Products Business Solutions Support Store ORB News Community Download Company Info Contact Info

 Main
Product Info
Product Details
Gallery
Support & Resources
Usage & Distribution
Futuremark » Products » SPMark » SPMarkJava JSR 184 » The Tests

SPMarkJava JSR 184 Tests

The professional's tool

SPMarkJava06 is a MIDP 2.0 / JSR 184 benchmark based on workloads that correlate with a blend of popular, real world applications and theoretical tests, developed in partnership with leading industry members. Designed to run on most Java API enabled mobile devices, SPMarkJava JSR 184 allows users to measure and compare Java app performance across their development boards and existing product lines. The benchmark incorporates unbiased, consistent and fair tests that generate accurate measurement data.

SPMarkJava JSR 184 Game Tests

  • High Detail 3D Game Test – Flying warship in fast pursuit of off-road vehicle
    • JSR 184 implementation required to run this test
    • Runs at full resolution for device, up to QVGA (320*240)
    • Approximately 5,000 polygons rendered per frame
    • Total of 21,380 triangles in the scene
    • 15 objects, 1 camera, 1 directional light affecting one object (warship)
    • 128x128 texture size
    • M3G options/hints in default run
      • imageFilter: IMAGE_FILTER_LINEAR
      • levelFilter: LEVEL_FILTER_NEAREST
      • shading: SHADING_SMOOTH
      • lighting: LIGHTING_ON
      • perspectiveCorrection: PERSPECTIVE_CORRECTION_ON
    • Performance measurement reported as Frames per second
    • This test is available only in the Professional Edition
  • Low Detail 3D Game Test – off-road vehicle driving around a desolated area
    • JSR 184 implementation required to run this test
    • Runs at full resolution for device, up to QVGA (320*240)
    • Approximately 3,000 polygons rendered per frame
    • Total of 12,139 triangles in the scene
    • 10 objects, 1 camera, 1 light (disabled in default setting)
    • 128x128 texture size
    • M3g options/hints in default run
      • imageFilter: IMAGE_FILTER_NEAREST
      • levelFilter: LEVEL_FILTER_BASE_LEVEL
      • shading: SHADING_SMOOTH
      • lighting: LIGHTING_OFF
      • perspectiveCorrection: PERSPECTIVE_CORRECTION_OFF
    • Performance measurement reported as Frames per second

SPMarkJava JSR 184 2D Game Test

  • Designed to replicate workload of simple game with bitmap graphics: Driving game with bird’s eye view
    • MIDP 2.0 required to run test
    • Test measures performance on CLDC 1.0 devices
    • javax.microedition.lcdui.game package classes used heavily including bitmap rendering and collision detection
    • Workload generated by adding independently scrolling bitmap layers with transparency o
    • ace track rendered as single, large, fully opaque TiledLayer object
    • Benchmark code operates as a fully playable, but limited game (Game available only in Advanced and Professional Editions)
    • Performance measurement reported as Frames per second

SPMarkJava JSR 184 Feature Tests

  • 3D Fill Rate test
    • Measures the 3D fill rate of the target device
    • Draws two 256*256 textures to a single triangle polygon
    • Uses two texture units if available
    • Rotation used as a visual effect
    • Performance measurement reported as thousand texels per second (Ktexels/s)
  • 3D Polygon Throughput test
    • Measures the amount of polygons drawn on the screen
    • Draws a set of 9 torus knots with 10 000 triangles on the screen
    • Polygons are lit
    • Performance measurement reported as thousand vertices processed per second
  • Video Processing
    • Measures video playback
    • Uses H.263 Codec and JSR 135 API
    • 176 x 144 resolution
    • 15 FPS frame rate
    • 40kBps bit rate
    • Measurement reported as “Pass/fail”
    • This test is available only in the Professional Edition
  • Image Processing
    • Portions of an image are encoded and decoded on screen
    • PNG image format (the format widely used in internal image storing in handheld devices)
    • Each test operation result is displayed
    • Geometric mean of both operations used to produce test results
    • The result is given as a geometric mean of encoding and decoding test results. The individual test results in kPixels processed per second are also available
  • JVM-Test
    • Theoretical test, consisting of series of subtests
    • Subtests measure specific JVM operations
    • Subtests vary from simple arithmetic operations to complicated but critical operations such as object creation
    • Different data type variations tested:
      • 32bit
      • 64bit
    • Subtests include:
      • Integer addition and subtraction in various data types
      • Integer multiplication and division in various data types
      • Conditional code test (if-else-statement)
      • Simple object creation test
      • java.lang.System.arraycopy(…) test – results in MB/s
      • Static method call
      • Array read access
      • Floating point addition/multiplication tests (these tests require CLDC 1.1)
      • Sine, cosine, tangent and square root calculations (these tests require CLDC 1.1)
    • Performance measurement reported as operations per second
    • Individual test results displayed
    • Overall JVM-Test score is a geometric mean of subtest scores
  • Battery Test
    • User selected tests run in looped sequence until battery is fully drained or the OS shuts down the Java application due to low battery.
    • Measurement reported as the Time to drain battery in seconds
    • Results not factored into overall SPMarkJava JSR 184 score
    • This test is available only in the Professional Edition


Privacy Statement Legal Advertise With Us Press Room Benchmark Development Program Partnerships Contact Info