program-table #210

naure · 2024-09-11T13:34:56Z

Issue #110.

kunxian-xia

This PR looks good to me except some nitpicks. However, the riscv_add example fails on my local machine.

ceno_zkvm/src/instructions/riscv/addsub.rs

hero78119

change are very neat!

(A TODO marks in later PR): we also need to extend MockProver to support program table lookup for better debug

hero78119 · 2024-09-12T07:40:17Z

ceno_zkvm/examples/riscv_add.rs

+        range_config.clone(),
+        &(),
+    );
+    zkvm_fixed_traces.register_table_circuit::<ProgramTableCircuit<E>>(


👍 since we register new circuit, to verify e2e test not broken, this PR probabaly go after #209

Makes sense, I’ll just wait and make it work with that test.

The program table is now integrated in the MockProver as well. Unittests work.

ceno_zkvm/examples/riscv_add.rs

ceno_zkvm/src/tables/program.rs

naure · 2024-09-13T11:54:02Z

This PR looks good to me except some nitpicks. However, the riscv_add example fails on my local machine.

I haven’t figured this out yet but I realized it happens to me on master as well.

➡ #222

naure · 2024-09-13T17:38:52Z

This is in a good enough shape to merge. Remaining work is in issues.

This is another approach to #223, fixing #222 / #210. Features: - Prepare for parallel proving with independent transcripts per thread. - Soundness with different transcripts in each thread. - Type-safe API which captures the transcript by value. There is no function to merge the threads back into one transcript, but that can be added if ever needed. --------- Co-authored-by: Aurélien Nicolas <[email protected]>

naure · 2024-09-14T11:06:26Z

Resolved merge conflicts and added support for SUB.

### Goal To make sumcheck protocol support different num_vars, aiming for - [x] minimal change to sumcheck protocol, make verifier remain the same. no extra meta data passed from prover, and what prover have just mle and it's eval size Besides this PR also remove some parallism in verifier since it's unnecessary for relative low cost. ### design rationale (Also comments in codebase for reference) To deal with different num_vars, we exploit a fact that for each product which num_vars < max_num_vars, for it evaluation value we need to times 2^(max_num_vars - num_vars) E.g. Giving multivariate poly $f(X) = f_1(X1) + f_2(X), X1 \in {F}^{n'}, X \in {F}^{n}, |X1| := n', |X| = n, n' <= n$ For i round univariate poly, $f^i(x)$ $f^i[0] = \sum_b f(r, 0, b), b \in [0, 1]^{n-i-1}, r \in {F}^{n-i-1}$ chanllenge get from prev rounds = $\sum_b f_1(r, 0, b1) + f_2(r, 0, b), |b| >= |b1|, |b| - |b1| = n - n'$ = $2^{(|b| - |b1|)} * \sum_{b1} f_1(r, 0, b1) + \sum_b f_2(r, 0, b)$ same applied on f^i[1] It imply that, for every evals in f_1, to compute univariate poly, we just need to times a factor 2^(|b| - |b1|) for it evaluation value ### benchmark benchmark with ceno_zkvm `riscv_add`, and gkr `keccak` both remain the same and no impact. You might see some redundancy coding style, but this is for retain the best performance. I tried other variants and it impact benchmark results ### scope Related to #109 #210 .... To address #126 #127 This enhance protocol features potiential can be used for `range table-circuit`, `init/final-memory`, `cpu-init/cpu-final halt` to make selector sumcheck support batching different num_instance witin.

This is another approach to #223, fixing #222 / #210. Features: - Prepare for parallel proving with independent transcripts per thread. - Soundness with different transcripts in each thread. - Type-safe API which captures the transcript by value. There is no function to merge the threads back into one transcript, but that can be added if ever needed. --------- Co-authored-by: Aurélien Nicolas <[email protected]>

_Issue #110._ --------- Co-authored-by: Aurélien Nicolas <[email protected]>

### Goal To make sumcheck protocol support different num_vars, aiming for - [x] minimal change to sumcheck protocol, make verifier remain the same. no extra meta data passed from prover, and what prover have just mle and it's eval size Besides this PR also remove some parallism in verifier since it's unnecessary for relative low cost. ### design rationale (Also comments in codebase for reference) To deal with different num_vars, we exploit a fact that for each product which num_vars < max_num_vars, for it evaluation value we need to times 2^(max_num_vars - num_vars) E.g. Giving multivariate poly $f(X) = f_1(X1) + f_2(X), X1 \in {F}^{n'}, X \in {F}^{n}, |X1| := n', |X| = n, n' <= n$ For i round univariate poly, $f^i(x)$ $f^i[0] = \sum_b f(r, 0, b), b \in [0, 1]^{n-i-1}, r \in {F}^{n-i-1}$ chanllenge get from prev rounds = $\sum_b f_1(r, 0, b1) + f_2(r, 0, b), |b| >= |b1|, |b| - |b1| = n - n'$ = $2^{(|b| - |b1|)} * \sum_{b1} f_1(r, 0, b1) + \sum_b f_2(r, 0, b)$ same applied on f^i[1] It imply that, for every evals in f_1, to compute univariate poly, we just need to times a factor 2^(|b| - |b1|) for it evaluation value ### benchmark benchmark with ceno_zkvm `riscv_add`, and gkr `keccak` both remain the same and no impact. You might see some redundancy coding style, but this is for retain the best performance. I tried other variants and it impact benchmark results ### scope Related to #109 #210 .... To address #126 #127 This enhance protocol features potiential can be used for `range table-circuit`, `init/final-memory`, `cpu-init/cpu-final halt` to make selector sumcheck support batching different num_instance witin.

Aurélien Nicolas added 2 commits September 11, 2024 15:34

program-table: initial program table

7ed82fc

program-table: fetch instruction from ADD circuit

adcfa59

naure requested a review from kunxian-xia September 11, 2024 15:41

Aurélien Nicolas added 3 commits September 11, 2024 20:15

program-table: Example passing with PC and opcode

7e29198

program-table: Record type

cab400f

program-table: Support registers and funct

3eeb0f7

naure force-pushed the program-table branch from 8dac4bc to 3eeb0f7 Compare September 11, 2024 19:51

naure marked this pull request as ready for review September 11, 2024 19:53

naure changed the title ~~program-table: initial program table~~ program-table Sep 11, 2024

program-table: put back example parameters

7e5c718

naure force-pushed the program-table branch from 1750502 to 7e5c718 Compare September 11, 2024 20:10

program-table: constants

b7ec526

naure force-pushed the program-table branch from 384236f to b7ec526 Compare September 11, 2024 20:29

This was linked to issues Sep 12, 2024

add program table circuit #110

Closed

register machine bytecode lookup #104

Closed

kunxian-xia added the table circuit label Sep 12, 2024

kunxian-xia assigned naure Sep 12, 2024

hero78119 self-requested a review September 12, 2024 06:39

kunxian-xia reviewed Sep 12, 2024

View reviewed changes

ceno_zkvm/src/instructions/riscv/addsub.rs Outdated Show resolved Hide resolved

hero78119 reviewed Sep 12, 2024

View reviewed changes

ceno_zkvm/src/tables/program.rs Show resolved Hide resolved

Merge branch 'master' into program-table

4739ec8

naure mentioned this pull request Sep 13, 2024

Decode immediates in program table #221

Closed

program-table: move constants

cffa410

Aurélien Nicolas added 2 commits September 13, 2024 17:25

program-table: zip != enumerate

b70aed7

program-table: Support for MockProver

5e99ca4

naure requested a review from kunxian-xia September 13, 2024 17:33

kunxian-xia mentioned this pull request Sep 14, 2024

Fix: the transcript should not depend on the proving order #223

Closed

naure mentioned this pull request Sep 14, 2024

Fix/transcript fork #224

Merged

Aurélien Nicolas added 3 commits September 14, 2024 12:43

Merge branch 'master' into program-table

a1f746a

program-table: fix after merge

4c19e86

program-table: support SUB in MockProver

a5dde7c

kunxian-xia approved these changes Sep 14, 2024

View reviewed changes

kunxian-xia merged commit a8174ae into master Sep 14, 2024
6 checks passed

kunxian-xia deleted the program-table branch September 14, 2024 13:42

hero78119 mentioned this pull request Sep 17, 2024

sumcheck protocol support mixed num_vars monomial form #235

Merged

1 task

hero78119 pushed a commit that referenced this pull request Sep 30, 2024

program-table (#210)

2c950a3

_Issue #110._ --------- Co-authored-by: Aurélien Nicolas <[email protected]>

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

program-table #210

program-table #210

naure commented Sep 11, 2024 •

edited

Loading

kunxian-xia left a comment •

edited

Loading

hero78119 left a comment

hero78119 Sep 12, 2024

naure Sep 13, 2024

naure Sep 13, 2024 •

edited

Loading

naure commented Sep 13, 2024 •

edited

Loading

naure commented Sep 13, 2024

naure commented Sep 14, 2024

program-table #210

program-table #210

Conversation

naure commented Sep 11, 2024 • edited Loading

kunxian-xia left a comment • edited Loading

Choose a reason for hiding this comment

hero78119 left a comment

Choose a reason for hiding this comment

hero78119 Sep 12, 2024

Choose a reason for hiding this comment

naure Sep 13, 2024

Choose a reason for hiding this comment

naure Sep 13, 2024 • edited Loading

Choose a reason for hiding this comment

naure commented Sep 13, 2024 • edited Loading

naure commented Sep 13, 2024

naure commented Sep 14, 2024

naure commented Sep 11, 2024 •

edited

Loading

kunxian-xia left a comment •

edited

Loading

naure Sep 13, 2024 •

edited

Loading

naure commented Sep 13, 2024 •

edited

Loading