Hi,
Has anyone tested vray on say two 12 core Opterons? I’m buying a new system so I really want to know if vray will use it well. Now I have two Opterons 275 (2 cores on 2.2ghz). I’m wondering if those 24 cores will be used well enough.
thanks for the info
hmm my experience with vray 1.5 is big disapointment with 16 cores AMD system (4*4)so i will wait to vray 2.0 to show how can manage mucho cores 
btw: why chaos cant compile vray for AMD cores too ? (some optimisations for amd cores can push things towards) …
Really? Can you give me some numbers? Like if you had 2 cores, instead of 8 times faster on 16 cores its…? Scalability is supposed to be vrays thing
ok i can do some test but which scene i can use ?? probably test scene right here on forum ??
From what I remember some of the initial multi-core and multi-processor AMD systems had severe issues with memory access and showed very bad performance. We could not attribute this to any particular portion of V-Ray, it seemed a general problem. This was a while ago, so things may have changed since then.
Intel machines on the other hand, seem to have less issues in that regard - there are lots of people running 24 cores now and that seems to work well.
Best regards,
Vlado
vlado : can i ask to compile vray with amd compiler ? only for me
btw : with my 16 core machine (4*4) i cant gain time cuts with more than 8 cores so i have to render 2 instances of max with different cameras to use my 16 core machine optimally …
AMD do not have their own compiler.
Best regards,
Vlado
So, anyone here using a bit newer Opterons that can give us some numbers?
Mine are working pretty well, even though they’re old, but I DID install some patch few months after I bought it, can’t recall now what was it about.
Maybe Chaos group could give us some tests They must have one Xeon and one Opteron system for testing 
2vlado : i know is off topic but what about this one : http://developer.amd.com/cpu/open64/Pages/default.aspx
2Crayox13 : so i will try test scene and we will see …
It is only for Linux.
Best regards,
Vlado
This is really concerning since I’m planing on going for 12 or 24 cores, depending if I’ll go for Xeon or Opteron system. What exactly was it tested on? MB, CPUs…etc.
For that benchmark (as well as other scenes) have you changed the “Number of passes” in the light cache settings to “24”? Otherwise vray will not use all of the cores during the lightcache calc..
Running on 120 threads (mixed core2duo and core i7) without problems. Only network bandwith becomes a bottleneck sometimes.
This obviosly is a pretty poor performance for AMD cpus…
@ Janosik: For your test do you mean 16 cpus or 16 cores? How many actual machines are you using? The DR performance will not be linear as 16 cores will not be twice as fast as 8 cores.
Of course it won’t be linear, but at least 70%?
Keep in mind that render times of < 1 min will not give very meaningful results; it would be best to test on scenes that take at least 5 to 10 minutes pure render time.
Best regards,
Vlado
ok Vlado ,so i can choose something from my production … or Crayox you can send to me your test scene with expected render time (5-10 min )
cpus i mean cores ofcourse
and yes i changed number of LC passes with numbers of cores i allow to compute vray test scene …
and my rig : tyan s4985-E , 4x barcelona B3 2,2 ghz , 12 gb of ram mixed 1 and 2 gb modules but still two of them for each socket …
Hehe we’ve been moved…maybe we’ve tapped on something…kidding
We could also use some scene I did, but I think this says it all. We’ll see what Chaos think.
