Re: [Qemu-devel] [PATCH RFC] virtio: put last seen used index into ring itself

["Michael S. Tsirkin" <mst@redhat.com>]

On Sun, May 23, 2010 at 07:03:10PM +0300, Avi Kivity wrote:
> On 05/23/2010 06:51 PM, Michael S. Tsirkin wrote:
>>>
>>>> So locked version seems to be faster than unlocked,
>>>> and share/unshare not to matter?
>>>>
>>>>        
>>> May be due to the processor using the LOCK operation as a hint to
>>> reserve the cacheline for a bit.
>>>      
>> Maybe we should use atomics on index then?
>>    
>
> This should only be helpful if you access the cacheline several times in  
> a row.  That's not the case in virtio (or here).

So why does it help?

> I think the problem is that LOCKSHARE and SHARE are not symmetric, so  
> they can't be directly compared.

In what sense are they not symmetric?

>> OK, after adding mb in code patch will be sent separately,
>> the test works for my workstation. locked is still fastest,
>> unshared sometimes shows wins and sometimes loses over shared.
>>
>> [root@qus19 ~]# ./cachebounce share 0 1
>> CPU 0: share cacheline: 6638521 usec
>> CPU 1: share cacheline: 6638478 usec
>>    
>
> 66 ns? nice.
>
>> [root@qus19 ~]# ./cachebounce share 0 2
>> CPU 0: share cacheline: 14529198 usec
>> CPU 2: share cacheline: 14529156 usec
>>    
>
> 140 ns, not too bad.  I hope I'm not misinterpreting the results.
>
> -- 
> error compiling committee.c: too many arguments to function