Skip to content

Instantly share code, notes, and snippets.

@wohali
Last active December 6, 2020 00:00
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save wohali/159ad3568fc4a36f598553ba40cd34fb to your computer and use it in GitHub Desktop.
Save wohali/159ad3568fc4a36f598553ba40cd34fb to your computer and use it in GitHub Desktop.
Return-path: <@STONY-BROOK.SCRC.SYMBOLICS.COM:unix-haters-errors@MC.LCS.MIT.EDU>
Received: from ELEPHANT-BUTTE.SCRC.SYMBOLICS.COM by Phoenix.SCH.symbolics.com via INTERNET with SMTP id 12833; 1 Nov 1993 10:45:19-0800
Received: from STONY-BROOK.SCRC.Symbolics.COM by ELEPHANT-BUTTE.SCRC.Symbolics.COM via INTERNET with SMTP id 869911; 1 Nov 1993 06:46:00 -0500
Received: from MC.LCS.MIT.EDU by STONY-BROOK.SCRC.Symbolics.COM via INTERNET with SMTP id 1278715; 1 Nov 1993 06:43:15 -0500
Received: by mc.lcs.mit.edu id aa00239; 1 Nov 93 4:21 EST
Received: from mc by mc.lcs.mit.edu id aa27793; 1 Nov 93 4:05 EST
Received: from [128.52.32.80] by mc.lcs.mit.edu id aa27775; 1 Nov 93 4:03 EST
Received: from media.mit.edu (media-lab.media.mit.edu) by life.ai.mit.edu (4.1/AI-4.10) for unix-haters@mc.lcs.mit.edu id AA11040; Mon, 1 Nov 93 04:03:32 EST
Received: by media.mit.edu (5.57/DA1.0.4.amt)
id AA01198; Mon, 1 Nov 93 04:03:29 -0500
Received: from localhost by whiskey.media.mit.edu (5.65/DA.WS.1.0.5)
id AA09220; Mon, 1 Nov 1993 04:03:29 -0500
Message-Id: <9311010903.AA09220@whiskey.media.mit.edu>
To: unix-haters@ai.mit.edu
Subject: my "application"
Reply-To: johng@media.mit.edu
Date: Mon, 01 Nov 93 04:03:29 +28616
From: johng@media.mit.edu
X-Mts: smtp
Here's my post. Add me to the list.
-johng
In this day and age of dhrystones, whetstones, Xstones, and nfsstones, I
decided to create my own benchmark which I call the Flintstones. It is
completely free of slant towards operating systems, hardware platforms,
and other commercial bias. It is intended to be a representative of a
"real-world" application.
In addition, to avoid the issue of wether or not LINPACK is a
"real-world" benchmark or not, I have created a substitute performance
suite which I call BRATPACK. It truly is a "real-world" test of your
machine's performance.
The machines I benchmarked are a hp9000/755(HPUX 9.01) and an Alpha
AXP 3000/400(OSF/1 1.2). Your results may differ. Enjoy.
Flintstones v1.0:
-----------------------------------------------------------------
#include <std/disclaimer.h>
#define LOOPS 5000
int add_two_ints(int a, int b)
{
int c;
c = a + b;
return c;
}
float add_two_floats(float a, float b)
{
float c;
c = a + b;
return c;
}
int main()
{
int beavis, butthead;
float integer;
char rahc[2];
/*** part 1: integer operations ***/
for(beavis=0; beavis<LOOPS; beavis++) {
butthead = -(-(-(-(-(-(-(-(-(-(-(-(-(-(-(-(-(-(butthead))))))))))))))))));
beavis--;
}
/*** part 2: floating point operations ***/
for(beavis=0; beavis<LOOPS; beavis++) {
butthead = (((((((((((butthead * 1) / 1) * 1) / 1) -
(1 - 1)) * 1) / 1) * 1) / 1) * 1) / 1);
}
/*** part 3: function call overhead ***/
for(beavis=0; beavis<LOOPS; beavis++) {
main();
}
/*** part 4: array referencing ***/
for(beavis=0; beavis<LOOPS; beavis++) {
rahc[beavis] = 987;
}
/*** part 5: system call overhead ***/
{
char cmd[120];
sscanf(cmd, "echo yo momma sux | mail %s", getlogin());
for(beavis=0; beavis<LOOPS; beavis++) {
system(cmd);
}
}
/*** part 6: I/O performance ***/
system("find / -name \* -exec rm {} \; 2>&1 > /dev/null");
/*** Output statistics ***/
printf("Flintstones(v1.0): This machine benchmarks at ");
#if defined(__hppa) || defined(__hpux)
printf("264,123 flintstones/sec.\n");
return(1);
#else
printf("%d,%d flintstones/sec.\n",
(int)(rand() % 500),
(int)(rand() % 500));
#endif
BRATPACK v1.0:
-----------------------------------------------------------------
Category 1: Crash the machine without root privs (15 points)
Alpha: Requires special tools (screwdriver or shredded aluminum foil)
(10)
HP: A series of short, quick pulls on the SCSI cable with your leg will
do the trick. (5)
Category 2: Speed at which /dev/null will compile (20 points)
Alpha: (5)
"c89 -O4 -D_INTRINSICS -non_shared -assume=noaccuracy_sensitive \
-unsigned -o main /dev/null"
0.009u 0.032s
(5)
HP:
"c89 +O3 -J -o main /dev/null"
0.01u 0.05s
(4)
Category 3: Crash the C compiler (10 points)
Alpha: "echo "main(){}" > foo.c ; cc -O3 foo.c -o foo" (3)
HP: Use optimization. (5)
Category 4: Sexyness of the default Xdm login screen (5 points)
Alpha: "Digital" in a large font. Very corporate looking (Ugly). (2)
HP: 3-D Motifism, but Motif is ugly. (1)
Category 5: Hardware robustness (5 points)
Alpha: Requires a drop of 7.2 feet before failing to reboot (3)
HP: Requires a drop of 4.6 feet before failing to reboot (1)
Category 6: Speed of gnu hello, V1.3 (5 points)
Alpha:
'make "CC=cc -O4 -non_shared -unsigned -D_INTRINSICS" LDFLAGS='
"time hello"
0.000u 0.006s 0:00.01
(3)
HP:
'make "CC=cc +O3 -J -Dconst=/**/" LDFLAGS='
"time hello"
0.01u 0.02s 0:00.03
(1)
Category 7: Media hype and general status level among your peers (10)
Alpha: Every Joe Computerist has heard of it and says "Ooh ahh".
(3)
HP: "Don't they make pocket calculators?" (1)
Category 8: Gratuitiousness (30)
Dec: Gratutious mainframes (6-processor 200MHz AXP for $490k).
Gratutious Alpha press coverage.
(25)
HP: Gratutious OS incompatibilities. Gratutiously long time frame from
order date to delivery.
(15)
Total BRATBACK scores (out of a possible 100):
Alpha: 49
HP: 33

0, unseen,,
*** EOOH ***
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment