Luke Kenneth Casson Leighton [Tue, 25 Oct 2022 19:00:16 +0000 (20:00 +0100)]
comments
Luke Kenneth Casson Leighton [Tue, 25 Oct 2022 18:58:24 +0000 (19:58 +0100)]
shadd pseudocode cleanup
Dmitry Selyutin [Tue, 25 Oct 2022 18:34:51 +0000 (21:34 +0300)]
pysvp64asm: support shadd/shadduw instructions
Dmitry Selyutin [Tue, 25 Oct 2022 18:34:30 +0000 (21:34 +0300)]
pysvp64asm: introduce more flexible Z23 wrapper
Dmitry Selyutin [Tue, 25 Oct 2022 18:16:25 +0000 (21:16 +0300)]
test_pysvp64dis: test shadd/shadduw instructions
Dmitry Selyutin [Mon, 24 Oct 2022 18:20:39 +0000 (21:20 +0300)]
bitmanip.mdwn: support shadd/shadduw instructions
Dmitry Selyutin [Mon, 24 Oct 2022 18:12:04 +0000 (21:12 +0300)]
minor_4.csv: support shadd/shadduw instructions
Luke Kenneth Casson Leighton [Mon, 24 Oct 2022 09:54:53 +0000 (10:54 +0100)]
add maxs. combined with cmp capability
https://bugs.libre-soc.org/show_bug.cgi?id=915
Luke Kenneth Casson Leighton [Sun, 23 Oct 2022 09:52:15 +0000 (10:52 +0100)]
use svshape2 instead of svindex for the 4th shape
(cycling through modulo4 shifts) as there is no
change of order, svindex wastes a regfile lookup
chacha20
Luke Kenneth Casson Leighton [Sat, 22 Oct 2022 16:37:04 +0000 (17:37 +0100)]
add extra pysvp64dis tests for divmod2du and maddedu
Luke Kenneth Casson Leighton [Sat, 22 Oct 2022 16:32:41 +0000 (17:32 +0100)]
argh, extremely annoying: 4-operand dsld/dsrd is not possible to
have EXTRA3, therefore it fails as a bigint operation. reverting
Luke Kenneth Casson Leighton [Sat, 22 Oct 2022 16:16:26 +0000 (17:16 +0100)]
remove redundant case_dsrd3
Luke Kenneth Casson Leighton [Sat, 22 Oct 2022 16:15:17 +0000 (17:15 +0100)]
bigint shuffle
* divmod2du moves to XO=58 (from XO=52)
* dsld/dsrd become Rc=1 and move to XO=52-55 in VA2-Form
* dsld/dsrd pseudocode no longer is overwrite with "sm" mode
* Z23 "sm" removed from fields.txt
Jacob Lifshay [Sat, 22 Oct 2022 00:46:58 +0000 (17:46 -0700)]
fix get_masked_reg and add test
Jacob Lifshay [Fri, 21 Oct 2022 23:37:14 +0000 (16:37 -0700)]
format code removing unused imports
Luke Kenneth Casson Leighton [Fri, 21 Oct 2022 12:29:07 +0000 (13:29 +0100)]
code-comments
Luke Kenneth Casson Leighton [Fri, 21 Oct 2022 12:00:41 +0000 (13:00 +0100)]
add 2nd outer loop, CTR 2 rounds, in chacha20 test
Luke Kenneth Casson Leighton [Fri, 21 Oct 2022 11:47:41 +0000 (12:47 +0100)]
move chacha20 to separate test, set/get masked regs to ISACaller
Luke Kenneth Casson Leighton [Fri, 21 Oct 2022 11:29:33 +0000 (12:29 +0100)]
move HASK, ROTL32, ROTL64, MASK32, into helper class
Luke Kenneth Casson Leighton [Fri, 21 Oct 2022 11:28:43 +0000 (12:28 +0100)]
use XLEN/2 for ROTL32 in fixedshift.mdwn
Luke Kenneth Casson Leighton [Thu, 20 Oct 2022 17:49:32 +0000 (18:49 +0100)]
comments
Luke Kenneth Casson Leighton [Thu, 20 Oct 2022 16:27:59 +0000 (17:27 +0100)]
add first chacha20 round test
Dmitry Selyutin [Wed, 19 Oct 2022 20:23:29 +0000 (23:23 +0300)]
sv_binutils_fptrans: fix registers generation
Dmitry Selyutin [Wed, 19 Oct 2022 18:23:15 +0000 (21:23 +0300)]
av.mdwn: fix missing bmask operand
Luke Kenneth Casson Leighton [Wed, 19 Oct 2022 10:49:51 +0000 (11:49 +0100)]
TODO, sort out remap indices order
Jacob Lifshay [Tue, 18 Oct 2022 05:49:42 +0000 (22:49 -0700)]
add test for scalar sv.maddedu
Jacob Lifshay [Tue, 18 Oct 2022 05:49:28 +0000 (22:49 -0700)]
add missing files to .gitignore
Dmitry Selyutin [Mon, 17 Oct 2022 18:52:59 +0000 (21:52 +0300)]
av.mdwn: fix Rc-augmented cprop instruction
Luke Kenneth Casson Leighton [Sun, 16 Oct 2022 11:24:50 +0000 (12:24 +0100)]
debug print correction
Luke Kenneth Casson Leighton [Sun, 16 Oct 2022 11:22:34 +0000 (12:22 +0100)]
sigh, have to use yield from on get_out_map()
Luke Kenneth Casson Leighton [Sun, 16 Oct 2022 11:08:56 +0000 (12:08 +0100)]
rewrite get_idx_out2 in ISACaller to split out
RS/out2 relationship
Luke Kenneth Casson Leighton [Sun, 16 Oct 2022 11:01:10 +0000 (12:01 +0100)]
rewrite get_idx_out in ISACaller to split out
RT/out relationship
Luke Kenneth Casson Leighton [Sun, 16 Oct 2022 10:30:06 +0000 (11:30 +0100)]
add unit test showing two svindex calls, found bugs,
needs resolving in ISACaller. REMAP application to RA/B/C/T/S is
not properly routing
Luke Kenneth Casson Leighton [Sun, 16 Oct 2022 10:27:24 +0000 (11:27 +0100)]
code-shuffle, rework get_idx_in() to separate out the in1/2/3 map
Luke Kenneth Casson Leighton [Fri, 14 Oct 2022 20:37:08 +0000 (21:37 +0100)]
whoops missed an update MEM(EA...) in pifixedstore
Dmitry Selyutin [Fri, 14 Oct 2022 19:17:35 +0000 (22:17 +0300)]
sv_binutils_fptrans: fix opcodes mode
Dmitry Selyutin [Fri, 14 Oct 2022 19:14:32 +0000 (22:14 +0300)]
power_insn: really skip sv. entries for PPC database
Dmitry Selyutin [Fri, 14 Oct 2022 19:07:07 +0000 (22:07 +0300)]
sv_binutils_fptrans: generate all permutations
Dmitry Selyutin [Thu, 13 Oct 2022 13:59:27 +0000 (16:59 +0300)]
pysvp64asm: fix coding style
Dmitry Selyutin [Fri, 7 Oct 2022 12:16:05 +0000 (15:16 +0300)]
power_insn: skip sv. instructions in PPC database
Dmitry Selyutin [Fri, 7 Oct 2022 12:15:09 +0000 (15:15 +0300)]
power_insn: fix AA match
Dmitry Selyutin [Fri, 7 Oct 2022 12:14:19 +0000 (15:14 +0300)]
power_insn: do not allow default records
Luke Kenneth Casson Leighton [Fri, 14 Oct 2022 13:14:53 +0000 (14:14 +0100)]
add max-with-getting-index-of vertical-first loop example
Konstantinos Margaritis [Fri, 14 Oct 2022 10:34:05 +0000 (10:34 +0000)]
small update in the max detection code
Luke Kenneth Casson Leighton [Fri, 14 Oct 2022 09:16:33 +0000 (10:16 +0100)]
SVP64RMModeDecode detects Post-Inc LDST-imm mode
Luke Kenneth Casson Leighton [Thu, 13 Oct 2022 06:45:46 +0000 (07:45 +0100)]
correct comments
Luke Kenneth Casson Leighton [Wed, 12 Oct 2022 21:47:48 +0000 (22:47 +0100)]
add in zeroing on test strncpy
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 20:14:34 +0000 (21:14 +0100)]
remove unneeded svstate from test
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 16:09:54 +0000 (17:09 +0100)]
add strncpy example - 6 instructions
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 14:17:45 +0000 (15:17 +0100)]
add sv.stwu/pi example in test_sv_load_store_postinc
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 13:11:04 +0000 (14:11 +0100)]
add ld/st-immediate "post-inc" mode support. unit test for LD
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 12:31:58 +0000 (13:31 +0100)]
add /pi to sv/trans/svp64.py and power_insns.py
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 12:30:56 +0000 (13:30 +0100)]
add new LD-Immediate Post constants
Konstantinos Margaritis [Fri, 14 Oct 2022 00:16:59 +0000 (00:16 +0000)]
first working version
Konstantinos Margaritis [Fri, 14 Oct 2022 00:16:37 +0000 (00:16 +0000)]
increase buffer size, fix svp64 address for r5
Luke Kenneth Casson Leighton [Wed, 12 Oct 2022 20:55:50 +0000 (21:55 +0100)]
add sv.divmod2du test, inverse of the sv.madded
using the same values
Luke Kenneth Casson Leighton [Wed, 12 Oct 2022 15:04:21 +0000 (16:04 +0100)]
comments clean-up on bigint big-mul case
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 16:09:28 +0000 (17:09 +0100)]
whoops ea not ra in pifixedstore.mdwn
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 14:05:58 +0000 (15:05 +0100)]
add Post-increment version of fixedstore.mdwn
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 16:09:08 +0000 (17:09 +0100)]
add asciidump option to Mem class
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 16:08:52 +0000 (17:08 +0100)]
whoops zero-error on masked-out
Konstantinos Margaritis [Tue, 11 Oct 2022 09:51:34 +0000 (09:51 +0000)]
WIP: add initial AV1 SVP64 porting
Konstantinos Margaritis [Tue, 11 Oct 2022 09:49:24 +0000 (09:49 +0000)]
move pypowersim_wrapper on its own
Luke Kenneth Casson Leighton [Tue, 11 Oct 2022 09:48:29 +0000 (10:48 +0100)]
add experimental post-increment fixedload pseudocode
Luke Kenneth Casson Leighton [Mon, 10 Oct 2022 19:29:43 +0000 (20:29 +0100)]
add elwidth overrides on Indexed REMAP, 8-bit example. reduces reg usage
Luke Kenneth Casson Leighton [Mon, 10 Oct 2022 19:02:59 +0000 (20:02 +0100)]
add elwidth overrides to get_idx_out2
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 23:35:34 +0000 (00:35 +0100)]
fix format in debug log
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 23:32:05 +0000 (00:32 +0100)]
forgot to add offset on GPR() get
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 23:11:15 +0000 (00:11 +0100)]
add elwidth overrides on destination (write) in ISACaller.
first two unit tests pass (sv.add/ew=8, sv.add/ew=32)
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 20:14:24 +0000 (21:14 +0100)]
split out base,offset in register decoding for elwidth overrides to work
previously, calculating the register number was fine, it was straight
64-bit reg indexed. however elwidths are *part-way* through registers
(packed) so need to compute the reg differently
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 19:42:17 +0000 (20:42 +0100)]
add 8-bit elwidth alu svp64 case
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 13:37:33 +0000 (14:37 +0100)]
add rfscv to major_19.csv, add test_pysvp64dis.py unit test
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 13:29:20 +0000 (14:29 +0100)]
drat
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 13:28:00 +0000 (14:28 +0100)]
add sc and scv support after moving from major.csv to extra.csv
this now involves a laborious brute-force search looking for anything
with an extra.csv path, in order to prioritise the (full) 32-bit
pattern-match over e.g. MAJOR XO=17.
attn should also work (but currently does not, no idea why, possibly
because it should actually be in major.csv?
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 12:26:09 +0000 (13:26 +0100)]
vector name "RSp" not recognised in sv.stq, added as example
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 12:23:03 +0000 (13:23 +0100)]
add stq to CSV files and unit test to test_pysvp64dis.py
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 12:14:13 +0000 (13:14 +0100)]
separate out DQ and DS to separate custom_immediates
D was unhappy about being a custom_field as an immediate.
better: create SignedImmediate class deriving from ImmediateOperand
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 12:11:33 +0000 (13:11 +0100)]
use new base-class EXTSOperand, derive from ImmediateOperand
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 11:45:30 +0000 (12:45 +0100)]
convert TargetAddrOperand to base class EXTSOperand
(about to do DQ/DS operand)
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 11:22:56 +0000 (12:22 +0100)]
add lq and CONST_DQ
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 10:11:28 +0000 (11:11 +0100)]
restore tests, accidentally disabled
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 10:09:39 +0000 (11:09 +0100)]
add CY operand to fields.txt, in Z23-Form
(missing from Power ISA v3.0B and v3.1 spec!)
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 10:04:05 +0000 (11:04 +0100)]
add XER bits to register enums
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 09:29:38 +0000 (10:29 +0100)]
add addex to csv and sv_analysis db. also needs CryIn.OV enum
added quick test_pysvp64dis.py test too
Luke Kenneth Casson Leighton [Sat, 8 Oct 2022 09:10:15 +0000 (10:10 +0100)]
misnamed instruction, lfiwzx
Luke Kenneth Casson Leighton [Fri, 7 Oct 2022 12:47:03 +0000 (13:47 +0100)]
more work on inssort. add useful reg-dump in ISACaller
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 21:41:04 +0000 (22:41 +0100)]
nope. failfirst needs to always save the result, but truncate VL *after*.
https://bugs.libre-soc.org/show_bug.cgi?id=936
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 19:19:42 +0000 (20:19 +0100)]
fix fail-first to exclude failed element in VLi=0 mode
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 17:34:38 +0000 (18:34 +0100)]
sort out CROPs fail-first in ISACaller. needed to take a copy of CR
for when sv.cmp (and other pseudocode) *overwrites* CR and it needs
restoring (when VLI=0). also needed to identify 3-bit and 5-bit
ffirst mode, and extract bottom 2 bits of BF
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 16:26:24 +0000 (17:26 +0100)]
make fail-first cope with sv.cmp which uses CR[BF]
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 14:59:54 +0000 (15:59 +0100)]
add insert sort svp64 test
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 14:58:11 +0000 (15:58 +0100)]
search for BF in registers to over-ride Vector lookup into CR Register
CR[32+BF...] is used, so it is more complex
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 13:36:23 +0000 (14:36 +0100)]
starting to add sv.cmp support and failfirst, had to add
SVMode to SVP64RMModeDecode to identify the different RM modes first
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 12:16:28 +0000 (13:16 +0100)]
add PredicateBaseRM decode to CR Ops Simple mode as well as ff=3-bit
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 11:57:57 +0000 (12:57 +0100)]
add vli mode to ff=5 CR ops
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 11:51:42 +0000 (12:51 +0100)]
whoops must only be PredicateBaseRM in CROpFF5RM
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 11:43:00 +0000 (12:43 +0100)]
add sv.cmp (ffirst-5) decode/encode asm support
* sv/trans/svp64.py needed a totally different ffirst handling
* CROpFF5RM needs to derive from FFPRRc0BaseRM and PredicateWidthBaseRM
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 11:18:30 +0000 (12:18 +0100)]
slightly different crops failfirst mode bits
Luke Kenneth Casson Leighton [Thu, 6 Oct 2022 10:53:59 +0000 (11:53 +0100)]
add sv.cmp and try fail-first test_pysvp64dist.py
Luke Kenneth Casson Leighton [Sun, 2 Oct 2022 11:38:32 +0000 (12:38 +0100)]
remove complaints about standard Cray-style Vectors for the past 40 years
being able to have VL set to zero dynamically at runtime.
it is not appropriate to have complaints about 40-year-old standard
canonical behaviour of Cray Vectors in source code comments
VL=0 being set dynamically at runtime based on an algorithm input
sets the operations to nop because that is expected behaviour.
to not have VL=0 would be catastrophically inconvenient: it would
require either Illegal-Instruction traps to be raised, or Condition
Codes to be set and followed up with instructions to test and
branch for the dynamic condition when RA was zero, or to pre-test
for RA or CTR pre-being-zero prior to entry into a loop.
Data-Dependent Fail-First would become irrevocably damaged as well
as it is possible for VL to be set to zero at that time (first
test fails)
sv.bc would also be irrevocably damaged as it would no longer
be possible in VLSET mode to have VL become truncated to zero
dynamically based on a Condition Code.
in all it is pretty catastrophic to the entire Cray-Vector paradigm
and would severely damage SimpleV to disallow VL=0 purely for
arbitrary "convenience"