Luke Kenneth Casson Leighton [Sat, 31 Jul 2021 18:38:07 +0000 (19:38 +0100)]
add SINGLE function to double2single, to replace manually-created version
Luke Kenneth Casson Leighton [Sat, 31 Jul 2021 18:37:48 +0000 (19:37 +0100)]
whoops, no ability to add comments in between functions in pseudocode
Luke Kenneth Casson Leighton [Sat, 31 Jul 2021 16:27:20 +0000 (17:27 +0100)]
add outer-inner RADIX2 iDCT unit test.
use FFT twin +/- MUL-ADD-SUB rather than the DCT +/- MUL-ADD-SUB
Dmitry Selyutin [Sat, 31 Jul 2021 13:48:08 +0000 (13:48 +0000)]
isa/bcd.mdwn: fix incorrect declaration
In pseudocode used in our markdown files, dc[16] does not mean that
we declare a variable of 16 bits; it only means that we access bit
16 of variable dc. Details:
https://bugs.libre-soc.org/show_bug.cgi?id=656#c15
Luke Kenneth Casson Leighton [Sat, 31 Jul 2021 13:28:25 +0000 (14:28 +0100)]
add SVP64 i-DCT unit test for inner butterfly, coefficients pre-computed
at present
Luke Kenneth Casson Leighton [Sat, 31 Jul 2021 13:19:46 +0000 (14:19 +0100)]
add i-DCT SVP64 unit test for outer butterfly
Luke Kenneth Casson Leighton [Sat, 31 Jul 2021 12:10:22 +0000 (13:10 +0100)]
add iDCT modes to interim svshape instruction pseudo-code
Luke Kenneth Casson Leighton [Sat, 31 Jul 2021 09:21:22 +0000 (10:21 +0100)]
corrections to iDCT demo printout
Dmitry Selyutin [Fri, 30 Jul 2021 18:49:56 +0000 (18:49 +0000)]
isa/bcd.mdwn: fix assignment operator
Dmitry Selyutin [Fri, 30 Jul 2021 18:33:03 +0000 (18:33 +0000)]
isa/bcd.mdwn: update addg6s pseudocode
Not only this pseudocode is wrong, since some actions are put inside
the loop, but also the algorithm as is cannot be translated due to
use of function call (carry_out). For more details, check these
discussions:
https://libre-soc.org/irclog/%23libre-soc.2021-07-28.log.html
https://bugs.libre-soc.org/show_bug.cgi?id=656
Dmitry Selyutin [Wed, 28 Jul 2021 17:24:16 +0000 (17:24 +0000)]
isatables: cbcdtd instruction
Dmitry Selyutin [Wed, 28 Jul 2021 17:24:06 +0000 (17:24 +0000)]
isatables: cdtbcd instruction
Luke Kenneth Casson Leighton [Fri, 30 Jul 2021 14:27:54 +0000 (15:27 +0100)]
got i-DCT yield schedule operational in fastdctlee.py test
Luke Kenneth Casson Leighton [Fri, 30 Jul 2021 10:15:37 +0000 (11:15 +0100)]
start adding i-dct schedule with debug-print, hard to tell what is going on
need to actually create a schedule-variant of i-DCT function, next
Luke Kenneth Casson Leighton [Thu, 29 Jul 2021 22:11:59 +0000 (23:11 +0100)]
random experimentation landed eventually on a "pass" of iDCT with
in-place half-swapping
Luke Kenneth Casson Leighton [Thu, 29 Jul 2021 21:58:02 +0000 (22:58 +0100)]
use coefficient table in iDCT
Luke Kenneth Casson Leighton [Thu, 29 Jul 2021 20:28:01 +0000 (21:28 +0100)]
move half-reverse to before outer butterfly in I-DCT
Luke Kenneth Casson Leighton [Thu, 29 Jul 2021 20:22:20 +0000 (21:22 +0100)]
get byte-swapping functional in inverse-dct proof-of-concept
Luke Kenneth Casson Leighton [Thu, 29 Jul 2021 19:52:29 +0000 (20:52 +0100)]
sigh, I-DCT had to reverse the order of middle loop to stop
overwrite-accumulation of iterative sum
Luke Kenneth Casson Leighton [Thu, 29 Jul 2021 17:12:53 +0000 (18:12 +0100)]
start on inverse dct, turning recursive to iterative
Luke Kenneth Casson Leighton [Wed, 28 Jul 2021 18:04:52 +0000 (19:04 +0100)]
start on inverse DCT, transforming to iterative
Luke Kenneth Casson Leighton [Wed, 28 Jul 2021 15:42:42 +0000 (16:42 +0100)]
got DCT LD-bit-rev demo operational in unit test
Luke Kenneth Casson Leighton [Wed, 28 Jul 2021 13:46:39 +0000 (14:46 +0100)]
fix LD/ST bitreverse with Matrix REMAP to instead be non-bitreversed.
slightly meaningless to bit-reverse on matrix LDs, but it is still possible
Luke Kenneth Casson Leighton [Wed, 28 Jul 2021 13:24:21 +0000 (14:24 +0100)]
argh, have LD-bitreverse select the offset from RA REMAP schedule
Luke Kenneth Casson Leighton [Wed, 28 Jul 2021 11:16:18 +0000 (12:16 +0100)]
add mode for half-swap, to be combined with LD-bit-reversed for loading DCT
data
Luke Kenneth Casson Leighton [Wed, 28 Jul 2021 11:03:51 +0000 (12:03 +0100)]
code comments
Luke Kenneth Casson Leighton [Tue, 27 Jul 2021 17:26:51 +0000 (18:26 +0100)]
fix test_power_decoder.py
Luke Kenneth Casson Leighton [Tue, 27 Jul 2021 16:20:07 +0000 (17:20 +0100)]
get DCT shortened table operational
Dmitry Selyutin [Tue, 27 Jul 2021 06:22:38 +0000 (06:22 +0000)]
isatables: addg6s instruction
Dmitry Selyutin [Mon, 26 Jul 2021 14:48:38 +0000 (14:48 +0000)]
power_enums: cbcdtd instruction
Dmitry Selyutin [Mon, 26 Jul 2021 14:48:25 +0000 (14:48 +0000)]
power_enums: cdtbcd instruction
Dmitry Selyutin [Mon, 26 Jul 2021 14:48:03 +0000 (14:48 +0000)]
power_enums: addg6s instruction
Luke Kenneth Casson Leighton [Tue, 27 Jul 2021 15:40:38 +0000 (16:40 +0100)]
adding reduced COS table DCT test
Luke Kenneth Casson Leighton [Tue, 27 Jul 2021 15:17:52 +0000 (16:17 +0100)]
add new DCT inner butterfly shorter COS-gen mode unit test
Luke Kenneth Casson Leighton [Tue, 27 Jul 2021 13:31:05 +0000 (14:31 +0100)]
fix new COSTABLE generator unit test,
cross-reference it to transcendentals scalar version
Luke Kenneth Casson Leighton [Tue, 27 Jul 2021 13:27:41 +0000 (14:27 +0100)]
fix up DCT modes for inner/outer butterfly,
add new costables schedule and set up REMAP correctly
Luke Kenneth Casson Leighton [Tue, 27 Jul 2021 13:20:57 +0000 (14:20 +0100)]
argh, LD/ST using DS has to be computed differently.
multiply DS by four, then compute SVP64 ELSTRIDE/UNITSTRIDE, then
*divide* DS by four afterwards
TODO, illegal instruction if the 2 LSBs are non-zero? does this ever occur?
Luke Kenneth Casson Leighton [Tue, 27 Jul 2021 13:19:31 +0000 (14:19 +0100)]
fix errors in detection of ffmadds (etc), enabling FFT twin-regs moed
when it should not have been. affected fcoss/fsins/fcfids
Luke Kenneth Casson Leighton [Tue, 27 Jul 2021 12:03:55 +0000 (13:03 +0100)]
clear persist bit if setvl explicitly called
Luke Kenneth Casson Leighton [Tue, 27 Jul 2021 11:35:55 +0000 (12:35 +0100)]
add new cos coefficient pre-computed and on-the-fly mode,
reorganise DCT modes due to needing more bits
Luke Kenneth Casson Leighton [Mon, 26 Jul 2021 15:22:48 +0000 (16:22 +0100)]
use ydimsz as sub-mode in DCT/FFT butterfly
Luke Kenneth Casson Leighton [Mon, 26 Jul 2021 15:22:26 +0000 (16:22 +0100)]
use std not stw in transcendentals ld/st-convert test
Luke Kenneth Casson Leighton [Mon, 26 Jul 2021 14:57:07 +0000 (15:57 +0100)]
add dct cos 8 table test
Luke Kenneth Casson Leighton [Sat, 24 Jul 2021 18:09:20 +0000 (19:09 +0100)]
add sv.fcoss SVP64Asm support
Luke Kenneth Casson Leighton [Sat, 24 Jul 2021 18:08:53 +0000 (19:08 +0100)]
add DS-Form support for sv.std
Luke Kenneth Casson Leighton [Sat, 24 Jul 2021 16:43:09 +0000 (17:43 +0100)]
added an extra SVP64 instruction, svstep, to replace setvl
"get end state" mode
Luke Kenneth Casson Leighton [Sat, 24 Jul 2021 12:06:47 +0000 (13:06 +0100)]
add experiment to convert int to float and multiply by PI etc.
for later use in DCT
Luke Kenneth Casson Leighton [Sat, 24 Jul 2021 10:43:23 +0000 (11:43 +0100)]
add ability to get current SVSHAPE indices into a register,
using setvl "Vertical First" test mode
Luke Kenneth Casson Leighton [Sat, 24 Jul 2021 09:59:28 +0000 (10:59 +0100)]
comments
Luke Kenneth Casson Leighton [Sat, 24 Jul 2021 07:31:16 +0000 (08:31 +0100)]
add DCT unit test combining DCT inner and outer butterfly
Luke Kenneth Casson Leighton [Sat, 24 Jul 2021 07:30:45 +0000 (08:30 +0100)]
make REMAP persistent (if persistence requested) even on svshape
Luke Kenneth Casson Leighton [Sat, 24 Jul 2021 06:59:10 +0000 (07:59 +0100)]
create schedule for calculating COS coefficient in DCT
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 17:29:43 +0000 (18:29 +0100)]
add DCT outer butterfly iterative overlapping ADD schedule
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 16:53:10 +0000 (17:53 +0100)]
add DCT outer butterfly svshape setup
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 15:57:31 +0000 (16:57 +0100)]
small inner DCT butterfly test, fix up order of fdmadds
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 14:41:54 +0000 (15:41 +0100)]
add DCT inner butterfly results test
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 13:48:53 +0000 (14:48 +0100)]
"fix" fdmadd DCT mul-add-sub unit test with values that will
not cause rounding. "good enough" for now
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 13:24:04 +0000 (14:24 +0100)]
add sv.fdmadds unit test
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 13:19:12 +0000 (14:19 +0100)]
add sv.fdmadds to SVP64Asm
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 13:17:59 +0000 (14:17 +0100)]
add DCT mul-add to CSV and enums
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 13:16:59 +0000 (14:16 +0100)]
add DCT variant of twin MUL-ADD. actually an add and a MUL-SUB
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 13:16:23 +0000 (14:16 +0100)]
add DCT butterfly mode into svremap
Luke Kenneth Casson Leighton [Fri, 23 Jul 2021 10:09:53 +0000 (11:09 +0100)]
set up submodes for SVSHAPE, to include DCT butterfly yielders
Luke Kenneth Casson Leighton [Thu, 22 Jul 2021 22:18:25 +0000 (23:18 +0100)]
split out 2nd dct outer butterfly scheduler
Luke Kenneth Casson Leighton [Thu, 22 Jul 2021 16:31:29 +0000 (17:31 +0100)]
half way through converting in-place dct to yield unit test
Luke Kenneth Casson Leighton [Thu, 22 Jul 2021 14:46:56 +0000 (15:46 +0100)]
add inner and outer yield version of DCT inner and out butterfly
Luke Kenneth Casson Leighton [Thu, 22 Jul 2021 14:24:36 +0000 (15:24 +0100)]
copy of halfrev2 algorithm updated
Luke Kenneth Casson Leighton [Thu, 22 Jul 2021 14:24:00 +0000 (15:24 +0100)]
simplification of halfrev2 algorithm (really neat)
Luke Kenneth Casson Leighton [Thu, 22 Jul 2021 14:12:43 +0000 (15:12 +0100)]
add REMAP DCT yield schedule function, TODO
Luke Kenneth Casson Leighton [Thu, 22 Jul 2021 14:12:09 +0000 (15:12 +0100)]
add hybrid LD-ST-bitreverse with REMAP as an experiment
Luke Kenneth Casson Leighton [Thu, 22 Jul 2021 12:17:06 +0000 (13:17 +0100)]
corrections to SVP64 LD/ST unit tests
Luke Kenneth Casson Leighton [Tue, 20 Jul 2021 21:04:45 +0000 (22:04 +0100)]
comments
Luke Kenneth Casson Leighton [Tue, 20 Jul 2021 21:03:44 +0000 (22:03 +0100)]
create cos table independent, outside of the inner loops
Luke Kenneth Casson Leighton [Tue, 20 Jul 2021 20:58:35 +0000 (21:58 +0100)]
cleanup
Luke Kenneth Casson Leighton [Tue, 20 Jul 2021 20:49:23 +0000 (21:49 +0100)]
add iterative list-reversing algorithm, replace recursive variant.
actually really simple (to implement in hardware)
Luke Kenneth Casson Leighton [Tue, 20 Jul 2021 16:04:22 +0000 (17:04 +0100)]
pre-reverse order of data indices in DCT so that *after* the inner
butterfly is done the data is in the correct order for the outer one.
this so that no data-swaps are needed
Luke Kenneth Casson Leighton [Tue, 20 Jul 2021 14:56:43 +0000 (15:56 +0100)]
temporary reordering after the DCT schedule is carried out, this removes
the need for *data* swaps.
Luke Kenneth Casson Leighton [Tue, 20 Jul 2021 13:08:06 +0000 (14:08 +0100)]
realised that SVSHAPE0-3 is not privileged
Luke Kenneth Casson Leighton [Tue, 20 Jul 2021 13:06:04 +0000 (14:06 +0100)]
add inner sub-loop testing from svstep Rc=1
Luke Kenneth Casson Leighton [Tue, 20 Jul 2021 09:01:08 +0000 (10:01 +0100)]
comments
Luke Kenneth Casson Leighton [Tue, 20 Jul 2021 08:58:09 +0000 (09:58 +0100)]
comments
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 19:45:25 +0000 (20:45 +0100)]
bit of a reorg, adding option to test end of inner loops of SVSTATE(s)
needed to pass the immediate to svstep as an option of which
SVSTATE0-3 to test
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 15:24:48 +0000 (16:24 +0100)]
do in-place swap
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 14:59:03 +0000 (15:59 +0100)]
annoying: missed out something in the unit test, not working yet
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 14:53:17 +0000 (15:53 +0100)]
simplify DCT code
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 14:50:23 +0000 (15:50 +0100)]
create coefficient table for DCT outside of loops
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 14:48:17 +0000 (15:48 +0100)]
update comments
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 14:13:25 +0000 (15:13 +0100)]
more comments
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 14:05:26 +0000 (15:05 +0100)]
update comments and license
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 13:47:59 +0000 (14:47 +0100)]
remove unneeded code
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 13:46:25 +0000 (14:46 +0100)]
no need for len(j) > 1 test, half of 1 is zero which stops swap anyway
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 13:44:20 +0000 (14:44 +0100)]
remove unneeded code
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 13:41:04 +0000 (14:41 +0100)]
swap the indices rather than the data in DCT top half: bizarrely this works!
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 12:25:51 +0000 (13:25 +0100)]
remove copy, use in-place with post-inner-loop swap
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 11:56:51 +0000 (12:56 +0100)]
add experimental order-reversing code (commented out) to DCT
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 11:08:38 +0000 (12:08 +0100)]
code comments
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 11:08:00 +0000 (12:08 +0100)]
whitespace cleanup
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 10:57:37 +0000 (11:57 +0100)]
move bit-reversing to before MULs in DCT
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 10:53:33 +0000 (11:53 +0100)]
reverse bit-order of in-place outer DCT butterfly
Luke Kenneth Casson Leighton [Mon, 19 Jul 2021 10:48:45 +0000 (11:48 +0100)]
finallygot the DCT outer butterfly correct