[Fwd: HTT/SMP servers instability on 5.3-STABLE]

[Colin J. Raven]

On Dec 10, Lucas Holt launched this into the bitstream:

>
>>> 
>> 
>> Well thanks for sharing those observations. Until I read what you said I 
>> *assumed* that there was a performance difference with HTT enabled. The 
>> specs seem to show that there *is* a theoretical difference, yet clearly 
>> according to your observations there just isn't any difference of earth 
>> shaking proportions.
>> 
>> 
>
>
> I've got a dual xeon 2.0 ghz workstation.  I've tried running with and 
> without HTT enabled on it.  My observations were that it seemed to help 
> slightly with certain tasks but in general degraded system performance.  For 
> example, if i build a world to upgrade the OS it takes at least 30 seconds 
> longer with HTT enabled.  I tried it multiple times to verify it.  I even 
> tried it with different values for the j flag.  (3 4 5 and 6)
>
> Similarly, if i run it in windows with HTT enabled I noticed that windows was 
> peppy but it was a disaster during gaming.  Of course that makes sense as 
> most of my games are not designed for multiple cpus.  (enemy territory, quake 
> 3, Doom 3,  etc)  I also realize i could play some of them in fbsd, except my 
> radeon 9600 xt isn't supported for 3d acceleration!
>
>
>
This is *most* surprising but extremely useful information, many thanks 
for taking the time to impart first hand experiences.