[1] Added prnymmr/dumpymmr, prnxmmr/dumpxmmr, prnmmxr/dumpmmxr. The reversed versions of the same routines, so that the SSE / AVX data appear as they are declared in you data sections. Useful for matrix and matrix transposition works.
Regards
S.A.R
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Just added "fpuset.zip" to the main distribution. Useful for almost all kind of floating-point purposes. This is all you need if you want to deal with FPU, MMX, SSE and AVX operations. It supports up to byte vector operations. Win64 only.
Also a small bug-fix to CPULIB (prnxmm/prnxmmr) affecting "cpu64w" (source and binary). BASELIB are not affected. Revision stays at 4.1.6.
See you next year.
Regards
S.A.R
Sandakan, Sabah, Malaysia
December 5th, 2018.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
A number of users have requested for full ABI64-compliant versions of CPULIB. So for my weekend project, I modified CPULIB, plus other additions and deprecations, to come up with CPU2.0.
So starting from next year, I will be providing long-term support for CPU2.0 only. Sorry at this time, it comes in binaries only. No source. And this version offers no 32-bit variants. Use CPULIB / BASELIB for 32-bit variants.
Regards
S.A.R
December 9th, 2018.
Last edit: Soffian Abdul Rasad 2018-12-09
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Added 24 new array-display routines, offering both forward (little-endian) and reversed views. These new routines are now used in all XMM/YMM/MMX viewing routines.
timer_start now return in XMM0 instead of RAX. Now after calling timer_stop(), you can view the time difference as double type via prndbl function.
Regards
S.A.R
December 12th, 2018
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
CPU2.0 Revision 1.0.4 [1] Added formatted versions of dumpymm/dummxmm/dumpmmx, for both forward and reversed variants. [2] Added some more "aprnxxxx" array routines to support formatted versions of [1] [3] Added prnhexbf, prnintdf, prnintwf, prnintuf,prnintf, prndblf and prnfltf to support [1] and [2] [4] Fixed a small bug in 64-bit Linux version of "dumpxmm". [5] Dropped "fpuset" library from circulation. All functions have been transferred to CPU2.0
Regards
S.A.R
December 18th, 2018
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I am creating this forum / topic to open up for any Q&A, discussions or anything, especially suggestions and bug reports.
Regards
S.A.R
Last edit: Soffian Abdul Rasad 2018-11-05
Just released Revision 4.1.2.
Added "mem_tofile" routine to allow memory content to be saved to a new file.
Just make sure you the "new file" name doesn't resemble any of your current files.
Enjoy.
Revision 4.1.3 released. (November 13, 2018)
Improvements to a number of routines
Regards
S.A.R
Revision 4.1.5 Release. (November 21, 2018)
[1] Added prnymmr/dumpymmr, prnxmmr/dumpxmmr, prnmmxr/dumpmmxr. The reversed versions of the same routines, so that the SSE / AVX data appear as they are declared in you data sections. Useful for matrix and matrix transposition works.
Regards
S.A.R
Revision 4.1.6 is out (December 3rd, 2018).
Major overhauls to all MMX/SSE/AVX viewing routines. They are now fully-featured so you can deal with MMX/SSE/AVX instructions up to byte granularity.
Use them for good purposes.
Regards
S.A.R
Just added "fpuset.zip" to the main distribution. Useful for almost all kind of floating-point purposes. This is all you need if you want to deal with FPU, MMX, SSE and AVX operations. It supports up to byte vector operations. Win64 only.
Also a small bug-fix to CPULIB (prnxmm/prnxmmr) affecting "cpu64w" (source and binary). BASELIB are not affected. Revision stays at 4.1.6.
See you next year.
Regards
S.A.R
Sandakan, Sabah, Malaysia
December 5th, 2018.
A number of users have requested for full ABI64-compliant versions of CPULIB. So for my weekend project, I modified CPULIB, plus other additions and deprecations, to come up with CPU2.0.
So starting from next year, I will be providing long-term support for CPU2.0 only. Sorry at this time, it comes in binaries only. No source. And this version offers no 32-bit variants. Use CPULIB / BASELIB for 32-bit variants.
Regards
S.A.R
December 9th, 2018.
Last edit: Soffian Abdul Rasad 2018-12-09
This is CPU2.0 Revision 1.0.1 (First revision)
Added 24 new array-display routines, offering both forward (little-endian) and reversed views. These new routines are now used in all XMM/YMM/MMX viewing routines.
timer_start now return in XMM0 instead of RAX. Now after calling timer_stop(), you can view the time difference as double type via prndbl function.
Regards
S.A.R
December 12th, 2018
CPU2.0 Revision 1.0.2
[1] Minor bug fix to "fpu_stack" (Win64 version)
[2] Improvements to selected routines
Regards
S.A.R
December 14th, 2018
CPU2.0 Revision 1.0.3
[1] Added "str_lengthd". Find a length of a string using a delimiter byte instead of null-terminated byte.
Regards
S.A.R
December 15th, 2018
CPU2.0 Revision 1.0.4
[1] Added formatted versions of dumpymm/dummxmm/dumpmmx, for both forward and reversed variants.
[2] Added some more "aprnxxxx" array routines to support formatted versions of [1]
[3] Added prnhexbf, prnintdf, prnintwf, prnintuf,prnintf, prndblf and prnfltf to support [1] and [2]
[4] Fixed a small bug in 64-bit Linux version of "dumpxmm".
[5] Dropped "fpuset" library from circulation. All functions have been transferred to CPU2.0
Regards
S.A.R
December 18th, 2018