openssl speed weirdness

[Garrett D'Amore]

Darren J Moffat wrote:
>
>
>
> Just as another data point:
> On the Sun Ray server i'm using just now (which is an 8 way 1.8GHz 
> UltraSPARC IV+) and on which the Fair Share Scheduler is being used (and 
> I don't have all shares and there are 43 other users logged in) I get:
>
> $ openssl speed -multi 64 blowfish
> blowfish cbc    295797.16k   214208.98k   285885.70k   486516.24k 
> 465664.16k
> $ openssl speed -multi 256 blowfish
> blowfish cbc    347912.25k   329789.53k   375919.05k   312966.65k 
> 496012.92k
>
>   

Hmm... I'd expect that behavior with TS.  But if I understand FSS 
correctly, I'd guess that you shouldn't get more shares.  Unless, 
perhaps, the *other* consumers have quited down, so that the total 
number of active shares in the second case is smaller than in the first?

Admittedly, I've never tried playing with FSS.

    -- Garrett