24/10/2022 - FracNEON updated with more results in LINUX stuff. All benchmarks are updated within one archive including source and now all calibrate and then loop for 10 seconds each for more precise timing measurement. Also non graphical versions are included.
25/06/2022 - FracNEON_sp_mt/dp_mt V1.0 and results added to LINUX stuff. Now that benchmark using multiple threads/cores
03/08/2021 - FracNEON_dp_opt V1.0 and results added to LINUX stuff. Both, single and double precision versions contain now also text-only versions
16/06/2021 - FracNEON_sp_opt V1.0 and results added to LINUX stuff
08/06/2021 - CodeCraft#4 compo link added to RISC OS section
08/06/2021 - Links to pouet releases added to x86 and RISC OS sections so you can always check there for my latest demoscene related releases. Other stuff will still be hosted directly here.
08/06/2021 - LINUX section added
10/12/2016 - x86 Stuff: Added "You are invaded" and "Kaliskop", two 256 byte intros for "Function" and "Demosplash" demo parties 2016.
10/05/2016 - x86 Stuff: Release of "Crystal Comet". A tiny 128 byte DOS intro released at "Outline" demo party 2016.
10/03/2016 - RISC OS : updated VFPBasicExamples with the new FracCollection code, updated !MemSpeed with the latest version and added !LoadStore within the same archive. Updated FixFrac with the new FracCollection code. Updated FireBench for 32bit colour modes.
10/03/2016 - Release of "FracCollection" for RISC OS. Collection and update of my old Mandelbrot fractal benchmarks.
16/01/2016 - Release of "!3dBall" for RISC OS. Another GLSL experiment ported to assembler.
16/09/2015 - Release of "fishy" and "aldeshul" for x86. Two tiny 256 byte DOS intros released at "Outline" and "Function" demo party this year.
05/03/2015 - Update of "!radix_sort" for RISC OS
25/11/2014 - Release of "!BasicASMv6v7" for RISC OS. Provides a work-around to assemble ARM v6/v7 opcodes within the Basic Assembler.
03/11/2014 - Update of "!radix_sort" for RISC OS
22/10/2014 - Release of "!radix_sort" for RISC OS
03/01/2014 - Created the "VFP/NEON" section, some hopefully useful information for VFP/NEON assembly coders.
15/12/2013 - Release of "GLSL without GLSL" for RISC OS
22/06/2012 - Release of "!FracZoom" for RISC OS
13/03/2012 - Release of "!iq_tunnel" and "PB_CPUdetect" for RISC OS and VFPBasic_Examples (FracNeon/FracVFP/BiliZo) updated for Pandaboard.
20/10/2011 - More results added for KMB V 0.53I. AMD Bulldozer and C-50 added.
29/08/2011 - Adaption of "BiLiZo" added to the VFPBasic_Examples.
09/08/2011 - VFPBasic_Examples added in the RISC OS section. Some applications to use with the VFP/NEON compatible BBC Basic Assembler done by TBA Software for the Beagle Board.
20/03/2011 - !MemSpeed added in the RISC OS section. Some applications for memory transfer speed testing, especially regarding the Beagle Board.
27/01/2011 - FracNEON updated to V0.2, 25% speed increase, now comparable code (okay, still single precision) to my x86 stuff.
22/01/2011 - More results added for KMB V 0.53I. Especially for the new Intel Sandy Bridge CPU's.
14/01/2011 - FracNEON and FracVFP added for RISC OS. Demonstrates the use of the NEON and VFP processor units of the Cortex-A8 CPU on the Beagle Board.
26/09/2010 - Added a special tuned version called "FixFrac_64BitMUL" of FixFrac for RISC OS, also to compare directly to my x86 results. The file is added in the same archive like the original version, also a ReadMe was added.
15/09/2010 - FireBench for RISC OS updated for better compatibility with the BeagleBoard
07/06/2009 - KMB V 0.53I - SSE4.1 version added, just a small change but gaining about 2% on SSE4.1 capable CPU's.
10/04/2009 - More results added.
11/01/2009 - KMB V 0.53I - Version released, minor changes regarding threading and result display and Windows issues.
13/11/2008 - More results added. Especially Intel Core i7 920.
04/09/2008 - More results added. Especially Via Nano and Intel Atom provide interesting data.
22/06/2008 - More results added.
01/06/2008 - KMB V 0.53H - Version released, another up to 40 % faster FPU version. SSE2 version changed also to 3 times loop unrolling, giving another plus of up to 14 %.
13/04/2008 - KMB V 0.53G - Version released, finally a 40 to 90 % faster FPU version. Good old thing isn't dead yet. SSE2 version unchanged.
02/04/2008 - KMB V 0.53F - Version released, almost 20 % faster than KMB V 0.53E by implementing 6 exits of the inner loop and doing the iteration count in integer.
03/02/2008 - More results added. Strange behaviour of the Hyperthreading CPU's.
17/01/2008 - KMB V 0.53E - Version, seems to be best for AMD and INTEL except Pentium M.
13/01/2008 - KMB V 0.53D - Version, another speed up and now also seperate AMD and INTEL optimized versions.
05/01/2008 - KMB V 0.53C - Version, major speedup by implementing lots of ideas from Xorpd/Quickman code.
08/12/2007 - First KMB - AMD Phenom Results - Major Update of the KMB Benchmark table (x86 stuff) and also added the results and link of Xorpd!s x64 OS version called KMB V 0.57.
11/07/2006 - KMB V 0.53 MT released, major bugfix for systems with more than 4 CPU cores. More results.
29/06/2006 - Major Update of the KMB Benchmark table (x86 stuff).
08/06/2006 - KMB V 0.52 MT released, supports now up to 16 threads, so multi cpu systems benefit a lot, also dual core hyper threading CPUs. Only neglectable speed loss for other CPUs.
21/05/2006 - Updated the KMB Benchmark table (x86 stuff) again. Lot's of KMB V 0.5 results replaced or taken out.
14/05/2006 - KMB V 0.51 MT released, only minor change, slightly differen results.
11/05/2006 - KMB V 0.5 MT Results updated, very interesting values for hyper-threading CPU's.
08/05/2006 - KMB V 0.5 MT released. Old versions and results obsolete. Now supports Multi-Threading, achieving almost double perfomrace on dual core CPUs !
30/04/2006 - Updated the KMB Benchmark table (x86 stuff) again.
24/04/2006 - Updated the KMB Benchmark table (x86 stuff) with lots of results. Also a INCLUDE directory for FASM to compile it successfully is available for download now.
17/04/2006 - Finaly online, including my x86 and Java code.
|
|
