Amiga.org
Amiga computer related discussion => Amiga Software Issues and Discussion => Topic started by: darksun9210 on November 01, 2019, 11:41:09 PM
-
Hey all, sorry I don't wish to duplicate threads, but now I *think* I've got the hardware sorted out, the software still seems to not be quite there - hence starting a mediator software thread. hope this is ok.
so. A4000/040. mediator. Radeon9250. gets the beat down from a CV64/3D. what the heck? "the most powerfull graphics card on classic systems". nope. not to me. feels chuggy and glitchy.
latest Elbox software update installed? yup.
latest Picasso, sorry "P96" software installed? yup.
i'm HOPING it's just a CPU issue, missing local ram, having to use mainboard memory.
If anyones got a similar setup, could you post your P96speed output please? I just need some confirmation that this is correct.
here's mine:-
(a3k = A3000, cyberstorm3/060, cybervision64/3D - A500 = A500, Vampire2 - A4k = A4000 as above)
operation A3k A500 *A4k*
RectFill 2889 4261 6722
RectFillPattern 2842 6158 5875
WritePixel 147969 301582 57870
WriteChunkyPixels 741 6158 403
WritePixelArray 730 6072 386
WritePixelLine 30072 36449 11630
DrawElipse 9874 24698 4920
DrawCircle 10743 25527 5975
Draw 14644 6768 1328
DrawHor/Ver 24624 16067 9811
ScrollRasterX 162 313 1259
ScrollRasterY 160 389 1330
Text 10351 12596 2849
BltBitMap 8660 8530 6229
BltBitMapRastPort 7058 7135 3368
BitMapScaling 75 548 66
Openwindow 103 172 20
MoveWindow 392* 1135* 43
SizeWindow 117* 261* 11
ConOutput 371 652 104
ScreenToFront 61 61 60
-
What version of pci.library are you running? They just released a new version a few days ago. I had a Radeon in mine, but I replaced it with a Voodoo 3, or I would test it. But I agree your numbers seem off. Then again you say the 4000 has an 040, but the other two have 060/080? That could be why the numbers are off. My mediator box has an 060/PPC.
-
aha. i'll give that a try. am running 9.11 of pci.library, and 2.23 of Radeon.card. thanks for the heads up on the new version, i hadn't spotted that.
i've even tried a flat install of OS3.1, cdrom (idefix) drivers, and then the mediator CD install, and the update .lha file CD2.5 (pre 31oct2019). and then update to P96 to v2.3.
running P96 speed generates slightly different numbers. some a little better, some a little worse. but effectively the same as i'm within a margin of error of a few points per thousand.
i managed to rectify a lot of my perceived user noticeable performance issues by maxing out the on board ram to 16meg from it's original 8meg. seems that crossing into zorro3 ram doubles down on putting a damper on performance.
i'd imagine you're right. not having access to true cpu local ram really hurts. that said, i figured something was right when the povray bench test was nearly two hours faster than previous, and a tiny bit faster than my a3k running the same test with the same cpu board. So i had high hopes P96 speed would be good. sadly no real change.
trouble is, local ram cpu boards are few and few between. if i get the time, i may swap the A3k and a4k cpu boards around and re-run the bench on both machines just as a double check