Poll: Is EE overclocking a useful feature?
You do not have permission to vote in this poll.
Yes
91.49%
43 91.49%
No
8.51%
4 8.51%
Total 47 vote(s) 100%
* You voted for this item. [Show Results]

Overclocking the EE (discussion) (testers wanted)
#51
Can you send me a build?, I believe I have an (obscure?) game that performs badly in PCSX2.
Reply

Sponsored links

#52
Done and done.
[Image: XTe1j6J.png]
Gaming Rig: Intel i7 6700k @ 4.8Ghz | GTX 1070 TI | 32GB RAM | 960GB(480GB+480GB RAID0) SSD | 2x 1TB HDD
Reply
#53
Just so you guys know, the current "50% overclock" setting is pretty wonky. I would limit testing to the 33% one for now. According to rama, the way I set up the 50% setting basically uncapped it completely which is why it has such a devastating performance hit.

As I said on Github, this comes from them actually understanding how things work versus me looking at the code, making a hypothesis, and changing stuff. Tongue2
[Image: XTe1j6J.png]
Gaming Rig: Intel i7 6700k @ 4.8Ghz | GTX 1070 TI | 32GB RAM | 960GB(480GB+480GB RAID0) SSD | 2x 1TB HDD
Reply
#54
Oh, I didn't realize this was going on here at the same time Tongue2
Reply
#55
(01-31-2015, 08:25 PM)rama Wrote: Oh, I didn't realize this was going on here at the same time Tongue2

Well, I've been repeating the important stuff. Since I sent out builds I want people up to date on what's going on with it.
[Image: XTe1j6J.png]
Gaming Rig: Intel i7 6700k @ 4.8Ghz | GTX 1070 TI | 32GB RAM | 960GB(480GB+480GB RAID0) SSD | 2x 1TB HDD
Reply
#56
Used all default emulation settings + MTVU
GS settings Direct3D11 1920x1080 + 8x MSAA

I'm pretty sure I can confirm Airblade is broke by the EE overclock, just black screen, nothing appears but maintains a high VFPS.

(I only used the 50% overclock)

edit:
Quote:I would limit testing to the 33% one for now.
Well damn LoL

Dark Cloud 1 varied from anything from a 20VFPS loss to a 50VFPS loss compared to without the overclock, using the intro rendered cutscene for accurate repeatability.

Internally wise(RFPS), I feel like with this game when walking around the town and in the dungeon(wise owl forest) that it does feel smoother somewhat but will need to wait on aforementioned tool to see if I'm just feeling a placebo effect Tongue2
<REDACTED>
Reply
#57
(01-31-2015, 08:31 PM)Fezzer Wrote: Dark Cloud 1 varied from anything from a 20fps loss to a 50fps loss compared to without the overclock, using the intro rendered cutscene for accurate repeatability.

Might wanna clarify when you say "loss" that you mean in GSdx title, not in game FPS.

e.g. VFPS vs RFPS.

VFPS is what we have in the title, RFPS is what the game is internally that we can't see right now. (At least, that's how I understand it. It's either that or vice versa)
[Image: XTe1j6J.png]
Gaming Rig: Intel i7 6700k @ 4.8Ghz | GTX 1070 TI | 32GB RAM | 960GB(480GB+480GB RAID0) SSD | 2x 1TB HDD
Reply
#58
^Ok, it's corrected.
<REDACTED>
Reply
#59
(01-31-2015, 08:47 PM)Fezzer Wrote: ^Ok, it's corrected.

Thanks. Yeah that's one of the two problems I see with this feature:

1. Most users don't understand the difference between VFPS and RFPS, they don't understand that the FPS in the title bar is not in game FPS
2. Many users will likely think it's a magic optimization to make games run faster. People hear "overclock" and think "faster."

Those reasons are why I definitely agree with rama and the others who say that it shouldn't be exposed by default. Otherwise we will have youtube guides advising people with Pentium 2s to ramp the EE up all the way for "moar speed"
[Image: XTe1j6J.png]
Gaming Rig: Intel i7 6700k @ 4.8Ghz | GTX 1070 TI | 32GB RAM | 960GB(480GB+480GB RAID0) SSD | 2x 1TB HDD
Reply
#60
I was thinking maybe the description text should be changed from "in-game" fps to "internal" fps when the EE slide is increased?

I think 'in-game' might catch out more people if they're unfamiliar with emulation.
<REDACTED>
Reply




Users browsing this thread: 1 Guest(s)