Forum Index  >>  Off Topic  >>  

1964 UltraFast - 60FPS N64 emulator

      

   Page 4 of 14   

1 year ago - Tuesday, February 08, 2011
Type: Member
Posts: 41
Joined: 2/7/2011
RetroRalph wrote:

I didn't mean VSYNC on your PC. I mean the internal emulator vsync to the "fake" TV. ie 50 or 60 VIs. On the real system you get a VI every 1/60th of a second regardless, if the console couldn't push out a frame in that 1/60th of a second then you're not getting a 1:1 translation of frame to sync.

You are not getting 60FPS average framerate in any normal emulator or on the real system. If it "appears" smooth enough for you then that is a separate issue but GoldenEye in particular very rarely hits 60FPS on the real console. So you cannot tell me that you're getting "full speed" without overclocking because to me full speed is getting 60 real frames per second, all the time.

I am not using VSYNC on my notebook, it is the vsync option in the Glide64 plugin.

RetroRalph wrote:

It's not a mockery because I am using the information you have given me and compared it to my own results on a CPU that is the same architecture as yours. There is no valid explanation as to why my core2 CPU can outperform yours in this emulator (or the original 1964) by over 1000% when it's only ~50% faster (taking into account more MHz and bus/memory speeds).

You're basically telling me a core2 CPU at 2.5GHz made in 2008 is getting less performance in normal 1964 than a 1.4GHz AMD CPU I used to own nearly 11 years ago? Can you see why I think there is something wrong with your system or configuration?

The info I have provided is not for benchmarking against your system and your comparison is a mockery from a developer standard.
1. what resolution are you using? windowed or fullscreen?
2. which scene in the game are you comparing and what fps or VI etc?
3. which plugin are you using to benchmark the game? sound, video, RSP etc.
4. which setting is checked or unchecked in the plugin and rom settings options?
etc etc etc etc

Now do you understand why I say it is a mockery?
Your previous comment mention I7 with My core2 @ 3.2GHz.
Now you mention 1.4GHz AMD CPU.
Next, are you saying it is a PIII or PIV ???

Let me give another result from a different perspective.
I just need to set different settings, I can get 57.x+ to 59.x+ fps when I go in-game and the transition between scene during cutscene is almost instant with no overclocking on my build. Also, CPU usage is not max.
If I use the same setting on your build, I will still get the same result except that fps fluctuation is wider when overclock (50.x+ to 59.x+) - animation is slightly slower(sound is affected & CPU is max) compared to my non-overclock build using the same setting.

Now you know why I say your comparison and statement of 1000% is a mockery?
I am not using a setting that will give me almost 60fps because it is the only way to see if overclocking will improve performance, remain the same or decrease performance.

Also, you mention about comparing your result from a CPU with same architecture as mine. So, you are saying your 1.4GHz AMD CPU has the same architecture as mine? (I will let other experts to determine the truth of it?)
Also, you mention your AMD outperform mine by comparing result but I wonder how you know which settings I use(as mentioned in the above questions) that you are so sure that you are comparing "apple to apple" and not "apple to orange" etc.

Just a final note.
I am sure what I am comparing because I run the same game with the same scene for comparison on the same machine with the same settings for all 3 builds. Maybe you are a "Mind-Reader".
Edited 12 times. Last edit by RetroRalph
1 year ago - Wednesday, February 09, 2011
Type: Member
Posts: 6
Joined: 2/7/2011
If anyone else happens to make it here with similar issues to those I found, here are my system specifications:

AMD 64 X2 Dual Core 4200 @ 2.20GHz
4.00 GB Crucial Ballistix RAM
Windows 7 64 bit
ATI Radeon HD 5570


Perhaps version 3 will enable this emulator to run at full speed with the dynamic recompiler.
Edited 1 times. Last edit by throwingmuse
1 year ago - Wednesday, February 09, 2011
Type: Administrator
Posts: 1336
Joined: 7/19/2009
Pokefan999 wrote:

I am not using VSYNC on my notebook, it is the vsync option in the Glide64 plugin.


I am not talking about VSYNC option for your monitor (which is what the glide plugin option is for). I don't know how I can make that more clear.

Pokefan999 wrote:

Let me give another result from a different perspective.
I just need to set different settings, I can get 57.x+ to 59.x+ fps when I go in-game and the transition between scene during cutscene is almost instant with no overclocking on my build. Also, CPU usage is not max.
If I use the same setting on your build, I will still get the same result except that fps fluctuation is wider when overclock (50.x+ to 59.x+) - animation is slightly slower(sound is affected & CPU is max) compared to my non-overclock build using the same setting.


Then you are mistaking FPS for VIs. No non overclocked emulator is getting 60 real FPS for any length of time in GoldenEye during that opening first level intro scene. GoldenEye is CPU limited in most scenes. Do not continue to argue this because it's the last time I'm going to tell you a fact. If you want to believe and talk about something other than reality then you're in the wrong place.

Pokefan999 wrote:

Also, you mention about comparing your result from a CPU with same architecture as mine. So, you are saying your 1.4GHz AMD CPU has the same architecture as mine? (I will let other experts to determine the truth of it?)
Also, you mention your AMD outperform mine by comparing result but I wonder how you know which settings I use(as mentioned in the above questions) that you are so sure that you are comparing "apple to apple" and not "apple to orange" etc.


I have multiple testing machines, and my testers also do. There is no "confusion" here.

My AMD 1.4Ghz gets more performance than your 2.5GHz core2 in 1964 with the default plugins. Let alone my core2 3.2GHz which has 1000% the performance of your setup in UltraFast. This is going on your results you posted in this thread.

That doesn't make sense does it? It shows you have something wrong with your system (dynamic underclocking, power settings, etc) or a configuration issue or you don't know how to describe properly what you're seeing.
1 year ago - Wednesday, February 09, 2011
Type: Administrator
Posts: 1336
Joined: 7/19/2009
throwingmuse wrote:

Perhaps version 3 will enable this emulator to run at full speed with the dynamic recompiler.


Are the dynarec settings in game options all set to default values? I'll try a few things for v3 anyhow.
1 year ago - Wednesday, February 09, 2011
Type: Member
Posts: 6
Joined: 2/7/2011
Found the issue!

I am finally able to get dynamic recompilation to work correctly IF I set the compiler to "No Check" for Self-Modifying Code rather than the default "Protected Memory."

After this change, I get consistent 30-35 REAL FPS in GoldenEye 007. I cannot get a consistent 60 REAL FPS. When I increase the overclocking, I get audio stuttering. So it becomes a balance. The best setting I've found is to underclock the game at 50MHz. The REAL FPS stay around 20-25 and the audio does not stutter.

Now I'll try some other video/audio plugins to see if I can get better performance knowing that Dynamic Recompilation works.

I wonder if this setting issue pertains to the dual core CPU?
Edited 3 times. Last edit by throwingmuse
1 year ago - Wednesday, February 09, 2011
Type: Administrator
Posts: 1336
Joined: 7/19/2009
throwingmuse wrote:

Found the issue!

I am finally able to get dynamic recompilation to work correctly IF I set the compiler to "No Check" for Self-Modifying Code rather than the default "Protected Memory."

After this change, I get consistent 30-35 REAL FPS in GoldenEye 007. I cannot get a consistent 60 REAL FPS. When I increase the overclocking, I get audio stuttering. So it becomes a balance. The best setting I've found is to underclock the game at 50MHz. The REAL FPS stay around 20-25 and the audio does not stutter.


You may want to try playing around with other settings then to see if you can find something else affecting your performance. That "check memory" shouldn't be severely affecting your speed like it is. You should be able to overclock with that CPU, see if v3 when released is any better for you.
1 year ago - Thursday, February 10, 2011
Type: Member
Posts: 41
Joined: 2/7/2011
Looks like your build works with Rice plugin only.
Using Rice plugin, I can overclock(x4) and I get REALFPS 60, good sound and CPU < max.
But your build is unable to work with Glide64 plugin although it doesn't matter to me because I can still get fullspeed.

Looks to me, there is problem with your overclocking because it should not work with Rice only unless you are telling me Glide64 plugin has problem or my system has problem.
1 year ago - Thursday, February 10, 2011
Type: Member
Posts: 41
Joined: 2/7/2011
Ok, I think I know how your overclocking works. I can get the Glide64 plugin to use your overclock(x4) for 60 REALFPS with good sound and CPU < max.
But your build has a nasty bug.
If I select the "audio setting" during gameplay without changing the actual setting, the REALFPS get stuck at the last known value which will be < 60 and gameplay sound will be affected when resume.
Select again and the game sound terrible until it hangs.
I do not know what you change to the audio code but it certainly works in original 1964.
So, there is no problem with my system and your machine is not 1000% faster than mine :)

Glide64 uses 2-5% more CPU than Rice on my machine.
You should try the latest Glide64 WIP.

I have said many times before I test with Glide64 and not Rice.
Anyway, Rice plugin(beta10) has too many bugs compared to Glide64.
Didn't you notice the cloud did not move at all.
Edited 4 times. Last edit by Pokefan999
1 year ago - Thursday, February 10, 2011
Type: Administrator
Posts: 1336
Joined: 7/19/2009
Pokefan999 wrote:

Looks like your build works with Rice plugin only.
Using Rice plugin, I can overclock(x4) and I get REALFPS 60, good sound and CPU < max.
But your build is unable to work with Glide64 plugin although it doesn't matter to me because I can still get fullspeed.


So why didn't you try Rice before? I said numerous times my performance with the plugins in the ZIP so you should have at least tried it to see if you could get higher performance than with Glide.

Pokefan999 wrote:

Looks to me, there is problem with your overclocking because it should not work with Rice only unless you are telling me Glide64 plugin has problem or my system has problem.


I have no problem with Glide on my computer and UltraFast (Napalm WX Release 1)



Glide appears to use a little more CPU, maybe 10-20% more of my I7 compared to Rice.
1 year ago - Thursday, February 10, 2011
Type: Administrator
Posts: 1336
Joined: 7/19/2009
Pokefan999 wrote:

Ok, I think I know how your overclocking works. I can get the Glide64 plugin to use your overclock(x4) for 60 REALFPS with good sound and CPU < max.


How? What was the magic setting? I told you all along it was likely a configuration issue didn't I? ;)

Pokefan999 wrote:

Anyway, Rice plugin(beta10) has too many bugs compared to Glide64.
Didn't you notice the cloud did not move at all.


It's not significantly different but yes I did notice that when using Glide. The truck also doesn't disappear when the camera moves in during the intro scene.

Problem with Glide is it needs a wrapper and almost brought down my system when I changed the video settings during a game (luckily ATI display driver rebooted itself). It's a shame the best video plugin is for some ancient Voodoo graphics card.
1 year ago - Thursday, February 10, 2011
Type: Member
Posts: 41
Joined: 2/7/2011
RetroRalph wrote:

Problem with Glide is it needs a wrapper and almost brought down my system when I changed the video settings during a game (luckily ATI display driver rebooted itself). It's a shame the best video plugin is for some ancient Voodoo graphics card.


I think the problem is your machine. I use a nVidia card not Voodo card, with wrapper and it works very well and fast(I don't have your problem). Another possibility is your Glide64 is not configured properly.
1 year ago - Thursday, February 10, 2011
Type: Member
Posts: 41
Joined: 2/7/2011
RetroRalph wrote:

I told you all along it was likely a configuration issue didn't I? ;)


If I remember correctly, you keep telling me that it is my system problem and you even prove your point by saying your machine is 1000% faster than mine.
Also, I told you you cannot compare your machine to mine based on the info supplied because it is not for benchmarking against your machine.
Also, I highlighted the needs to compare with the same settings(gave examples too) etc which you do not know which means your comparison is not "apple to apple".
But you didn't believe me.
It is working now because I test with different settings until it works which I have no source to check or settings requirement stated for your build. ;)
Edited 2 times. Last edit by Pokefan999
1 year ago - Thursday, February 10, 2011
Type: Administrator
Posts: 1336
Joined: 7/19/2009
Pokefan999 wrote:

If I remember correctly, you keep telling me that it is my system problem and you even prove your point by saying your machine is 1000% faster than mine.


You are mistaken.

RetroRalph wrote:

Except there is something wrong with your system. A system that is only 25% faster than yours (maybe 50% once you factor in bus/memory) beats your system by over 1000% . If you think that this correlation is normal and there is no configuration error here you are mistaken. So when you say "real FPS drops when I overclock" it's irrelevant. You have a problem with your system or configuration.


See the bolded parts? I basically said it is unbelievable your relatively good system can't handle UltraFast and pointed out numerous times it was either a configuration error or something else with your system. When I was saying my AMD 1.4GHz gets more speed than you it's showing you there is something wrong with what you're doing. I wasn't "bragging" that I have a faster computer that is 11 years old.

Instead of wasting my and your time with pointless posts talking about how crap UltraFast is compared to your build, how you don't need 60FPS and even if you did your "build" gets 60FPS, you should have tested different settings with your custom plugins. Which is exactly what I've been telling you to do for a few days now.
1 year ago - Thursday, February 10, 2011
Type: Member
Posts: 41
Joined: 2/7/2011
RetroRalph wrote:

Except there is something wrong with your system. A system that is only 25% faster than yours (maybe 50% once you factor in bus/memory) beats your system by over 1000% . If you think that this correlation is normal and there is no configuration error here you are mistaken. So when you say "real FPS drops when I overclock" it's irrelevant. You have a problem with your system or configuration.

See the bolded parts? I basically said it is unbelievable your relatively good system can't handle UltraFast and pointed out numerous times it was either a configuration error or something else with your system. When I was saying my AMD 1.4GHz gets more speed than you it's showing you there is something wrong with what you're doing. I wasn't "bragging" that I have a faster computer that is 11 years old.

Instead of wasting my and your time with pointless posts talking about how crap UltraFast is compared to your build, how you don't need 60FPS and even if you did your "build" gets 60FPS, you should have tested different settings with your custom plugins. Which is exactly what I've been telling you to do for a few days now.


Hahaha, I can easily extract previous comments to prove you wrong but I know you will delete my post just as you did previously when thing is not in your favour. So, I am not going to do that.

Oh, you are most likely to delete this post too. ;)
1 year ago - Thursday, February 10, 2011
Type: Administrator
Posts: 1336
Joined: 7/19/2009
Pokefan999 wrote:

Hahaha, I can easily extract previous comments to prove you wrong but I know you will delete my post just as you did previously when thing is not in your favour. So, I am not going to do that.


I delete posts that don't make sense or add unnecessary bulk to a thread. For instance after a few days of me telling you exactly what your problem is, you tell us today that the problem is a configuration error. Hurrah! If only you listened on the first day how you would have saved yourself some embarrassment. The only thing stopping me from thinking you're a troll is the fact you may not speak english as a first language and have some difficulties understanding me, but we will see I guess.

   Page 4 of 14   



RetroCopy: Making emulation easy and fun.
  |   RetroCopy © 2011