Makes me wonder how your design would work ina 64 Bit slot set at 32Bit with a 66Mhz Bus, have you tried that did that make any difference?
the max bus clock tested is 44mhz so far. that board designed for a 5v pci slot, thus 33mhz bus clock. here is 80mhz clocked pcx2 on 44phz pci bus, (26.3fps): https://www.facebook.com/zxc64.hw/posts/1047333155748915
on stock clocks pcx2 gives 21.6 fps in that bench iirc.
on 80mhz clock and 33mhz bus pcx2 gives about 24.8fps, so 11mhz boost in bus speed leads to about 1.5fps