I wouldn't be surprised, software rendering uses the CPU only so whatever the CPU can do it's likely to be a lot less than the gfx card, even with a low-mid-range card.
I can't report an exact comparison because my ibook is so slow anyway, it can't do more than a few frames per second of even the simplest graphics in OpenGL.
On the ibook, DrawPixmap is the same speed as DrawImage. I think that's because there is very limited video ram, so it actually keeps ALL textures in main memory, so drawing a pixmap from main memory to main memory is no faster or slower than doing it with an image. I also found on my faster iMac which has much faster gfx card for OpenGL hardware acceleration, that DrawPixmap on that machine is 10 times faster and DrawImage is 100 times faster, than the ibook's speed. That's a 1GHz G4. So you might find depending on the cpu and gfx card that there can be quite significant differences, in my case around 10 times performance drop in software. You just can't do what OpenGL can do, in software, and expect it to be anywhere near as fast as hardware accelerated highly graphics-processing specific GPU's.
I can't report an exact comparison because my ibook is so slow anyway, it can't do more than a few frames per second of even the simplest graphics in OpenGL.
On the ibook, DrawPixmap is the same speed as DrawImage. I think that's because there is very limited video ram, so it actually keeps ALL textures in main memory, so drawing a pixmap from main memory to main memory is no faster or slower than doing it with an image. I also found on my faster iMac which has much faster gfx card for OpenGL hardware acceleration, that DrawPixmap on that machine is 10 times faster and DrawImage is 100 times faster, than the ibook's speed. That's a 1GHz G4. So you might find depending on the cpu and gfx card that there can be quite significant differences, in my case around 10 times performance drop in software. You just can't do what OpenGL can do, in software, and expect it to be anywhere near as fast as hardware accelerated highly graphics-processing specific GPU's.