CopyMem Quick & Small released!

[SpeedGeek]

CopyMem Quick & Small v1.6
Parts of patch install code by Dirk Busse 1999
Enhanced patch code by SpeedGeek 2015

INTRODUCTION:
CMQ&S is a much smaller Copymem + Copymemquick patch. It's
main goal is to give a good speed increase without all the bloated
code found in the other CMQ patches. It does not attempt to be the
fastest CMQ patch, but at 92 bytes for the current version, it's
almost certainly one of the fastest in it's class.

FEATURES:
- Installs one of the smallest CMQ patches for 68020+ Amigas
- Safely exits if the patch is already installed (e.g. a good patch
program should really avoid patching itself)

REQUIREMENTS:
- Amiga with 68020+

NOTES:
Most other CMQ patches got bloated because their authors
wanted their patch to be faster than any previously released
patch. This means even bigger "unrolled" loops, redundant
testing for small & medium size copies and extra code to
handle misaligned transfers. Unfortunately, Testit from
COPMQR28 makes a disproportional number of these
"Worst Case" tests and other coders have fallen into the trap.
Motorola optimized it's 68K CPUs to improve average case
performance at the expense of worst case performance.
Need I say any more?

[daxb]

Some TestIt results for the fun. Attached results for various copymem paches if you care. If matthey read this you might have a look at "copymem_test_CopyMem040.txt". You might can explain the last result because it differs much compared to CopyMem060:

Copying 1 bytes 2190000 times (long -> long offset)
Old CopyMem     :  0.88 secs
New CopyMem     :  0.82 secs (- 5.7%)

SpeedGeek stuff:

10. System:C> CMQ&S040 
10. System:C> CopyMemQuicker 
CopyMemQuicker 2.8 (12.6.94)
Copyright © 1992 Arthur Hagen.
10. System:C> TestIt 
This test will compare the old CopyMem/CopyMemQuick routines with
the new ones you have installed.  A great variety of tests will be
run, and this might take some time, especially if your system has a
slow processor.

Initiating test (please be patient...)

Copying 65536 bytes 565 times (long -> long offset)
Old CopyMem     :  2.03 secs
New CopyMem     :  2.07 secs (+ 2.0%)
Old CopyMemQuick:  2.03 secs
New CopyMemQuick:  2.04 secs (+ 0.5%)

Copying 65536 bytes 147 times (long -> long+1 offset)
Old CopyMem     :  0.60 secs
New CopyMem     :  0.79 secs (+31.7%)

Copying 65536 bytes 413 times (long -> even offset)
Old CopyMem     :  1.68 secs
New CopyMem     :  1.47 secs (-12.5%)

Copying 65536 bytes 147 times (long -> even+1 offset)
Old CopyMem     :  0.60 secs
New CopyMem     :  0.81 secs (+35.0%)

Copying 65536 bytes 147 times (long+1 -> long offset)
Old CopyMem     :  0.54 secs
New CopyMem     :  0.81 secs (+50.0%)

Copying 65536 bytes 382 times (long+1 -> long+1 offset)
Old CopyMem     :  1.42 secs
New CopyMem     :  1.42 secs (+ 0.0%)

Copying 65536 bytes 147 times (long+1 -> even offset)
Old CopyMem     :  0.54 secs
New CopyMem     :  0.79 secs (+46.3%)

Copying 65536 bytes 501 times (long+1 -> even+1 offset)
Old CopyMem     :  1.87 secs
New CopyMem     :  1.94 secs (+ 3.7%)

Copying 65536 bytes 501 times (even -> long offset)
Old CopyMem     :  1.84 secs
New CopyMem     :  2.03 secs (+ 9.8%)

Copying 65536 bytes 147 times (even -> long+1 offset)
Old CopyMem     :  0.54 secs
New CopyMem     :  0.81 secs (+48.1%)

Copying 65536 bytes 382 times (even -> even offset)
Old CopyMem     :  1.42 secs
New CopyMem     :  1.42 secs (+ 0.0%)

Copying 65536 bytes 147 times (even -> even+1 offset)
Old CopyMem     :  0.54 secs
New CopyMem     :  0.79 secs (+44.4%)

Copying 65536 bytes 147 times (even+1 -> long offset)
Old CopyMem     :  0.54 secs
New CopyMem     :  0.81 secs (+50.0%)

Copying 65536 bytes 413 times (even+1 -> long+1 offset)
Old CopyMem     :  1.54 secs
New CopyMem     :  1.62 secs (+ 4.5%)

Copying 65536 bytes 147 times (even+1 -> even offset)
Old CopyMem     :  0.54 secs
New CopyMem     :  0.81 secs (+48.1%)

Copying 65536 bytes 564 times (even+1 -> even+1 offset)
Old CopyMem     :  2.10 secs
New CopyMem     :  2.10 secs (+ 0.0%)

Copying 1024 bytes 33900 times (long -> long offset)
Old CopyMem     :  0.54 secs
New CopyMem     :  0.70 secs (+29.6%)
Old CopyMemQuick:  0.53 secs
New CopyMemQuick:  0.69 secs (+30.2%)

Copying 1024 bytes 9400 times (long -> long+1 offset)
Old CopyMem     :  0.21 secs
New CopyMem     :  0.52 secs (+142.8%)

Copying 1024 bytes 24000 times (even -> even offset)
Old CopyMem     :  0.62 secs
New CopyMem     :  0.50 secs (-17.7%)

Copying 128 bytes 196000 times (long -> long offset)
Old CopyMem     :  0.50 secs
New CopyMem     :  0.98 secs (+94.0%)
Old CopyMemQuick:  0.45 secs
New CopyMemQuick:  0.86 secs (+91.1%)

Copying 128 bytes 155000 times (even -> even offset)
Old CopyMem     :  0.59 secs
New CopyMem     :  0.82 secs (+37.3%)

Copying 19 bytes 588000 times (long -> long offset)
Old CopyMem     :  0.87 secs
New CopyMem     :  0.93 secs (+ 5.7%)

Copying 18 bytes 622000 times (long -> long offset)
Old CopyMem     :  0.79 secs
New CopyMem     :  0.95 secs (+19.0%)

Copying 17 bytes 663000 times (long -> long offset)
Old CopyMem     :  0.94 secs
New CopyMem     :  0.99 secs (+ 5.3%)

Copying 16 bytes 956000 times (long -> long offset)
Old CopyMem     :  1.12 secs
New CopyMem     :  1.44 secs (+27.7%)
Old CopyMemQuick:  0.86 secs
New CopyMemQuick:  0.86 secs (- 0.0%)

Copying 8 bytes 1060000 times (long -> long offset)
Old CopyMem     :  0.98 secs
New CopyMem     :  1.33 secs (+34.7%)
Old CopyMemQuick:  0.69 secs
New CopyMemQuick:  0.69 secs (+ 0.0%)

Copying 4 bytes 1430000 times (long -> long offset)
Old CopyMem     :  1.11 secs
New CopyMem     :  1.07 secs (- 2.7%)
Old CopyMemQuick:  0.71 secs
New CopyMemQuick:  0.75 secs (+ 4.2%)

Copying 1 bytes 2190000 times (long -> long offset)
Old CopyMem     :  0.77 secs
New CopyMem     :  0.82 secs (+ 6.5%)

Total timing:
-------------
Old routines    :  32.81 secs
New routines    :  37.60 secs
Total slowdown  :  14.60 %

[SpeedGeek]

See CMQ&B040 thread for MOVE16 Testit results.

[SpeedGeek]

** NEWS UPDATE **

CMQ&S040 v1.6 released

v1.6 minor change
- source address compare code misqualified Move16 on 8 byte offset
(This is fixed now but the 4 byte offset still doesn't work for some reason)

[HanSolo]

Your patch needs to be run before setpatch or no matter ? (I am using 68060.library and mmu.library by Tomas Richter)

[daxb]

From txt file:

- Add it to your startup-sequence (after Setpatch)
OR
- Execute it from the shell

[SpeedGeek]

Originally Posted by HanSolo

Your patch needs to be run before setpatch or no matter ? (I am using 68060.library and mmu.library by Tomas Richter)

I run the patch before Setpatch as one of the tests I perform before releasing it but that's not the recommended use. Commodore really intended Setpatch to be the first patch that's run in your s-s and there are some good reasons for this.

Sometimes, it's necessary to run another patch before Setpatch (e.g. you have non-autoconfig Fast RAM to be added, or you use vbak2091 which needs to be run before any NSD config stuff). But none of the above apply to CMQ&S so the recommended use is the preferred use.

[Retrofan]

Hi SpeedGeek.I've just seen this. I am actually using CopyMem by Matt Hey. Does it substitute it?

[SpeedGeek]

Hi Retrofan,

CMQ&S is an alternative CMQ patch. Whether or not it would be a good substitute for any other patch is a matter of opinion and really depends on what your looking for.

[Retrofan]

Thanks a lot SpeedGeek. Yes, I've been reading and it seems there isn't a complete unanimity about it's convenience yet and also the benefits can depend on the hardware.

Thanks anyway for working on it.

[SpeedGeek]

** 2ND NEWS UPDATE **

CMQ&S v1.6 released
v1.6 minor change
- fixed install code which could (but seldom ever did) trash a few bytes
of memory past the end of the patch

[SpeedGeek]

** 3RD NEWS UPDATE **

Deleted post.

[SpeedGeek]

** 4TH NEWS UPDATE **

CMQ&S040 has been removed and all related posts deleted. CMQ&S was originally developed as "Proof of Concept" code for low memory systems. Since 040+ systems typically have much more memory than 020-030 systems it's really not practical for such systems. CMQ&B040 should be more practical in this case.

Here is the link:

http://eab.abime.net/showthread.php?t=76777