@@ -85,8 +85,8 @@ located in a module [zksync core witness]. However, for the new proof system, th
8585new location called [ separate witness binary] .
8686
8787Inside this new location, after the necessary data is fetched from storage, the witness generator calls another piece of
88- code from [ zkevm_test_harness witness] named ` run_with_fixed_params ` . This code is responsible for creating the
89- witnesses themselves (which can get really HUGE).
88+ code from [ zkevm_test_harness witness] named ` run_with_fixed_params ` . This code is responsible for creating the witnesses
89+ themselves (which can get really HUGE).
9090
9191## Generating the Proof
9292
@@ -96,9 +96,9 @@ The main goal of this step is to take an operation (for example, a calculation c
9696into smaller pieces. Then, we represent this information as a special mathematical expression called a polynomial.
9797
9898To construct these polynomials, we use something called a ` ConstraintSystem ` . The specific type that we use is called
99- zkSNARK, and our custom version of it is named bellman. You can find our code for this in the [ bellman repo] .
100- Additionally, we have an optimized version that's designed to run faster on certain types of hardware (using CUDA
101- technology), which you can find in the [ bellman cuda repo] .
99+ zkSNARK, and our custom version of it is named bellman. You can find our code for this in the [ bellman repo] . Additionally,
100+ we have an optimized version that's designed to run faster on certain types of hardware (using CUDA technology), which you
101+ can find in the [ bellman cuda repo] .
102102
103103An [ example ecrecover circuit] might give you a clearer picture of what this looks like in practice.
104104
@@ -107,9 +107,9 @@ heavy calculations, we use GPUs to speed things up.
107107
108108### Where is the Code
109109
110- The main code that utilizes the GPUs to create proofs is located in a repository named [ heavy_ops_service repo] . This
111- code combines elements from the [ bellman cuda repo] that we mentioned earlier, along with a huge amount of data
112- generated by the witness, to produce the final proofs.
110+ The main code that utilizes the GPUs to create proofs is located in a repository named [ heavy_ops_service repo] . This code
111+ combines elements from the [ bellman cuda repo] that we mentioned earlier, along with a huge amount of data generated by the
112+ witness, to produce the final proofs.
113113
114114## What Does "Verify Proof on L1" Mean
115115
0 commit comments