Project

General

Profile

Actions

Emulator Issues #3967

closed

F-Zero GX unstable in Dual Core Mode

Added by nitro322 almost 14 years ago.

Status:
Fixed
Priority:
Normal
Assignee:
-
Category:
UI
% Done:

0%

Operating system:
N/A
Issue type:
Bug
Milestone:
Regression:
No
Relates to usability:
No
Relates to performance:
No
Easy:
No
Relates to maintainability:
No
Regression start:
Fixed in:

Description

What's the problem?
When attempting to start F-Zero GX, the Dolphin will hang with a black screen if Dual Core support is enabled. If I disable that option, the game boots fine.

Dolphin version with the problem (as it appears in the title bar, Ex.: "R
4779" or "R 6403M"):
r6892

(optional) Dolphin version that does not have the problem:
The last I tried was around r6502, and I'm pretty sure it worked then.

Operating system and version: Gentoo Linux
32-bit or 64-bit: 64-bit

Game ID (as it appears in game properties, Ex.: "GZ2P01" or "RSBE01"):
GFZE01

Build command-line (not on Windows):
cmake -C /tmp/portage/games-emulation/dolphin-emu-9999_p6892/temp/gentoo_common_config.cmake -DCMAKE_INSTALL_PREFIX=/usr -DDOLPHIN_WC_REVISION=6892 -DCMAKE_INSTALL_PREFIX=/usr/games -Dprefix=/usr/games -Ddatadir=/usr/share/games/dolphin-emu -Dplugindir=/usr/games/lib64/dolphin-emu -DDISABLE_WX=OFF -DENCODE_FRAMEDUMPS=ON -DCMAKE_BUILD_TYPE=Gentoo -DCMAKE_INSTALL_DO_STRIP=OFF -DCMAKE_USER_MAKE_RULES_OVERRIDE=/tmp/portage/games-emulation/dolphin-emu-9999_p6892/temp/gentoo_rules.cmake /tmp/portage/games-emulation/dolphin-emu-9999_p6892/work/dolphin-emu-9999_p6892

Was the ISO a plain dump from disc, compressed and/or scrubbed?
Compressed ISO

Actions #1

Updated by tez759 almost 14 years ago

Yeah, I am having this problem too.

If you keep Dual Core enabled and disable Idle Skipping, it will boot, but expect to lose FPS.

Actions #2

Updated by marcel.werner3 almost 14 years ago

Works finde here with r6892. dual core and idle skipping enabled.
maybe it's a linux issue? I'm using Win7 x64

Actions #3

Updated by tmkkrak almost 14 years ago

Dx11 doesn't work...
Dx9 it's ok
But is massive slowdown with OpenCL on

Actions #4

Updated by fooldrainer almost 14 years ago

im using linux and it plays fine for me

Actions #5

Updated by nitro322 almost 14 years ago

Huh. The different reports are interested. I just tested OpenCL, but, at least for me, the problem is definitely with the Dual Core option. OpenCL doesn't appear to make a difference, but it will always fail if Dual Core is enabled. I also did try disabling Idle Skipping before posting my original report, but that doesn't seem to make a difference either (at least for me).

Actions #6

Updated by tmkkrak almost 14 years ago

My config:
Ati 5470, i5 2,26 4 gb ram
win 7 x64, r6888 and r6894
Dual core: ok
Idle Skipping: ok
dx 9 - ok
dx 11 -doesn't work
dx 9 + opencl - slowdown

Actions #7

Updated by sandro.o.maia almost 14 years ago

FZero needs mmu to be enabled, if i not wrong. I can play with both single / dual core modes but only when i check mmu .....

Actions #8

Updated by nitro322 almost 14 years ago

The MMU requirement is confirmed here, but still no-go on dual core. More specifically:

no Dual Core, MMU - success
no Dual Core, no MMU - fail
Dual Core, MMU - fail
Dual Core, No MMU - fail

Actions #9

Updated by tmkkrak almost 14 years ago

Update:
Dx11 isn't working when the option is copy to texture...
in my case mmu doesn't matter

Actions #10

Updated by tmkkrak almost 14 years ago

Update 2:
mms hoever does matter with dual core :)

Actions #11

Updated by wespipes69 almost 14 years ago

I can boot it with DC and MMU BUT only one the 2nd attempt. Upon launching it the first time, it just hangs on a black screen. If I stop and try again, it'll launch fine. Weird.

Actions #12

Updated by marcosvitali almost 14 years ago

wespipes69: the same here soemtime the first time doesnot work :S

Actions #13

Updated by lkeeler.flipout over 13 years ago

My machine:
Intel Core2Quad Q6600
Win 7 x64
AMD HD6870
2Gb ram.

I can run it using OpenGl, D9 and D11, anyways D11 is slow as hell and OpenGL gets random resets.
Deactivating dual core just slowdown the game, i can play it percfectly with it.

Actions #14

Updated by skidau over 13 years ago

Is this issue still occurring?

Actions #15

Updated by wespipes69 over 13 years ago

Works for me fine now.

Actions #16

Updated by skidau over 13 years ago

  • Status changed from New to Fixed
Actions #17

Updated by wespipes69 over 13 years ago

Ok Skid, sorry I was wrong - I was testing with DX9 only. There was a lot of back and forth in this thread so I'd go ahead and disregard most of that at this point.

So here's wher we're at as of r7338:
DX9 + Dual Core = Boots 90% of the time (just sometimes not on the first)
DX11\OpenGL + Dual Core = Never boots

In Single Core mode, there's zero booting issues. Please reactivate this bug as it is still valid. Thanks!

Actions #18

Updated by skidau over 13 years ago

  • Status changed from Fixed to New

Would you please try to find when this issue started occurring? I would guess that it is sometime after r7185 (i.e. the game should be working in r7185).

Actions #19

Updated by wespipes69 over 13 years ago

DX11 became completely unusable for this game in DC mode between r7158 (works) and r7167 (can't boot).

That's all I can narrow it down to for sure. Looks like it wasn't that big fifo comment that broke it but there's certainly a few culprits around there. :)

Actions #20

Updated by matthew.p.mueller over 13 years ago

Based on the last two comments I tested some revisions with Linux 64 bit and dual core enabled.

r7157-r7164: displays several gfx fifo "unknown opcode" errors, but then works
r7165-r7167: never boots
(r7185: build fail)
r7187: never boots

Actions #21

Updated by wespipes69 over 13 years ago

r7165 it is then! :) Looking at the comments for the commit, not surprising.

Actions #22

Updated by wespipes69 over 13 years ago

Should this be assigned to marcosvitali then? And can we change the title to include the rev that broke this (r7165) and how it mainly affects DX11.

DX9 still needs to be restarted a few times to boot, but I think that's how it's alwyas been.

Actions #23

Updated by nitro322 over 13 years ago

FYI (since there's so much focus on DX11), this is still a problem under Linux, and is still, on my system at least, directly related to the Dual Core option. As I posted in my initial report, F-Zero GX will always hang on loading with Dual Core enable, and always boot fine with Dual Core disabled. I just tested again with r7554 and still get the same behavior.

Even if the DX11 driver is somehow involved, please don't overlook the Dual Core issue that was the original cause of this bug report. Thanks.

Actions #24

Updated by rasert over 13 years ago

I have the same problem here.

I'm running Dolphin 3.0 on Mac OS X Lion 10.7.0, so here is the thing:

OpenGL + Dual Core = FAIL
OpenGL + Single Core = SUCCESS

I have Windows 7 SP1 running over BootCamp here and I have Dolphin 3.0 installed in Windows. There I can run the game with the following settings:

DirectX 9 + Dual Core = SUCCESS

Actions #25

Updated by wespipes69 about 13 years ago

Still no change as of 3.0-183. Please see comment 24 for a couple things to maybe do for this report.

To sum up everything at this point:

  • DX9 + Dual Core = Boots 90% of the time (just sometimes not on the first)
  • DX11\OpenGL + Dual Core = Never boots
  • Single core = No boot issues
  • Broken by r7165
Actions #26

Updated by jodr666 about 13 years ago

With 3.0-199 x64:

DX9+Dualcore = Works fine. I get sound stuttering on some tracks.
DX11+Dualcore = It boots but after it does i get random FPS speeds plus the music not playing correctly.

Actions #27

Updated by wespipes69 about 13 years ago

Using 3.0-201, I still cannot get DX11+DC to boot at all - even if I start/stop emulation multiple times and in different sessions. It just never starts.

But I did observe something else that's good, when using DX9, the game always starts on the very first time now. Tried it like 25 times. Before it always needed to be restarted once or twice. So that's some nice improvement. Maybe something to do with the cache improvements lately?

Actions #28

Updated by wespipes69 over 12 years ago

Now, the game seems to boot just fine, on any backend, every time.

There's an issue with restarts (on all backends) and DX11 is incredibly slow (like 1-5 fps) on a Sandy CPU. Something screwy there. But those are the only two issues left that I see.

Actions #29

Updated by wespipes69 over 12 years ago

Update: DX11 run a lot better nowadays but still not full speed (~70%).

Stability issues remain (worse with DX11). Maybe this should be tracked in new bugs?

Actions #30

Updated by Autoran1 about 12 years ago

All backends work fine with DC DX11 is slower itself
rare random reboots still presented

Actions #31

Updated by aavindraa almost 12 years ago

It seems to be working now? I think dual core should be enabled by default in the config for this game now. It's really slow in one core mode.

Actions #32

Updated by gamedevistator almost 12 years ago

The game works in dual core, but when you start the game, it restart once when it gets to the start screen and then form there on it works, but its really slow at some parts of the game with DX11 even on hle it gets really lagy. I dont know why this game is so hardware demanding. I can play most of game at 60fps with 1080p resolution and IR at 4x.
Specs:
Intel Core2 Quad Q8400 2.66Ghz
6gb DDR3 1333Mhz
Asus Radeon HD 6870 1gb GDDR5 OC 915mhz
Win 8 Pro 64 bit

Actions #33

Updated by Billiard26 almost 12 years ago

  • Status changed from New to Accepted
  • Issue type set to Bug
  • Category set to ui
Actions #35

Updated by kostamarino almost 12 years ago

I don't know whether it is a good idea or not to change it since dual core is still unstable, i wasn't able to finish a tournament without the game freezing-restarting every time i tried it while with single core is 100% stable. My guess is that this issue will be replaced with issues of F-zero being unstable-freezing.

Actions #37

Updated by skidau almost 12 years ago

  • Status changed from Accepted to Fixed

Fixed by r0e2c3f3483e2.

Actions

Also available in: Atom PDF