Return-Path: Received: from p3plsmtpa01-07.prod.phx3.secureserver.net ([72.167.82.87] verified) by media-motion.tv (CommuniGate Pro SMTP 4.2.10) with SMTP id 4745071 for AE-List@media-motion.tv; Wed, 13 Jun 2012 04:09:42 +0200 Received: (qmail 19974 invoked from network); 13 Jun 2012 02:12:22 -0000 Received: from unknown (173.247.4.230) by p3plsmtpa01-07.prod.phx3.secureserver.net (72.167.82.87) with ESMTP; 13 Jun 2012 02:12:21 -0000 Subject: Re: [AE] (OT) An example of a PC version of a Mac Pro with today's tech for AE References: From: Greg Balint Content-Type: text/plain; charset=us-ascii X-Mailer: iPhone Mail (9A405) In-Reply-To: Message-Id: Date: Tue, 12 Jun 2012 22:12:49 -0400 To: After Effects Mail List Content-Transfer-Encoding: quoted-printable Mime-Version: 1.0 (1.0) Stephen, I suspect things like drive speed writing and reading, codec speed to write e= ach frame, and other bottlenecks can change the speed and efficiency of HT.=20= If there's footage being used, HT could see lesser productivity vs non HT if= the bottleneck is how fast AE can bring in each frame anyway. You end up wa= iting on those frames to be brought in before the rest of the rendering can t= ake place. 200 cores or threads wouldn't make as much difference if AE stil= l needs to pull each frame and it takes a while to do.=20 I seem like I'm in the know. I'm pretty sure of the things I'm talking about= , but I could be proven wrong. I'm not a complete expert on how HTT and AE p= lay together.=20 ////Greg Balint ///Art Director / Motion Graphics Designer delRAZOR.com/ On Jun 12, 2012, at 9:24 PM, "Stephen van Vuuren" w= rote: > I do know AE does not always handle lots of threads well - but is Adobe > saying it's HTT to blame or overall MP management. AE should not care - th= e > OS and CPU should try to handle it. >=20 > I'm not a programmer but I know MP can't work with Raytracer, many > time-based plugins. Is this related to HTT performance - AE is making call= s > somehow to the physical cores that can't be processed? Or asking for CPU > resources that cannot be split across the virtual cores (which supposedly i= t > cannot identify)? >=20 > Any suggestions on a test to properly figure this out? I see performance > improvement on renders with HTT turned on - but not every time and with > every project. And like Greg B. - the information out about HTT on AE is n= ot > in agreement with what Chris M. and Steve F. are posting now. >=20 > stephen van vuuren > 336.202.4777 >=20 > http://www.sv2dcp.com/ > http://www.sv2studios.com/ > http://www.outsideinthemovie.com/ >=20 > A film is - or should be - more like music than like fiction. It should be= a > progression of moods and feelings. The theme, what's behind the emotion, t= he > meaning, all that comes later. > -Stanley Kubrick >=20 > -----Original Message----- > From: After Effects Mail List [mailto:AE-List@media-motion.tv] On Behalf O= f > Greg Balint > Sent: Tuesday, June 12, 2012 8:08 PM > To: After Effects Mail List > Subject: Re: [AE] (OT) An example of a PC version of a Mac Pro with today'= s > tech for AE >=20 > This is just confusing the heck out of me, because I'm pretty sure I > understood it all until today.=20 >=20 > I most certainly see a big difference rendering in AE with MP turned on an= d > Hyperthreading enabled. AE has no clue the difference between a 4 core > system with HT on or an 8 core with no HT.=20 >=20 > As long as I have the ram to power the "virtual" cores, AE doesn't care if= > they are virtual cores or not.=20 >=20 > They are two different things, but AE's Multiprocessing most certainly tak= es > advantage of all threads available to it.=20 >=20 > ////Greg Balint > ///Art Director / Motion Graphics Designer delRAZOR.com/ >=20 > On Jun 12, 2012, at 7:30 PM, Chris Meyer wrote: >=20 >> Aside from that, you can enable multiprocessing, which launches multiple > copies of After Effects in the background, assigning one copy (process) to= > one core. This particular trick needs physical cores to actually do any > work. >=20 > +---End of message---+ > To unsubscribe send any message to >=20 >=20 >=20 > +---End of message---+ > To unsubscribe send any message to >=20