Return-Path: Received: from smtpoutwbe08.prod.mesa1.secureserver.net ([208.109.78.210] verified) by media-motion.tv (CommuniGate Pro SMTP 4.2.10) with SMTP id 4739814 for AE-List@media-motion.tv; Fri, 08 Jun 2012 19:58:31 +0200 Received: (qmail 30171 invoked from network); 8 Jun 2012 18:01:03 -0000 Received: from unknown (HELO localhost) (72.167.218.144) by smtpoutwbe08.prod.mesa1.secureserver.net with SMTP; 8 Jun 2012 18:01:03 -0000 Received: (qmail 13169 invoked by uid 99); 8 Jun 2012 18:01:03 -0000 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" X-Originating-IP: 75.83.206.115 User-Agent: Workspace Webmail 5.6.18 Message-Id: <20120608110102.89259f98f291d771214680e8971719cb.dbd1bf89c6.wbe@email04.secureserver.net> From: "Robert W. Walker" To: "After Effects Mail List" Subject: RE: [AE] Why a raytraced =?UTF-8?Q?renderer=3F=20-=20Same=20config=2C=20w?= =?UTF-8?Q?ay=20different=20results?= Date: Fri, 08 Jun 2012 11:01:02 -0700 Mime-Version: 1.0 Interesting: our MacPro 3,1 (2.8ghz) running AE 11.01 on OS 10.7.4,=0A32gb= RAM with a Quadro 4000 just rendered the scene in 17:23. That same=0Aconf= iguration on your machine took 45:23 according to the results=0Aspreadsheet= ? =0A=0ARobert W. Walker=0ALos Angeles=0A=0A=0A> -------- Original Messa= ge --------=0A> Subject: Re: [AE] Why a raytraced renderer?=0A> From: rende= rnyc =0A> Date: Fri, June 08, 2012 1:12 pm=0A> To: "Af= ter Effects Mail List" =0A> =0A> =0A> Im not sure = what you mean by explicity testing the GPU.=0A> I didnt build this for benc= hmarking purposes but rather had it laying=0A> around so just started testi= ng some cards=0A> =0A> The comp is extruded shape layers with some lights, = shadows, reflections,=0A> motion blur and DoF being raytraced.=0A> Its not = relying on any outside footage or really doing anything else other=0A> than= rendering the 3d objects=0A> its heavier than just some text, but if thats= all that "you" will be using=0A> the raytracer for than that would be a be= tter benchmark.=0A> =0A> =0A> =0A> =0A> =0A> =0A> =0A> On Fri, Jun 8, 2012 = at 1:00 PM, Teddy Gage wrote:=0A> =0A> > ah, just to = clarify, the robot project took 48 min vs 13 - the AEBENCHCS6=0A> > is clos= er to 5 minutes on a gtx 580; I have to look up the exact times if=0A> > an= yone's interested=0A> >=0A> >=0A> > On Fri, Jun 8, 2012 at 12:58 PM, Teddy = Gage wrote:=0A> >=0A> >> Danny, I'm not convinced you= r scene is explicitly testing the GPU, have=0A> >> you tried the benchmark = file I posted a while ago? I'd be curious to see=0A> >> your results on it.= I actually did a migration test and I got the exact=0A> >> same render tim= es in two different machines with the same GPU. Not so with=0A> >> this rob= ot project...=0A> >>=0A> >> http://www.teddygage.com/AEBENCHCS6/=0A> >>=0A>= >> And Greg (Balint), using the CPU engine for raytrace, the above benchma= rk=0A> >> project took 48 minutes vs. 13 minutes with GPU on, on a highly o= verclocked=0A> >> core i7 3930K @ 4.7 ghz=0A> >>=0A> >> One thing is clear,= the CS6 engine is optimized solely for GPUs. This is=0A> >> significantly = slower than any CPU renderer I use like mental ray or Vray,=0A> >> for a sc= ene an order of magnitude less complicated. However with GPU on it=0A> >> i= s certainly comparable / possibly faster for some situations.=0A> >>=0A> >>= =0A> >>=0A> >> On Fri, Jun 8, 2012 at 12:35 PM, Greg Balint wrote:=0A> >>=0A> >>> Wow. Alright, thanks!=0A> >>>=0A> >>>=0A> >>> //= //Greg Balint=0A> >>> ///Art Director / Motion Graphics Designer=0A> >>> de= lRAZOR.com/=0A> >>>=0A> >>> On Jun 8, 2012, at 12:30 PM, rendernyc wrote:=0A> >>>=0A> >>> did a little test with lloyd the other= day that had some extuded text=0A> >>> with a reflective floor. 2008 macpr= o CPU render was abotu 15mins. he turned=0A> >>> on his 8800GT that is not = supported and it rendered in about 7 mins. on a=0A> >>> q4000 the same scen= e took 37 seconds=0A> >>>=0A> >>>=0A> >>>=0A> >>> On Fri, Jun 8, 2012 at 12= :26 PM, rendernyc wrote:=0A> >>>=0A> >>>> CPU based r= ender?=0A> >>>>=0A> >>>>=0A> >>>> On Fri, Jun 8, 2012 at 12:24 PM, Greg Bal= int wrote:=0A> >>>>=0A> >>>>> Danny, could you do a tes= t for me on the same project file with the=0A> >>>>> 580 turned off? Just t= o see how much the Nvidia cards are helping the=0A> >>>>> situation?=0A> >>= >>>=0A> >>>>> ////Greg Balint=0A> >>>>> ///Art Director / Motion Graphics D= esigner=0A> >>>>> delRAZOR.com/=0A> >>>>>=0A> >>>>> On Jun 8, 2012, at 12:1= 6 PM, rendernyc wrote:=0A> >>>>>=0A> >>>>> if ANYONE = has one of the cards tested already or a different one and=0A> >>>>> wouldn= t mind popping off a render please email me so we can get more results=0A> = >>>>>=0A> >>>>> danny=0A> >>>>>=0A> >>>>> On Fri, Jun 8, 2012 at 12:06 PM, = Andrew Embury wrote:=0A> >>>>>=0A> >>>>>> Fantastic!=0A>= >>>>>>=0A> >>>>>> Thank you ever so much for doing this. This is truly phe= nomenal of=0A> >>>>>> the user group to come together to make this all poss= ible.=0A> >>>>>>=0A> >>>>>> Cheers.=0A> >>>>>>=0A> >>>>>> - Andrew=0A> >>>>= >>=0A> >>>>>>=0A> >>>>>> On Fri, Jun 8, 2012 at 11:59 AM, rendernyc wrote:=0A> >>>>>>=0A> >>>>>>> The 480 uses more power than th= e 5XX=0A> >>>>>>> im expecting the 570 to be about the same speed as the 48= 0 while=0A> >>>>>>> using less power=0A> >>>>>>>=0A> >>>>>>> waiting for so= me tests on that one=0A> >>>>>>>=0A> >>>>>>>=0A> >>>>>>> On Fri, Jun 8, 201= 2 at 11:57 AM, rendernyc wrote:=0A> >>>>>>>=0A> >>>>>>= >> we had the 580 at 15:07=0A> >>>>>>>> so 480 was about 8% slower=0A> >>>>= >>>>=0A> >>>>>>>>=0A> >>>>>>>>=0A> >>>>>>>> On Fri, Jun 8, 2012 at 11:39 AM= , Andrew Embury wrote:=0A> >>>>>>>>=0A> >>>>>>>>> For th= e performance to cost ratio....WOW! That's phenomenal!=0A> >>>>>>>>>=0A> >>= >>>>>>> Now, the 580 is cutting that time by 3 minutes is that correct?=0A>= >>>>>>>>>=0A> >>>>>>>>> Cheers.=0A> >>>>>>>>>=0A> >>>>>>>>> - Andrew=0A> >= >>>>>>>>=0A> >>>>>>>>>=0A> >>>>>>>>> On Fri, Jun 8, 2012 at 11:36 AM, rende= rnyc wrote:=0A> >>>>>>>>>=0A> >>>>>>>>>> 480 in a 12 c= ore macpro was 16:21=0A> >>>>>>>>>>=0A> >>>>>>>>>>=0A> >>>>>>>>>> On Fri, J= un 8, 2012 at 11:31 AM, Andrew Embury >>>>>>>>>> > w= rote:=0A> >>>>>>>>>>=0A> >>>>>>>>>>> What about the 480? Did that not score= between the 580 and 680?=0A> >>>>>>>>>>>=0A> >>>>>>>>>>> I'd be very inter= ested to see that.=0A> >>>>>>>>>>>=0A> >>>>>>>>>>> Cheers.=0A> >>>>>>>>>>>= =0A> >>>>>>>>>>> - Andrew=0A> >>>>>>>>>>>=0A> >>>>>>>>>>>=0A> >>>>>>>>>>>= =0A> >>>>>>>>>>> On Fri, Jun 8, 2012 at 11:28 AM, rendernyc wrote:=0A> >>>>>>>>>>>=0A> >>>>>>>>>>>> interesting=0A> >>>>>>>>>>>> = thats why we need more results...=0A> >>>>>>>>>>>>=0A> >>>>>>>>>>>> the 680= tested in a z800 rendered in 17:34=0A> >>>>>>>>>>>> a 580 in an 2008 mac p= ro was 15:07=0A> >>>>>>>>>>>>=0A> >>>>>>>>>>>>=0A> >>>>>>>>>>>> On Fri, Jun= 8, 2012 at 11:20 AM, Teddy Gage <=0A> >>>>>>>>>>>> teddygage@gmail.com> wr= ote:=0A> >>>>>>>>>>>>=0A> >>>>>>>>>>>>> GTX 680 - 2 GB VRAM - clock speed 1= 096 MHZ=0A> >>>>>>>>>>>>>=0A> >>>>>>>>>>>>> core i7 3930K (6 logical cores)= overclocked to 4.7 gHZ=0A> >>>>>>>>>>>>>=0A> >>>>>>>>>>>>> render time : 1= 3 min 5 seconds =3D 780 seconds=0A> >>>>>>>>>>>>>=0A> >>>>>>>>>>>>>=0A> >>>= >>>>>>>>>> GTX 580 - 3 GB VRAM - clock speed 866 MHZ=0A> >>>>>>>>>>>>>=0A> = >>>>>>>>>>>>> core i7 940 (4 cores) oc'd to 3.7 gHZ=0A> >>>>>>>>>>>>>=0A> >= >>>>>>>>>>>> render time 13 min 50 seconds =3D 830 seconds=0A> >>>>>>>>>>>>= >=0A> >>>>>>>>>>>>>=0A> >>>>>>>>>>>>> So these results seem to imply there = is a disk and / or CPU=0A> >>>>>>>>>>>>> bottleneck in your benchmark proje= ct, or there is something else going on.=0A> >>>>>>>>>>>>> Maybe RAM plays = a factor. Because in my own benchmark tests with a=0A> >>>>>>>>>>>>> differ= ent project, I get different results, where the GTX 580 is faster,=0A> >>>>= >>>>>>>>> not the 680 in the faster machine. Interesting...=0A> >>>>>>>>>>>= >>=0A> >>>>>>>>>>>>>=0A> >>>>>>>>>>>>> On Fri, Jun 8, 2012 at 11:07 AM, ren= dernyc <=0A> >>>>>>>>>>>>> rendernyc@gmail.com> wrote:=0A> >>>>>>>>>>>>>=0A= > >>>>>>>>>>>>>> email sent=0A> >>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>> thanks=0A= > >>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>> On Fri, Jun 8, 2012 = at 10:26 AM, Dave Bittner <=0A> >>>>>>>>>>>>>> dave@pixelworkshop.com> wrot= e:=0A> >>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>>> I'm running a flashed 470 from eB= ay. I'll run the test and=0A> >>>>>>>>>>>>>>> send you the results.=0A> >>>= >>>>>>>>>>>>=0A> >>>>>>>>>>>>>>> On Jun 7, 2012, at 5:41PM, rendernyc wrote= :=0A> >>>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>>> > yup, been testing lots of diff = cards with the raytracer.=0A> >>>>>>>>>>>>>>> > want to run a test on your = 570 for the chart?=0A> >>>>>>>>>>>>>>> >=0A> >>>>>>>>>>>>>>> > have the 285= , 480, 580, 680, q4000 and q4800 so far=0A> >>>>>>>>>>>>>>> >=0A> >>>>>>>>>= >>>>>> > was just saying that the only card someone from adobe=0A> >>>>>>>>= >>>>>>> could "officially" talk about is the currently shipping q4000=0A> >= >>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>>> +---End of message--= -+=0A> >>>>>>>>>>>>>>> To unsubscribe send any message to <=0A> >>>>>>>>>>>= >>>> ae-list-off@media-motion.tv>=0A> >>>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>>=0A= > >>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>> --=0A> >>>>>>>>>>>>>= > danny princz=0A> >>>>>>>>>>>>>>=0A> >>>>>>>>>>>>>> exposedideas.com=0A> >= >>>>>>>>>>>>>=0A> >>>>>>>>>>>>>=0A> >>>>>>>>>>>>>=0A> >>>>>>>>>>>>>=0A> >>>= >>>>>>>>>> --=0A> >>>>>>>>>>>>> Animator & Editor=0A> >>>>>>>>>>>>> www.ted= dygage.com=0A> >>>>>>>>>>>>> Brooklyn=0A> >>>>>>>>>>>>>=0A> >>>>>>>>>>>>>= =0A> >>>>>>>>>>>>=0A> >>>>>>>>>>>>=0A> >>>>>>>>>>>> --=0A> >>>>>>>>>>>> dan= ny princz=0A> >>>>>>>>>>>>=0A> >>>>>>>>>>>> exposedideas.com=0A> >>>>>>>>>>= >>=0A> >>>>>>>>>>>=0A> >>>>>>>>>>>=0A> >>>>>>>>>>=0A> >>>>>>>>>>=0A> >>>>>>= >>>> --=0A> >>>>>>>>>> danny princz=0A> >>>>>>>>>>=0A> >>>>>>>>>> exposedid= eas.com=0A> >>>>>>>>>>=0A> >>>>>>>>>=0A> >>>>>>>>>=0A> >>>>>>>>=0A> >>>>>>>= >=0A> >>>>>>>> --=0A> >>>>>>>> danny princz=0A> >>>>>>>>=0A> >>>>>>>> expos= edideas.com=0A> >>>>>>>>=0A> >>>>>>>=0A> >>>>>>>=0A> >>>>>>>=0A> >>>>>>> --= =0A> >>>>>>> danny princz=0A> >>>>>>>=0A> >>>>>>> exposedideas.com=0A> >>>>= >>>=0A> >>>>>>=0A> >>>>>>=0A> >>>>>=0A> >>>>>=0A> >>>>> --=0A> >>>>> danny = princz=0A> >>>>>=0A> >>>>> exposedideas.com=0A> >>>>>=0A> >>>>>=0A> >>>>=0A= > >>>>=0A> >>>> --=0A> >>>> danny princz=0A> >>>>=0A> >>>> exposedideas.com= =0A> >>>>=0A> >>>=0A> >>>=0A> >>>=0A> >>> --=0A> >>> danny princz=0A> >>>= =0A> >>> exposedideas.com=0A> >>>=0A> >>>=0A> >>=0A> >>=0A> >> --=0A> >> An= imator & Editor=0A> >> www.teddygage.com=0A> >> Brooklyn=0A> >>=0A> >>=0A> = >=0A> >=0A> > --=0A> > Animator & Editor=0A> > www.teddygage.com=0A> > Broo= klyn=0A> >=0A> >=0A> =0A> =0A> -- =0A> danny princz=0A> =0A> exposedideas.c= om=0A