Joost Rekveld

just moved to max 4.6 on an intel machine, and found this problem 

with the 'wrap' and 'fold' operator in jit.op:

works fine in max/msp 4.5.6 jitter 1.5.2 on a G4 dual 1Ghz

gives 'nan' and -2147483648 and such values in max/msp 4.6 jitter 

also I have an issue to do with the 'atan2' operator suddenly giving 

me 'nan' values where there were none before, but I haven't been able 

to pin that one down, it might be something earlier in my patch 

for what i'm doing (many jitter operations on big matrices) the 2 x 

2.66 Ghz macpro seems about 4 times as fast as the 2 x 1 Ghz G4, so I 

can now start making things 4 times as complicated.. :)

just moved to max 4.6 on an intel machine, and found this problem  
with the 'wrap' and 'fold' operator in jit.op:

works fine in max/msp 4.5.6 jitter 1.5.2 on a G4 dual 1Ghz
gives 'nan' and -2147483648 and such values in max/msp 4.6 jitter  
1.6.2 on MacPro dual 2.66 Ghz

also I have an issue to do with the 'atan2' operator suddenly giving  
me 'nan' values where there were none before, but I haven't been able  
to pin that one down, it might be something earlier in my patch  
messing things up..

for what i'm doing (many jitter operations on big matrices) the 2 x  
2.66 Ghz macpro seems about 4 times as fast as the 2 x 1 Ghz G4, so I  
can now start making things 4 times as complicated.. :)

jit-op-op-wrap-and-fold-bug

On Oct 3, 2006, at 6:34 AM, Joost Rekveld wrote:

> just moved to max 4.6 on an intel machine, and found this problem 

> with the 'wrap' and 'fold' operator in jit.op:

> works fine in max/msp 4.5.6 jitter 1.5.2 on a G4 dual 1Ghz

> gives 'nan' and -2147483648 and such values in max/msp 4.6 jitter 

Thanks for the report. I can verify and will try to fix for 1.6.3.

> just moved to max 4.6 on an intel machine, and found this problem  
> with the 'wrap' and 'fold' operator in jit.op:
> works fine in max/msp 4.5.6 jitter 1.5.2 on a G4 dual 1Ghz
> gives 'nan' and -2147483648 and such values in max/msp 4.6 jitter  
> 1.6.2 on MacPro dual 2.66 Ghz

jit.op @op wrap and fold bug ?