Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 5 Dec 2019 10:44:36 +0100 (CET)
From:      Wojciech Puchar <wojtek@puchar.net>
To:        Darius Mihai <dariusmihaim@gmail.com>
Cc:        Wojciech Puchar <wojtek@puchar.net>, "Rodney W. Grimes" <freebsd-rwg@gndrsh.dnsmgr.net>, freebsd-hackers@freebsd.org
Subject:   Re: bhyve+windows 7 multicore performance
Message-ID:  <alpine.BSF.2.20.1912051043510.83609@puchar.net>
In-Reply-To: <CAPj=67vMMNoG%2BYEATOD6p=cw83egvUh9SzvJLsw_rsfcprywdA@mail.gmail.com>
References:  <201912031529.xB3FTMnY035614@gndrsh.dnsmgr.net> <alpine.BSF.2.20.1912041720570.5885@puchar.net> <CAPj=67vMMNoG%2BYEATOD6p=cw83egvUh9SzvJLsw_rsfcprywdA@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help


On Wed, 4 Dec 2019, Darius Mihai wrote:

> 
> 
> On Wed, Dec 4, 2019, 18:22 Wojciech Puchar <wojtek@puchar.net> wrote:
>       >>>
>       >>> Try 4 cores, and drop the priority boost, you may be causeing
>       >>
>       >> Well i need more cores than 4 for that VM.
>       >> I will try even number (10) and no nice.
>       >
>       > If 4 cores performs better than 10 cores why would you want
>
>       performs better per thread. Not as total.
> 
>
>       removing -P option improved performance a bit. Well - a large bit.
> 
> 
> If I remember correctly, windows runs PAUSE very often when idling,
> so having many cores that stay idle may slow down the system
> since -P forces context switches when that instruction is
> executed. 
> 
> Darius
Possibly. But for sure - this made things faster a lot. Actually close to 
bare metal performance.



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?alpine.BSF.2.20.1912051043510.83609>