From 3ed947fb30727aab198e164853cf3dcfc960429e Mon Sep 17 00:00:00 2001
From: antirez <antirez@gmail.com>
Date: Sat, 29 Mar 2014 12:12:44 +0100
Subject: [PATCH] HLLCOUNT 3x faster taking fast path for default params.

---
 src/hyperloglog.c | 56 ++++++++++++++++++++++++++++++++++++++---------
 1 file changed, 46 insertions(+), 10 deletions(-)

diff --git a/src/hyperloglog.c b/src/hyperloglog.c
index 50983c33..951f8bc5 100644
--- a/src/hyperloglog.c
+++ b/src/hyperloglog.c
@@ -290,16 +290,52 @@ uint64_t hllCount(uint8_t *registers) {
         initialized = 1;
     }
 
-    /* Compute SUM(2^-register[0..i]). */
-    for (j = 0; j < REDIS_HLL_REGISTERS; j++) {
-        unsigned long reg;
+    /* Compute SUM(2^-register[0..i]).
+     * Redis default is to use 16384 registers 6 bits each. The code works
+     * with other values by modifying the defines, but for our target value
+     * we take a faster path with unrolled loops. */
+    if (REDIS_HLL_REGISTERS == 16384 && REDIS_HLL_BITS == 6 && 1) {
+        uint8_t *r = registers;
+        unsigned long r0, r1, r2, r3, r4, r5, r6, r7, r8, r9,
+                      r10, r11, r12, r13, r14, r15;
+        for (j = 0; j < 1024; j++) {
+            /* Handle 16 registers per iteration. */
+            r0 = r[0] & 63; if (r0 == 0) ez++;
+            r1 = (r[0] << 6 | r[1] >> 2) & 63; if (r1 == 0) ez++;
+            r2 = (r[1] << 4 | r[2] >> 4) & 63; if (r2 == 0) ez++;
+            r3 = ((r[2] << 2) | (r[3] >> 6)) & 63; if (r3 == 0) ez++;
+            r4 = (r[3] | r[4] >> 8) & 63; if (r4 == 0) ez++;
+            r5 = (r[3] << 6 | r[4] >> 2) & 63; if (r5 == 0) ez++;
+            r6 = (r[4] << 4 | r[5] >> 4) & 63; if (r6 == 0) ez++;
+            r7 = (r[5] << 2 | r[6] >> 6) & 63; if (r7 == 0) ez++;
+            r8 = (r[6] | r[7] >> 8) & 63; if (r8 == 0) ez++;
+            r9 = (r[6] << 6 | r[7] >> 2) & 63; if (r9 == 0) ez++;
+            r10 = (r[7] << 4 | r[8] >> 4) & 63; if (r10 == 0) ez++;
+            r11 = (r[8] << 2 | r[9] >> 6) & 63; if (r11 == 0) ez++;
+            r12 = (r[9] | r[10] >> 8) & 63; if (r12 == 0) ez++;
+            r13 = (r[9] << 6 | r[10] >> 2) & 63; if (r13 == 0) ez++;
+            r14 = (r[10] << 4 | r[11] >> 4) & 63; if (r14 == 0) ez++;
+            r15 = (r[11] << 2 | r[12] >> 6) & 63; if (r15 == 0) ez++;
 
-        HLL_GET_REGISTER(reg,registers,j);
-        if (reg == 0) {
-            ez++;
-            E += 1; /* 2^(-reg[j]) is 1 when m is 0. */
-        } else {
-            E += PE[reg]; /* Precomputed 2^(-reg[j]). */
+            /* Additional parens will allow the compiler to optimize the
+             * code more with a loss of precision that is not very relevant
+             * here (floating point math is not commutative!). */
+            E += (PE[r0] + PE[r1]) + (PE[r2] + PE[r3]) + (PE[r4] + PE[r5]) +
+                 (PE[r6] + PE[r7]) + (PE[r8] + PE[r9]) + (PE[r10] + PE[r11]) +
+                 (PE[r12] + PE[r13]) + (PE[r14] + PE[r15]);
+            r += 12;
+        }
+    } else {
+        for (j = 0; j < REDIS_HLL_REGISTERS; j++) {
+            unsigned long reg;
+
+            HLL_GET_REGISTER(reg,registers,j);
+            if (reg == 0) {
+                ez++;
+                E += 1; /* 2^(-reg[j]) is 1 when m is 0. */
+            } else {
+                E += PE[reg]; /* Precomputed 2^(-reg[j]). */
+            }
         }
     }
     /* Muliply the inverse of E for alpha_m * m^2 to have the raw estimate. */
@@ -376,7 +412,7 @@ void hllCountCommand(redisClient *c) {
 
     if (o == NULL) {
         /* No key? Cardinality is zero since no element was added, otherwise
-         * we would have a key as HHLADD creates it as a side effect. */
+         * we would have a key as HLLADD creates it as a side effect. */
         addReply(c,shared.czero);
     } else {
         /* Key exists, check type */