horrible glTexSubImage3D performance in linux

anyone have any theory as to why i would be getting awful performance with glTexSubImage3D in linux?

i have a volume editing app where i apply various (small) kernel filters to a volume (1x1x1 - 5x5x5), then use tex sub image to update (instead of downloading the whole thing again).

i think i remember a time when it was pretty fast, but haven’t used the code in a long while. now it takes almost 1 second to update a 5x5x5 region and between 1 and 2 tenths to update a 1x1x1 region. i can’t figure it out.

here’s my driver version:
NVRM version: NVIDIA Linux x86 NVIDIA Kernel Module 1.0-6111 Tue Jul 27 07:55:38 PDT 2004
GCC version: gcc version 3.3.3 20040412 (Red Hat Linux 3.3.3-7)

i’m on FC2 with a Quadro4 700 GoGL.

thanks for any insight…

just an update if anyone is iterested:

turns out there is a possible driver bug when any dimension passed to glTexSubImage3D is 1. problem occurs on multiple cards, operating systems, and drivers. nvidia is looking into it.