Change default backend from simd_backend to u32_backend

Cargo.toml

@@ -5,16 +5,15 @@ polynomials = "0.2"
 [dependencies.curve25519-dalek]
 default-features = false
 version = "3.0"
-features = ["simd_backend", "nightly", "serde", "alloc"]
+features = ["nightly", "serde", "alloc"]
 
 [dependencies.blake2]
 default-features = false
 version = "0.9"
-features = ["simd_asm"]
 
 [dependencies.merlin]
 default-features = false
-version = "2"
+ersion = "2"
 
 [dependencies.serde]
 default-features = false
@@ -26,8 +25,9 @@ optional = true
 criterion = "0.3"
 
 [features]
-default = ["std", "serde"]
+default = ["std", "serde", "curve25519-dalek/u32_backend"]
 std = []
+simd_backend = ["curve25519-dalek/simd_backend", "blake2/simd_asm"]
 
 [[bench]]
 name = "prove_and_verify"

README.md

@@ -6,11 +6,11 @@ proof system [[link][arcturus-paper]].
 Arcturus enables efficient proof and verification of confidential transactions with very large
 anonymity sets. A correct proof provides the following guarantees:
 1) The signer posesses the signing key for each spent output in the ring.
-1) The sum of spent inputs matches the sum of newly minted outputs.
+1) The sum of spent inputs matches the sum of newly minted outputs.<sup>[1](#usage-notes)</sup>
 1) Each spent input is accompanied with a unique, deterministic, linking tag to detect double
-   spends.
+   spends.<sup>[2](#usage-notes)</sup>
 1) The transaction input and output values are hidden (aka confidential).
-1) The transaction inputs and signing keys are hidden in a large anonymity set.
+1) The transaction inputs and signing keys are hidden in a large anonymity set.<sup>[3](#usage-notes)</sup>
 
 # ⚠️  Security Warning
 This crate is a work in progress and has not been independently audited!
@@ -89,7 +89,7 @@ assert!(gens.verify(&mut t, &ring[..], &proofs[..]).is_ok());
 Benchmarks are run using [criterion.rs][criterion-crate].
 ```sh
 export RUSTFLAGS="-C target_cpu=native"
-cargo bench
+cargo bench --no-default-features --features "std simd_backend serde"
 ```
 
 # Contributing

benches/prove_and_verify.rs

@@ -49,7 +49,6 @@ fn proof_time(c: &mut Criterion) {
     for cfg in [
         (16, 2, 3),
         (16, 3, 3),
-        (16, 4, 3),
     ]
     .iter()
     {
@@ -84,7 +83,6 @@ fn verification_time(c: &mut Criterion) {
     for cfg in [
         (16, 2, 3),
         (16, 3, 3),
-        (16, 4, 3),
     ]
     .iter()
     {