[experiments] Thunk the eliminator cases #327
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This allows us to remove the special case for `bool_rect`, and is a step towards targeting `vm_compute` rather than `lazy`.
andres-erbsen
approved these changes
Mar 20, 2018
JasonGross
added a commit
to JasonGross/fiat-crypto
that referenced
this pull request
Feb 21, 2022
Probably if the lists are the same lengths, then we want to compare them element-wise rather than all at once. It's way too verbose to keep expanding them, so we only do that when lists are not the same length. We now get error messages such as ``` Unable to unify: [inr [378, 381, 384]] == [inr [101, 106, 108]] Could not unify the values at index 0: [mit-plv#378, mit-plv#381, mit-plv#384] != [mit-plv#101, mit-plv#106, mit-plv#108] index 0: mit-plv#378 != mit-plv#101 (slice 0 44, [mit-plv#377]) != (slice 0 44, [mit-plv#98]) index 0: mit-plv#377 != mit-plv#98 (add 64, [mit-plv#345, mit-plv#375]) != (add 64, [#57, mit-plv#96]) index 0: mit-plv#345 != #57 (slice 0 44, [mit-plv#337]) != (slice 0 44, [#44]) index 0: mit-plv#337 != #44 (add 64, [#41, mit-plv#334]) != (add 64, [#25, #41]) index 1: mit-plv#334 != #25 (mul 64, [#1, mit-plv#331]) != (mul 64, [#0, #1, #22]) [(add 64, [mit-plv#329, mit-plv#329])] != [#0, (const 20, [])] [(add 64, [(mul 64, [#7, mit-plv#328]), (mul 64, [#7, mit-plv#328])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [#0, mit-plv#327])]), (mul 64, [(const 2, []), (add 64, [#0, mit-plv#327])])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, mit-plv#326])])]), (mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, mit-plv#326])])])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, (const 4, [])])])]), (mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, (const 4, [])])])])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [(old 64 0, []), (mul 64, [(old 64 0, []), (const 4, [])])])]), (mul 64, [(const 2, []), (add 64, [(old 64 0, []), (mul 64, [(old 64 0, []), (const 4, [])])])])])] != [(old 64 0, []), (const 20, [])] ``` The second to last line is generally the one to look at; the last line adds a bit more detail to it. Perhaps we should instead list out the values of indices rather than expanding one additional level?
JasonGross
added a commit
to JasonGross/fiat-crypto
that referenced
this pull request
Feb 21, 2022
Probably if the lists are the same lengths, then we want to compare them element-wise rather than all at once. It's way too verbose to keep expanding them, so we only do that when lists are not the same length. We now get error messages such as ``` Unable to unify: [inr [378, 381, 384]] == [inr [101, 106, 108]] Could not unify the values at index 0: [mit-plv#378, mit-plv#381, mit-plv#384] != [mit-plv#101, mit-plv#106, mit-plv#108] index 0: mit-plv#378 != mit-plv#101 (slice 0 44, [mit-plv#377]) != (slice 0 44, [mit-plv#98]) index 0: mit-plv#377 != mit-plv#98 (add 64, [mit-plv#345, mit-plv#375]) != (add 64, [#57, mit-plv#96]) index 0: mit-plv#345 != #57 (slice 0 44, [mit-plv#337]) != (slice 0 44, [#44]) index 0: mit-plv#337 != #44 (add 64, [#41, mit-plv#334]) != (add 64, [#25, #41]) index 1: mit-plv#334 != #25 (mul 64, [#1, mit-plv#331]) != (mul 64, [#0, #1, #22]) [(add 64, [mit-plv#329, mit-plv#329])] != [#0, (const 20, [])] [(add 64, [(mul 64, [#7, mit-plv#328]), (mul 64, [#7, mit-plv#328])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [#0, mit-plv#327])]), (mul 64, [(const 2, []), (add 64, [#0, mit-plv#327])])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, mit-plv#326])])]), (mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, mit-plv#326])])])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, (const 4, [])])])]), (mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, (const 4, [])])])])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [(old 64 0, []), (mul 64, [(old 64 0, []), (const 4, [])])])]), (mul 64, [(const 2, []), (add 64, [(old 64 0, []), (mul 64, [(old 64 0, []), (const 4, [])])])])])] != [(old 64 0, []), (const 20, [])] ``` The second to last line is generally the one to look at; the last line adds a bit more detail to it. Perhaps we should instead list out the values of indices rather than expanding one additional level?
JasonGross
added a commit
that referenced
this pull request
Feb 22, 2022
Probably if the lists are the same lengths, then we want to compare them element-wise rather than all at once. It's way too verbose to keep expanding them, so we only do that when lists are not the same length. We now get error messages such as ``` Unable to unify: [inr [378, 381, 384]] == [inr [101, 106, 108]] Could not unify the values at index 0: [#378, #381, #384] != [#101, #106, #108] index 0: #378 != #101 (slice 0 44, [#377]) != (slice 0 44, [#98]) index 0: #377 != #98 (add 64, [#345, #375]) != (add 64, [#57, #96]) index 0: #345 != #57 (slice 0 44, [#337]) != (slice 0 44, [#44]) index 0: #337 != #44 (add 64, [#41, #334]) != (add 64, [#25, #41]) index 1: #334 != #25 (mul 64, [#1, #331]) != (mul 64, [#0, #1, #22]) [(add 64, [#329, #329])] != [#0, (const 20, [])] [(add 64, [(mul 64, [#7, #328]), (mul 64, [#7, #328])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [#0, #327])]), (mul 64, [(const 2, []), (add 64, [#0, #327])])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, #326])])]), (mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, #326])])])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, (const 4, [])])])]), (mul 64, [(const 2, []), (add 64, [#0, (mul 64, [#0, (const 4, [])])])])])] != [#0, (const 20, [])] [(add 64, [(mul 64, [(const 2, []), (add 64, [(old 64 0, []), (mul 64, [(old 64 0, []), (const 4, [])])])]), (mul 64, [(const 2, []), (add 64, [(old 64 0, []), (mul 64, [(old 64 0, []), (const 4, [])])])])])] != [(old 64 0, []), (const 20, [])] ``` The second to last line is generally the one to look at; the last line adds a bit more detail to it. Perhaps we should instead list out the values of indices rather than expanding one additional level?
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This allows us to remove the special case for
bool_rect, and is astep towards targeting
vm_computerather thanlazy.Note that this does not, in fact, seem to by itself allow
vm_computeto work in reasonable time; thelazyinbase_51_addtakes about 0.5 seconds, while usingvm_computetakes over a minute.