>Same results as selecting "NoBlitter=No
this is the setting that should use and then Hardware Blit work.
when set NoBlitter=YES then the P96 bitblit test should go slow.
Oops! Sorry, I meant that the 2 p96speed slow blit ops are excactly as slow if I set NoBlitter=Yes, meaning that when it's goofing off, the blitter isn't being used on these 2 (while the rest work OK). Naturally, NoBlitter=No is normally set.
On my a4000 I bootup with a 1024x768x16 screenmode.
When I run the blitbitmap tests I get:
BlitBitMap() - 5415
BlitBitMapRastPort() - 4540
Thanks for checking. I have to "reset" the screen first (like changing fonts, etc) to get the faster numbers. I gave up tracking it down for now... tried just about everything. I don't think it's affecting anything else, or related to the slow scrolling problem after all - I can confirm a 1024x768x16 Netsurf window will scroll quickly, if enough VRAM is free.
BTW - using MuRedox here. I can't detect a speed improvement either, but it's definitely doing something.
Typical "MuRedox SHOWPATCHEDINSTRS" output while running Netsurf (which might as well be Chinese to me):
04f228cc : 4c3c 2401 0000 0000 mulu.l #$0,d1:d2 ;extended opcode
04f228cc : 4c3c 6c04 0000 0000 muls.l #$0,d4:d6 ;extended opcode
04f228cc : 4c3c 1403 0000 0000 mulu.l #$0,d3:d1 ;extended opcode
04f228cc : 4c3c 5c01 0000 0000 muls.l #$0,d1:d5 ;extended opcode
04f228cc : 4c3c 1c03 0000 0000 muls.l #$0,d3:d1 ;extended opcode
04f228cc : 4c12 3c00 muls.l (a2),d0:d3 ;extended opcode
etc
I agree that ISP and location matters. On rare occasions, amigaworld will load in under 20 seconds, while usually it's 25-35s. Fastest amiga.org was 19.8s.