nativ wrote:have you tried running gembench on an atari running EMU TOS ? does that work?
nativ wrote:as Chris says the supporting hardware will not be there running it on an amiga running EmuTOS to report much..
nativ wrote:I am not sure how complete EMUtos on amiga actually is..
GemBench only tests GEM and CPU, so I don't see what could prevent it to run on EmuTOS for non-Atari hardware. Anyway, as GemBench only targets Atari hardware, I will not insist. There are other good benchmarks.
exxos wrote:It likely fails as the assembly code turns off interrupts for the keyboard, floppy polling etc.
nativ wrote:why not contact the EMUTOS amiga dev team and query it with them.. or the GEMbench author and see if he can produce a limited version for you to test with...
nativ wrote:failing that.. write your own EMU TOS compliant testing software/benchmark
exxos wrote:No idea what other test could be added though its just a thought really as the FPU test can't be done on most machines (other than the falcon) anyway.
So I am really wondering if the FPU test in GB6 should just be taken out and some other more useful test put there instead. No idea what other test could be added though its just a thought really as the FPU test can't be done on most machines (other than the falcon) anyway.
Atari030 wrote:exxos wrote:OK, so put it another way then, what do you guys use the FPU for ?
Filling empty sockets on motherboards. The CT2 does my OCD in.
EmpireAndrew wrote:I'd like to see FPU tests for my TT so that I could compare it before and after overclock...
Code: Select all
GEMBENCH 606 Build 24
TOS 4.00 AES 3.99 TOS 2.06 AES 3.20
BLiTTER Disabled BLiTTER Enabled
RES 1088*736*256 RES 640*200*4
GEMDOS v0.20 GEMDOS v0.20
Falcon 68030 STE 68000
LineF FPU inst. FPU Not present
NVDI 5.03 present NVDI not present
CACHE D=ON I=ON CACHE D=OFF I=OFF
Ran from ST-RAM Ran from ST-RAM
GEM Dialog Box: 7.340 179%
VDI Text: 1.010 1448%
VDI Text Effects: 1.240 1738%
VDI Small Text: 2.580 876%
VDI Graphics: 5.140 542%
GEM Window: 12.235 81%
Integer Division: 1.820 1728%
Float Math: 2.350 0%
RAM Access: 23.300 155%
ROM Access: 11.540 646%
Blitting: 9.675 37%
VDI Scroll: 7.655 125%
Justified Text: 3.825 303%
VDI Enquire: 6.955 297%
ALT-RAM: 10.745 676%
Ektus wrote:GemBench 6 build 24 works okay on my CT2A Falcon. Operating System is MagiC 6.20. The GemBench window is fixed in the middle of the screen. It doesn't like multitasking, won't come to the front again once hidden by other windows.
Ektus wrote: It fails to start on my CT63+CTPCI Falcon. Error message: "Arithmetic overflow at line 514 in file E:\GB6\GUI24.BAS". That's on the graphics card in 1920x1200@16M colours.
Ektus wrote:On that machine, GemBench 4.03 throws an Exception #2 immediately after start.
Ektus wrote:In the following bench results, the "REFERENCE" and "SYSTEMS" columns appear switched.
exxos wrote:I think 16million colours doesn't have a bit plane number ? Can't really remember though I think "true colour" doesn't use planes ? So assume it came back as a stupid huge number and caused the error your saw.
Cyprian wrote:exxos wrote:I think 16million colours doesn't have a bit plane number ? Can't really remember though I think "true colour" doesn't use planes ? So assume it came back as a stupid huge number and caused the error your saw.
bitplanes are used only in 1-8 bit color mode. 15bit - 32bit modes use chunky pixel (two/three/four consecutive bytes define pixel color).
lp wrote:32-bit gfx has an extra (unused) byte for an alpha channel. Simple fix is if you see 32-bit mode, report it as 24-bit color space, since that's what it actually yields. ie total_colors=shl(1,24)
lp wrote:As for the multi-tasking issues, it's basically terrible to be blunt. Allowing it to run is silly since it's not hard to detect a tasking OS. My advice is, if you are not going to support multi-tasking operating systems then you should check the aes global array entry for maximum current applications and show and alert box and exit. I can do this in about 4 lines of gfa code.
See global(1): http://toshyp.atari.org/en/008002.html#AES_20bindings
Users browsing this forum: Rustynutt and 2 guests