Results 1 to 2 of 2

Thread: CINEBENCH 2003

  1. #1
    Join Date
    Aug 2000
    Location
    Concord, CA
    Posts
    7,056

    Default

    Maxon’s new CINEBENCH 2003 http://www.maxon.net/pages/download/benchmarks.html is a free benchmarking tool for Windows and MacOS based on the 3D software CINEMA 4D R8. The tool is set to deliver accurate benchmarks by testing not only a computer's raw processing speed but also all other areas that affect system performance such as OpenGL, multithreading, multiprocessors and Intel's new HT Technology (Pentium 4 HyperThreading and ONLY accurate with Windows XP Professional). CINEBENCH 2003 includes render tasks that test the performance of up to 16 multiprocessors on the same computer as well as software-only shading tests and OpenGL shading tests on huge numbers of animated polygons that will push any computer to its limits.

    Why CB 2003 does not appear to support AltiVec I don’t know. I do know that Maxon’s real application, CINEMA 4D R8, does support AltiVec for some operations. If some PeeCees can do HyperThreading in this test app, why no AltiVec support for Macs? Does not seem like a benchmark level playing field once again. Maybe a Mac would look too good. I can say that CINEBENCH 2003 will task your Mac to its fullest and future Macs for at least awhile. My fastest Mac takes over 8 minutes to complete the tests, my slowest about 16 minutes even without a multiple processor 2xCPU test. It’s worth a spin, and as the docs say, the results will tell you whether you have your Mac setup correctly for this kind of work.

    My machines tested:
    G4-1.25G DP MDD (not FW800), GF4ti, 1.5GB RAM, OS9.2.2 and 10.2.4
    G4-800DP (QuickSilver), GF3, 1.5GB RAM, OS9.2.2 and 10.2.4
    PTP-Sonnet G4-800/800/200, ProFormance3+ 32MB, 1.0GB RAM, OS9.1
    PTP-XLR8 G3-540/270, ProFormance3+ 32MB, 1.0GB RAM, OS9.1

    All tests run at 1152x870xMillions on my Nokia Multigraph 445Xpro 21” monitor. CB refers to the CINEBENCH units of measure, greater is faster. I have shortened the column descriptions to fit.

    The Tests:

    Processor__CPUs___OS__GFxBoard_1xCPU_2xCPU_2xCPU_S peedup__C4D__OGL-SW-L_OGL-HW-L_OpenGL_Speedup_Render¬?Time
    ------------------------------------------------------------------------------------------------------------------------------------------------
    G4-1.25G_DP_2___9.2.2__GF4ti___118CB_224CB_____1.90__ ________147CB__411CB____860CB_____5.85(HW-L)_____498.44¬?sec
    G4-1.25G_DP_2__10.2.4__GF4ti___120CB_212CB_____1.77__ ________149CB__416CB____858CB_____5.76(HW-L)_____498.47¬?sec
    G4-800DP____2___9.2.2__GF3______75CB¬?138CB_____1.86_ __________94CB__269CB____574CB_____6.12(HW-L)_____786.43 sec
    G4-800DP____2__10.2.4__GF3______76CB¬?135CB_____1.78_ __________95CB__269CB____560CB_____5.88(HW-L)_____780.09 sec
    PTP/G4-800/_1____9.1__PF3+32MB__71CB¬?-----______----____________73CB__205CB____169CB_____2.81(SW-L)____*759.10¬?sec*
    800/200
    PTP/G3-540/_1____9.1__PF3+32MB__54CB¬?-----______----____________62CB__168CB____137CB_____2.70(SW-L)____*955.08 sec*
    270

    *These two tests have no rendering time added for the 2xCPU rendering.*
    BTW, even using Option-space for a fixed width space, I could not get the columns to align. Sorry for how badly this looks using the underline character for spacing.

    Notes on The Graphics Benchmarks
    --------------------------------
    The graphics tests are divided into various shading and lighting tests.

    1."1xCPU" renders a large highly detailed 3D scene with one processor.

    2."2xCPU" renders the same scene with multiple processors, in my case 2.

    3."C4D" (shading) benchmarks uses CINEMA 4D's own software shading engine/algorithm.

    4."OGL SW-L" tests the OpenGL acceleration in combination with CINEMA 4D's software based lighting. The test runs with the help of OpenGL. The lighting is handled by CINEMA 4D's software algorithms. Shading and transformation is carried out by the graphics card. Older graphics cards may run these tests faster than the "OpenGL HW-L" benchmark.

    5."OGL HW-L" benchmarks the graphics card's OpenGL acceleration as well as its possible hardware lighting acceleration. These tests run entirely with the help of the graphics card’s OpenGL accelerator. CINEMA 4D merely transfers the position of the light sources and geometry. Modern graphics cards will run this test much faster than the software based algorithms.

    6."OpenGL Speedup" measures the speed increase ratio between CINEMA 4D's software calculations and OpenGL. Values greater than 1 mean that OpenGL acceleration has taken place. Values less than 1 mean that CINEMA 4D redraws the scene faster than the graphics card's OpenGL implementation.

    7.My vintage PTP, while obviously slower than the G4s, with either a G3/G4 processor card, will do these tests and the PF3+ 32MB graphics card benefits from five OpenGL libraries as well as its own Formac OpenGL library. The sixth MacOS OpenGL library, OpenGLRendererATI, should not be used with the Formac card, just as one might suspect. BTW, I used the OS9.1 libraries. Later OS OpenGL libraries, such as OS9.2.2, might be more mature and do better. The rendered output is just as good as the later Macs. But the card will not do virtually any 2D/3D in OS X accelerated, nor do I have X installed on the PTP. Formac discontinued support for the PF3 cards prior to X. k

    [This message has been edited by kaye (edited 09 March 2003).]

  2. #2
    Join Date
    Jan 2002
    Location
    NW Montana
    Posts
    8,197

    Default

    Thanks, Kaye. - Randy

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •