| Commit message (Collapse) | Author | Age |
| |
|
|
|
|
| |
src/Compilers/Z/ArithmeticSimplifierInterp.v
|
| |
|
|
|
|
| |
removing lemma wordToNat_wzero is ok because it's already in bbv
|
|
|
|
|
|
|
| |
This notation system is fragile and kludgy.
This discovered from @davidben's
https://github.com/mit-plv/fiat-crypto/pull/289/commits/ff0fb38346dde67abef982d6305595216d18519b#r159793723
|
|
|
|
|
| |
We do this by adding notations for addcarryx and subborrow for all of
the smaller-than-max-bitwidth sizes of arguments.
|
|
|
|
| |
This handles bullet 3 of #288
|
|
|
|
| |
This handles bullet point 1 of #288
|
|
|
|
|
|
|
|
|
|
|
| |
This closes #286
This is actually a +1,-1 diff in the python script generating the
notations, plus running it and rebuilding:
```diff
-ADD_CARRY_SUB_BORROW_SIZES = (32, 64, 128, 51)
+ADD_CARRY_SUB_BORROW_SIZES = (32, 64, 128, 25, 26, 51)
```
|
|
|
|
|
|
| |
This makes it easier to add support for more kinds of addcarryx, etc,
and also add `: expr_scope` to work around changes from
https://github.com/coq/coq/pull/873
|
| |
|
|
|
|
|
| |
This will hopefully pave the way for not needing to prove Wf anywhere in
the bounds pipeline.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
After | File Name | Before || Change | % Change
----------------------------------------------------------------------------------------------
19m16.05s | Total | 21m25.28s || -2m09.23s | -10.05%
----------------------------------------------------------------------------------------------
4m01.34s | Specific/X25519/C64/ladderstep | 4m59.49s || -0m58.15s | -19.41%
2m48.52s | Specific/solinas32_2e255m765_13limbs/femul | 3m12.95s || -0m24.42s | -12.66%
2m23.70s | Specific/solinas32_2e255m765_12limbs/femul | 2m44.11s || -0m20.41s | -12.43%
3m09.62s | Specific/NISTP256/AMD64/femul | 3m22.52s || -0m12.90s | -6.36%
0m36.32s | Specific/X25519/C64/femul | 0m39.50s || -0m03.17s | -8.05%
0m30.13s | Specific/X25519/C64/fesquare | 0m32.24s || -0m02.11s | -6.54%
0m35.40s | Specific/NISTP256/AMD64/feadd | 0m37.21s || -0m01.81s | -4.86%
0m31.50s | Specific/X25519/C64/freeze | 0m33.24s || -0m01.74s | -5.23%
0m24.99s | Specific/X25519/C64/fecarry | 0m26.31s || -0m01.32s | -5.01%
0m22.65s | Specific/X25519/C64/fesub | 0m23.72s || -0m01.07s | -4.51%
0m45.75s | Specific/solinas32_2e255m765_13limbs/Synthesis | 0m45.58s || +0m00.17s | +0.37%
0m39.59s | Specific/NISTP256/AMD64/fesub | 0m40.09s || -0m00.50s | -1.24%
0m36.92s | Specific/solinas32_2e255m765_12limbs/Synthesis | 0m36.64s || +0m00.28s | +0.76%
0m28.51s | Specific/NISTP256/AMD64/feopp | 0m29.46s || -0m00.94s | -3.22%
0m25.50s | Specific/NISTP256/AMD64/fenz | 0m26.41s || -0m00.91s | -3.44%
0m20.93s | Specific/X25519/C64/feadd | 0m21.41s || -0m00.48s | -2.24%
0m12.55s | Specific/NISTP256/AMD64/Synthesis | 0m12.54s || +0m00.01s | +0.07%
0m10.37s | Specific/X25519/C64/Synthesis | 0m10.30s || +0m00.06s | +0.67%
0m07.18s | Compilers/Z/Bounds/Pipeline/Definition | 0m07.22s || -0m00.04s | -0.55%
0m01.72s | Compilers/Z/Bounds/Pipeline/ReflectiveTactics | 0m01.58s || +0m00.13s | +8.86%
0m01.67s | Specific/Framework/SynthesisFramework | 0m01.72s || -0m00.05s | -2.90%
0m01.19s | Compilers/Z/Bounds/Pipeline | 0m01.04s || +0m00.14s | +14.42%
|
|
|
|
|
| |
Also add a dummy option about renaming binders, to be used in an
upcoming commit.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Hopefully this will help speed up some primes, like:
```
tactic local total calls max
────────────────────────────────────────┴──────┴──────┴───────┴─────────┘
─synthesize_with_carry ----------------- 0.0% 100.0% 1 4158.848s
└Pipeline.refine_reflectively_gen ------ 0.0% 99.0% 1 4117.688s
└ReflectiveTactics.do_reflective_pipelin 0.0% 98.7% 1 4106.804s
└ReflectiveTactics.solve_side_conditions 0.0% 98.7% 1 4106.580s
├─UnifyAbstractReflexivity.unify_transf 73.7% 74.5% 8 2283.572s
├─ReflectiveTactics.handle_boundedness_ 0.0% 18.8% 1 781.616s
│└intros [a b] ------------------------ 18.4% 18.4% 29 33.640s
└─ReflectiveTactics.do_reify ---------- 0.0% 5.4% 1 225.672s
└Reify.do_reify_abs_goal ------------- 4.7% 4.7% 2 196.164s
└Reify.do_reifyf_goal ---------------- 4.7% 4.7% 270 102.004s
src/Specific/solinas32_2e512m569/fesquare (real: 5415.91, user: 5115.86, sys: 27.59, mem: 33162744 ko)
```
and
```
tactic local total calls max
────────────────────────────────────────┴──────┴──────┴───────┴─────────┘
─Pipeline.refine_reflectively_gen ------ 0.0% 97.8% 1 437.088s
├─ReflectiveTactics.do_reflective_pipel 0.0% 94.0% 1 420.396s
│└ReflectiveTactics.solve_side_conditio 0.0% 94.0% 1 420.232s
│ ├─ReflectiveTactics.handle_boundednes 0.0% 63.3% 1 283.096s
│ │└intros [a b] ---------------------- 61.9% 61.9% 39 10.288s
│ ├─ReflectiveTactics.do_reify -------- 0.1% 15.8% 1 70.528s
│ │ ├─Reify.do_reify_abs_goal --------- 12.6% 12.7% 2 56.572s
│ │ │└Reify.do_reifyf_goal ------------ 12.5% 12.5% 180 30.120s
│ │ │└eexact -------------------------- 3.9% 3.9% 180 0.132s
│ │ └─Reify.Reify_rhs_gen ------------- 0.2% 2.9% 1 12.836s
│ └─UnifyAbstractReflexivity.unify_tran 12.6% 14.5% 8 23.836s
└─Glue.refine_to_reflective_glue' ----- 0.0% 3.7% 1 16.692s
└Glue.zrange_to_reflective ----------- 0.0% 3.7% 1 16.372s
└Glue.zrange_to_reflective_goal ------ 0.4% 2.9% 1 13.084s
└pose proof (pf : Interpretation. 2.5% 2.5% 1 11.144s
─synthesize ---------------------------- 0.0% 2.2% 1 9.988s
└IntegrationTestTemporaryMiscCommon.do_r 0.0% 2.2% 1 9.904s
└change G' ----------------------------- 2.2% 2.2% 1 9.812s
src/Specific/solinas32_2e521m1/femul (real: 1165.84, user: 878.68, sys: 9.32, mem: 11218492 ko)
```
and
```
tactic local total calls max
────────────────────────────────────────┴──────┴──────┴───────┴─────────┘
─Pipeline.refine_reflectively_gen ------ 0.0% 97.6% 1 118.496s
├─ReflectiveTactics.do_reflective_pipel 0.0% 93.7% 1 113.844s
│└ReflectiveTactics.solve_side_conditio 0.0% 93.7% 1 113.772s
│ ├─ReflectiveTactics.handle_boundednes 0.0% 52.8% 1 64.160s
│ │└intros [a b] ---------------------- 50.9% 50.9% 29 2.420s
│ ├─ReflectiveTactics.do_reify -------- 0.2% 23.7% 1 28.840s
│ │ ├─Reify.do_reify_abs_goal --------- 17.8% 17.9% 2 21.792s
│ │ │└Reify.do_reifyf_goal ------------ 17.6% 17.7% 135 10.952s
│ │ │└eexact -------------------------- 5.4% 5.4% 135 0.064s
│ │ └─Reify.Reify_rhs_gen ------------- 0.4% 5.3% 1 6.452s
│ └─UnifyAbstractReflexivity.unify_tran 14.0% 16.4% 8 8.544s
│ └unify (constr) (constr) ----------- 2.0% 2.0% 6 1.460s
└─Glue.refine_to_reflective_glue' ----- 0.0% 3.8% 1 4.644s
└Glue.zrange_to_reflective ----------- 0.0% 3.7% 1 4.472s
└Glue.zrange_to_reflective_goal ------ 0.6% 2.8% 1 3.392s
└pose proof (pf : Interpretation. 2.1% 2.1% 1 2.564s
─synthesize_with_carry ----------------- 0.0% 2.4% 1 2.968s
└IntegrationTestTemporaryMiscCommon.do_r 0.0% 2.4% 1 2.920s
└change G' ----------------------------- 2.4% 2.4% 1 2.868s
src/Specific/solinas32_2e255m765/femul (real: 307.37, user: 255.97, sys: 0.81, mem: 4050732 ko)
```
```
A bit of testing suggests that it is, indeed, much better:
```
tactic local total calls max
────────────────────────────────────────┴──────┴──────┴───────┴─────────┘
─Pipeline.refine_reflectively_gen ------ 0.0% 95.5% 1 60.100s
├─ReflectiveTactics.do_reflective_pipel 0.0% 88.7% 1 55.764s
│└ReflectiveTactics.solve_side_conditio 0.0% 88.3% 1 55.556s
│ ├─ReflectiveTactics.do_reify -------- 0.0% 48.3% 1 30.412s
│ │└Reify.Reify_rhs_gen --------------- 1.2% 47.8% 1 30.048s
│ │ ├─Reify.do_reify_abs_goal --------- 36.0% 36.3% 2 22.804s
│ │ │└Reify.do_reifyf_goal ------------ 35.4% 35.6% 117 11.708s
│ │ │└eexact -------------------------- 10.9% 10.9% 117 0.080s
│ │ ├─rewrite H ----------------------- 2.9% 2.9% 1 1.820s
│ │ └─prove_interp_compile_correct ---- 0.0% 2.7% 1 1.724s
│ │ └rewrite ?EtaInterp.InterpExprEta 2.5% 2.5% 1 1.584s
│ └─ReflectiveTactics.solve_post_reifie 0.1% 40.0% 1 25.144s
│ └UnifyAbstractReflexivity.unify_tran 32.0% 37.4% 8 10.584s
│ └unify (constr) (constr) ----------- 4.5% 4.5% 6 1.504s
└─Glue.refine_to_reflective_glue' ----- 0.0% 6.9% 1 4.336s
└Glue.zrange_to_reflective ----------- 0.0% 6.4% 1 4.056s
└Glue.zrange_to_reflective_goal ------ 1.3% 4.9% 1 3.104s
└pose proof (pf : Interpretation. 3.5% 3.5% 1 2.192s
─synthesize ---------------------------- 0.0% 4.5% 1 2.800s
└IntegrationTestTemporaryMiscCommon.do_r 0.0% 4.3% 1 2.720s
└change G' ----------------------------- 4.2% 4.2% 1 2.652s
src/Specific/solinas32_2e255m765_13limbs/femul (real: 178.02, user: 176.01, sys: 0.38, mem: 1713136 ko)
```
and
```
tactic local total calls max
────────────────────────────────────────┴──────┴──────┴───────┴─────────┘
─Pipeline.refine_reflectively_gen ------ 0.0% 95.8% 1 48.612s
├─ReflectiveTactics.do_reflective_pipel -0.0% 89.1% 1 45.204s
│└ReflectiveTactics.solve_side_conditio 0.0% 88.8% 1 45.040s
│ ├─ReflectiveTactics.do_reify -------- 0.0% 50.7% 1 25.720s
│ │└Reify.Reify_rhs_gen --------------- 1.4% 50.1% 1 25.400s
│ │ ├─Reify.do_reify_abs_goal --------- 37.3% 37.5% 2 19.012s
│ │ │└Reify.do_reifyf_goal ------------ 36.6% 36.9% 108 9.628s
│ │ │└eexact -------------------------- 11.4% 11.4% 108 0.072s
│ │ ├─prove_interp_compile_correct ---- 0.0% 2.9% 1 1.492s
│ │ │└rewrite ?EtaInterp.InterpExprEta 2.6% 2.6% 1 1.344s
│ │ └─rewrite H ----------------------- 2.9% 2.9% 1 1.460s
│ └─ReflectiveTactics.solve_post_reifie 0.1% 38.1% 1 19.320s
│ └UnifyAbstractReflexivity.unify_tran 30.2% 35.2% 8 7.924s
│ └unify (constr) (constr) ----------- 4.2% 4.2% 6 1.060s
└─Glue.refine_to_reflective_glue' ----- 0.0% 6.7% 1 3.408s
└Glue.zrange_to_reflective ----------- 0.0% 6.2% 1 3.140s
└Glue.zrange_to_reflective_goal ------ 1.3% 4.7% 1 2.388s
└pose proof (pf : Interpretation. 3.2% 3.2% 1 1.644s
─synthesize ---------------------------- 0.0% 4.2% 1 2.136s
└IntegrationTestTemporaryMiscCommon.do_r 0.0% 4.1% 1 2.072s
└change G' ----------------------------- 4.0% 4.0% 1 2.008s
src/Specific/solinas32_2e255m765_12limbs/femul (real: 153.02, user: 151.06, sys: 0.37, mem: 1557920 ko)
```
|
| |
|
| |
|
|
|
|
| |
Now we no longer bundle the side-condition solver with the reifier
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
|
| |
Builds, but haven't tested the output
This closes #265
|
| |
|
| |
|
| |
|
| |
|
|
|
|
|
| |
It's the final argument, not the second-to-final argument, that needs to
be a particular size
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|
| |
|