These jumps are always forwards, and it's more efficient in the VM to
decode an unsigned argument. These opcodes are already optimised versions
of the sequence "dup-top pop-jump-if-x pop" so it doesn't hurt generality
to optimise them further.
Signed-off-by: Damien George <damien@micropython.org>
This commit introduces changes:
- All jump opcodes are changed to have variable length arguments, of either
1 or 2 bytes (previously they were fixed at 2 bytes). In most cases only
1 byte is needed to encode the short jump offset, saving bytecode size.
- The bytecode emitter now selects 1 byte jump arguments when the jump
offset is guaranteed to fit in 1 byte. This is achieved by checking if
the code size changed during the last pass and, if it did (if it shrank),
then requesting that the compiler make another pass to get the correct
offsets of the now-smaller code. This can continue multiple times until
the code stabilises. The code can only ever shrink so this iteration is
guaranteed to complete. In most cases no extra passes are needed, the
original 4 passes are enough to get it right by the 4th pass (because the
2nd pass computes roughly the correct labels and the 3rd pass computes
the correct size for the jump argument).
This change to the jump opcode encoding reduces .mpy files and RAM usage
(when bytecode is in RAM) by about 2% on average.
The performance of the VM is not impacted, at least within measurment of
the performance benchmark suite.
Code size is reduced for builds that include a decent amount of frozen
bytecode. ARM Cortex-M builds without any frozen code increase by about
350 bytes.
Signed-off-by: Damien George <damien@micropython.org>
The bytecode state variables mp_showbc_code_start and mp_showbc_constants
have been removed and made local variables passed into the various
functions.
As part of this, the DECODE_PTR macro is fixed so it extracts the relevant
pointer from the child_table (a regression introduced in
f2040bfc7e).
Signed-off-by: Damien George <damien@micropython.org>
Background: .mpy files are precompiled .py files, built using mpy-cross,
that contain compiled bytecode functions (and can also contain machine
code). The benefit of using an .mpy file over a .py file is that they are
faster to import and take less memory when importing. They are also
smaller on disk.
But the real benefit of .mpy files comes when they are frozen into the
firmware. This is done by loading the .mpy file during compilation of the
firmware and turning it into a set of big C data structures (the job of
mpy-tool.py), which are then compiled and downloaded into the ROM of a
device. These C data structures can be executed in-place, ie directly from
ROM. This makes importing even faster because there is very little to do,
and also means such frozen modules take up much less RAM (because their
bytecode stays in ROM).
The downside of frozen code is that it requires recompiling and reflashing
the entire firmware. This can be a big barrier to entry, slows down
development time, and makes it harder to do OTA updates of frozen code
(because the whole firmware must be updated).
This commit attempts to solve this problem by providing a solution that
sits between loading .mpy files into RAM and freezing them into the
firmware. The .mpy file format has been reworked so that it consists of
data and bytecode which is mostly static and ready to run in-place. If
these new .mpy files are located in flash/ROM which is memory addressable,
the .mpy file can be executed (mostly) in-place.
With this approach there is still a small amount of unpacking and linking
of the .mpy file that needs to be done when it's imported, but it's still
much better than loading an .mpy from disk into RAM (although not as good
as freezing .mpy files into the firmware).
The main trick to make static .mpy files is to adjust the bytecode so any
qstrs that it references now go through a lookup table to convert from
local qstr number in the module to global qstr number in the firmware.
That means the bytecode does not need linking/rewriting of qstrs when it's
loaded. Instead only a small qstr table needs to be built (and put in RAM)
at import time. This means the bytecode itself is static/constant and can
be used directly if it's in addressable memory. Also the qstr string data
in the .mpy file, and some constant object data, can be used directly.
Note that the qstr table is global to the module (ie not per function).
In more detail, in the VM what used to be (schematically):
qst = DECODE_QSTR_VALUE;
is now (schematically):
idx = DECODE_QSTR_INDEX;
qst = qstr_table[idx];
That allows the bytecode to be fixed at compile time and not need
relinking/rewriting of the qstr values. Only qstr_table needs to be linked
when the .mpy is loaded.
Incidentally, this helps to reduce the size of bytecode because what used
to be 2-byte qstr values in the bytecode are now (mostly) 1-byte indices.
If the module uses the same qstr more than two times then the bytecode is
smaller than before.
The following changes are measured for this commit compared to the
previous (the baseline):
- average 7%-9% reduction in size of .mpy files
- frozen code size is reduced by about 5%-7%
- importing .py files uses about 5% less RAM in total
- importing .mpy files uses about 4% less RAM in total
- importing .py and .mpy files takes about the same time as before
The qstr indirection in the bytecode has only a small impact on VM
performance. For stm32 on PYBv1.0 the performance change of this commit
is:
diff of scores (higher is better)
N=100 M=100 baseline -> this-commit diff diff% (error%)
bm_chaos.py 371.07 -> 357.39 : -13.68 = -3.687% (+/-0.02%)
bm_fannkuch.py 78.72 -> 77.49 : -1.23 = -1.563% (+/-0.01%)
bm_fft.py 2591.73 -> 2539.28 : -52.45 = -2.024% (+/-0.00%)
bm_float.py 6034.93 -> 5908.30 : -126.63 = -2.098% (+/-0.01%)
bm_hexiom.py 48.96 -> 47.93 : -1.03 = -2.104% (+/-0.00%)
bm_nqueens.py 4510.63 -> 4459.94 : -50.69 = -1.124% (+/-0.00%)
bm_pidigits.py 650.28 -> 644.96 : -5.32 = -0.818% (+/-0.23%)
core_import_mpy_multi.py 564.77 -> 581.49 : +16.72 = +2.960% (+/-0.01%)
core_import_mpy_single.py 68.67 -> 67.16 : -1.51 = -2.199% (+/-0.01%)
core_qstr.py 64.16 -> 64.12 : -0.04 = -0.062% (+/-0.00%)
core_yield_from.py 362.58 -> 354.50 : -8.08 = -2.228% (+/-0.00%)
misc_aes.py 429.69 -> 405.59 : -24.10 = -5.609% (+/-0.01%)
misc_mandel.py 3485.13 -> 3416.51 : -68.62 = -1.969% (+/-0.00%)
misc_pystone.py 2496.53 -> 2405.56 : -90.97 = -3.644% (+/-0.01%)
misc_raytrace.py 381.47 -> 374.01 : -7.46 = -1.956% (+/-0.01%)
viper_call0.py 576.73 -> 572.49 : -4.24 = -0.735% (+/-0.04%)
viper_call1a.py 550.37 -> 546.21 : -4.16 = -0.756% (+/-0.09%)
viper_call1b.py 438.23 -> 435.68 : -2.55 = -0.582% (+/-0.06%)
viper_call1c.py 442.84 -> 440.04 : -2.80 = -0.632% (+/-0.08%)
viper_call2a.py 536.31 -> 532.35 : -3.96 = -0.738% (+/-0.06%)
viper_call2b.py 382.34 -> 377.07 : -5.27 = -1.378% (+/-0.03%)
And for unix on x64:
diff of scores (higher is better)
N=2000 M=2000 baseline -> this-commit diff diff% (error%)
bm_chaos.py 13594.20 -> 13073.84 : -520.36 = -3.828% (+/-5.44%)
bm_fannkuch.py 60.63 -> 59.58 : -1.05 = -1.732% (+/-3.01%)
bm_fft.py 112009.15 -> 111603.32 : -405.83 = -0.362% (+/-4.03%)
bm_float.py 246202.55 -> 247923.81 : +1721.26 = +0.699% (+/-2.79%)
bm_hexiom.py 615.65 -> 617.21 : +1.56 = +0.253% (+/-1.64%)
bm_nqueens.py 215807.95 -> 215600.96 : -206.99 = -0.096% (+/-3.52%)
bm_pidigits.py 8246.74 -> 8422.82 : +176.08 = +2.135% (+/-3.64%)
misc_aes.py 16133.00 -> 16452.74 : +319.74 = +1.982% (+/-1.50%)
misc_mandel.py 128146.69 -> 130796.43 : +2649.74 = +2.068% (+/-3.18%)
misc_pystone.py 83811.49 -> 83124.85 : -686.64 = -0.819% (+/-1.03%)
misc_raytrace.py 21688.02 -> 21385.10 : -302.92 = -1.397% (+/-3.20%)
The code size change is (firmware with a lot of frozen code benefits the
most):
bare-arm: +396 +0.697%
minimal x86: +1595 +0.979% [incl +32(data)]
unix x64: +2408 +0.470% [incl +800(data)]
unix nanbox: +1396 +0.309% [incl -96(data)]
stm32: -1256 -0.318% PYBV10
cc3200: +288 +0.157%
esp8266: -260 -0.037% GENERIC
esp32: -216 -0.014% GENERIC[incl -1072(data)]
nrf: +116 +0.067% pca10040
rp2: -664 -0.135% PICO
samd: +844 +0.607% ADAFRUIT_ITSYBITSY_M4_EXPRESS
As part of this change the .mpy file format version is bumped to version 6.
And mpy-tool.py has been improved to provide a good visualisation of the
contents of .mpy files.
In summary: this commit changes the bytecode to use qstr indirection, and
reworks the .mpy file format to be simpler and allow .mpy files to be
executed in-place. Performance is not impacted too much. Eventually it
will be possible to store such .mpy files in a linear, read-only, memory-
mappable filesystem so they can be executed from flash/ROM. This will
essentially be able to replace frozen code for most applications.
Signed-off-by: Damien George <damien@micropython.org>
This commit removes all parts of code associated with the existing
MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE optimisation option, including the
-mcache-lookup-bc option to mpy-cross.
This feature originally provided a significant performance boost for Unix,
but wasn't able to be enabled for MCU targets (due to frozen bytecode), and
added significant extra complexity to generating and distributing .mpy
files.
The equivalent performance gain is now provided by the combination of
MICROPY_OPT_LOAD_ATTR_FAST_PATH and MICROPY_OPT_MAP_LOOKUP_CACHE (which has
been enabled on the unix port in the previous commit).
It's hard to provide precise performance numbers, but tests have been run
on a wide variety of architectures (x86-64, ARM Cortex, Aarch64, RISC-V,
xtensa) and they all generally agree on the qualitative improvements seen
by the combination of MICROPY_OPT_LOAD_ATTR_FAST_PATH and
MICROPY_OPT_MAP_LOOKUP_CACHE.
For example, on a "quiet" Linux x64 environment (i3-5010U @ 2.10GHz) the
change from CACHE_MAP_LOOKUP_IN_BYTECODE, to LOAD_ATTR_FAST_PATH combined
with MAP_LOOKUP_CACHE is:
diff of scores (higher is better)
N=2000 M=2000 bccache -> attrmapcache diff diff% (error%)
bm_chaos.py 13742.56 -> 13905.67 : +163.11 = +1.187% (+/-3.75%)
bm_fannkuch.py 60.13 -> 61.34 : +1.21 = +2.012% (+/-2.11%)
bm_fft.py 113083.20 -> 114793.68 : +1710.48 = +1.513% (+/-1.57%)
bm_float.py 256552.80 -> 243908.29 : -12644.51 = -4.929% (+/-1.90%)
bm_hexiom.py 521.93 -> 625.41 : +103.48 = +19.826% (+/-0.40%)
bm_nqueens.py 197544.25 -> 217713.12 : +20168.87 = +10.210% (+/-3.01%)
bm_pidigits.py 8072.98 -> 8198.75 : +125.77 = +1.558% (+/-3.22%)
misc_aes.py 17283.45 -> 16480.52 : -802.93 = -4.646% (+/-0.82%)
misc_mandel.py 99083.99 -> 128939.84 : +29855.85 = +30.132% (+/-5.88%)
misc_pystone.py 83860.10 -> 82592.56 : -1267.54 = -1.511% (+/-2.27%)
misc_raytrace.py 21490.40 -> 22227.23 : +736.83 = +3.429% (+/-1.88%)
This shows that the new optimisations are at least as good as the existing
inline-bytecode-caching, and are sometimes much better (because the new
ones apply caching to a wider variety of map lookups).
The new optimisations can also benefit code generated by the native
emitter, because they apply to the runtime rather than the generated code.
The improvement for the native emitter when LOAD_ATTR_FAST_PATH and
MAP_LOOKUP_CACHE are enabled is (same Linux environment as above):
diff of scores (higher is better)
N=2000 M=2000 native -> nat-attrmapcache diff diff% (error%)
bm_chaos.py 14130.62 -> 15464.68 : +1334.06 = +9.441% (+/-7.11%)
bm_fannkuch.py 74.96 -> 76.16 : +1.20 = +1.601% (+/-1.80%)
bm_fft.py 166682.99 -> 168221.86 : +1538.87 = +0.923% (+/-4.20%)
bm_float.py 233415.23 -> 265524.90 : +32109.67 = +13.756% (+/-2.57%)
bm_hexiom.py 628.59 -> 734.17 : +105.58 = +16.796% (+/-1.39%)
bm_nqueens.py 225418.44 -> 232926.45 : +7508.01 = +3.331% (+/-3.10%)
bm_pidigits.py 6322.00 -> 6379.52 : +57.52 = +0.910% (+/-5.62%)
misc_aes.py 20670.10 -> 27223.18 : +6553.08 = +31.703% (+/-1.56%)
misc_mandel.py 138221.11 -> 152014.01 : +13792.90 = +9.979% (+/-2.46%)
misc_pystone.py 85032.14 -> 105681.44 : +20649.30 = +24.284% (+/-2.25%)
misc_raytrace.py 19800.01 -> 23350.73 : +3550.72 = +17.933% (+/-2.79%)
In summary, compared to MICROPY_OPT_CACHE_MAP_LOOKUP_IN_BYTECODE, the new
MICROPY_OPT_LOAD_ATTR_FAST_PATH and MICROPY_OPT_MAP_LOOKUP_CACHE options:
- are simpler;
- take less code size;
- are faster (generally);
- work with code generated by the native emitter;
- can be used on embedded targets with a small and constant RAM overhead;
- allow the same .mpy bytecode to run on all targets.
See #7680 for further discussion. And see also #7653 for a discussion
about simplifying mpy-cross options.
Signed-off-by: Jim Mussared <jim.mussared@gmail.com>
Fixes the following (the line numbers match commit 0e87459e2b):
../../extmod/crypto-algorithms/sha256.c:49:19: runtime error: left shif...
../../extmod/moduasyncio.c:106:35: runtime error: member access within ...
../../py/binary.c:210:13: runtime error: left shift of negative value -...
../../py/mpz.c:744:16: runtime error: negation of -9223372036854775808 ...
../../py/objint.c:109:22: runtime error: left shift of 1 by 31 places c...
../../py/objint_mpz.c:374:9: runtime error: left shift of 4611686018427...
../../py/objint_mpz.c:374:9: runtime error: left shift of negative valu...
../../py/parsenum.c:106:14: runtime error: left shift of 46116860184273...
../../py/runtime.c:395:33: runtime error: left shift of negative value ...
../../py/showbc.c:177:28: runtime error: left shift of negative value -...
../../py/vm.c:321:36: runtime error: left shift of negative value -1```
Testing was done on an amd64 Debian Buster system using gcc-8.3 and these
settings:
CFLAGS += -g3 -Og -fsanitize=undefined
LDFLAGS += -fsanitize=undefined
The introduced TASK_PAIRHEAP macro's conditional (x ? &x->i : NULL)
assembles (under amd64 gcc 8.3 -Os) to the same as &x->i, since i is the
initial field of the struct. However, for the purposes of undefined
behavior analysis the conditional is needed.
Signed-off-by: Jeff Epler <jepler@gmail.com>
This patch compresses the second part of the bytecode prelude which
contains the source file name, function name, source-line-number mapping
and cell closure information. This part of the prelude now begins with a
single varible length unsigned integer which encodes 2 numbers, being the
byte-size of the following 2 sections in the header: the "source info
section" and the "closure section". After decoding this variable unsigned
integer it's possible to skip over one or both of these sections very
easily.
This scheme saves about 2 bytes for most functions compared to the original
format: one in the case that there are no closure cells, and one because
padding was eliminated.
The start of the bytecode prelude contains 6 numbers telling the amount of
stack needed for the Python values and exceptions, and the signature of the
function. Prior to this patch these numbers were all encoded one after the
other (2x variable unsigned integers, then 4x bytes), but using so many
bytes is unnecessary.
An entropy analysis of around 150,000 bytecode functions from the CPython
standard library showed that the optimal Shannon coding would need about
7.1 bits on average to encode these 6 numbers, compared to the existing 48
bits.
This patch attempts to get close to this optimal value by packing the 6
numbers into a single, varible-length unsigned integer via bit-wise
interleaving. The interleaving scheme is chosen to minimise the average
number of bytes needed, and at the same time keep the scheme simple enough
so it can be implemented without too much overhead in code size or speed.
The scheme requires about 10.5 bits on average to store the 6 numbers.
As a result most functions which originally took 6 bytes to encode these 6
numbers now need only 1 byte (in 80% of cases).
From the beginning of this project the RAISE_VARARGS opcode was named and
implemented following CPython, where it has an argument (to the opcode)
counting how many args the raise takes:
raise # 0 args (re-raise previous exception)
raise exc # 1 arg
raise exc from exc2 # 2 args (chained raise)
In the bytecode this operation therefore takes 2 bytes, one for
RAISE_VARARGS and one for the number of args.
This patch splits this opcode into 3, where each is now a single byte.
This reduces bytecode size by 1 byte for each use of raise. Every byte
counts! It also has the benefit of reducing code size (on all ports except
nanbox).
POP_BLOCK and POP_EXCEPT are now the same, and are always followed by a
JUMP. So this optimisation reduces code size, and RAM usage of bytecode by
two bytes for each try-except handler.
This patch allows the following code to run without allocating on the heap:
super().foo(...)
Before this patch such a call would allocate a super object on the heap and
then load the foo method and call it right away. The super object is only
needed to perform the lookup of the method and not needed after that. This
patch makes an optimisation to allocate the super object on the C stack and
discard it right after use.
Changes in code size due to this patch are:
bare-arm: +128
minimal: +232
unix x64: +416
unix nanbox: +364
stmhal: +184
esp8266: +340
cc3200: +128
The system printf is no longer used by the core uPy code. Instead, the
platform print stream or DEBUG_printf is used. Using DEBUG_printf in the
showbc functions would mean that the code can't be tested by the test
suite, so use the normal output instead.
This patch also fixes parsing of bytecode-line-number mappings.
With the previous patch combining 3 emit functions into 1, it now makes
sense to also combine the corresponding VM opcodes, which is what this
patch does. This eliminates 2 opcodes which simplifies the VM and reduces
code size, in bytes: bare-arm:44, minimal:64, unix(NDEBUG,x86-64):272,
stmhal:92, esp8266:200. Profiling (with a simple script that creates many
list/dict/set comprehensions) shows no measurable change in performance.
Fixes#1684 and makes "not" match Python semantics. The code is also
simplified (the separate MP_BC_NOT opcode is removed) and the patch saves
68 bytes for bare-arm/ and 52 bytes for minimal/.
Previously "not x" was implemented as !mp_unary_op(x, MP_UNARY_OP_BOOL),
so any given object only needs to implement MP_UNARY_OP_BOOL (and the VM
had a special opcode to do the ! bit).
With this patch "not x" is implemented as mp_unary_op(x, MP_UNARY_OP_NOT),
but this operation is caught at the start of mp_unary_op and dispatched as
!mp_obj_is_true(x). mp_obj_is_true has special logic to test for
truthness, and is the correct way to handle the not operation.
This allows the mp_obj_t type to be configured to something other than a
pointer-sized primitive type.
This patch also includes additional changes to allow the code to compile
when sizeof(mp_uint_t) != sizeof(void*), such as using size_t instead of
mp_uint_t, and various casts.
Main changes when MICROPY_PERSISTENT_CODE is enabled are:
- qstrs are encoded as 2-byte fixed width in the bytecode
- all pointers are removed from bytecode and put in const_table (this
includes const objects and raw code pointers)
Ultimately this option will enable persistence for not just bytecode but
also native code.
Previous to this patch each time a bytes object was referenced a new
instance (with the same data) was created. With this patch a single
bytes object is created in the compiler and is loaded directly at execute
time as a true constant (similar to loading bignum and float objects).
This saves on allocating RAM and means that bytes objects can now be
used when the memory manager is locked (eg in interrupts).
The MP_BC_LOAD_CONST_BYTES bytecode was removed as part of this.
Generated bytecode is slightly larger due to storing a pointer to the
bytes object instead of the qstr identifier.
Code size is reduced by about 60 bytes on Thumb2 architectures.
Previous to this patch, a big-int, float or imag constant was interned
(made into a qstr) and then parsed at runtime to create an object each
time it was needed. This is wasteful in RAM and not efficient. Now,
these constants are parsed straight away in the parser and turned into
objects. This allows constants with large numbers of digits (so
addresses issue #1103) and takes us a step closer to #722.
Eg, "() + 1" now tells you that __add__ is not supported for tuple and
int types (before it just said the generic "binary operator"). We reuse
the table of names for slot lookup because it would be a waste of code
space to store the pretty name for each operator.
This is consistent with how BC_JUMP was handled before. We never show jumps
destinations relative to jump instrucion itself, only relative to beginning
of function. Another useful way to show them as absolute (real memory
address), and this change makes result expected and consistent with how
BC_JUMP is shown.
There is a lot potential in compress bytecodes and make more use of the
coding space. This patch introduces "multi" bytecodes which have their
argument included in the bytecode (by addition).
UNARY_OP and BINARY_OP now no longer take a 1 byte argument for the
opcode. Rather, the opcode is included in the first byte itself.
LOAD_FAST_[0,1,2] and STORE_FAST_[0,1,2] are removed in favour of their
multi versions, which can take an argument between 0 and 15 inclusive.
The majority of LOAD_FAST/STORE_FAST codes fit in this range and so this
saves a byte for each of these.
LOAD_CONST_SMALL_INT_MULTI is used to load small ints between -16 and 47
inclusive. Such ints are quite common and now only need 1 byte to
store, and now have much faster decoding.
In all this patch saves about 2% RAM for typically bytecode (1.8% on
64-bit test, 2.5% on pyboard test). It also reduces the binary size
(because bytecodes are simplified) and doesn't harm performance.
This saves a lot of RAM for 2 reasons:
1. For functions that don't have default values, var args or var kw
args (which is a large number of functions in the general case), the
mp_obj_fun_bc_t type now fits in 1 GC block (previously needed 2 because
of the extra pointer to point to the arg_names array). So this saves 16
bytes per function (32 bytes on 64-bit machines).
2. Combining separate memory regions generally saves RAM because the
unused bytes at the end of the GC block are saved for 1 of the blocks
(since that block doesn't exist on its own anymore). So generally this
saves 8 bytes per function.
Tested by importing lots of modules:
- 64-bit Linux gave about an 8% RAM saving for 86k of used RAM.
- pyboard gave about a 6% RAM saving for 31k of used RAM.
Damien George