dispersedalgorithms
3 years ago- removal of "Helvetica" for better PDF/A conformance pdfa
spio [Wed, 10 Aug 2016 07:17:02 +0000 (09:17 +0200)]
- removal of "Helvetica" for better PDF/A conformance

4 years ago- variant which skips the failed runs instead of pretending they were zero newruby bomba/newruby origin/newruby
Josef Spillner [Mon, 27 Apr 2015 20:51:47 +0000 (22:51 +0200)]
- variant which skips the failed runs instead of pretending they were zero

4 years ago- visually improved plot
Josef Spillner [Mon, 27 Apr 2015 20:40:52 +0000 (22:40 +0200)]
- visually improved plot

4 years ago- new plot for aggregate runtimes
Josef Spillner [Mon, 27 Apr 2015 20:35:57 +0000 (22:35 +0200)]
- new plot for aggregate runtimes

4 years ago- additional measurement of aggregate functions
Josef Spillner [Mon, 27 Apr 2015 19:25:03 +0000 (21:25 +0200)]
- additional measurement of aggregate functions

4 years ago- further experimentation now also with dispersion and with debugging turned off
Josef Spillner [Mon, 27 Apr 2015 14:12:58 +0000 (16:12 +0200)]
- further experimentation now also with dispersion and with debugging turned off

4 years ago- first experimental measurement results in an overall comparison
Josef Spillner [Mon, 27 Apr 2015 06:16:56 +0000 (08:16 +0200)]
- first experimental measurement results in an overall comparison

4 years ago- marking of a bug which results in 10 FAILED issues in the systematictesting script
Josef Spillner [Sun, 26 Apr 2015 22:55:35 +0000 (00:55 +0200)]
- marking of a bug which results in 10 FAILED issues in the systematictesting script

4 years ago- make OPE sort of work on Debian 8.0 which has two competing Ruby versions but ope...
Josef Spillner [Sun, 26 Apr 2015 22:24:11 +0000 (00:24 +0200)]
- make OPE sort of work on Debian 8.0 which has two competing Ruby versions but ope-rb.rb isn't found in 2.1.0

4 years ago- OPE benchmarking in Ruby with associated plot
Josef Spillner [Tue, 21 Apr 2015 22:08:05 +0000 (00:08 +0200)]
- OPE benchmarking in Ruby with associated plot

4 years ago- plotfile
Josef Spillner [Tue, 21 Apr 2015 19:39:46 +0000 (21:39 +0200)]
- plotfile

4 years ago- alternative benchmark script for the direct C programming interfaces
Josef Spillner [Tue, 21 Apr 2015 19:33:36 +0000 (21:33 +0200)]
- alternative benchmark script for the direct C programming interfaces

4 years ago- joint multiplots to save some precious space
Josef Spillner [Tue, 21 Apr 2015 18:49:13 +0000 (20:49 +0200)]
- joint multiplots to save some precious space

4 years ago- modification to make minimal example work, despite still being unable to extract...
Josef Spillner [Tue, 21 Apr 2015 18:21:59 +0000 (20:21 +0200)]
- modification to make minimal example work, despite still being unable to extract the final result as a plain ulong number

4 years ago- inclusion of fastexpander measurements into the plot
Josef Spillner [Tue, 21 Apr 2015 16:28:24 +0000 (18:28 +0200)]
- inclusion of fastexpander measurements into the plot

4 years ago- fast C re-implementation of the fragment expander
Josef Spillner [Tue, 21 Apr 2015 16:14:27 +0000 (18:14 +0200)]
- fast C re-implementation of the fragment expander

4 years ago- modularisation of the RLE method
Josef Spillner [Tue, 21 Apr 2015 14:16:31 +0000 (16:16 +0200)]
- modularisation of the RLE method
- new plot on the runtime when compressing one million numbers with varying P

4 years ago- addition of relative compression ratio degradation as interesting metric
Josef Spillner [Tue, 21 Apr 2015 11:59:08 +0000 (13:59 +0200)]
- addition of relative compression ratio degradation as interesting metric

4 years ago- clarifications on the mechanisms of this script
Josef Spillner [Tue, 21 Apr 2015 11:19:49 +0000 (13:19 +0200)]
- clarifications on the mechanisms of this script

4 years ago- improved/consolidated 3D plots
Josef Spillner [Tue, 21 Apr 2015 09:38:21 +0000 (11:38 +0200)]
- improved/consolidated 3D plots

4 years agoMerge branch 'algorithms'
Josef Spillner [Sun, 19 Apr 2015 21:13:48 +0000 (23:13 +0200)]
Merge branch 'algorithms'

4 years ago- square root implementation origin/algorithms
Josef Spillner [Thu, 2 Apr 2015 17:34:36 +0000 (19:34 +0200)]
- square root implementation

4 years ago- analysis of Pythagoras triplets, which apparently won't work in dispersed mode
Josef Spillner [Thu, 2 Apr 2015 16:47:19 +0000 (18:47 +0200)]
- analysis of Pythagoras triplets, which apparently won't work in dispersed mode

4 years ago- dispersed multiplication across 3 nodes
Josef Spillner [Thu, 2 Apr 2015 16:37:50 +0000 (18:37 +0200)]
- dispersed multiplication across 3 nodes

4 years ago- analysis of dispersed prime numbers
Josef Spillner [Thu, 2 Apr 2015 15:53:56 +0000 (17:53 +0200)]
- analysis of dispersed prime numbers

4 years ago- syntactic sugar: PERMANENT becomes FOREVER bomba/pqueries origin/pqueries
Josef Spillner [Sun, 29 Mar 2015 18:50:03 +0000 (20:50 +0200)]
- syntactic sugar: PERMANENT becomes FOREVER

4 years ago- bogus workaround which is apparently needed to prevent crashes for the bsc-notify...
Josef Spillner [Sun, 29 Mar 2015 17:42:23 +0000 (19:42 +0200)]
- bogus workaround which is apparently needed to prevent crashes for the bsc-notify script

4 years ago- more robust repeated average executions with sporadic intermediate column content...
Josef Spillner [Sun, 29 Mar 2015 12:41:02 +0000 (14:41 +0200)]
- more robust repeated average executions with sporadic intermediate column content deletions

4 years ago- more ubiquitous use of the WHERE clause so that local aggregates work
Josef Spillner [Sun, 29 Mar 2015 11:29:43 +0000 (13:29 +0200)]
- more ubiquitous use of the WHERE clause so that local aggregates work

4 years ago- more consistent naming
Josef Spillner [Wed, 18 Mar 2015 22:43:26 +0000 (23:43 +0100)]
- more consistent naming

4 years ago- extension of the search script so that it now performs a combinatorial search for...
Josef Spillner [Wed, 18 Mar 2015 22:26:49 +0000 (23:26 +0100)]
- extension of the search script so that it now performs a combinatorial search for each word >=5 characters appearing in the text
- word length/precision correlation analysis

4 years ago- first implementation of a half-fuzzy + entirely heuristic search over XOR-redundant...
Josef Spillner [Wed, 18 Mar 2015 21:22:44 +0000 (22:22 +0100)]
- first implementation of a half-fuzzy + entirely heuristic search over XOR-redundant data

4 years ago- further analysis of redundancy distribution
Josef Spillner [Wed, 18 Mar 2015 20:36:54 +0000 (21:36 +0100)]
- further analysis of redundancy distribution

4 years ago- analysis of statistical properties of simple XOR redundancy
Josef Spillner [Wed, 18 Mar 2015 19:34:11 +0000 (20:34 +0100)]
- analysis of statistical properties of simple XOR redundancy

4 years ago- more adaptive removals
Josef Spillner [Mon, 9 Mar 2015 19:31:41 +0000 (20:31 +0100)]
- more adaptive removals

4 years ago- corrected composition of remote 16-bit dispersed + encrypted integers/floating...
Josef Spillner [Mon, 9 Mar 2015 18:08:15 +0000 (19:08 +0100)]
- corrected composition of remote 16-bit dispersed + encrypted integers/floating point numbers
-

4 years ago- use stable/convergent remote column naming, using the column's hashsum, so that...
Josef Spillner [Sun, 8 Mar 2015 19:27:55 +0000 (20:27 +0100)]
- use stable/convergent remote column naming, using the column's hashsum, so that multiple clients get to see the same data

4 years ago- introduction of query object class for somewhat more compact query handling
Josef Spillner [Wed, 25 Feb 2015 17:44:05 +0000 (18:44 +0100)]
- introduction of query object class for somewhat more compact query handling
- first experimental support for PERMANENT keyword after a query to make it a standing query (aka. event stream processing operator)

4 years ago- first experimental inotify support code
Josef Spillner [Wed, 25 Feb 2015 17:06:12 +0000 (18:06 +0100)]
- first experimental inotify support code
- this breaks pickle and therefore needs to move from stealthdb.py to stealthdb

4 years ago- bug fix for the not automatically propagated imports
Josef Spillner [Wed, 25 Feb 2015 13:58:21 +0000 (14:58 +0100)]
- bug fix for the not automatically propagated imports

4 years ago- first inotify test script
Josef Spillner [Wed, 25 Feb 2015 13:58:15 +0000 (14:58 +0100)]
- first inotify test script

4 years agoMerge remote-tracking branch 'daamobdav/32bit'
Josef Spillner [Wed, 4 Feb 2015 15:39:04 +0000 (16:39 +0100)]
Merge remote-tracking branch 'daamobdav/32bit'

4 years ago- printf format specifiers should use %z for size_t from sizeof()
Josef Spillner [Wed, 4 Feb 2015 15:36:17 +0000 (16:36 +0100)]
- printf format specifiers should use %z for size_t from sizeof()
- installation script should bail out on the first error

4 years ago- proof-of-concept WSGI adapter for three methods under GAE
Josef Spillner [Sat, 31 Jan 2015 15:23:09 +0000 (16:23 +0100)]
- proof-of-concept WSGI adapter for three methods under GAE

4 years ago- experimental runtime flavour: WSGI
Josef Spillner [Sat, 31 Jan 2015 14:09:44 +0000 (15:09 +0100)]
- experimental runtime flavour: WSGI

4 years ago- crude bugfix: do not crash on INT/SMALLINT with n=1
Josef Spillner [Fri, 23 Jan 2015 12:38:47 +0000 (13:38 +0100)]
- crude bugfix: do not crash on INT/SMALLINT with n=1

4 years ago- bugfix: exception wrapper around locateNS()
Josef Spillner [Fri, 23 Jan 2015 12:04:59 +0000 (13:04 +0100)]
- bugfix: exception wrapper around locateNS()

4 years ago- weird shuffling of the 16-bits special casing; now it applies only to numbers,...
Josef Spillner [Fri, 23 Jan 2015 11:45:11 +0000 (12:45 +0100)]
- weird shuffling of the 16-bits special casing; now it applies only to numbers, not strings, and only in debug mode as long as we don't know which bitsize was used

4 years ago- bugfix: s-suffix
Josef Spillner [Fri, 23 Jan 2015 10:40:23 +0000 (11:40 +0100)]
- bugfix: s-suffix

4 years ago- implementation of naive log-start sampling to experiment with its effects
Josef Spillner [Fri, 23 Jan 2015 09:22:27 +0000 (10:22 +0100)]
- implementation of naive log-start sampling to experiment with its effects

4 years ago- conversion of the singular optimisation goal specification into a list of potential...
Josef Spillner [Fri, 23 Jan 2015 08:42:18 +0000 (09:42 +0100)]
- conversion of the singular optimisation goal specification into a list of potentially conflicting goals

4 years ago- separation of stealthdb-cloud backend script into service interfaces for cloud...
Josef Spillner [Wed, 21 Jan 2015 21:03:49 +0000 (22:03 +0100)]
- separation of stealthdb-cloud backend script into service interfaces for cloud storage and computation

4 years ago- externalised verbosity control
Josef Spillner [Wed, 21 Jan 2015 20:39:30 +0000 (21:39 +0100)]
- externalised verbosity control

4 years ago- somewhat chattier cloud backend which logs all requests by default
Josef Spillner [Wed, 21 Jan 2015 15:10:05 +0000 (16:10 +0100)]
- somewhat chattier cloud backend which logs all requests by default

4 years ago- stealthdb-cloud --ip should take 127.0.0.1 into account as last resort
Josef Spillner [Wed, 21 Jan 2015 14:55:26 +0000 (15:55 +0100)]
- stealthdb-cloud --ip should take 127.0.0.1 into account as last resort
- improved complex test setup script which starts stealthdb by default (as it is often needed) and redirects empty qdbus return values to /dev/null

4 years ago- some more trial and error to find out about the non-working state: bits need to...
Josef Spillner [Tue, 13 Jan 2015 22:19:02 +0000 (23:19 +0100)]
- some more trial and error to find out about the non-working state: bits need to be high for paillier but low for paillierfastencint
- convenience python3-ported snapshot of the current paillier python libraries

4 years ago- bugfix, sort of ported from the 'c' flavour: use modmult instead of just %
Josef Spillner [Tue, 13 Jan 2015 21:54:20 +0000 (22:54 +0100)]
- bugfix, sort of ported from the 'c' flavour: use modmult instead of just %

4 years ago- a couple of additional tests for the integer-sized native implementation
Josef Spillner [Tue, 13 Jan 2015 21:41:16 +0000 (22:41 +0100)]
- a couple of additional tests for the integer-sized native implementation

4 years ago- naming conversion: fastenc -> fastencint
Josef Spillner [Tue, 13 Jan 2015 21:25:49 +0000 (22:25 +0100)]
- naming conversion: fastenc -> fastencint

4 years ago- rescued code from before the int->u-l-l conversion, keep it somehow for performance...
Josef Spillner [Tue, 13 Jan 2015 21:20:00 +0000 (22:20 +0100)]
- rescued code from before the int->u-l-l conversion, keep it somehow for performance reasons

4 years ago- slightly adapted testcase which still does not work
Josef Spillner [Tue, 13 Jan 2015 14:57:06 +0000 (15:57 +0100)]
- slightly adapted testcase which still does not work

4 years ago- first sample code to try out libpaillier
Josef Spillner [Tue, 13 Jan 2015 14:49:49 +0000 (15:49 +0100)]
- first sample code to try out libpaillier

4 years ago- more measuring and plotting, this time including the Bobo Raspberry-π cluster
Josef Spillner [Tue, 13 Jan 2015 13:08:41 +0000 (14:08 +0100)]
- more measuring and plotting, this time including the Bobo Raspberry-π cluster

4 years ago- plotfiles for the performance of both implementations on the Rumba system
Josef Spillner [Tue, 13 Jan 2015 10:27:17 +0000 (11:27 +0100)]
- plotfiles for the performance of both implementations on the Rumba system

4 years ago- first measurement results
Josef Spillner [Tue, 13 Jan 2015 09:54:07 +0000 (10:54 +0100)]
- first measurement results

4 years ago- bugfix: do not treat 2-digit numbers as tuples
Josef Spillner [Sat, 10 Jan 2015 22:50:37 +0000 (23:50 +0100)]
- bugfix: do not treat 2-digit numbers as tuples

4 years ago- experimental implementation of fast mass data transport endpoint complementing...
Josef Spillner [Sat, 10 Jan 2015 12:53:01 +0000 (13:53 +0100)]
- experimental implementation of fast mass data transport endpoint complementing cloud service endpoints

4 years ago- first attempt on implementing the MEDIAN() function
Josef Spillner [Fri, 9 Jan 2015 22:08:46 +0000 (23:08 +0100)]
- first attempt on implementing the MEDIAN() function

4 years ago- reinstating the crude workaround which at least works now for the special case...
Josef Spillner [Fri, 9 Jan 2015 21:27:15 +0000 (22:27 +0100)]
- reinstating the crude workaround which at least works now for the special case of SUM(INT) with k=2

4 years ago- sort-of working implementation: variable-width integers can now be retrieved, but...
Josef Spillner [Fri, 9 Jan 2015 14:50:20 +0000 (15:50 +0100)]
- sort-of working implementation: variable-width integers can now be retrieved, but not yet used for calculation

4 years ago- in theory, the bitwidth is now determined adaptively as a multiple-of-eighth of...
Josef Spillner [Fri, 9 Jan 2015 14:01:52 +0000 (15:01 +0100)]
- in theory, the bitwidth is now determined adaptively as a multiple-of-eighth of the input number width

4 years ago- reversal of the large-int splitting workaround due to it not passing the testsuite
Josef Spillner [Fri, 9 Jan 2015 13:15:48 +0000 (14:15 +0100)]
- reversal of the large-int splitting workaround due to it not passing the testsuite

The real solution would be to make use of the new bitwidth argument to the bitsplitter, barring any variable-width issues when aggregating over columns

4 years ago- adaptation of Python bitsplitter interface and implementation to additional bitwidt...
Josef Spillner [Fri, 9 Jan 2015 13:04:00 +0000 (14:04 +0100)]
- adaptation of Python bitsplitter interface and implementation to additional bitwidth parameter in the C library

4 years agoMerge branch 'mixedbitsizes' into testfix origin/testfix
Josef Spillner [Fri, 9 Jan 2015 12:51:45 +0000 (13:51 +0100)]
Merge branch 'mixedbitsizes' into testfix

4 years agoMerge branch 'master' into testfix
Josef Spillner [Fri, 9 Jan 2015 12:51:31 +0000 (13:51 +0100)]
Merge branch 'master' into testfix

4 years ago- exported diagram
Josef Spillner [Fri, 9 Jan 2015 08:22:17 +0000 (09:22 +0100)]
- exported diagram

4 years ago- option to export CSV data
Josef Spillner [Thu, 8 Jan 2015 22:23:36 +0000 (23:23 +0100)]
- option to export CSV data
- reference CSV logfiles which show the linear behaviours

4 years ago- proof-of-concept code which shows a saving potential for operating only on the...
Josef Spillner [Thu, 8 Jan 2015 22:10:58 +0000 (23:10 +0100)]
- proof-of-concept code which shows a saving potential for operating only on the MSB fragment stream

4 years ago- small yet insufficient bugfix: only convert when there is a comparison value to...
Josef Spillner [Tue, 6 Jan 2015 13:41:48 +0000 (14:41 +0100)]
- small yet insufficient bugfix: only convert when there is a comparison value to convert as well

The remaining bug is that sometimes, values look numeric but are strings (e.g. with replication), and therefore any comparison will be lexicographic instead o fnumeric

4 years ago- experimental version of the expander and corresponding diagram plotting (without...
Josef Spillner [Fri, 2 Jan 2015 19:39:17 +0000 (20:39 +0100)]
- experimental version of the expander and corresponding diagram plotting (without dgrid3d to avoid wrong interpolation)

4 years ago- improved algorithm: higher parameterisation robustness, more symbolic formula repre...
Josef Spillner [Fri, 2 Jan 2015 08:12:25 +0000 (09:12 +0100)]
- improved algorithm: higher parameterisation robustness, more symbolic formula representation

4 years ago- finished generator application with superformula for non-signal variability determi...
Josef Spillner [Thu, 1 Jan 2015 20:02:13 +0000 (21:02 +0100)]
- finished generator application with superformula for non-signal variability determination

4 years ago- apparently well-working intermediate version
Josef Spillner [Thu, 1 Jan 2015 19:57:28 +0000 (20:57 +0100)]
- apparently well-working intermediate version

4 years ago- working, but not optimal expansion (it misses a couple of possibilities to introduc...
Josef Spillner [Thu, 1 Jan 2015 16:37:47 +0000 (17:37 +0100)]
- working, but not optimal expansion (it misses a couple of possibilities to introduce more zeros)

4 years ago- more systematic testing and quorum breach detection
Josef Spillner [Thu, 1 Jan 2015 14:17:37 +0000 (15:17 +0100)]
- more systematic testing and quorum breach detection

4 years ago- proof-of-concept 1-bit fragment expansion
Josef Spillner [Thu, 1 Jan 2015 13:53:20 +0000 (14:53 +0100)]
- proof-of-concept 1-bit fragment expansion

4 years ago- proof-of-concept code on how to build a chunkable OPE scheme
Josef Spillner [Wed, 31 Dec 2014 21:54:08 +0000 (22:54 +0100)]
- proof-of-concept code on how to build a chunkable OPE scheme

4 years ago- correction of axis label
Josef Spillner [Wed, 31 Dec 2014 18:05:32 +0000 (19:05 +0100)]
- correction of axis label

4 years ago- much improved experimental determination of compression quality ratio
Josef Spillner [Wed, 31 Dec 2014 17:58:33 +0000 (18:58 +0100)]
- much improved experimental determination of compression quality ratio

4 years ago- graph production based on the CSV file
Josef Spillner [Wed, 31 Dec 2014 17:29:47 +0000 (18:29 +0100)]
- graph production based on the CSV file

4 years ago- experiment execution code to determine dependence of runlength compression ratio...
Josef Spillner [Wed, 31 Dec 2014 17:13:43 +0000 (18:13 +0100)]
- experiment execution code to determine dependence of runlength compression ratio on chunksize

4 years ago- proof-of-concept prototype for a stochastic process random number sequence generato...
Josef Spillner [Wed, 31 Dec 2014 16:40:42 +0000 (17:40 +0100)]
- proof-of-concept prototype for a stochastic process random number sequence generator, and a PCX-like runlength encoder for the resulting sequence

4 years ago- plotting variant b/w with tighter z-scale origin/mixedbitsizes
Josef Spillner [Wed, 31 Dec 2014 12:37:18 +0000 (13:37 +0100)]
- plotting variant b/w with tighter z-scale

4 years ago- updated from the current branch implementation
Josef Spillner [Wed, 31 Dec 2014 11:40:55 +0000 (12:40 +0100)]
- updated from the current branch implementation
- blacklist is now unnecessary, no more crashes on joining (although still content errors)

4 years agoMerge branch 'vbr' into mixedbitsizes
Josef Spillner [Wed, 31 Dec 2014 09:52:32 +0000 (10:52 +0100)]
Merge branch 'vbr' into mixedbitsizes

4 years ago- improved plotting instructions bomba/vbr origin/vbr
Josef Spillner [Wed, 31 Dec 2014 09:51:42 +0000 (10:51 +0100)]
- improved plotting instructions

4 years ago- first measurements of the older (vbr-branch) bitsplitter
Josef Spillner [Wed, 31 Dec 2014 09:38:20 +0000 (10:38 +0100)]
- first measurements of the older (vbr-branch) bitsplitter

4 years ago- new application to systematically determine the splitting performance
Josef Spillner [Wed, 31 Dec 2014 07:56:09 +0000 (08:56 +0100)]
- new application to systematically determine the splitting performance
- bugfix to determine buffer corruption when splitting files

4 years ago- some few additional corner cases; bump to 87% OKs
Josef Spillner [Tue, 30 Dec 2014 23:23:51 +0000 (00:23 +0100)]
- some few additional corner cases; bump to 87% OKs

4 years ago- special-casing of 2:9 and less special-casing of 2:7
Josef Spillner [Tue, 30 Dec 2014 23:12:16 +0000 (00:12 +0100)]
- special-casing of 2:9 and less special-casing of 2:7