RetroPie forum home
    • Recent
    • Tags
    • Popular
    • Home
    • Docs
    • Register
    • Login
    Please do not post a support request without first reading and following the advice in https://retropie.org.uk/forum/topic/3/read-this-first

    VICE Performance Decrease since 2.4 -> 2.4.30

    Scheduled Pinned Locked Moved Help and Support
    viceperformance
    16 Posts 3 Posters 3.5k Views
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • vbsV
      vbs
      last edited by

      If I am right then VICE got recently updated from version 2.4 to 2.4.30.
      I have the feeling that performance got much worse in 2.4.30. I tested on Pi3 and on x86. I noticed it on the cracker intro of Madness (MADNESS__1986__RAINBOW_ART2.D64). That intro runs smooth on 2.4 with about 40% CPU on Pi3. On 2.4.30 it maxes out CPU to 100% and stutters. I had to disable SidEngine to make it run smooth. Without SidEndinge CPU was back to 40%.

      Is it me (maybe messing up config) or did someone else notice?

      1 Reply Last reply Reply Quote 0
      • BuZzB
        BuZz administrators
        last edited by

        The x86 and rpi3 defaults are different afair. But the update could have affected performance elsewhere. You should make sure you have the new default config in place as old config may cause issues.

        To help us help you - please make sure you read the sticky topics before posting - https://retropie.org.uk/forum/topic/3/read-this-first

        1 Reply Last reply Reply Quote 0
        • vbsV
          vbs
          last edited by

          Yes true, the config was changed too. But still, I think performance is significantly worse with identical configs. But maybe it is only related to SidEngine (which is disabled by default both in 2.4 and 2.4.30 if I am not mistaken).

          1 Reply Last reply Reply Quote 0
          • BuZzB
            BuZz administrators
            last edited by

            It's possible it got slower (and more accurate etc), but when I tested, I used some pretty complicated demos on the rpi3 and performance was fine. Make sure you have doublesize/double scan set to 0 also.

            Did you test with the new default config on the rpi3 ?

            To help us help you - please make sure you read the sticky topics before posting - https://retropie.org.uk/forum/topic/3/read-this-first

            1 Reply Last reply Reply Quote 0
            • BuZzB
              BuZz administrators
              last edited by

              BTW, if you are talking about performance with resid sid emulation on - it will be significantly slower. use the fast sid emulation for performance.

              To help us help you - please make sure you read the sticky topics before posting - https://retropie.org.uk/forum/topic/3/read-this-first

              1 Reply Last reply Reply Quote 0
              • BuZzB
                BuZz administrators
                last edited by

                BTW which crack intro is it ? (So I can test) - The Nostalgia crack or another ?

                To help us help you - please make sure you read the sticky topics before posting - https://retropie.org.uk/forum/topic/3/read-this-first

                1 Reply Last reply Reply Quote 0
                • BuZzB
                  BuZz administrators
                  last edited by BuZz

                  2 of the demos I tested with http://csdb.dk/release/?id=133934 / http://csdb.dk/release/?id=72550

                  To help us help you - please make sure you read the sticky topics before posting - https://retropie.org.uk/forum/topic/3/read-this-first

                  1 Reply Last reply Reply Quote 0
                  • vbsV
                    vbs
                    last edited by

                    It is by group REMEMBER. It shows big blue letters with a green bar moving up and and down.
                    I mean the option "SidEngine". Was fast in 2.4 with value "1" and now it seems slow with same setting.

                    I will test this again tomorrow more systematically ;)

                    1 Reply Last reply Reply Quote 0
                    • BuZzB
                      BuZz administrators
                      last edited by BuZz

                      If you are talking about the Resid engine - I think they switched to using resid-fp which is slower (and more accurate).

                      the remember intro on my rpi3 with the resid engine on uses 80% cpu and sounds lovely.

                      the other thing that changed in the retropie config is default frequency on rpi2/3 from 22050 to 44100 (rpi1 still has a default of 22050). But the fast sid engine is still default on rpi.

                      To help us help you - please make sure you read the sticky topics before posting - https://retropie.org.uk/forum/topic/3/read-this-first

                      vbsV 1 Reply Last reply Reply Quote 0
                      • A
                        AmigaGamer
                        last edited by

                        @vbs check in your sound settings to see what the output driver is set to.
                        If its SDL change it to ALSA and see how you get on.

                        I had big slowdowns with the SDL driver, with alsa my RPI3 hasnt gone over 30% on that cracktro

                        BuZzB 1 Reply Last reply Reply Quote 0
                        • vbsV
                          vbs @BuZz
                          last edited by

                          @BuZz
                          Yesterday I thought the option SidEngine was a bool value but it turns out that it selects different engines. And you are right, it seems to be related to RESID which has gotten significantly slower it seems. With FASTSID the different is only subtle.

                          @AmigaGamer
                          Yes, thank you, but alsa was the default for me.

                          BuZzB 1 Reply Last reply Reply Quote 0
                          • BuZzB
                            BuZz administrators @vbs
                            last edited by

                            @vbs the intro you mentioned is still full speed here on rpi3 though - so on top of the resid engine being slower you may have another issue.

                            To help us help you - please make sure you read the sticky topics before posting - https://retropie.org.uk/forum/topic/3/read-this-first

                            1 Reply Last reply Reply Quote 0
                            • BuZzB
                              BuZz administrators @AmigaGamer
                              last edited by

                              @AmigaGamer it's 30% with fast Sid emulation. 80% with resid here on rpi3. More complicated demos seem to hover at around 80% too, so I might change the engine to resid on rpi3 by default as it sounds a lot nicer than the fast sid emulation.

                              To help us help you - please make sure you read the sticky topics before posting - https://retropie.org.uk/forum/topic/3/read-this-first

                              1 Reply Last reply Reply Quote 0
                              • A
                                AmigaGamer
                                last edited by

                                @buzz Agree on both the above points. The cracktro with resid is sitting comfortably at 77% for me.

                                A switch to resid on rpi3 undoubtebly sounds better/more authentic. I just wonder if the increased cpu overhead will affect in other areas.
                                Eg load times/warp mode performance.
                                Ill be staying with REsid for a while and see how i get on.
                                FWIW my opinion right now is FASTsid is the safer bet for the default install. The support forum is less likely to fill up with questions about audio fidelity than it is for slowdowns/buffer overflow errors!

                                @vbs only other things i can think of are check the following
                                sound buffer 100 msec
                                fragment size medium
                                frequency 44100

                                in system/sid settings make sure you dont have an "extra" sid enabled

                                1 Reply Last reply Reply Quote 0
                                • vbsV
                                  vbs
                                  last edited by

                                  Guys, sorry I probably was not clear: I currently don't have a real problem or at least I think it runs normally for me. I just noticed that decrease in performance. When I set SidEngine to 1 then for me also it runs smooth with about 80% CPU.
                                  But the performance decrease is explainable if they changed something in the SidEngine which is more costly now.
                                  But in the end it means it won't be possible to have RESID and the CRT effect at the same time, right?

                                  1 Reply Last reply Reply Quote 0
                                  • A
                                    AmigaGamer
                                    last edited by

                                    @vbs makes sense that you had slowdown now you mention the CRT filters. The filters adding approx 20% load so combined with resid thats going to cause slowdown. Testing the Edge of Disgrace demo with both switched on is causing stuttering/frame drop to 40fps

                                    1 Reply Last reply Reply Quote 0
                                    • First post
                                      Last post

                                    Contributions to the project are always appreciated, so if you would like to support us with a donation you can do so here.

                                    Hosting provided by Mythic-Beasts. See the Hosting Information page for more information.