From: W. Felix Handte <w@felixhandte.com>
Date: Mon, 13 Dec 2021 20:46:41 +0000 (-0500)
Subject: Increment Step by 1 not 2
X-Git-Tag: v1.5.1~1^2~16^2~1
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=82a49c88f917c73b56780398fee65665a6e2a25e;p=thirdparty%2Fzstd.git

Increment Step by 1 not 2

I couldn't find a good way to spread `ip0` and `ip1` apart when we accelerate
due to incompressible inputs. (The methods I tried slowed things down quite a
bit.)

Since we aren't splaying ip0 and ip1 apart (which would be like `0_1_2_3_`, as
opposed to the `01__23__` we were actually doing), it's a big ambitious to
increment `step` by 2. Instead, let's increment it by 1, which has the benefit
sliiightly improving compression. Speed remains pretty much unchanged.
---

diff --git a/lib/compress/zstd_fast.c b/lib/compress/zstd_fast.c
index 852f4286f..2bae2f70e 100644
--- a/lib/compress/zstd_fast.c
+++ b/lib/compress/zstd_fast.c
@@ -234,19 +234,19 @@ _start: /* Requires: ip0 */
         hash0 = hash1;
         hash1 = ZSTD_hashPtr(ip2, hlog, mls);
 
+        /* advance to next positions */
+        ip0 = ip1;
+        ip1 = ip2;
+        ip2 = ip0 + step;
+        ip3 = ip1 + step;
+
         /* calculate step */
         if (ip2 >= nextStep) {
+            step++;
             PREFETCH_L1(ip1 + 64);
             PREFETCH_L1(ip1 + 128);
-            step += 2;
             nextStep += kStepIncr;
         }
-
-        /* advance to next positions */
-        ip0 = ip1;
-        ip1 = ip2;
-        ip2 = ip0 + step;
-        ip3 = ip1 + step;
     } while (ip3 < ilimit);
 
 _cleanup: