I guess you are right. I just did an adhoc test between 6.0, 7.0, and 8.0. I too found that 7.0 is slower then 6.0, 8.0 is much slower then 7.0.

>Greg,
>Unfortunately it's worse in 8.0 SP1. I know Calvin made some changes and I read about NTI in his logs however I just tested (multiple times - same code in that thread) under VFP7 and 8. VFP8 is over 50% slower if m. is not used. If m. is used stil slower than VFP7 but ignorable.
>
>AMD Athlon 2500+, 512Mb
>
>VFP7: with m. 4.1 secs, w/o mdot 13.3 secs
>VFP8: with m. 4.6 secs, w/o mdot 20.1 secs
>(best scores)
>
>PS:
>VFP6 still holds to be the winner:
>VFP6: with m. under 4.1 secs, w/o mdot 4.2 secs
>
>Cetin
>
>>As this may be true for 7.0. I know that Calvin Hsia did a major overhaul to the variables name space in 8.0. Form my performance testing; I saw that 8.0 was considerably faster then previous versions. I do not know if those changes directly effected the M. feature.
>>
>>>If you have time:
>>>
>>>Message #646222 && A little snowball
>>>Thread #647967 && Got bigger
>>>Thread #649450 && Avalanche:)
>>>
>>>Cetin
>>>
>>>>I guess I missed that dialog.
>>>>