Accelerad 0.7 beta

Changes in Version 0.7

• The ray tracing engine is updated to OptiX™ 6.0.0 using CUDA 10.0 with support for RTX graphics cards.
• AcceleradRT is included in the Windows distribution of Accelerad.
• Added support for instances.
• Reduced memory overhead for meshes.
• Added support for reinhartb.cal, disk2square.cal, klems_half.cal, klems_quarter.cal, and all rfluxmtx h options.
• Fixed program name checking in rfluxmtx and genBSDF to avoid errors on some machines.
• Added support for rtrace -orRxX outputs.
• Number of allowed antimatter materials increased to 32.
• Various improvements to error handling.
• Fixed compatibility issue with Honeybee standard error formatting.
• Fixed bug in which ray tracing fails for empty scenes.
• Fixed bug in which end of file message appeared erroneously in rcontrib.
• Fixed bug in which some contribution indices might not be recorded in rcontrib.
• Fixed bug in ambient super sampling.
• Fixed bug in aft clipping plane for scenes with transparent surfaces.
• Adjustable color scale added to AcceleradRT.
• Radiance and irradiance visualization added to AcceleradRT.
• Walking mode added to AcceleradRT.
• Fixed bug that caused sky to render too bright in AcceleradRT.
• Fixed bug in DGP calculation when no task area is selected in AcceleradRT.
• Use of -g parameter to set GPU stack size is deprecated and has no effect.
• Use of -t parameter to set timeout callback is deprecated and has no effect.
• Support discontinued for SM 2.X (Fermi) and SM 3.X (Kepler) GPUs.
• Support discontinued for Mac due to lack of available drivers.
• Support discontinued for VCA.

Accelerad 0.6 beta

Changes in Version 0.6

• Added Accelerad versions of rcontrib, rfluxmtx, and genBSDF programs.
• Added support for Russian roulette ray extinguishing with negative -lr arguments. Using Russian roulette may have a negative effect on speedup and require a larger -g argument.
• Added support for src_phi2 and src_phi4 to be used in brightdata calls with source.cal.
• Slight performance improvement by ignoring materials that are defined but not used.
• Fixed bug in shadows in antimatter regions.
• Fixed bug in ambient calculations at low angles.
• Fixed bug that allowed ray weights greater than one in irradiance calculations.

Accelerad 0.5 beta

Changes in Version 0.5

• The ray tracing engine is updated to OptiX™ 3.9.1 final release using CUDA 7.5, which provides support for the Pascal architecture.
• Includes bug fixes from Radiance version 5.0.a.11.
• Added support for antimatter, although first string argument must be void.
• Added support for ambient inclusion and exclusion (-ae, -aE, -ai, -aI).
• Added support for .cal files created with GenCumulativeSky.
• Added support for Mark Stock’s utah.cal.
• Fixed bug in coordinate mapping for data from .dat files.
• Fixed bug in rendering glass with nothing behind it.
• Fixed bug in rendering shadows of trans materials.
• Fixed bug in rpict images with odd numbers in dimensions.

Accelerad 0.4 beta

Changes in Version 0.4

• The ray tracing engine is updated to OptiX™ 3.8.0 final release using CUDA 7.0.
• Includes bug fixes from Radiance version 5.0 final release.
• Added support for ambient supersampling (-as) so long as -aa > 0.
• Added support for Monte Carlo random seeds (-u).
• Added support for motion blur in rpict (-pm).
• Added support for depth of field blur in rpict (-pd).
• Added -ag parameter for faster final gather step for debugging.
• Added support for cone, cup, cylinder, tube, and ring objects in .rad files.
• Added support for illum materials with non-opaque or void string arguments.
• Added support for corr, boxcorr, and cylcorr functions from source.cal.
• Improved handling of normals modified by texfunc objects.
• Brightdata from source.cal can now be applied as a material to sources.
• Ignored triangulation errors with a warning.
• Printed an extra line in rpict to handle a bug found in IES VE.
• Fixed bug in display of GPU global memory size.
• Added support for Linux and Mac.

Accelerad 0.3 beta

Version 0.3 beta includes the following changes:

• The ray tracing engine is updated to OptiX™ 3.8.0 beta.
• Includes bug fixes from Radiance version 5.0a.
• Major speed improvements for ambient calculations in rpict and rtrace.
• Sphere and bubble objects in .rad files are supported.
• Alias objects in .rad files are supported.
• Function objects no longer need to be direct parents of the objects they modify, but still only one function object is supported per modified object.
• Non-square images are less likely to produce errors.
• Renamed -am argument to -an to avoid conflict with new Radiance 5.0a parameter.
• The -x and -y arguments are no longer required for rtrace, but they do result in better performance.
• The number of ambient divisions specified by -ad is no longer limited.
• The number of TCC-enabled graphics cards that can be used is no longer limited.
• The GPU stack size is reduced and defaults to 4096 bytes.
• Improved error reporting.
• Improved documentation in the readme file.

Accelerad 0.2 beta

This is a beta release of Accelerad for Windows.

• The ray tracing engine is updated to OptiX™ 3.7.0 beta 3, which provides support for CUDA 6.5 and SM 3.7 GPUs, including the Tesla K80.
• Users are no longer required to download and install CUDA.
• Includes bug fixes from Radiance version 4.3.a.2.
• Radiance mesh objects are supported.
• Changed new -at parameter to work independently of -aa and -ar.
• Fixed a bug that prevented caching of ambient values for certain specular parameters.
• Fixed a bug in Gaussian transmission for trans materials.
• Changed error messages and return values to match those of Radiance.
• Improved documentation in the readme file.

Accelerad 0.1 beta

This is a beta release of Accelerad for Windows. Separate installation of CUDA 6.0 is required.

• The ray tracing engine is updated to OptiX™ 3.6.3, which provides improved support for Maxwell-based SM 5.2 GPUs, including the GeForce GTX 980.
• Includes bug fixes from Radiance version 4.2.2.
• Enabled an experimental algorithm for parallel irradiance caching in rtrace.
• Implemented -I and -ld arguments from rtrace.
• Improved handling of not-a-number (NaN) errors in both GPU and standard implementations of rpict.
• Fixed a bug in Gaussian transmission for trans materials.
• Added ability to assign geometric transformations to skies defined with skybright.cal or perezlum.cal.
• Fixed a bug that caused out-of-resources error (cuda error 7) on some machines.
• Fixed a bug that caused stack overflow errors in some cases for geometry located far from the origin.
• Improved documentation in the readme file.

Accelerad 0 beta

This is the first beta release of Accelerad for Windows. Separate installation of CUDA 6.0 is required.