03-13-2014, 05:16 PM (This post was last modified: 03-13-2014, 05:18 PM by Archanfel.)
(03-12-2014, 11:15 PM)ssshadow Wrote: Interesting. If this is an emulator problem I want to know. I will be investigating performance for recent builds under the conditions you described when I'm back at my PC on Friday afternoon European time. But please report if you find a way to reliably reproduce the issue.
Unfortunately I've been unable to get it run at good speed again, I'll keep trying though and update this thread. I thought it might be because I run in classic windows mode rather than "Aero" but I've ruled that out. All I can think of is either the AMD display drivers causing an issue or it's just an issue with the emulator itself. Would be interesting to see performance figures from people with similar systems (i5 QuadCore and a GCN core graphics card (AMD 7 Series and up) or maybe any semi-modern AMD card that uses latest drivers (5/6 series))
03-13-2014, 08:07 PM (This post was last modified: 03-13-2014, 08:08 PM by Cole Roddy.)
(03-13-2014, 05:16 PM)Archanfel Wrote:
(03-12-2014, 11:15 PM)ssshadow Wrote: Interesting. If this is an emulator problem I want to know. I will be investigating performance for recent builds under the conditions you described when I'm back at my PC on Friday afternoon European time. But please report if you find a way to reliably reproduce the issue.
Unfortunately I've been unable to get it run at good speed again, I'll keep trying though and update this thread. I thought it might be because I run in classic windows mode rather than "Aero" but I've ruled that out. All I can think of is either the AMD display drivers causing an issue or it's just an issue with the emulator itself. Would be interesting to see performance figures from people with similar systems (i5 QuadCore and a GCN core graphics card (AMD 7 Series and up) or maybe any semi-modern AMD card that uses latest drivers (5/6 series))
Did you tried the "fresh folder" solution that I did sugested you to do? As I said before, I had the same problem you had, but could not replicate the issue, otherwise I would file a bug report. So after that, I made a new folder and I started over.
It's a bug in RPCS3, not in your AMD card. I have a nvidia card, just so you know
(03-13-2014, 08:07 PM)Cole Roddy Wrote: Did you tried the "fresh folder" solution that I did sugested you to do? As I said before, I had the same problem you had, but could not replicate the issue, otherwise I would file a bug report. So after that, I made a new folder and I started over.
It's a bug in RPCS3, not in your AMD card. I have a nvidia card, just so you know
Yes I tried the fresh folder, fresh settings, fresh everything - several times and no change.
Anyway, just tried running it from my original setup location and it was running at correct speed again. Tried running it after and it was back to slooowww mode. The difference being this time I got a screenshot of it running at the faster speeds which I've attached. We can clearly see that the CPU is being used a lot more on all four cores and that the GPU is at 3d clocks this time. However the GPU didn't boost to 3d clocks for a while as you can see from the GPU-Z Sensor graph and while it was still at 2d clocks the game was running at 30FPS, so while the GPU clocks obviously have a bearing it is something else in the emulator I assume causing this issue. So yes, I guess we can rule out the GPU but I don't think this emulator likes my PC setup heheh
Oh, forgot to say we can also see the GPU load is at 19% while it's running good whereas it was 0% running slow (occasionally jumping to 5%). Something is obviously causing a bottleneck somewhere.
03-13-2014, 10:39 PM (This post was last modified: 03-13-2014, 10:40 PM by Cole Roddy.)
I noticed low CPU usage too when it happened to me. It puzzled me, but as I could not replicate the issue, I can't help much. I just rebooted my rig, recompiled the build and moved on. The issue never happened again, unfortunely. Are you using the compiled versions avaliable here at the forums, Archanfel? Or are you using your own compiled builds? I wish I could help, but without been able to replicate the issue, it's a dead end for me.^^''
(03-13-2014, 10:39 PM)Cole Roddy Wrote: I noticed low CPU usage too when it happened to me. It puzzled me, but as I could not replicate the issue, I can't help much. I just rebooted my rig, recompiled the build and moved on. The issue never happened again, unfortunely. Are you using the compiled versions avaliable here at the forums, Archanfel? Or are you using your own compiled builds? I wish I could help, but without been able to replicate the issue, it's a dead end for me.^^''
I was using the builds compiled at emucr but have switched to the builds here. I don't bother compiling myself as my knowledge of C++ is limited, that and I don't have Visual Studio. Only programming I do/have done is in Turbo Pascal/Delphi.
Thanks for the efforts though! I'm sure the developers can locate whatever piece of code is causing this though I suppose it isn't that great a deal if it's only really affecting me lol pity we can't replicate it at will
Update! The analog stick required to move block does work under the condition that another key is held in at the same time. So if I hold in X (mapped to j) I can used page up/page down to move blocks, and move the cursor around with WASD. As soon as I release X, things stop working.
Also, home, page up, etc, is horrible to use on a laptop keyboard Here is a new video where we actually play the game: http://youtu.be/JrCRs39RAr8
Asus N55SF, i7-2670QM (~2,8 ghz under typical load), GeForce GT 555M (only OpenGL)
03-16-2014, 10:56 AM (This post was last modified: 03-16-2014, 10:57 AM by Cole Roddy.)
(03-16-2014, 10:45 AM)ssshadow Wrote: Update! The analog stick required to move block does work under the condition that another key is held in at the same time. So if I hold in X (mapped to j) I can used page up/page down to move blocks, and move the cursor around with WASD. As soon as I release X, things stop working.
Also, home, page up, etc, is horrible to use on a laptop keyboard Here is a new video where we actually play the game: http://youtu.be/JrCRs39RAr8
If you look at the code, that's how it's supposed to work. I even pointed that portion of the code some posts ago (page 7). It was coded 2 months ago. I guess I should've played the game in my demo video instead of just showing it running.
heheh
pity we can't replicate it at will
