diff --git a/evm/Cargo.toml b/evm/Cargo.toml
index e942a9f9..2d1e2850 100644
--- a/evm/Cargo.toml
+++ b/evm/Cargo.toml
@@ -6,6 +6,8 @@ edition = "2021"
 
 [dependencies]
 anyhow = "1.0.40"
+blake = "2.0.2"
+env_logger = "0.9.0"
 eth_trie_utils = "0.4.0"
 ethereum-types = "0.14.0"
 hex = { version = "0.4.3", optional = true }
diff --git a/evm/src/cpu/kernel/aggregator.rs b/evm/src/cpu/kernel/aggregator.rs
index 01ba419b..25e25119 100644
--- a/evm/src/cpu/kernel/aggregator.rs
+++ b/evm/src/cpu/kernel/aggregator.rs
@@ -41,7 +41,6 @@ pub(crate) fn combined_kernel() -> Kernel {
         include_str!("asm/halt.asm"),
         include_str!("asm/hash/blake/compression.asm"),
         include_str!("asm/hash/blake/g_functions.asm"),
-        include_str!("asm/hash/blake/initial_state.asm"),
         include_str!("asm/hash/blake/iv.asm"),
         include_str!("asm/hash/blake/ops.asm"),
         include_str!("asm/hash/blake/permutations.asm"),
diff --git a/evm/src/cpu/kernel/asm/hash/blake/compression.asm b/evm/src/cpu/kernel/asm/hash/blake/compression.asm
index f06df538..44b10ffb 100644
--- a/evm/src/cpu/kernel/asm/hash/blake/compression.asm
+++ b/evm/src/cpu/kernel/asm/hash/blake/compression.asm
@@ -1,14 +1,14 @@
 %macro blake_initial_state
-    %blake_iv(7)
-    %blake_iv(6)
-    %blake_iv(5)
-    %blake_iv(4)
-    %blake_iv(3)
-    %blake_iv(2)
-    %blake_iv(1)
+    %blake_iv_i(7)
+    %blake_iv_i(6)
+    %blake_iv_i(5)
+    %blake_iv_i(4)
+    %blake_iv_i(3)
+    %blake_iv_i(2)
+    %blake_iv_i(1)
     // stack: IV_1, IV_2, IV_3, IV_4, IV_5, IV_6, IV_7
     PUSH 0x01010040 // params: key = 00, digest_size = 64 = 0x40
-    %blake_iv(0)
+    %blake_iv_i(0)
     XOR
     // stack: IV_0 ^ params, IV_1, IV_2, IV_3, IV_4, IV_5, IV_6, IV_7
 %endmacro
@@ -19,7 +19,8 @@
     %mload_kernel_general
     // stack: num_blocks
     %mul_const(128)
-    // stack: num_bytes
+    %increment
+    // stack: num_bytes+1
 %endmacro
 
 %macro blake_message_addr
@@ -30,7 +31,7 @@
 global blake_compression:
     %blake_initial_state
     // stack: t_0, t_1, h_0, h_1, h_2, h_3, h_4, h_5, h_6, h_7
-    %stack: () -> (0, 0, 0)
+    %stack () -> (0, 0, 0)
     // stack: cur_block = 0, t_0 = 0, t_1 = 0, h_0, h_1, h_2, h_3, h_4, h_5, h_6, h_7
 
 
diff --git a/evm/src/cpu/kernel/asm/hash/blake/permutations.asm b/evm/src/cpu/kernel/asm/hash/blake/permutations.asm
index 6f444c3b..da278ec2 100644
--- a/evm/src/cpu/kernel/asm/hash/blake/permutations.asm
+++ b/evm/src/cpu/kernel/asm/hash/blake/permutations.asm
@@ -1,4 +1,4 @@
-permutation_1_constants:
+global permutation_1_constants:
     BYTES 14
     BYTES 10
     BYTES 4
@@ -16,7 +16,7 @@ permutation_1_constants:
     BYTES 5
     BYTES 3
 
-permutation_2_constants:
+global permutation_2_constants:
     BYTES 11
     BYTES 8
     BYTES 12
@@ -34,7 +34,7 @@ permutation_2_constants:
     BYTES 9
     BYTES 4
 
-permutation_3_constants:
+global permutation_3_constants:
     BYTES 7
     BYTES 9
     BYTES 3
@@ -52,7 +52,7 @@ permutation_3_constants:
     BYTES 15
     BYTES 8
 
-permutation_4_constants:
+global permutation_4_constants:
     BYTES 9
     BYTES 0
     BYTES 5
@@ -70,7 +70,7 @@ permutation_4_constants:
     BYTES 3
     BYTES 13
 
-permutation_5_constants:
+global permutation_5_constants:
     BYTES 2
     BYTES 12
     BYTES 6
@@ -88,7 +88,7 @@ permutation_5_constants:
     BYTES 1
     BYTES 9
 
-permutation_6_constants:
+global permutation_6_constants:
     BYTES 12
     BYTES 5
     BYTES 1
@@ -106,7 +106,7 @@ permutation_6_constants:
     BYTES 8
     BYTES 11
 
-permutation_7_constants:
+global permutation_7_constants:
     BYTES 13
     BYTES 11
     BYTES 7
@@ -124,7 +124,7 @@ permutation_7_constants:
     BYTES 2
     BYTES 10
 
-permutation_8_constants:
+global permutation_8_constants:
     BYTES 6
     BYTES 15
     BYTES 14
@@ -142,7 +142,7 @@ permutation_8_constants:
     BYTES 10
     BYTES 5
 
-permutation_9_constants:
+global permutation_9_constants:
     BYTES 10
     BYTES 2
     BYTES 8
diff --git a/evm/src/cpu/kernel/tests/hash.rs b/evm/src/cpu/kernel/tests/hash.rs
index b7b2e5c2..52c8d9dd 100644
--- a/evm/src/cpu/kernel/tests/hash.rs
+++ b/evm/src/cpu/kernel/tests/hash.rs
@@ -1,6 +1,7 @@
 use std::str::FromStr;
 
 use anyhow::Result;
+use blake::{hash as blake_hash};
 use ethereum_types::U256;
 use rand::{thread_rng, Rng};
 use ripemd::{Digest, Ripemd160};
@@ -23,6 +24,13 @@ fn ripemd(input: Vec<u8>) -> U256 {
     U256::from(&hasher.finalize()[..])
 }
 
+/// Standard Blake implementation.
+fn blake(input: Vec<u8>) -> U256 {
+    let mut result = [0; 32];
+    blake_hash(256, &input, &mut result).unwrap();
+    U256::from(result)
+}
+
 fn make_random_input() -> Vec<u8> {
     // Generate a random message, between 0 and 9999 bytes.
     let mut rng = thread_rng();
@@ -88,3 +96,8 @@ fn test_sha2() -> Result<()> {
 fn test_ripemd() -> Result<()> {
     test_hash("ripemd_stack", &ripemd)
 }
+
+#[test]
+fn test_blake() -> Result<()> {
+    test_hash("blake", &blake)
+}