Dr. Fiemost
93bb291203
Fix UBSan error
2017-01-22 19:08:31 +01:00
Dr. Fiemost
dff94ab870
Initialize remaining struct members
2017-01-21 14:17:39 +01:00
Dr. Fiemost
379a588484
Initialize allocated memory
2017-01-21 14:16:26 +01:00
Dr. Fiemost
ed250b1a62
Avoid use of deprecated function
2017-01-21 14:13:06 +01:00
Dr. Fiemost
aed0d9a3c3
Fix https://github.com/mist64/perfect6502/issues/3
2017-01-17 20:53:44 +01:00
Michael Steil
4142ac152e
moved nodes names into netlist_6502.h
2014-12-01 16:42:48 +01:00
Michael Steil
0269c1bdf1
cleanups
2014-12-01 14:03:16 +01:00
Michael Steil
76d798db79
split code into generic netlist simulation and 6502-specific code
2014-12-01 13:53:46 +01:00
Michael Steil
44cde717a1
cleanup
2014-12-01 13:29:08 +01:00
Michael Steil
c93f229339
cleanup
2014-12-01 13:10:28 +01:00
Michael Steil
4b70feb863
cleanup
2014-12-01 12:48:41 +01:00
Michael Steil
79d5e114ef
cleanups
2014-12-01 12:44:03 +01:00
Michael Steil
c7ede71e52
actually de-dup transistors!
2014-12-01 12:31:14 +01:00
Michael Steil
e49ca0be5f
cleanups
2014-12-01 12:23:47 +01:00
Michael Steil
b31b545032
malloc'ed all state instead of having a gigantic struct
...
state_t is now 208 bytes on x86_64 instead of 26 MB. Startup is a little slower though because of ~7000 malloc calls.
2014-12-01 12:10:21 +01:00
Michael Steil
1c9eb78154
cleaned up node names
2014-12-01 11:48:01 +01:00
Michael Steil
fc3082a259
updated copyright
2014-12-01 11:01:11 +01:00
Michael Steil
c4f7cfb449
The state pointer is no longer global, but passed down to every function.
...
...and surprisingly, it's not any slower. if anything, it's 1% faster.
2014-12-01 11:00:30 +01:00
Michael Steil
b8af4edbbe
moved all globals into a single global struct -> 3% speedup with clang :)
2014-11-29 00:55:06 +01:00
Michael Steil
b3f4c56541
added README
2014-11-28 08:31:36 +01:00
Michael Steil
9a8e2c396f
fixed line endings
2014-11-28 07:57:38 +01:00
Michael Steil
8161e5e600
added object files to .gitignore
2014-11-28 07:56:51 +01:00
Michael Steil
b2cce88620
optimizations (~25%) and cleanup
...
Optimizations:
1. When a transistor is turned on, only add one of the nodes it
controls to listout, otherwise group calculation would be done twice
for this group. This simplifies recalcNodeList(). (This required
the following change: Instead of adding the nodes that control
transistors that have changed to listout, put the nodes that are
switched by transistors on listout.)
2. The group value is calculated while collecting nodes, and the enum
group_contains_value makes sure we don't check for weaker value
indicators than have already been established.
3. listout is now a set, no node will be visited twice for group
calculation, within one interation.
4. When setting input pins of the package to a value, only recalc the
node if it actually changes.
5. Toggling a transistor is done by assigning it the group value
instead of the inverse of its value.
Cleanup:
* Merge the concept of listin with the concept of nodes that change
from the outside and trigger a recalc.
2014-11-28 07:04:07 +01:00
Michael Steil
bdbe8eff59
...
2010-10-20 07:49:51 +00:00
Michael Steil
63ce225449
added apple1basic
2010-10-12 14:29:11 +00:00
Michael Steil
f424b464f2
cleanup
2010-10-08 04:09:56 +00:00
Michael Steil
48e5756c38
cleanup
2010-10-08 03:58:09 +00:00
Michael Steil
8d948b5f64
indirect lostout
2010-10-08 03:22:54 +00:00
Michael Steil
d9ba3fc769
experimental...
2010-10-07 05:20:49 +00:00
Michael Steil
c0d191a03f
added concept of dependants - same speed, but might
...
lead to more optimization opportunities
2010-10-07 04:39:40 +00:00
Michael Steil
6e8fd7f401
debug code; eliminate more duplicate transistors (with c1 and c2 reversed)
2010-10-07 02:19:44 +00:00
Michael Steil
07d31a1529
cleanup
2010-10-07 00:28:02 +00:00
Michael Steil
44f7f48a9c
fixed "broken transistors" test
2010-10-06 16:53:40 +00:00
Michael Steil
217ba8f681
rename
2010-10-06 16:34:57 +00:00
Michael Steil
f8fe8d9d46
made test work independently
2010-10-06 16:33:26 +00:00
Michael Steil
9318b580e4
huh, missing file?
2010-10-06 16:33:04 +00:00
Michael Steil
a2ad35412c
separated emulator and runtime
2010-10-06 15:58:13 +00:00
Michael Steil
c53fa8ff9c
cleanup
2010-10-06 04:32:54 +00:00
Michael Steil
7c58dfd703
memory read and write on the same edge now; cleanup
2010-10-06 03:00:20 +00:00
Michael Steil
4b95933e67
bitmap size fix -> 10% faster
2010-10-06 02:13:38 +00:00
Michael Steil
1b22b7fec9
rename
2010-10-05 18:20:02 +00:00
Michael Steil
4189c81c70
cleanup, restore missing loop exit case
2010-10-05 18:13:48 +00:00
Michael Steil
57d2ce2c91
cleanup
2010-10-05 18:01:13 +00:00
Michael Steil
44ea268beb
don't track "floating" any more. much easier, same speed
2010-10-05 17:54:51 +00:00
Michael Steil
12e8208234
cleanup
2010-10-05 17:48:01 +00:00
Michael Steil
446774eaa3
cleanup
2010-10-05 16:54:12 +00:00
Michael Steil
eee151c3ec
nicer getNodeValue() interface, same speed
2010-10-05 16:46:27 +00:00
Michael Steil
90dfd9791a
cleanup
2010-10-05 16:27:33 +00:00
Michael Steil
3ef6b7636a
cleanup
2010-10-05 16:18:28 +00:00
Michael Steil
f101c89980
emulator / tester work
2010-09-30 02:39:11 +00:00