Skip to content
Snippets Groups Projects

Repository graph

You can move around the graph by using the arrow keys.
Select Git revision
  • ADD_CSV_BRANCH
  • ADD_CSV_BRANCH_STABLE
  • cvsHEAD
  • fsl-3_2
  • fsl-3_3-branch
  • master default protected
  • newnew
  • osx-109-branch
  • sge-branch
  • streamliner
  • universal
  • universal_branch
  • 2202.11
  • 2202.10
  • 2202.9
  • 2202.8
  • 2202.7
  • 2202.6
  • 2202.5
  • 2202.4
  • 2202.3
  • 2202.2
  • 2202.1
  • 2202.0
  • 2111.0
  • 2103.0
  • 2101.0
  • 2012.0
  • 2011.0
  • 2010.0
  • 2007.0
  • 2001.0
32 results
Created with Raphaël 2.2.011Apr6Mar28Feb2218Jan17Dec161413107327Nov9526Oct234Sep330Aug17326Jul231413225Jun1229May28251715Mar623Feb30Jan271121Dec11Nov917Oct11108622Jul2118141171Jun27May20181310618Apr1117Feb158721Jan1410Dec119Nov10Sep4Aug24Jun14May1030Apr2211Mar18Feb28Jan222115Dec130Nov27262524232017326Oct2216151328Sep28Aug26181712107631Jul27212018171611911Jun19May8518Apr19Mar145317Feb1616Jan913Dec28Nov1324Oct211313Sep1229Aug282018131211621Jul11130Jun27May30Apr2523221553217Mar65329Feb27252420185429Jan8220Dec1914121130Nov281531Oct2410314Sep22Aug2120158762131Jul3029272625242320191811106420Jun14118730May1410932130Apr2026Mar8522Feb18916Jan13Dec6130Nov2827929Sep1987628Aug2226Jul16Jun2May29Apr26251165427Mar262422109820Feb12Jan1128Nov820Oct1Sep23Aug1816109827Jul18151424Jun17161410927May26133Mar13Jan1072Dec18Nov28Oct201918151230Sep292317Changed to have a compiled version. Don't need to compile each run timeDeleted dim_blocks.h: Deprecated, now in options/options.hChanged to use Dynamic Shared MemoryDeleted bedpostx_multigpu_SGE because now the MultiGPU version in SGE will be controlled by fsl_sub and NGPU parameter in bedpostx_gpuChanged to use Dynamic Shared MemoryChanged the way of divide set of voxels (before by slices). Now in X parts each one divided in subparts of a maximum number of voxels that are sent to the GPU in a iteretive fashionDeleted splitter_gpu.cc splitter_multigpu.cc because now the division of slices and voxels is make inside scripts and xfibres_gpu.ccNow accept in the input full tensors (9 elements), not necessarily symmetricChange to don't repeat all slices in case of fail in oneChanges in grad-nonlin functionsmake faster mix parameters after repeat a vox because f0make faster some methods that use thrust librarymake faster mix parameters after repeat a vox because f0changed the way of round some math operationsnew version Levenberg: several threads per voxelchanged the pow(x,2) function with x*xnew version Levenberg: several threads per voxelRemove warningsChanges in grad-nonlin functionsUse ARD for dstdAdd hyperprior parameters to the output*** empty log message ***add checking the set of the queues used in SGEThis commit was manufactured by cvs2svn to create tag 'fsl-5_0_3'.fsl-5_0_3 fsl-5…fsl-5_0_3 fsl-5_0_4bug bix: invwarp refadd save also slices with 0 voxelsrepair bug when slices have 0 voxelsrepair bug with -g parameterchange the name of the command execute for print it in the file commands.txtchange the order for get the Device number after allocating some data in the GPUinclude boost directory for the CUDA compilationfix bugadded Copyrightadded the possibility of compile Bedpostx CUDA versionVersion CUDA of xfibresMethod (called from Bedpostx) that call the xfibres CUDA version for only 1 sliceMethod (called from Bedpostx) that call the xfibres CUDA version for each slice (Only 1 GPU used)Implementation of the new Sample class prepared for save multi VoxelsDefinition of the new Sample class prepared for save multi VoxelsDefinition of the main methods