Revision c22889e76cb9b7fd8a4710d9bf53e827aaa907e4 authored by Shuhei Kadowaki on 24 August 2021, 04:21:29 UTC, committed by Shuhei Kadowaki on 26 October 2021, 14:45:33 UTC
Currently our constant-prop' heuristics work in the following way:
1. `const_prop_entry_heuristic`
2. `const_prop_argument_heuristic` & `const_prop_rettype_heuristic`
3. `force_const_prop` custom heuristic & `!const_prop_function_heuristic`
4. `MethodInstance` specialization and `const_prop_methodinstance_heuristic`

This PR changes it so that the step 1. now works like:

1. `force_const_prop` custom heuristic & `const_prop_entry_heuristic`

and the steps 2., 3. and 4. don't change

This change particularly allows us to more forcibly constant-propagate
for `getproperty` and `setproperty!`, and inline them more, e.g.:
```julia
mutable struct Foo
    val
    _::Int
end

function setter(xs)
    for x in xs
        x.val = nothing # `setproperty!` can be inlined with this PR
    end
end
```

It might be useful because now we can intervene into the constant-prop'
heuristic in a more reliable way with the `aggressive_constprop` interface.

I did the simple benchmark below, and it looks like this change doesn't
cause the latency problem for this particular example:
```zsh
~/julia master aviatesk@amdci2 6s
❯ ./usr/bin/julia -e '@time using Plots; @time plot(rand(10,3))'
  3.708500 seconds (7.28 M allocations: 506.128 MiB, 3.45% gc time, 1.13% compilation time)
  2.817794 seconds (3.45 M allocations: 195.127 MiB, 7.84% gc time, 53.76% compilation time)

~/julia avi/forceconstantprop aviatesk@amdci2 6s
❯ ./usr/bin/julia -e '@time using Plots; @time plot(rand(10,3))'
  3.622109 seconds (7.02 M allocations: 481.710 MiB, 4.19% gc time, 1.17% compilation time)
  2.863419 seconds (3.44 M allocations: 194.210 MiB, 8.02% gc time, 53.53% compilation time)
```
1 parent da71d29
Raw File
timing.c
// This file is a part of Julia. License is MIT: https://julialang.org/license

#include <inttypes.h>
#include "julia.h"
#include "julia_internal.h"
#include "options.h"
#include "stdio.h"

#ifdef __cplusplus
extern "C" {
#endif

#ifdef ENABLE_TIMINGS
#include "timing.h"

#ifndef HAVE_TIMING_SUPPORT
#error Timings are not supported on your compiler
#endif

static uint64_t t0;
uint64_t jl_timing_data[(int)JL_TIMING_LAST] = {0};
const char *jl_timing_names[(int)JL_TIMING_LAST] =
    {
#define X(name) #name
        JL_TIMING_OWNERS
#undef X
    };

void jl_print_timings(void)
{
    uint64_t total_time = cycleclock() - t0;
    uint64_t root_time = total_time;
    for (int i = 0; i < JL_TIMING_LAST; i++) {
        root_time -= jl_timing_data[i];
    }
    jl_timing_data[0] = root_time;
    for (int i = 0; i < JL_TIMING_LAST; i++) {
        if (jl_timing_data[i] != 0)
            fprintf(stderr, "%-25s : %5.2f %%   %" PRIu64 "\n", jl_timing_names[i],
                    100 * (((double)jl_timing_data[i]) / total_time), jl_timing_data[i]);
    }
}

void jl_init_timing(void)
{
    t0 = cycleclock();
}

void jl_destroy_timing(void)
{
    jl_ptls_t ptls = jl_current_task->ptls;
    jl_timing_block_t *stack = ptls->timing_stack;
    while (stack) {
        _jl_timing_block_destroy(stack);
        stack = stack->prev;
    }
}

jl_timing_block_t *jl_pop_timing_block(jl_timing_block_t *cur_block)
{
    _jl_timing_block_destroy(cur_block);
    return cur_block->prev;
}

void jl_timing_block_start(jl_timing_block_t *cur_block)
{
    _jl_timing_block_start(cur_block, cycleclock());
}

void jl_timing_block_stop(jl_timing_block_t *cur_block)
{
    _jl_timing_block_stop(cur_block, cycleclock());
}

#else

void jl_init_timing(void) { }
void jl_destroy_timing(void) { }

#endif

#ifdef __cplusplus
}
#endif
back to top