xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: George Dunlap <george.dunlap@eu.citrix.com>
To: xen-devel@lists.xensource.com
Cc: george.dunlap@eu.citrix.com
Subject: [PATCH 5 of 8] xenalyze: Rework math to remove two 64-bit divisions
Date: Thu, 26 Jan 2012 17:21:02 +0000	[thread overview]
Message-ID: <4b3639bd32550d40bfc4.1327598462@elijah> (raw)
In-Reply-To: <patchbomb.1327598457@elijah>

abs_cycles_to_time is called on every record for dump mode; it has
four 64-bit divisions (well, 3 divisions and 1 mod), which map to
library functions on a 32-bit platform.  A simple rework of the math
can eliminate two of those.

Signed-off-by: George Dunlap <george.dunlap@eu.citrix.com>

diff -r 0663e3e8f69d -r 4b3639bd3255 xenalyze.c
--- a/xenalyze.c	Thu Jan 26 17:16:59 2012 +0000
+++ b/xenalyze.c	Thu Jan 26 17:17:19 2012 +0000
@@ -1976,16 +1976,19 @@ void cpumask_union(cpu_mask_t *d, const 
 /* -- Time code -- */
 
 void cycles_to_time(unsigned long long c, struct time_struct *t) {
-    t->time = ((c) * 1000) / (opt.cpu_hz / 1000000 );    
-    t->s = t->time / 1000000000;                        
-    t->ns = t->time % 1000000000;                       
+    t->time = ((c - P.f.first_tsc) * 1000 * 1000000) / opt.cpu_hz;
+    t->s = t->time / 1000000000;
+    t->ns = t->time - (t->s * 1000000000);
 }
 
 void abs_cycles_to_time(unsigned long long ac, struct time_struct *t) {
     if(ac > P.f.first_tsc) {
-        t->time = ((ac - P.f.first_tsc) * 1000) / (opt.cpu_hz / 1000000 );    
-        t->s = t->time / 1000000000;                        
-        t->ns = t->time % 1000000000;
+        /* t->time = ((ac - P.f.first_tsc) * 1000) / (opt.cpu_hz / 1000000 );     */
+        /* t->s = t->time / 1000000000;                         */
+        /* t->ns = t->time % 1000000000; */
+        t->time = ((ac - P.f.first_tsc) * 1000 * 1000000) / opt.cpu_hz;
+        t->s = t->time / 1000000000;
+        t->ns = t->time - (t->s * 1000000000);
     } else {
         t->time = t->s = t->ns = 0;
     }

  parent reply	other threads:[~2012-01-26 17:21 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-01-26 17:20 [PATCH 0 of 8] xenalyze: Optimizations George Dunlap
2012-01-26 17:20 ` [PATCH 1 of 8] xenalyze: Improve record-sorting algorithm George Dunlap
2012-01-26 17:20 ` [PATCH 2 of 8] xenalyze: Remove spurious dump_header construction George Dunlap
2012-01-26 17:21 ` [PATCH 3 of 8] xenalyze: Remove --dump-cooked George Dunlap
2012-01-26 17:21 ` [PATCH 4 of 8] xenalyze: Enable -O2 optimization level George Dunlap
2012-01-26 17:21 ` George Dunlap [this message]
2012-01-26 17:21 ` [PATCH 6 of 8] xenalyze: Eliminate unnecessary cycles_to_time calculation George Dunlap
2012-01-26 17:21 ` [PATCH 7 of 8] xenalyze: Introduce more efficient read mechanism George Dunlap
2012-01-26 17:21 ` [PATCH 8 of 8] xenalyze: Get rid of redundant hvm dump_header George Dunlap

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4b3639bd32550d40bfc4.1327598462@elijah \
    --to=george.dunlap@eu.citrix.com \
    --cc=xen-devel@lists.xensource.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).