X-POP3-Rcpt:	mlehmann AT universe DOT sgh-net DOT de
21 Jan 1998 10:19:29 +0100 (CET) :
From:	Holger Burbach <holly AT cthulhu DOT rhein-neckar DOT de>
X-Sender:	holly AT uggae DOT rhein-neckar DOT de
To:	Jack Duan <jduan AT sprintmail DOT com>
cc:	Brian Makin <makinb AT ctc DOT com>, beastium-list AT Desk DOT nl
Subject:	Re: PGCC optimizing AMD K6?
Message-ID:	<Pine.LNX.3.96.980121100814.5217A-100000@uggae.rhein-neckar.de>
MIME-Version:	1.0
Sender:	Marc Lehmann <pcg AT goof DOT com>
Status:	RO
Lines:	50

Yes it does. Use the -mamdk6 command line option.

I've run the Byte Benchmark (nbench-byte-2.1) with gcc, egcs, and pgcc to
see what execution improvements can be expected.

All tests were executed on the same platform:

K6 200MHz, 512kb 2nd-level Cache, linux-2.1.79, glibc-2.0.5

===========================================================================
                      (0)  (1)  (2)  (3)  (4)  (5)  (6)  (7)  (8)  (9) (10)
===========================================================================
NUMERIC SORT         0.84 0.84 0.79 0.84 0.84 0.84 0.79 0.88 0.82 0.72 0.91
STRING SORT          0.75 0.77 0.78 0.78 0.82 0.83 0.82 0.84 0.80 0.78 0.82
BITFIELD             0.87 0.88 0.87 0.89 0.87 0.94 0.86 0.91 0.87 0.83 0.90
FP EMULATION         0.81 0.71 0.74 0.80 0.75 0.76 0.85 0.78 0.82 0.89 0.88
FOURIER              1.23 1.23 1.20 1.20 1.20 1.21 1.22 1.23 1.22 1.17 1.22
ASSIGNMENT           1.08 0.83 0.84 1.10 0.93 0.93 1.21  ++   ++  1.18 1.19
IDEA                 0.87 0.69 0.70 0.91 0.68 0.74 0.89 0.67 0.74 0.98 0.94
HUFFMAN              0.81 0.79 0.82 0.78 0.77 0.76 0.78 0.88 0.87 0.83 0.93
NEURAL NET           0.83 0.89 0.86 0.89 0.91 0.88 0.92  **   **   **   **
LU DECOMPOSITION     1.11 1.13 1.12 1.25 1.19 1.19 1.34 1.15 1.17 1.35 1.34
===========================================================================
MEMORY INDEX         0.89 0.82 0.83 0.91 0.87 0.90 0.95 0.91 0.87 0.92 0.96
INTEGER INDEX        0.83 0.75 0.76 0.83 0.76 0.77 0.83 0.80 0.81 0.85 0.91
FLOATING-POINT INDEX 1.04 1.07 1.05 1.11 1.09 1.08 1.15 1.12 1.13 1.17 1.18

(0)  gcc-2.7.2.3 -O6 -m486 -funroll-loops
(1)  gcc-2.8.0   -O6 -m486 -funroll-loops
(2)  gcc-2.8.0   -O6 -mpentium -funroll-loops
(3)  gcc-2.8.0   -O6 -mpentiumpro -funroll-loops
(4)  egcs-980115 (haifa sched) -O6 -m486 -funroll-loops
(5)  egcs-980115 (haifa sched) -O6 -mpentium -funroll-loops
(6)  egcs-980115 (haifa sched) -O6 -mpentiumpro -funroll-loops
(7)  pgcc-980115 (haifa sched) -O6 -mpentiumpro -funroll-loops
(8)  pgcc-980115 (haifa sched) -O6 -mpentiumpro -funroll-loops
(9)  pgcc-980115 (haifa sched) -O6 -mpentiumpro -funroll-loops
(10) pgcc-980115 (haifa sched) -O6 -mamdk6 -funroll-loops

++ segmentation fault
** non-terminating

These problems are not K6 specific. They do occur on 486 and pentium, too.
 
-- 
"Nobody will ever need more than 640k RAM!" (Bill Gates)
"Windows 95 needs atleast 8 MB RAM."        (Bill Gates)
"Nobody will ever need Windows 95."         (logical conclusion)     

- Raw text -