Products Product Overview What's New Product Demos Licensing & Pricing Annual Upgrade Plan Product FAQ Companion Products Houdini Store Solutions Go Procedural Visual Effects Lighting and Rendering Game Art Community Gallery Houdini Forum Events Customer Stories Community Links Contests Houdini Jobs Smart 3D Asset Store Mail List Learning Getting Started Tutorials Houdini Training Learning Materials Houdini Education Schools Support Customer Support System Requirements Documentation FAQs Journals Submit Bugs and RFEs Download Free Learning Edition Daily Builds Company About Side Effects Demo Reels Houdini in Film News Houdini in The News Press Careers @ Sidefx From the President Contact Information

FAQ       Search       Memberlist       Usergroups       Register       Profile       Check private messages       Log in      Benchmarks on 6 core Phenom? Goto page 1, 2  Next          Houdini Forum Index -> Technical Discussion View previous topic :: View next topic   Author Message pbowmar Houdini Guru Joined: 03 Jul 2002 Posts: 3999 Location: Vancouver, BC, Canada Posted: Thu Oct 21, 2010 1:44 pm GMT    Post subject: Benchmarks on 6 core Phenom? Hi, Does anyone use a 6 core AMD Phenom 3.2ghz machine? I'm trying to do some benchmarks on real Mantra rendering vs our Intel machines and I would very much appreciate if I can send someone an IFD file (or .hip file) to test render. It would be completely self contained so you should be able to load and click render Cheers, Peter B Back to top twod Developer Joined: 16 Aug 2002 Posts: 2319 Location: Side Effects Posted: Thu Oct 21, 2010 2:54 pm GMT    Post subject: I do not have one, but Tom's Hardware has a fairly decent CPU comparison chart: http://www.tomshardware.com/charts/desktop-cpu-charts-2010/benchmarks,112.html Mantra isn't included, but Cinebench and 3D Studio Max renderers are. Might help if all else fails The AMD hexacore 3.2GHz CPU performs about like a i7 960 in both cases. Back to top Pagefan Houdini Master Joined: 24 Jul 2005 Posts: 519 Posted: Thu Oct 21, 2010 3:32 pm GMT    Post subject: Peter, I am getting one next week and a a friend of mine has one running so if you mail me the IFD then I am happy enough to pass it on and make him do some benchmark tests Back to top Solitude Houdini Adept Joined: 27 Mar 2009 Posts: 206 Location: Vancouver Posted: Thu Oct 21, 2010 3:36 pm GMT    Post subject: Please post up the results when you're done -- I'm considering a new machine soon, would love to see the results. Back to top SYmek Houdini Guru Joined: 13 Mar 2003 Posts: 1241 Posted: Thu Oct 21, 2010 5:00 pm GMT    Post subject: Great idea. Peter, if you could post also ifd file, please. So we could make even more comparision tests. Back to top pbowmar Houdini Guru Joined: 03 Jul 2002 Posts: 3999 Location: Vancouver, BC, Canada Posted: Fri Oct 22, 2010 2:32 pm GMT    Post subject: Cool, thanks guys. Mark, that's a great benchmark site, the issue is that the CPU I have at work isn't on the list (it's an E5440 quad core, x2 CPUs) and I really like "real world" benchmarks since the numbers don't tell me how it will feel However, certainly price/performance is with AMD on this one. Shame I can't find a dual-CPU AM3+ mobo and have 12 cores I'll get that IFD up on the weekend! Cheers, Peter B Back to top twod Developer Joined: 16 Aug 2002 Posts: 2319 Location: Side Effects Posted: Fri Oct 22, 2010 4:14 pm GMT    Post subject: AMD's dual motherboards are strictly socket F or G34, not AM2/3/+ (I believe). A 12-core Mangy-cours CPU clocks out at 2.4GHz, and isn't suitable for workstation work. The i7 930 would be the closest to your E5440, though it only has 8MB of cache rather than 12, and is 60Mhz slower. Most benchmarks show that increasing cache size at that level has only a very small effect on performance. Back to top pbowmar Houdini Guru Joined: 03 Jul 2002 Posts: 3999 Location: Vancouver, BC, Canada Posted: Fri Oct 22, 2010 6:31 pm GMT    Post subject: Ah OK thanks Mark. So price/performance between the AMD and the Intel is pretty much identical... Back to top pbowmar Houdini Guru Joined: 03 Jul 2002 Posts: 3999 Location: Vancouver, BC, Canada Posted: Sun Oct 24, 2010 6:51 pm GMT    Post subject: OK here's the IFD file along with the resultant image (or as I call it, "art") On my dual core Athlon 4400+ it was 44 minutes. This was Suse 11.2, using "gcc 4.3" build. On my dual core laptop with T5550 1.83Ghz Core Duo, it was 60 minutes. This was Vista 32bit. Note that I used 11.0.542 to enable the new option to disable displacement shaders (applied to the ROP). I'm including the .hip file in case people want to take a look. However, please use this IFD file in 11.0.542 for the benchmark to ensure complete apples to apples, if at all possible. If not possible any info is OK too. If you're on Linux or OSX, open a shell with Houdini set up and simply do: mantra < benchmark.ifd Note that it will write a file called "benchmark.pic" in the same directory. On Windows, from the Start Menu, go to the Houdini version and open a Command Line Tools DOS window. Type the drive letter of the drive the .ifd file is on, hit enter, then "cd" to the directory with the IFD. Once there, do: mantra < benchmark.ifd just like on Linux. Same thing, a benchmark.pic will be written to the same directory. I'd love to hear about anyone's findings! I'll try this at work on 4 and 8 cores of our work machines, plus the OSX machines we just bought. On Thursday or Friday when I'm back at work, at least Cheers, Peter B benchmark.jpg  Description: "art"  Filesize:  26.87 KB  Viewed:  8935 Time(s) benchmark.zip  Description: The .hip file Download  Filename:  benchmark.zip  Filesize:  133.16 KB  Downloaded:  1553 Time(s) benchmark.ifd.zip  Description: The IFD file Download  Filename:  benchmark.ifd.zip  Filesize:  126.26 KB  Downloaded:  678 Time(s) Back to top Wolfwood Village Idiot Joined: 27 Feb 2003 Posts: 3601 Location: North of Canada Posted: Sun Oct 24, 2010 10:45 pm GMT    Post subject: Quad Core - Intel(R) Core(TM) i7 CPU 920 @ 2.67GHz, Hyperthreading enabled. OpenSuse 11.3, Houdini 11.0.542 (gcc.4.4 build) 8.5 minutes here. Code: cat benchmark.ifd | mantra ... stuff ... Render Time: 66:10.60u 25.41s 8:28.20r _________________ if(coffees<2,round(float),float) Back to top Alanw Houdini Expert Joined: 06 Aug 2007 Posts: 320 Posted: Wed Oct 27, 2010 1:08 pm GMT    Post subject: Howdy, I gave this a shot on two machines. Both running Houdini 11.0.542. The Linux box uses the gcc 4.4 build. Mac OSX 10.6 Snow Leopard / 2.8 GHz Core i5 4GB RAM (quad core, doesn't have hyperthreading) Code: Render Time: 43:37.86u 7.86s 11:44.66r Memory:  142.71 MB of 10.58 GB arena size Gentoo Linux 2.6.34-gentoo-r12 gcc 4.4.4 Dual Intel(R) Xeon(R) CPU E5520 @ 2.27GHz 24GB RAM with Hyperthreading on (16 threads) Code: Render Time: 78:01.39u 25.34s 5:01.32r Memory:  119.58 MB of 120.52 MB arena size. VM Size: 1.13 GB _________________ www.alan-warren.com Back to top mark Developer Joined: 03 Jul 2002 Posts: 1503 Location: Toronto, Ontario Posted: Wed Oct 27, 2010 8:56 pm GMT    Post subject: Code: Ubuntu 8.04 x86_64:     Intel(R) Core(TM) i7 CPU 920  @ 2.67GHz (HT enabled)     Render Time: 65:36.89u 29.61s 8:26.58r     Memory:  73.80 MB of 74.33 MB arena size. VM Size: 565.63 MB Ubuntu 8.04 x86_64:     Intel(R) Core(TM) i5 CPU 750  @ 2.67GHz     Render Time: 45:04.50u 15.40s 11:38.54r     Memory:  55.27 MB of 56.71 MB arena size. VM Size: 276.78 MB     Ubuntu 8.04 i686:     Intel(R) Core(TM) i3 CPU 540  @ 3.07GH (HT enabled)     Render Time: 3.62u 69:43.21s 17:49.99r     Memory:  45.16 MB of 46.54 MB arena size. VM Size: 81.13 MB Maybe we should post this on a wiki somewhere? P.S. Those times on the i3 look suspicious -- it looks like the user and system times might be swapped. Back to top splitpoint Socialite Joined: 22 May 2007 Posts: 57 Location: Colorado Posted: Thu Oct 28, 2010 2:12 pm GMT    Post subject: Windows 7 64bit on a dual Opteron 2376 2.3ghz running 11.0.469: Render Time: 9:16.814u 1.107s 9:22.64r Memory: 306.93 MB. VM Size: 401.54 MB Back to top JColdrick Houdini Guru Joined: 23 Jul 2002 Posts: 4143 Location: Toronto Posted: Thu Oct 28, 2010 3:19 pm GMT    Post subject: Apologies, but I don't have access to that exact build, but I rendered the same results with both 11.0.527 and 11.0.547: a little over 5.5 minutes. Code: Render Time: 44:36.43u 17.67s 5:42.73r I'm running the same software as Peter, SUSE 11.2, the gcc 4.3 build, and the hardware is: Intel(R) Core(TM) i7 CPU 975 @ 3.33GHz with hyperthreading turned on. Cheers, J.C. Back to top pbowmar Houdini Guru Joined: 03 Jul 2002 Posts: 3999 Location: Vancouver, BC, Canada Posted: Thu Oct 28, 2010 3:32 pm GMT    Post subject: Hey John, How many cores on that? My work benchmarks are (gcc 4.1, Centos 5.2, Houdini 11.0.538) Intel(R) Xeon(R) CPU E5440 @ 2.83GHz 8 cores no HT Code: Render Time: 51:34.91u 17.73s 6:31.98r 4 cores no HT Code: 50:59.78u 16.21s 12:50.20r Cheers, Peter B Back to top JColdrick Houdini Guru Joined: 23 Jul 2002 Posts: 4143 Location: Toronto Posted: Thu Oct 28, 2010 3:47 pm GMT    Post subject: "Eight" cores(quotes courtesy Hyperthreading Inc ). That seems somewhat comparable with your results, yeah? My clock speed is a little faster... Cheers, J.C. P.S. the full blurb on the cpu is "Intel i7 Quad Core Enhanced Performance V2" P.P.S. Well, no it doesn't seem comparable. You're running two cpus, I'm running one with HT. According to the scuttlebutt I keep reading here, I'm supposed to be penalized more heavily, yes? This is why I've been mystified by all the negative talk about HT. Back to top jason_iversen Houdini Guru Joined: 02 Sep 2002 Posts: 5284 Location: r+h, los angeles Posted: Fri Oct 29, 2010 2:37 am GMT    Post subject: mark wrote: P.S. Those times on the i3 look suspicious -- it looks like the user and system times might be swapped. I never know which is which, really. I just look for the biggest number _________________ jason iversen, fx supervisor @ r+h, http://www.odforce.net https://plus.google.com/103473736257525043693 Back to top pbowmar Houdini Guru Joined: 03 Jul 2002 Posts: 3999 Location: Vancouver, BC, Canada Posted: Fri Oct 29, 2010 11:38 am GMT    Post subject: My work benchmarks now include on OSX 10.6 8core 2.4ghz Intel (not sure exactly the CPU) Code: 5:40 which is very fast IMO Sadly, we had a 12 core (dual 6core) test machine in that has since been returned. My tests at the time showed an almost 50% speed increase over our normal (spec'd above) 8 core machines. Anybody able to test the 6 core Phenom 3.2ghz? This is all great, but none of these faster ones can I afford Cheers, Peter B Back to top Korhon Socialite Joined: 16 Jul 2007 Posts: 98 Location: Oslo Posted: Wed Nov 03, 2010 10:57 am GMT    Post subject: pbowmar`s benchmark scene. Render Time: 2:44.331u 0.873s 2:51.27r Memory: 559.67 MB of 590.99 MB arena size. VM Size: 682.80 MB Wohoo! Gotta love cpu!! Dual Intel Xeon X5680 3.33ghz (12cores) 24gb ram _________________ www.gimpville.no Back to top pbowmar Houdini Guru Joined: 03 Jul 2002 Posts: 3999 Location: Vancouver, BC, Canada Posted: Wed Nov 03, 2010 11:36 am GMT    Post subject: If only I could even come close to affording that Back to top twod Developer Joined: 16 Aug 2002 Posts: 2319 Location: Side Effects Posted: Thu Nov 04, 2010 3:10 pm GMT    Post subject: I ran the benchmark on my Early 2008 Mac pro with dual 2.8GHz Xeons and 8GB FB-DIMM memory, but using Ubuntu 9.10. (8 cores, 8 threads, no turbo boost - based on the core2 series) Code: 8 threads Render Time: 50:45.00u 21.94s 6:26.09r      51:07t Memory:  76.00 MB of 76.61 MB arena size. VM Size: 567.92 MB 7 threads Render Time: 50:37.75u 17.20s 7:18.25r      50:55t Memory:  67.20 MB of 67.96 MB arena size. VM Size: 491.50 MB 6 threads Render Time: 50:23.38u 20.45s 8:28.46r      50:45t Memory:  65.17 MB of 66.11 MB arena size. VM Size: 420.94 MB 5 threads Render Time: 50:15.28u 20.73s 10:08.71r      50:36t Memory:  60.92 MB of 61.88 MB arena size. VM Size: 349.36 MB 4 threads Render Time: 50:07.26u 18.94s 12:37.92r      50:26t Memory:  56.02 MB of 56.99 MB arena size. VM Size: 275.89 MB 3 threads Render Time: 49:47.93u 16.97s 16:42.40r      50:04t Memory:  49.98 MB of 50.79 MB arena size. VM Size: 266.77 MB 2 threads Render Time: 49:32.92u 10.19s 24:52.12r      49:43t Memory:  47.50 MB of 48.89 MB arena size. VM Size: 196.95 MB 1 thread Render Time: 48:35.57u 13.49s 48:48.81r      48:48t Memory:  50.40 MB of 52.38 MB arena size. VM Size: 68.52 MB The result is near-perfect linear scaling with real cores. I've attached some graphs of this. (The new ##:##t column is the sum of the u & s numbers, total CPU time). - 'render_time' is the time in seconds that the various threads took to render. The lower the better. - 'scaling' is the speedup over a single thread (ideally 8x at 8 threads), with the ideal curve plotted. The closer the actual curve is to the ideal curve, the better. - 'contention' should really be named the threading overhead. It's the amount of time wasted with thread setup, cleanup, synchronization, waiting, etc. It is expressed in a percentage of the single-threaded time (ie, 0-4.5%). The lower the better. - 'memory' is the amount of virtual memory used, as reported by the memory command, expressed in a multiple of the single-threaded usage (so '4' means it took 4x the memory). The lower the better. Because overhead for this case is very low and scaling near-ideal, SMT (hyperthreaded) cores will improve performance. An SMT core will improve performance as long as the overall thread overhead doesn't exceed the SMT speedup (which is generally 0-40% over a single-issue core). The Core i series hyperthreading is actually quite good; it has enough memory bandwidth to feed the CPU, enough cache to keep thrashing down, and enough resources within the cores themselves to keep the CPU busy -- unlike the old Pentium hyperthreading which had issues on all those fronts. Oh, and Peter -- if price is an issue, you just got lucky. times.png  Description: Render time in seconds, 1-8 threads  Filesize:  9.4 KB  Viewed:  8436 Time(s) scaling.png  Description: Multithreaded Scaling, 1-8 threads  Filesize:  11.29 KB  Viewed:  8436 Time(s) contention.png  Description: Thread Overhead, in % of total single-threaded time  Filesize:  7.1 KB  Viewed:  8436 Time(s) memory.png  Description: Memory Usage, in multiples of single-threaded use  Filesize:  7.44 KB  Viewed:  8436 Time(s) Back to top pbowmar Houdini Guru Joined: 03 Jul 2002 Posts: 3999 Location: Vancouver, BC, Canada Posted: Mon Dec 13, 2010 2:09 am GMT    Post subject: So, I bought a 1090T (6 core 3.2ghz AMD Phenom II) and after a week of "which component is dead?" hell (it was one stick of RAM) I have some interesting results... Benchmark.ifd from this thread, using 11.0.581 x86_64 gcc4.4 on Suse 11.3: 7:44 Identical machine, dual-booted to Windows 7 64bit, 11.0.581 64bit build: 9:15 I am quite surprised, I had heard Windows 7 had improved its effeciency, but apparently not... First attempt to overclock failed (Mantra crashed) so I will likely have to either a) fiddle or b) give up and live with a stock CPU. I suspect b) is more likely Cheers, Peter B Back to top splitpoint Socialite Joined: 22 May 2007 Posts: 57 Location: Colorado Posted: Mon Dec 13, 2010 2:28 pm GMT    Post subject: Thanks for the followup Peter. I was thinking of picking up a couple of those 1090t's for rendering. It's nice to know that the performance is pretty much identical to my Opteron machine. Al Back to top goldleaf Houdini Adept Joined: 18 Sep 2007 Posts: 179 Posted: Fri Dec 17, 2010 3:17 am GMT    Post subject: I just finished building my own Phenom II X6 system, with a 2.8Ghz 1055T and 8GB Ram. Here are my results: Code: Render Time: 53:44.60u 32.26s 9:24.50r Memory:  62.64 MB of 63.21 MB arena size. VM Size: 418.01 MB For how much I paid, this is amazing. _________________ I'm o.d.d. Back to top pbowmar Houdini Guru Joined: 03 Jul 2002 Posts: 3999 Location: Vancouver, BC, Canada Posted: Fri Dec 17, 2010 3:48 am GMT    Post subject: Oh, I did get overclocking working. 15% overclock gives me: 6:44 Woot! Back to top Display posts from previous: All Posts1 Day7 Days2 Weeks1 Month3 Months6 Months1 Year Oldest FirstNewest First         Houdini Forum Index -> Technical Discussion All times are GMT - 5 Hours Goto page 1, 2  Next Page 1 of 2   Jump to: Select a forum Houdini----------------AnnouncementsGeneral DiscussionTechnical DiscussionHoudini ApprenticeLearning HoudiniWork in ProgressThe Orbolt Smart 3D Asset Store  You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You cannot attach files in this forum You can download files in this forum phpBB component v1.2.4 RC3 by Adam van Dongen. Based on phpBB © 2001, 2002 phpBB Group