Fixed the documentation

[Mograsim.git] / tests / net.mograsim.logic.model.am2900.tests / src / net / mograsim / logic / model / am2900 / gcd-mpm-doc.txt

diff --git a/tests/net.mograsim.logic.model.am2900.tests/src/net/mograsim/logic/model/am2900/gcd-mpm-doc.txt b/tests/net.mograsim.logic.model.am2900.tests/src/net/mograsim/logic/model/am2900/gcd-mpm-doc.txt
index 3fd0abc..49e7137 100644 (file)
--- a/tests/net.mograsim.logic.model.am2900.tests/src/net/mograsim/logic/model/am2900/gcd-mpm-doc.txt
+++ b/tests/net.mograsim.logic.model.am2900.tests/src/net/mograsim/logic/model/am2900/gcd-mpm-doc.txt
@@ -2,8 +2,8 @@ Register allocation:
 
 R0: a1 (Euclid)
 R2: a2 (Euclid)
 
 R0: a1 (Euclid)
 R2: a2 (Euclid)

-R1: b (Euclid)
-Q: c (copy of b) / result
+R1: b (Euclid) / result
+Q: c (copy of b)

 
 Most of the time, a1 holds the actual value of a, and a2 holds some garbage. But in the main subtract-and-half loop, it sometimes is the other way around for speed reasons.
 
 
 Most of the time, a1 holds the actual value of a, and a2 holds some garbage. But in the main subtract-and-half loop, it sometimes is the other way around for speed reasons.