[openib-general] Re: performance counters in /sys

[Sean Hefty]

Mark Seger wrote:
> As I said in my base note I'm currently reading from /proc while some 
> sets of counters are better organized than others, I can still access 
> them relatively efficiently.  While I could certainly "get by" reading 
> one variable per file, I do worry about  the overhead as the sampling 
> frequency goes down.  This will also be a problem as the number of 
> counters and devices grow.  The suggestion about using perfquery would 
> certainly work, but I'd also be concerned about the overhead in running 
> it at smaller sampling intervals.

I could be mistaken here, but I believe that these are hardware based 
counters that are accessed by generating a Performance Management MAD to 
retrieve that desired value(s).  (It looks like one MAD is generated for 
each individual counter that is read when accessed through /proc.)  I'm not 
sure that you can get much more efficient that using something similar to 
perfquery, which retrieves all values using a single MAD, and uses the 
standard IB method of reading the counters.  Additionally, you should be 
able to access these counters from across the fabric.

- Sean