|
|
|||||||
| About Us | Register | FAQ | Members List | Calendar | Mark Forums Read |
![]() |
|
|
LinkBack | Thread Tools | Display Modes |
|
|
#41 (permalink) | |
|
Emu author
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Sep 2006
Location: VisualC
Posts: 686
|
Quote:
Anyway, whenever you release it I'll profile it properly and compare with the official 2D core to see the perfomance differences, and commit it to the official CVS. You'll understand that I need to review changes to the official CVS even if they come from a reputed coder, I don't want to offend you
__________________
Emulator development blog Last edited by shashClp; February 11th, 2008 at 20:54. |
|
|
|
|
|
|
#42 (permalink) |
|
PCSX2 Coder
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Jan 2004
Location: Plymouth, UK
Posts: 9,744
|
I have to agree with the profiling comments. What youve described about commenting ou tthe 2d core and the music ran faster, this is bound to happen. Everything thats being used takes processing power, so cutting that processing power out will free it up for other parts to use, but it doesnt mean that code is slow, it just means its working
![]() its like in a game on the pc, saying AA is slow cos you put it on AAx4 and your frame rate reduces, but you turn it off and it goes loads faster, its nothing to do with AA performance, its more than likely that performance is at a peak with the AA, its the fact its not being used when you turn it off. IF it does make a difference over the official 2d core it would be good to see it run. im all for the "every little helps" frame of mind, much like saving up pennies for a rainy day :P
__________________
http://www.generalemu.net/ Intel Core 2 Quad Q6600 @ 3.4Ghz (425x8), eVGA 8800GTX 768mb, 996Gb HDs, 2Gb Corsair DDR2-800 @ DDR2-1020 5-5-5-18 Dual Channel, 14605 3dMark 06 Points CPU-Z Link 3dMark06 Score Link Dont PM me for help, use the forums, thats what its for! |
|
|
|
|
|
#43 (permalink) | ||
|
Eric Cartmenez
![]() ![]() ![]() Join Date: Jul 2006
Location: Brazil
Posts: 359
|
Quote:
Quote:
|
||
|
|
|
|
|
#48 (permalink) | |
|
Emu author
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Sep 2006
Location: VisualC
Posts: 686
|
Quote:
__________________
Emulator development blog |
|
|
|
|
|
|
#49 (permalink) |
|
Plugin author
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Jul 2007
Location: Nulle part
Posts: 413
|
It seems DesMuMe has a wrong bios. In effect, in New Super Mario Bros, the mini-games don't work. When you select them, the game calls funcs in bios. I've constated that the values shown by the disassembler are diferent from the values shown by NeonDS. And the mini-games work with NeonDS. So I think DesMuMe has a wrong bios.
|
|
|
|
|
|
#50 (permalink) | |
|
Emu author
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Sep 2006
Location: VisualC
Posts: 686
|
Quote:
__________________
Emulator development blog |
|
|
|
|
|
|
#51 (permalink) |
|
Plugin author
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Jul 2007
Location: Nulle part
Posts: 413
|
yes. Are the mini-games working with your DesMuMe ? I've seen you can play New Super Mario Bros with it. If you can play the mini-games, how did you get them to work ?
If not, this HLEd bios could be the problem. PS : I'm currently working on implementing RTC. |
|
|
|
|
|
#52 (permalink) |
|
Emu author
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Sep 2006
Location: VisualC
Posts: 686
|
Just checked, they don't work. I'll debug later, and see what's the problem.
__________________
Emulator development blog |
|
|
|
|
|
#54 (permalink) |
|
Emu author
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Sep 2006
Location: VisualC
Posts: 686
|
Just looked a bit at the issue, it seems it's related to a DMA failing or not copying data. Afaik, it has nothing to do with the bios, as it's jumping where it should :P
EDIT: It CAN'T be a DMA, because the uninitialized area it the ARM9 ITCM. For some unknown reason, whatever should be copied there, isn't, so it's failing
__________________
Emulator development blog Last edited by shashClp; February 17th, 2008 at 20:25. |
|
|
|
|
|
#56 (permalink) |
|
Emu author
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Sep 2006
Location: VisualC
Posts: 686
|
No, read my edit, due to the uninitialized area being the ITCM, it CAN'T be a DMA, as the DMA can't access ITCM (atleast that's what I read in the gbadev forums). Here's a reference, I'll test it ASAP. Else, just search gbadev forums for the ITCM for more info :P
EDIT: I would bet that's it's something with the ITCM not being initialized correctly, atleast that's what's differing with Ideas and NeonDS. Well have to figure it out (preferred one), or ask to the NeonDS author to add an Step/Breakpoint functionality (lazy option) :P
__________________
Emulator development blog Last edited by shashClp; February 18th, 2008 at 07:21. |
|
|
|
|
|
#57 (permalink) |
|
Plugin author
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Jul 2007
Location: Nulle part
Posts: 413
|
Ah okay, I'll see why ARM9 ITCM is not properly copied. I've provided a snapshot, you are right, I've loaded the game and updated the disassemblers, the ARM7 one shows opcodes, whereas the ARM9 one shows lines of "0x00000000 andeq r0, r0, r0" and nothing other, so ITCM is not initialized.
|
|
|
|
|
|
#59 (permalink) |
|
Emu author
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() Join Date: Sep 2006
Location: VisualC
Posts: 686
|
Oh, hell, I'll test my version today adding a proper patch for that, I'll tell you the results. By the way, I'll update the CVS this night with some interesting fixes on the CPU, which I guess you'll already have.
Thanks for the information
__________________
Emulator development blog |
|
|
|
![]() |
| Thread Tools | |
| Display Modes | |
|