Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.sys.apple2.programmer > #863 > unrolled thread

DHRYSTONE 1.1 for Aztec C65 3.2b available for download

Back to article view | Back to comp.sys.apple2.programmer

Contents

  DHRYSTONE 1.1 for Aztec C65 3.2b available for download "Bill Buckels" <bbuckels@mts.net> - 2013-09-12 21:23 -0500
    Re: DHRYSTONE 1.1 for Aztec C65 3.2b available for download "Bill Buckels" <bbuckels@mts.net> - 2013-09-14 21:18 -0500
    Re: DHRYSTONE 1.1 for Aztec C65 3.2b available for download "Bill Buckels" <bbuckels@mts.net> - 2013-09-14 21:35 -0500
      Re: DHRYSTONE 1.1 for Aztec C65 3.2b available for download Steven Hirsch <snhirsch@gmail.com> - 2013-09-15 08:56 -0400
        Re: DHRYSTONE 1.1 for Aztec C65 3.2b available for download "Bill Buckels" <bbuckels@mts.net> - 2013-09-15 20:40 -0500
    Re: DHRYSTONE 1.1 for Aztec C65 3.2b available for download bill.rowe.ottawa@gmail.com - 2013-09-29 08:07 -0700
    Re: DHRYSTONE 1.1 for Aztec C65 3.2b available for download "Bill Buckels" <bbuckels@mts.net> - 2013-10-04 05:29 -0500
    Re: DHRYSTONE 1.1 for Aztec C65 3.2b available for download "Bill Buckels" <bbuckels@mts.net> - 2013-10-04 20:22 -0500
      Re: DHRYSTONE 1.1 for Aztec C65 3.2b available for download Antoine Vignau <antoine.vignau@laposte.net> - 2013-10-04 19:28 -0700

#863 — DHRYSTONE 1.1 for Aztec C65 3.2b available for download

DHRYSTONE for ProDOS 8 written in Aztec C65 3.2b is now available for 
download complete with disk image and source at the following link:

http://www.aztecmuseum.ca/extras/DHRY.zip

This version of DHRYSTONE requires the No-Slot Clock. I have only tested 
this in AppleWin where I get a result of 39 DHRYSTONES/second.

The DHRYSTONE Version 1.1 that I have used was last updated the same year 
that Aztec C65 3.2b was released. If you have the AppleX 2013 update 
installed, unzip this into the programs directory and it should build 
without problem. If you run this in AppleWin's fastest speed, or on a real 
Apple II with an accelerator you will need to expand the timing loop define 
because it has been written to run at stock 1 MHz speed.

I have no DHRYSTONE results for CC65 for the Apple II for comparison.

The only stats I can find are targetted at the Commodore where cc65 runs 28% 
faster than Aztec C65 3.2b (cc65 stats according to the following link):

http://www.cc65.org/mailarchive/2009-07/6616.html

So I am comparing Apples to Bananas at this point. I hope to have a 
DHRYSTONE for AztecC64 a little later in the week.

Now about CP/M 80...

The DHRYSTONE results for the Aztec C CPM/80 compiler do not indicate the 
platform so I don't know how the Softcard or the Applicard will compare. I 
have both here for my real Apple //e. I also don't know how my real C64 with 
the CP/M cart or my C128 will do. Also the DHRY results for CP/M 80 are for 
an older compiler. So it seems I have some work to do to gather these stats. 
Not sure about clock access either but on the Apple II I will need to do 
some research to figure-out if CP/M's memory space inteferes with my access 
to the NSC. The clock driver binary can be called from mixed-mode 6502 and 
Z80 if I can...

Otherwise, help...

Bill

PS - If anyone sees Egan let him know I finally got around to it.
 

[toc] | [next] | [standalone]

#866

"Bill Buckels" <bbuckels@mts.net> wrote:

>I have no DHRYSTONE results for CC65 for the Apple II for comparison. The 
>only stats I can find are targetted at the Commodore where cc65 runs 28% 
>faster than Aztec C65 3.2b (cc65 stats according to the following link):
>http://www.cc65.org/mailarchive/2009-07/6616.html
>So I am comparing Apples to Bananas at this point. I hope to have a 
>DHRYSTONE for AztecC64 a little later in the week.

I now have the results for Aztec C II Vers. 1.05h 6502  (C) 1982, 1983 by 
Manx Software Systems MS-DOS (Windows XP) Cross-Compiler running in WinVICE 
Version 1.22 at 41 DHRYSTONES/SEC .

This is already faster than Commodore 64 6510-1MHz C64 ROM  C Power 2.8  at 
36 DHRYSTONES/SEC  by 13%.

I will be posting a complete working program with d64 disk image to the 
Aztec C Website in the next few hours. If WinVice can be believed, the C64 
results for Aztec C II  are 4 DHRYSTONES/SEC over the older Apple IIe 
65C02-1.02Mhz DOS 3.3  Aztec CII v1.05i equivalent compiler  which achieved 
37 DHRYSTONES/SEC (2 DHRYSTONES/SEC slower than the newer version 3.2b 
achieved.) I will confirm this (in time of course), but if the same CII 
compiler is 10% slower on the Apple II than the C64, that may also hold true 
with the cc65 compiler on the Apple II.

I am speculating at this point because as I say I have no stats for cc65 for 
the Apple II (just for the C64) but if my theory is correct, and the same 
ratio of 37 Apple II to 41 C64 applies to cc65, then cc65 should come-in at 
45 DHRYSTONES/SEC on the Apple II as opposed to Aztec C65 3.2b's 39 
DHRYSTONES/SEC. I'll need to leave it to the cc65 people to sort that one 
out. There was some comment about optimizations as well (see link above), 
and the fact that 5 points may be lost without them which may even bring 
cc65 down to 40 DHRYSTONES/SEC on the Apple II (again speculating). Without 
verifiable results it's hard to tell. What version of DHRY was that 
anyway???

Considering that Aztec C65 also offers floating point support and inline 
assembly, and the ability to read files in DOS 3.3, as well as the ability 
to create command line programs for its unix-like shells for both DOS 3.3 
and ProDOS...

So as it stands now, the older Aztec C for the Commodore 64 which also 
provides floating point support and well-documented overlays etc. still 
achieves 82% DHRYSTONES/SEC of cc65 posted results on the C64 (not sure 
about that 5 point business either) so it may be closer to 91%. I wonder too 
if I will achieve better results with the C64 when I complete my port to the 
newer Aztec CG65 3.2c compiler from Paul Santa-Maria?

I also found it interesting that even though I have absolutely no problems 
compiling DHRY for Aztec C for the C64:

http://www.cc65.org/mailarchive/2009-10/7171.html

It's a poor workman that blames his tools.

BTW, there won't be any cc65 results from me.... I've got my hands-full with 
Aztec C.  And Aztec C's documentation is too complete, and its environment 
too feature rich for me to bother with something else on the Apple II after 
using Aztec C for such a long time with good results.

I don't stumble through documentation either; I take time to read it. I must 
be getting old, because I feel like saying "If it was good enough for 
Grandpa, it's good enough for me". As for memcpy, I guess I could have used 
the Aztec C equivalent but didn't need it. I used an appropriate version of 
DHRY...

Bill

[toc] | [prev] | [next] | [standalone]

#867

"Bill Buckels" <bbuckels@mts.net> wrote:

>So I am comparing Apples to Bananas at this point. I hope to have a 
>DHRYSTONE for AztecC64 a little later in the week.

Now available as advertised (Bananas are done, so back to Apples):

http://www.aztecmuseum.ca/extras/DHRYC64.zip

>If anyone sees Egan let him know I finally got around to it.

I wonder what he's been up to... maybe he can run some Z80 DHRY tests.  I 
also gotta find-out about that clock on the Apple II when applicard or 
SoftCard is in use.

Bill

[toc] | [prev] | [next] | [standalone]

#868

On 09/14/2013 10:35 PM, Bill Buckels wrote:

> I also gotta find-out about that clock on the Apple II when applicard or
> SoftCard is in use.


Bill, what are you asking?  Do you want to know how to access the clock from 
Applicard or Softcard CP/M?  There should be Applicard drivers for just about 
all A2 clock cards (and No-Slot, which is what I have in my systems).

Support for Softcard is probably not going to be easy, although I believe AE 
CP/AM could use the AE clock cards (don't quote me on that - it's been a long 
time).

Steve

[toc] | [prev] | [next] | [standalone]

#869

"Steven Hirsch" <snhirsch@gmail.com> wrote:

>Bill, what are you asking?  Do you want to know how to access the clock 
>from Applicard or Softcard CP/M?  There should be Applicard drivers for 
>just about all A2 clock cards (and No-Slot, which is what I have in my 
>systems).

Good to know. Thanks Steve.

>Support for Softcard is probably not going to be easy, although I believe 
>AE CP/AM could use the AE clock cards (don't quote me on that - it's been a 
>long time).

I won't quote you:) but I won't take your words lightly either. I may be 
able to use the NSC's relocatable clock driver, and switch to 6502 mode and 
back to Z80 mode.

Bil 

[toc] | [prev] | [next] | [standalone]

#879

On Thursday, September 12, 2013 10:23:33 PM UTC-4, Bill Buckels wrote:
> DHRYSTONE for ProDOS 8 written in Aztec C65 3.2b is now available for 
> download complete with disk image and source at the following link:
> http://www.aztecmuseum.ca/extras/DHRY.zip

Hi Bill: Would it be possible for you to post the generated assembly language for the Dhrystone program?

I'm working on a compiler for another vintage 8 bit processor(COSMAC 1802) and clocking in at 13 dhrystones/sec.  I'd like to look at the generated code for the 6502 to see whether the delta is down to the processor or the compiler.

Thanks

[toc] | [prev] | [next] | [standalone]

#885

"Bill Buckels wrote:

>DHRYSTONE for ProDOS 8 written in Aztec C65 3.2b is now available... I get 
>a result of 39 DHRYSTONES/second.

Aztec C65 compilers produce executable programs for 2 targets:

- "RAW" (SA -Stand-Alone) executables (DOS 3.3 and ProDOS)
- PCODE (Pseudo-Code) executables that run only in the Apple II Aztec C 
Shells

In addition to the "RAW" ProDOS 8 DHRYSTONE SYS program test noted above, I 
have compiled and run DHRYSTONE PCODE (pseudo-code) versions written for 
both Aztec C65 Unix-like SHELL's for the Apple II (also using the NSC as a 
timer)... the results are identical for either Shell version and are almost 
exactly 50% of the speed of the "RAW" ProDOS 8 results.:

1. DOS 3.3 APPLE ][ SHELL 2.4 (1983) -  19.5 DHRYSTONES/second.
2. ProDOS 8 Aztec C Shell v199j 6-27-86 - 19.5 DHRYSTONES/second.

Aztec C's Unix-like Shell for the Apple II came in 2 different releases. 
These two shells are quite different; the earlier release is a DOS 3.3 "RAW" 
BRUNable BIN program called "SHELL". The later release is a ProDOS 8 SYS 
program called "SHELL.SYSTEM.

SHELL 2.4 was released as part of the CII compiler for DOS 3.3. Although the 
Aztec C65 documentation suggests that the newer ProDOS compiler C65 3.2b can 
produce PCODE executables that run in the DOS 3.3 shell, this is apparently 
not correct. PCODE executables produced with the newer compiler are not 
(entirely) compatible with the older shell.

The DHRY programs noted above were produced with 2 their two different 
respective versions of the C65 compiler that the Shells came with, yet they 
are identical.

Both Shells offer commandline arguments and shell scripting, smaller 
executables than "RAW" programs and their own command sets which are quite 
useful for writing utilities that run on the Apple II. And both shells allow 
you to mix "RAW" code (including assembly language calls to the Apple II ROM 
routines and ML routines like the NSC driver) with PCODE for time critical 
operations. Since the run-time library source for "RAW" code is available 
for Aztec C65, you can also optimize standard library routines selectively 
for time critical operations and include these as needed.

The ProDOS 8 Shell is much more robust than the DOS 3.3 Shell (which is also 
not too shabby when compared to Apple II BASIC). The ProDOS Shell far 
surpasses anything AppleSoft (BASIC.SYSTEM) has to offer when it comes to a 
"real command line"... the only thing that is really strongly missing is 
(sadly) commandline history. If you use the AppleWin emulator you can of 
course paste commands from the Windows Clipboard directly into the Shell's 
command-line to avoid repetitive typing to somewhat compensate for this...

If you are willing to "take a hit" on the performance of non-time critical 
parts of a Shell utility program for the convenience and functionality of a 
"Unix-like" environment, and you mix "RAW" code into your Shell utilities 
for time critical operations, it is reasonable to conclude that your 
execution speed for Shell programs will equal (or almost equal) "RAW" speed 
which comes-in at double the speed (based on the DHRY benchmark).

Results will of course vary depending on your program design and the raw 
code you insert into your Shell programs and so-forth.

As with all these DHRY results, and other benchmarking I am currently doing: 
when complete I will provide a webpage on the Aztec C website with links and 
a list of benchmarks.

Regards,

Bill Buckels

[toc] | [prev] | [next] | [standalone]

#909

Bill Buckels wrote:

>DHRYSTONE for ProDOS 8 written in Aztec C65 3.2b is now available... I get 
>a result of 39 DHRYSTONES/second.

I have just completed an additional test for DHRYSTONE 1.1 for ProDOS 8 
written in Aztec C65 3.2b with my real Apple //e with an 8 MHZ Zip Chip 
Accelerator:

206 DHRYSTONES/second.

- By comparison to DHRY of  197 for 8086-8Mhz RMX86 V6 Intel C-86 V2.0.
- By comparison to DHRY of 259 for IBM PC/XT 8088-4.77Mhz COHERENT 2.3.43 
Mark Williams.

While a Macintosh512 68000-7.7Mhz Mac ROM O/S DeSmet(C ware) has a DHRY of 
625 (about triple the speed), Aztec C65's performance on my Apple //e still 
doesn't seem too shabby for my little old 8 bit machine.

But thinking that hardware optimization may still be a little too slow for 
some tasks, and virtual optimization might then be in order, I also tested 
under AppleWin using the fastest custom speed settings on my  2.6 GHZ Intel 
G620 and the DHRY is 4166 (which is probably fast enough for anything that 
might need doing with this old compiler).

Regards,

Bill

PS - Sorry if all of this this stretches credibility, but I couldn't resist 
doing these last tests just for fun. And for the record, I never did own an 
Apple //e without an acceleretor.

[toc] | [prev] | [next] | [standalone]

#910

Very interesting comparisons, thank you Bill
av

[toc] | [prev] | [standalone]

Back to top | Article view | comp.sys.apple2.programmer

csiph-web