added by hand ex

added integer modulo operator

  Changed the precedence of the infix power operator to bind tighter than unary negation.
  Also added a table to the manual that summarizes the syntax of expression forms.

set setting

byhand

added byhand examples

unary negatation precedence is too high

Modified query pool code for resetting after an update run

Additional changes for explicit queries

Made changes to the front-end for explicit queries

lifted substitution

update TODO list

dev branch

clean up print statements

iterate distribute summation

removed det loop-fusion

 ein-to-low scalar for matrices

don't generate an empty output check for grids

project on i*j*k*l

compiles hessian mode

extend CONS

lift det field

lift field with constant index

update TODO with glk wish list

update TODO

write function in examples

added nrrd name

  another try at fixing the bookkeeping code.

  fix typo

add rest of d3s files

mip

name nrrd file in examples

  Fixing the counting of active/stable/dead strands and the counting of steps.

  update TODO list

  A couple of bug fixes:
    1. correctly handle the situation where the output/snapshot is empty
    2. use the correct strand-status test for OutputGet functions

sync examples

update TODO list

  update TODO

adding C2 tent, so that curvature of trilinearly interpolated values can be shown

shaded rendering of torus with two different kernels

one way to make a torus dataset

move evaluate call()

correct syntax for field evaluation

moved vr_taxi

update simple3d

fem

instructions

updated examples

call to curl/hess and added example

added storeVec to ir

removed curl3

 firedrake friendly

double dot between arbitrary tensors and fields

added examples to fem

added call to inside()

  Build improvements for documentation; merged some doc changes from vis12; fixed error
  reporting/return status when running a Diderot program.

  merging in some build changes from vis12-cl; cleanup error reporting in generated code.

fixing potential race conditions

tab expansion

add information about compiler to generated code

  Make sure that workgroups are at least 32-lanes wide.

add information about compiler to generated code

  Make configure script more robust in searching for OpenCL

  Some minor changes to accomodate using oclgrind on Diderot programs.
  We now accept the user's device specification, even when it is not
  a GPU, and "inline" OpenCL functions are now marked "static inline".

merge vis12+fem+ein

representation of mesh is based on vis branch, see fem branch

  Fix race condition of OpenCL (no-BSP) scheduler.  Also marked all targets of
  atomic operations volatile.

added various support for mesh rep. from start to finish

added meshPtr to IR

creating a copy of branch of vis12

removing branch

removing branch

creating a copy of branch of vis12

creating a copy of branch of vis12

put mesh in ein

 Creating branch to understand FEM fields

some refactoring in tree-il, move sum function location

  We seem to need to include "-pthread" in the linking phase on Linux; not sure
  how it used to work (maybe a change in the compiler?)

  merging changes from vis12 branch into OpenCL branch

  correct error in regression test

  merging in changes from OpenCL branch and adding some more support
  for the initially method

create a local copy of data sources for rtest2

various load|store changes

comment out most vector support

does not use vectors, benchmarks all faster

change to function arguement types

clean up normalization rules

evecs evals 2x2

  added "continue" statement inside the update method.  The effect is to exit
  the method, but to leave the strand active (unlike die and stabilize).

attempt at evecs 3x3

lift curl expression

  make CL version consistent with others and DIE

more collection hacking

  more collection hacking

  more collection hacking

  Working on collection output

  working on collection output support

  temporary workaround for collections

  fixed type error in CheckHighIL

change norm,normalize and added trig

  Increased GRAB_SZ for OpenCL target; added step limit command-line option.

high-to-mid subexpression limitation with mkops change

high-to-mid subexpression limitation

neg tensor change

hack

  more initialization fixes

  working on inputs

  porting some more input support from trunk

  working on fixing initialization of inputs

  formatting changes

without common subexpression

common subexpression

  fix names of matrix functions (remove "f" suffix)

  scheduling fixes for OpenCL

  added FIXME comment

  convert regression tests to make all images inputs.  Fixed issue with runtime
  flags for regression testing

  bug fixes for images in OpenCL target

  OpenCL scheduling fixes

  OpenCL bug fixes

  add missing code generation for inputs

organize derivative-ein

organize derivative-ein

  minor edits

  minor edits

  Fixed confusion between tensor[1] and tensor[] when testing Inside of 1D field.

more fixes to unused variable tracking

  add missing setLoc for functions

approximating scivis project 2

  more initialization fixes

  fix shutdown

  add missing initialization code to Initially kernel

  adding barriers

  working on OpenCL issues

lifted sine,cosine,arcsine,arccosine

working on OpenCL scheduling

back porting to OpenCL 1.1

  working on OpenCL issues

  fix name of image functions on OpenCL

  added missing #ifdef

remember to initialize the number of workers

  backport to OpenCL 1.1

  adding better error reporting for OpenCL

  change atom_ to atomic_

  fix type error

  added support for getting binary representation of program

  add missing image initialization code

  working on OpenCL issues

  Working on getting OpenCL version right

  fixing CL_VERSION info

  add missing compiler flags for OpenCL

  fix typo

  fix typo

  working on OpenCL issues

  working on OpenCL issues

  more OpenCL fixes

  working on OpenCL code generation issues

  working on OpenCL branch

  minor changes to make code more consistent with trunk

sqrt fields

trying to make idiomatic definition of FA

fixes

correct use of identity[3]

tweaks

work-around for inside bug

tweaks to FA definition

didnt mean to add this

didnt mean to add this

for debugging

trying to get FA isosurfaces working

didnt mean to add this

didnt mean to commit executable

  finishing up name reorg

  Merging in name reorganization changes from OpenCL branch

  new file

  Finish rationalization of names

  working on reorg of names

decCnt in high-il

  working on a rationalization of names

change det 3x3

change det 3x3

for tensor fields

det 3x3 field

  working on vis12-cl reorg

  some more documentation

Adding a test for stabalizing in the global function

Added the ability to include stabilize in global

  fix header comment

  detabbing

  detabbing

  Hack to avoid freeing images multiple times when border control is being used.

  remove generated files

  add configure support for rtest2

  add comment about converting output

  bug fixing for border control

  adding runtime code for clamp, mirror, and wrap

  working on border control

  working on border control

supported ops

matrix fields

load->image, det(), S*T

  change "load" to "image" in test programs

working now

bug fixed

merge in inputs from trunk

commit to sync directories

first attempt

  detabbing

  updates to OpenCL branch from vis12 branch

  scheduling fix

generic norm

  Fix bug in generated input functions

  update vr-lite-cam/main.c to new initialization protocol

  remove unused variable

  fix a bug where the border control code was confused by input images

  merging changes from vis12 branch into OpenCL branch

  detabbing

  tweaks to synchronization code

  fixing a bug in the initialization of the parallel world

  adding some missing timing code

  finish implementation of nan tensor support

  adding some runtime support for border control

  fix syntax error

  merging the input-handling code from the OpenCL branch.

  add missing case

mid to low clean up

Adding stabalize feature to global block

derivative division

general outer product

  merging support for Lamont's syntax into vis12 branch

  merging some vis12 branch code into Lamont's branch

  code formatting

Bug Fix: Allows the position variable to be a output variable

Bug Fix: The indentities to reduction operations for real numbers should use the CL.mkFlt to get the correct real values

  push version flag to all active branches

  porting changes from vis12 branch

  detabbing

  removing unused modules

  working on OpenCL

  working on border control

  add bug040

|matrix field| and speed up

  working on border controls

  border controls test

  beginning to add support for NaN tensors.

  working on border control

  working on border control

  image for testing border control

  working on border control for images

  implementation toto notes

  add missing checkIL pass after mid-to-low translation

  add missing IL checking pass

  change test to work on RGB images

fixing missing nrrd file

  pushing the ImageDim operator through the ILs

  working on image border controls

fix notation to follow conventions

update bugreport

moves test files

added run scripts

sync branches

ops

Adding a variance and standard dev test

add output

fix earlier typo

adding operators to SUPPORTED_OPS and some code-style fixes

adding operators to SUPPORTED_OPS and some code-style fixes

  moved resolved bugs from open to resolved.

  fixed a bug in addPred.

  reduce the size of bug039.diderot

  changed field continuity to avoid type errors

question about allowed continuity order

new for canny

for testing canny

testing ability to differentiate canny function

as of r2912 this bug gone

further info

unable to compile sampling isosurfaces of mean curvature

deleted lines have moved to other tests

?

for testing ability to sample curvature quantities, without differentiation

for flowline curvature testing

adding test for flowline curvature

norm, or norm squared, of hessian

for testing

sync branches

sigh

  add documentation of supported operators

  typechecking for inner product

example of inner prod

  use IntRedBlackMap for index maps

move summation around

update todo list

bug report

p

bug report

new info

oops, not sure what changed

not sure what's going on here crviso3d.diderot

moving this to new crviso3d.diderot

examples of trying to get total curvature

uncaught exception

updates on what works and not

tweaks

for testing

still tweaking

further work on isolating issues

GLKs experiments with 3D vector fields

added tensor +/- fields ops

clarifying what is and isn't working now

  remove deprecated source code from simplify phase

removed print statements

trying more things

trying more things

curl2

experiments in 3D

sync folders

comment about isocontour of curl

trying to do isocontours of 2D curl

rename vmiso2d viso2d

isosurfaces of vector magnitude

re-centering where seeding happens, and using fewer local variables

notes

renaming bugs

remove bug that was already fixed

a little more interesting

added consistency checking of CFG edges

fixed bug in IL.CFG.deleteNode

bug039

bug039

new data and test

  Not really a bug in the compiler (or very minor), so I'm removing it
  instead of moving it to the closed directory.

maybe same bug as bug037, but this time trigger has been isolated

  added missing handler for TypeError exception

fewer Diderot program errors, but still getting the compiler error

fewer Diderot program errors, but still getting the compiler error

trying to get a 2D synthetic function sampler working

  adding image border-controls to basis

a

s

  added size function to basis; still needs implementation in later phases

Added variance back to compiler

added sqrt,pow, and examples

moved split around, added norm to typechecker, added sqrt to ein

missed files

added rtest2

  merged in output redirection support from vis12 branch.

sync data

moved over more regression tests

added regression tests

rewrote mid-to-low and added tests/EIN

typo in mkOps-fieldproduct, rewrite deltas

added examples

Added runtime argument for global threshold for pthreads

Readded spatial timing to runtime arguments for pthreads

Updated single script to generate when needed

Updating boids.diderot file

Few changes to the boids update script

Updating boids benchmark test

Fixed a few bugs in the parallel kd-tree example

added field outer/inner/cross product and tensor slice

moved mkzero to handle ein, clears non-probed fields

added norm

code cleanup

added 2-d cross product, new rep. of 2-d curl

moved dy and dz in image loading

aligned is false

testing

  Create vis15 branch for Charisee's merge

edit split-ein

  Finish support for redirecting output

  Added short/long option forms

  Working on output redirection

  fix printing of "if then else if ..."

  working on output redirection

  working on support for output redirection

  update TODO list

lerp and clamp

moved differentation kernels

imgaddr computation moved to mid-to-low

changed types for imgAddr and imgLd

  detabbing

  porting bug fixes from vis12 branch

  Fixed bug010 (unreachable control-flow nodes) by adding edges from EXIT nodes to JOINS
  and masking out the RHS of PHI nodes.

  new unreachable code pruning; checks for missing strands/initially (the parser prevents
  these errors currently, but future language evolution may change that)

  Added Slide operator to Mid and Low ILs

  update Notes

  detabbing

  porting changes from vis12 branch

  Fixing some lingering issues with globals and inputs.

  Fixed some regressions

  new test cases

  fixing some regressions

  cleanup and detabbing

  Fix for bug 019 (allowing richer initializers for input globals).

  port changes from vis12 branch

  Changes from vis12-cl branch; detabbing

  adding file from vis12-cl branch

  porting improvements from vis12 branch

  remove C file from rtest

  Fix bug in shutdown fragment; proper tracking of HasGlobals property

  Better cleanup for docs; add promote-fn.sml from vis12 branch

  detabbing; some changes from vis12-cl branch

  merge in new globals code from vis12 branch

  Fixed bug 031

  new version of the Variable Analysis phase.  This mostly fixes bug 031, but
  there are still some C code generation issues that may (or may not) be
  related, so I'm not closing it yet.

  fixed a bug in the global promotion code

  working on globals

  Working on globals; split input initialization from the rest of
  global initialization in the SSA IL.

  Working on globals

  working on globals

  Working on representing globals as their own type in the SSA and Tree ILs.

field problem

Fixed minor problem with workers not getting the correct start and end parameters

  working on global translation

  Working on splitting out the representation of globals from local variables

Projection function

Updated parallel spatial code with medians of medians algorithm

copy function()

  copyright update

  merging in changes from vis12 branch

  merging some changes from the vis12-cl branch

  bug013 is now reported as an error

  Fix for bug013; it is now a type error to declare non-value-type strand variables.

  fix global-variable prefix for inputs

  Fix bug018 in a different place.

  Working on the fallout from fixing bug018

  Fixed bug018 (conflicts with C reserved words).  Also fixed issues for programs
  that do not define global or input variables.

  get rid of redundant name (globalTy) for globalsTy in code generator

  Fix for bug018

  Add prefixes to input, globals and strand-state variables to avoid conflicts with
  C keywords (bug018).

  update copyrights

remove redundant check for program properties

  Fix #ifdef

  added additional program properties StrandArray and HasStabilizeMethod

  porting changes from vis12-cl branch

  patch configure.ac for OpenCL header path

  Working on OpenCL

  Add world clock to sequential and SMP versions

  working on OpenCL support

  Input should be classified as an impure operation

  detab

  working on OpenCL support

  added output helper function fragment

  Working on OpenCL support

  working on OpenCL support

  working on OpenCL support

  Working on OpenCL support

 working on OpenCL support

  added missing data sets for regression tests

  Adding OpenCL scheduler stuff

  working on OpenCL scheduling

  bug fix in OpenCL temnplate code

  Fix name of GetSizesKern kernel

  InitiallyScope is now implemented in OpenCL

  Fixed a couple of bugs: image freeing code was broken and execution stage
  tracking was incomplete.

  adding additiona regression tests from vis12 branch

added projection function name

  merge changes from vis12 branch

  Some minor changes from the stage branch

  Porting changes from vis12 branch

  working on OpenCL

  working on OpenCL support

  Working on OpenCL support

  Working on OpenCL support

  Working on OpenCL support

  Changed the interface to code generation for freeing globals.  We now pass the list of TreeIL
  global variables to the Target.Program.free function.

  Working on OpenCL

  Working on OpenCL library support

  Make clinfo compile without depending on the Diderot library implementation.

  Working on OpenCL support

  Working on OpenCL support

  working on OpenCL support

  working on OpenCL support

  Working on OpenCL support

  added execution stages and checks that proper initialization has been completed.

  working on initialization

  Add check for OpenCL 1.2

  added some discussion of how libraries are initialized

  adding some implementaiton notes

  Working on OpenCL support

  Splitting c-util/gen-inputs.sml in to generic and target-specific bits.

  porting bug fixes from vis12 branch

  tab expansion

  Fix bug022 (empty sequence not supported yet)

  bug032 was fixed by the changes imported from the vis12-cl branch

  tab expansion

  Merging changes from vis12-cl branch.

  Added heron as a regression test (tests loading of sequences)

  Fixed assessment script (the expected number of elements was off)

 Working on OpenCL

  add new fragments

  new fragments

  Finish removing "Status_t" type

  Working on OpenCL support

  Working on OpenCL support

  porting some changes from the vis12 branch

  Better error checking for input initialization; removed Status_t type (just use bool
  for return status with true ==> error, false ==> ok)

  Merging some changes from the vis12-cl branch and fixing an issue with the
  dynamic sequence inputs.

  Working on OpenCL support

  Working on OpenCL support

  Working on OpenCL support

  Working on OpenCL support

  OpenCL support

  working on OpenCL support

  working on OpenCL support

  Working on OpenCL

  Working on OpenCL support

  more work on OpenCL support

  working on OpenCL support

  Working on OpenCL support

change to vector ops

  Woeking on OpenCL support

changed getPieces()

Cons Vector 

added trimVec

clean up tree-il

lift Save(tmp)=>store statement

lift Save(tmp)=>store statement

change bindings for E_tmp

mux, tmpmux

  Fix for bug009; this change requires updating to SML/NJ 110.77.

  fix comment

  Port updates from vis12 branch to vis12-cl

  more accurate description of command-line options when "-h" is given

cleanup tree-il exp/stmts

update

  add erfc to the supported math functions

  Add exp and erf as math functions.

updating info on how to process output, and adding orthographic projection option

printing matrices

  Bug 033 is also fixed

  Fixed bug034 (issues with unused variable tracking)

new bugs

data for new bugs

MkVec2, takes multiple expressions

represent tensors with arrays

Change global types to arrays and print them

changed inlcude

mkVec as an ASSN

made changes to ld and mkvec funcitons

added opr to c-target

added decl

moved getTypes and getOpr functions to the target

  backporting from trunk to staging

  porting changes from vis12 branch to staging

  tab expansion

  Fixed normalization of 3D curl.

  additional debugging support.

  fix bookkeeping for unused variable elimination

  Added unused variable warnigns

  documentation updates

  Working on unused variable warnings

  Starting to add support for detecting unused and shadowed variables.

  update TODO

  working on OpenCL

  Working on OpenCL support

  Adding bug032

opr to clang

porting changes from vis12 branch

  Remove extra newline when printing statement comments.

port fix for bug001 from vis12 branch

port fix for bug001 from vis12 branch

  fix code formatting

  revolved bug001 (unused inputs getting removed) by marking the Input operation
  as not pure.

add Types and oprators as sets in tree-il

  Merging bug fixes and other changes from vis12 branch into staging branch

  minor formatting edits

  tab conversion

  various changes

cFn added

  add output variable

  Fix typo in NOTES

scan for types

Tree-IL fixed printer and GlobalVar problem

hi

spltiVec

moved target functions

moved some things around

Wraped the spatial timing code in ifndefs for spatial communicaiton

Wraped the spatial timing code in ifndefs for spatial communicaiton

added load/store vec

  working on OpenCL

Fixed Bug with multiple reductions in the same expression

Major changes to lift reduction variable into its own block 

Major changes to lift reduction variable into its own block 

added tree-il expressions and types

Adding work for reduction

embedded summation split

clean up mid-to-low

InnerProduct, DoubleDot:

  indentation fix

Inner product

Quotient rule

  Working on OpenCL support

added files

code cleanup

  added some bug fixes and other edits from the trunk

Added Lift Constructor

  working on OpenCL support

  Add missing file to sources.cm

  Fixing errors in make rules for fragments

  Cleanup build ruules, sources, etc. for OpenCL target fragment files.

  port bug fix from vis12 branch to vis12-cl branch

  Fixed bug030 (problem with inlining functions whose bodies were conditionals)

  porting changes from vis12 branch

  update NOTES with new bugs (030 and 031)

needed for bug031

another bug

another bug

  Resolved bug 023

  update bug NOTES

  Working on a fix for bug023

normalize

Outer product

Outer product

Multiply Fields

Made presort parallel

Made presort parallel

minor fix to configuration files

minor fix to configuration files

minor fix to configuration files

Made spatial workers not wait if work is ready for them

Added field product

Stopped printing the tree in spatail view

Added back the parallel merge sort

Fixed a bug when not resetting the numRunning field when using a library for parallel KDTree building

Removed parallel merge sort from spatial build

Improve the spatial build by 2x

Made sorting happen initially during the spatial datastructure building

Potential fix to spatial speed up

Adding the single.sh script to the boids benchmark

Commiting Work for Spatial

Working on improving the spatial data structures

Fixed error with spatial data structure not getting deterministic results

Fixed error with spatial data structure not getting deterministic results

Fixed bug with queryPool not allocating correctly and fixed deadlock of spatail data structure

Adding script for producing boids benchmark

Fixed bug not initializing the global node correclty when performing global reductions

Commiting current version of new bug

Added to Eqfn

Adding Progress on fixing parallel new fix

Commiting the new strand tests

Code Clean up

Value numbering

Value numbering

submitting new test case

Fixed a bug where all workers were executing the global code when reductions do not exist

Fixed the problem with pthread spatial data structures stalling

Adding the boids benchmark

Fixed global reduction block

Curl regression tests

  update notes

  editing bug notes

  notes on bugs

uncomment dot product

Updating benchmark results

Adding files for sum-reduction benchmark

Updated Parallel Reduce code with removing one point of barrier synchornization

Fixed error in run script

Fixed parallel global reduction issue

Fixed a bug maybe?

type checker

Bug Fixes to parallel reduce code

Updating reduction testing

type checker

diff Fields

Change dot-times op

Change dot-times op

change to image ops

Adding in the reduction test code

Adding reduction code fragment

Fixed bugs in parallel reduction code

eintypes->mid-iltypes

fixed type problems

fixed type problems

added gentypes to mid-il and split to High-to-mid il

Added type Checker

ein to gen-ein functions and tests

 gen-ein

added kernels to gen

regular check-in

generate code

  Importing changes from vis12 branch

  minor fix for dynamic sequences

  some minor fixes for dynamic sequences

Committing temp global phase code

Added Tests

add app-ein.sml to high-il

Added divergence & hessian

Adding new work

 divide field

Current Global Reduction state with seperation into phases

mid-il fixes

old version

old version

remove generic rep, change Summation index, and add tests

add type checking

changes to high-to-mid

  merging in new SimpleAST changes

 ein

  porting changes from trunk

  Merging in changes from trunk

changes to expand Conv

  tab expansion

  Some minor cleanup to new simplify code.

  merging in new SimpleAST representation from stage

  merging in new SimpleAST changes from trunk.

Fixed bug with not getting the min or max overloaded variables

Committing updated global reduction code, removed spatial grid code (no variance)

Committing updated global reduction code, removed spatial grid code (no variance)

add expand-conv

probe.sml

  Fix comment

working in einTest so won'te compile

  Fix configuration (and INSTALL) to require SML/NJ 110.75.

  Added "make install" target and updated INSTALL documentation.

  Add test for "unu dnorm"

Updated Spatial Communication on Pthreads and fixing/adding global reductions for sequential and pthreads

Updated Spatial Communication on Pthreads and fixing/adding global reductions for sequential and pthreads

normalize test

  Fix compiler issues

Change Sum term

value numbering?

  Merging in changes to support relocation of installation from vis12 branch.

  Working on manpages

  Working on man pages

  Working on man pages

  Working on manpages

changed mu to alpha

  Modify compiler to dynamically determine the intall location of the Diderot
  libraries and include files.  This feature will allow the diderot compiler
  to be installed in places other than where it is built.

  Added support for DIDEROT_TEEM_DIR environment variable.

  more manpage stuff

  manpage for diderotc command; using asciidoc.

  remove debug timing from forward DFA.

  Replace dnorm with the new "unu dnorm" command from teem.  We also dynamically
  search for teem so that the location of teem at runtime can be different from
  the location of teem at compile time.

  removed OpenCL specific VSUB macros, since they are not used.

design changes

added deltas

Changed index type

Refactoring Sequential KDTree

Adding the parallel kdtree code

Fixes to code for Boids example

Fixed boids code

Finally, removing my old development repo

  refactoring

Adding KDTree code for c-tagert and parallel c-target

Updating boids test case

changes

  fix type error

Add dim to Fields

add operators

test

Fix printing

FIx printing

  more work on OpenCL support

  Working on OpenCL support

  Working on OpenCL support in vis12-cl branch.

  Create branch for adding OpenCL to vis12 branch.  Based on r2401 of the vis12 branch.

  minor edits

Tester

addchecks