From 955dcb6a209717c2912ecde552eb11c6678e7f23 Mon Sep 17 00:00:00 2001
From: Chris Duncan <chris@zoso.dev>
Date: Wed, 8 Jan 2025 12:08:33 -0800
Subject: [PATCH] Inline G mix 1 of round 0.

---
 src/lib/nano-pow/shaders/gpu-compute.ts | 89 ++++++++++++++++++++++++-
 1 file changed, 87 insertions(+), 2 deletions(-)

diff --git a/src/lib/nano-pow/shaders/gpu-compute.ts b/src/lib/nano-pow/shaders/gpu-compute.ts
index f710de6..0fcf82e 100644
--- a/src/lib/nano-pow/shaders/gpu-compute.ts
+++ b/src/lib/nano-pow/shaders/gpu-compute.ts
@@ -289,9 +289,94 @@ fn main(
 
 
 
-	/* ROUND 0.2 */
+	/**
+	* r=0, i=2, a=v[2-3], b=v[10-11], c=v[18-19], d=v[26-27]
+	*/
+
+	// a = a + b;
+	o0 = v2 + v10;
+	o1 = v3 + v11;
+	if (v2 > 0xFFFFFFFFu - v10) {
+		o1 = o1 + 1u;
+	}
+	v2 = o0;
+	v3 = o1;
+
+	// a = a + m[sigma[r][2*i+0]];
+	o0 = v2 + m4;
+	o1 = v3 + m5;
+	if (v2 > 0xFFFFFFFFu - m4) {
+		o1 = o1 + 1u;
+	}
+	v2 = o0;
+	v3 = o1;
+
+	// d = rotr64(d ^ a, 32);
+	xor0 = v26 ^ v2;
+	xor1 = v27 ^ v3;
+	v26 = xor1;
+	v27 = xor0;
+
+	// c = c + d;
+	o0 = v18 + v26;
+	o1 = v19 + v27;
+	if (v18 > 0xFFFFFFFFu - v26) {
+		o1 = o1 + 1u;
+	}
+	v18 = o0;
+	v19 = o1;
+
+	// b = rotr64(b ^ c, 24);
+	xor0 = v10 ^ v18;
+	xor1 = v11 ^ v19;
+	v10 = (xor0 >> 24u) ^ (xor1 << 8u);
+	v11 = (xor1 >> 24u) ^ (xor0 << 8u);
+
+	// a = a + b;
+	o0 = v2 + v10;
+	o1 = v3 + v11;
+	if (v2 > 0xFFFFFFFFu - v10) {
+		o1 = o1 + 1u;
+	}
+	v2 = o0;
+	v3 = o1;
+
+	// a = a + m[sigma[r][2*i+1]];
+	o0 = v2 + m6;
+	o1 = v3 + m7;
+	if (v2 > 0xFFFFFFFFu - m6) {
+		o1 = o1 + 1u;
+	}
+	v2 = o0;
+	v3 = o1;
+
+	// d = rotr64(d ^ a, 16)
+	xor0 = v26 ^ v2;
+	xor1 = v27 ^ v3;
+	v26 = (xor0 >> 16u) ^ (xor1 << 16u);
+	v27 = (xor1 >> 16u) ^ (xor0 << 16u);
+
+	// c = c + d;
+	o0 = v18 + v26;
+	o1 = v19 + v27;
+	if (v18 > 0xFFFFFFFFu - v26) {
+		o1 = o1 + 1u;
+	}
+	v18 = o0;
+	v19 = o1;
+
+	// b = rotr64(b ^ c, 63)
+	xor0 = v10 ^ v18;
+	xor1 = v11 ^ v19;
+	v10 = (xor1 >> 31u) ^ (xor0 << 1u);
+	v11 = (xor0 >> 31u) ^ (xor1 << 1u);
+
+
+
+	/**
+	* ROUND 0.2
+	*/
 
-	G(&v2, &v3, &v10, &v11, &v18, &v19, &v26, &v27, m4, m5, m6, m7);
 	G(&v4, &v5, &v12, &v13, &v20, &v21, &v28, &v29, m8, m9, 0u, 0u);
 	G(&v6, &v7, &v14, &v15, &v22, &v23, &v30, &v31, 0u, 0u, 0u, 0u);
 	G(&v0, &v1, &v10, &v11, &v20, &v21, &v30, &v31, 0u, 0u, 0u, 0u);
-- 
2.34.1