do 
anyway. You probably want to suck the data back in the python class hierarchy 
and manipulate it there. I think the ideal situation would be to pickle the 
objects and not use sql, however that was much slower. The slowest (and 
largest) was having a sql table of stat,x,y,value columns which meant reading a 
large array took forever. 


- Ali


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://reviews.gem5.org/r/1646/#review3936
-----------------------------------------------------------


On Jan. 15, 2013, 10:36 a.m., Andreas Hansson wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://reviews.gem5.org/r/1646/
> -----------------------------------------------------------
> 
> (Updated Jan. 15, 2013, 10:36 a.m.)
> 
> 
> Review request for Default.
> 
> 
> Description
> -------
> 
> Changeset 9499:bc23f2c316fc
> ---------------------------
> stats: Store vector stats using doubles and compress with zlib
> 
> This patch changes any arrays of values to be stored as an array of doubles,
> rather than floats in the SQL database. This is required as floats lose too 
> much
> accuracy. For example, if the stats are read from the database, and injected
> back into gem5's stats system, then formulas can be recalculated. If floats 
> are
> used, these formulas evaluate to be different from those originally calculated
> when creating the SQL database.
> 
> As doubles take up twice the space of a float (8 Bytes vs 4 Bytes) the SQL
> database becomes larger. The end result is that the database is larger than 
> the
> text based output without compression. Therefore, as the vector storage is
> already not human readable we compress this field using zlib. zlib has been in
> the python standard library since version 1.5.1. so it is already covered in
> the gem5 build prerequisites.
> 
> 
> Diffs
> -----
> 
>   src/python/m5/stats/sql.py PRE-CREATION 
> 
> Diff: http://reviews.gem5.org/r/1646/diff/
> 
> 
> Testing
> -------
> 
> 
> Thanks,
> 
> Andreas Hansson
> 
>

_______________________________________________
gem5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/gem5-dev

Reply via email to