Will 4.5GHz on a Haswell CPU be ennough to get fullspeed on R&C?

[karasuhebi]

I couldn't tell you for sure if what's happening is actually intentional frame-skipping by the engine or if it's frame dropping or maybe just longer frame timings. We'll know for sure once (or if) PCSX2 gets an actual FPS counter. It definitely happens to R&C though. Like I said, try VU Cycle Stealing '2' and '3' and you'll really notice it.

[dogen]

I couldn't tell you for sure if what's happening is actually frame-skipping or if it's just longer frame timings. We'll know for sure once (or if) PCSX2 gets an actual FPS counter. It definitely happens to R&C though. Like I said, try VU Cycle Stealing '2' and '3' and you'll really notice it.

You're right. I was mixed up there for a minute. Cycle rate is fine in this game, vu stealing messes it up.

[karasuhebi]

No worries man, I understand. Part of the problem is that having either of these speedhacks set too high will always have a chance to slow down a game, specially if a game is very EE-heavy. That's because they both give the EE less cycles to work with. That's why if you set EE Cyclerate to '3' while leaving VU Cycle Stealing to '0', you'll probably still get 60FPS on R&C.

[Clank]

In R&C VU cycle stealing works fine. For whatever reason EE cyclerate actually causes the "fake fps" frameskipping in those games.

Only mostly fine, there are a couple places in each game where VU stealing set at 2 caused them to lock up for me.

[karasuhebi]

Yeah I wouldn't recommend anything higher than '1' for VU Cycle Stealing in those games. Even at the very beginning of the game if I set VU Cycle Stealing to '2' I can already see some choppiness.

BTW dogen, if you set EE Cyclerate to '3' and VU Cycle Stealing to '1', you'll get choppy gameplay sometimes as well. So don't think VU Cycle Stealing set to '1' is completely devoid from slowing down the game. It all depends on how much the game is needing to use the EE at a certain point in time.

[dogen]

Yeah I wouldn't recommend anything higher than '1' for VU Cycle Stealing in those games. Even at the very beginning of the game if I set VU Cycle Stealing to '2' I can already see some choppiness.

BTW dogen, if you set EE Cyclerate to '3' and VU Cycle Stealing to '1', you'll get choppy gameplay sometimes as well. So don't think VU Cycle Stealing set to '1' is completely devoid from slowing down the game. It all depends on how much the game is needing to use the EE at a certain point in time.

Yeah, these games are pretty sensitive to that setting.

[karasuhebi]

Lowering the threads to 3 I loose 5fps max(same place without moving)

So with 3 extra rendering threads and only default speedhacks + MTVU, you get about 58FPS? (61 - 3 = 58)

Would you mind checking what you get with 2 extra rendering threads and no cycle hacks? That'll be my last test, I promise. XD I just want to be extra sure that I can get 60FPS in that game with just a bit more clock speed, which seems to be the case according to your results. This will finally answer this thread's original question

You have to be 100% sure that VU Cycle Stealing is breaking the game\making pcsx2 to crash on that game to not use it since that option as I said,it gives you huge speed boost

Check out my last post on the 1st page for an explanation as to why game-breaking and/or PCSX2-crashing are not the only reasons why using VU Cycle Stealing is not desirable.

dogen: You can check it out too if you'd like. I added a bit to it. See the part that says EDIT on the fourth paragraph.

Edit:Btw I more often see it little above 120 when I use 4 thread than when I use 5

That seems to be more or less the same effect I see on my computer. if I use more than 50% of my threads as extra rendering threads, I don't get as much of a speed up. In my case it's 3 or more, in your case is 5 or more. Interesting stuff!

[vsub]

Would you mind checking what you get with 2 extra rendering threads and no cycle hacks? That'll be my last test, I promise. XD I just want to be extra sure that I can get 60FPS in that game with just a bit more clock speed, which seems to be the case according to your results. This will finally answer this thread's original question

Not using certain hack on any game just because it breaks some games or reports false fps results(it doesn't look that way in R&C)sounds silly to me but it's your choose...making your cpu do more work,increasing the cpu temp and power usage and getting much lower fps or use a hack that will reduce all of that but increase your fps by a lot.

As soon as I can control Ratched and without moving
4 Threads - 60fps
3 Threads - 55fps
2 Threads - 51fps
1 Thread - 34fps
0 Threads - 28fps

No difference between AVX2 and SSE4 or maybe SSE4 is slower by 1-2 fps.

[karasuhebi]

Thanks for the info man, I appreciate it. Looks like with two threads I'm about 10FPS away from full speed. Let's hope the extra 600 or so megahertz helps the game run at full speed.

As far as VU Cycle Stealing goes, it definitely affects games in a negative way, and the R&C series was already confirmed to be among them in this very thread. Cycle hacks in general have a bad rep so I'd rather not use them unless I need to. And the code for VU Cycle Stealing is definitely not perfect at only skipping idle EE cycles (confirmed by pseudonym himself), so I'm avoiding it as much as I can.

[Nobbs66]

Thanks for the info man, I appreciate it. Looks like with two threads I'm about 10FPS away from full speed. Let's hope the extra 600 or so megahertz helps the game run at full speed.

As far as VU Cycle Stealing goes, it definitely affects games in a negative way, and the R&C series was already confirmed to be among them in this very thread. Cycle hacks in general have a bad rep so I'd rather not use them unless I need to. And the code for VU Cycle Stealing is definitely not perfect at only skipping idle EE cycles (confirmed by pseudonym himself), so I'm avoiding it as much as I can.

I told you what the hack did on steam yesterday. And Blyss told me based on the source code comments