You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: TODO.md
+23Lines changed: 23 additions & 0 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -27,6 +27,25 @@
27
27
- Opti WHIR https://github.com/tcoratger/whir-p3/issues/303 and https://github.com/tcoratger/whir-p3/issues/306
28
28
- Avoid committing to the 3 index columns, and replace it by a sumcheck? Using this idea, we would only commit to PC and FP for the execution table. Idea by Georg (Powdr). Do we even need to commit to FP then?
29
29
30
+
- About the ordering of the variables in sumchecks, currently we do as follows:
About "the packed pcs" (similar to SP1 Jagged PCS, slightly less efficient, but simpler (no sumchecks)):
31
50
- The best strategy is probably to pack as much as possible (the cost increasing the density = additional inner evaluations), if we can fit below a power of 2 - epsilon (epsilon = 20% for instance, tbd), if the sum of the non zero data is just above a power of 2, no packed technique, even the best, can help us, so we should spread aniway (to reduce the pressure of inner evaluations)
32
51
- About those inner evaluations, there is a trick: we need to compute M1(a, b, c, d, ...) then M2(b, c, d, ...), then M3(c, d, ...) -> The trick = compute the "eq(.) for (b, c, d), then dot product with M3. Then expand to eq(b, c, d, ...), dot product with M2. Then expand to eq(a, b, c, d, ...), dot product with M1. The idea is that in this order, computing each "eq" is easier is we start from the previous one.
@@ -68,6 +87,10 @@ But we reduce proof size a lot using instead (TODO):
68
87
└────────────────────────┘└──────────┘└─┘
69
88
```
70
89
90
+
## Security:
91
+
92
+
Fiat Shamir: add a claim tracing feature, to ensure all the claims are indeed checked (Lev)
93
+
71
94
## Not Perf
72
95
73
96
- Whir batching: handle the case where the second polynomial is too small compared to the first one
0 commit comments