Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 12 Apr 1996 11:49:38 -1000 (HST)
From:      "David Langford" <langfod@dihelix.com>
To:        hackers@freebsd.org
Subject:   Interesting pgcc results
Message-ID:  <199604122149.LAA24980@caliban.dihelix.com>

next in thread | raw e-mail | index | archive | help
I have been trying to do some benchmarking with my new computer
(PC Importers Deep Green PCI 486 motherboard adn AMD 5x86 160Mhz CPU)

Here are three BYTEbench reports all compiled with pgcc.

1) first one is with my normal stock cc -m486 kernel 
2) second is with a pgcc -mpentium -O3 kernel
3) third is with a pgcc -mpentium -O3 kernel and entire system (world)
   built with pgcc -mpentium -O2 - I will send more info on this compile later.

#3 is the most confusing- the two benchmarks I have run seem to indicate
   that the system is a little slower this way :-/

Also if you notice the dates there have been some changes in the current
kernel over this time. 

-  BYTE UNIX Benchmarks (Version 3.11)
-  System -- caliban
-  Start Benchmark Run: Sun Apr  7 09:31:27 HST 1996
-   3 interactive users.
- Arithmetic Test (type = double)               2541.7    13924.6        5.5
- Dhrystone 2 without register variables       22366.3   164463.7        7.4
- Execl Throughput Test                           16.5      319.0       19.3
- File Copy  (30 seconds)                        179.0     2107.0       11.8
- Pipe-based Context Switching Test             1318.5    12342.8        9.4
- Shell scripts (8 concurrent)                     4.0       36.7        9.2
-                                                                  =========
-      SUM of  6 items                                                  62.5
-      AVERAGE                                                          10.4


Kernel built with pgcc:

-  BYTE UNIX Benchmarks (Version 3.11)
-  System -- caliban
-  Start Benchmark Run: Thu Apr 11 18:31:10 HST 1996
-   2 interactive users.
- Arithmetic Test (type = double)               2541.7    13922.4        5.5
- Dhrystone 2 without register variables       22366.3   164075.4        7.3
- Execl Throughput Test                           16.5      379.7       23.0
- File Copy  (30 seconds)                        179.0     2370.0       13.2
- Pipe-based Context Switching Test             1318.5    15169.2       11.5
- Shell scripts (8 concurrent)                     4.0       39.7        9.9
-                                                                  =========
-      SUM of  6 items                                                  70.5
-      AVERAGE                                                          11.7



Kernel AND everything else in the system built with pgcc:

-  BYTE UNIX Benchmarks (Version 3.11)
-  System -- caliban
-  Start Benchmark Run: Fri Apr 12 09:57:01 HST 1996
-   3 interactive users.
- Arithmetic Test (type = double)               2541.7    13979.3        5.5
- Dhrystone 2 without register variables       22366.3   164505.3        7.4
- Execl Throughput Test                           16.5      364.4       22.1
- File Copy  (30 seconds)                        179.0     2194.0       12.3
- Pipe-based Context Switching Test             1318.5    12293.0        9.3
- Shell scripts (8 concurrent)                     4.0       40.0       10.0
-                                                                  =========
-      SUM of  6 items                                                  66.5
-      AVERAGE                                                          11.1

--
  David Langford
  langfod@dihelix.com
FreeBSD caliban 2.2-CURRENT FreeBSD 2.2-CURRENT #0: Fri Apr 12 08:12:34 HST 1996



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199604122149.LAA24980>