Remove the restriction to use efb copies only once as partial update

[mimimi085181]

I'm not entirely sure what is happening, but this optimisation is causing an issue in Sonic Riders: Zero Gravity. Apparently the issue would also be fixed by PR#3747, but this PR should also fix similar issues.

Games that use partial updates might get slower with this, so some performance regression testing would be nice. Games like New Super Mario Bros, DK Returns and Silent Hill. Testing with high graphics settings makes sense, since this would mostly end up in more work for the GPU.

---

This change is

[JMC47]

Fixes Sonic Riders/Sonic Riders Zero Gravity on EFB Copies to Texture

[delroth]

5.0 candidate?

[JMC47]

Both games worked on EFB Copies to Texture in 4.0, so I think we should consider it. But, these games aren't totally broken without this, so, if we find any reason not to merge it then it's not an absolute make or break feature.

[mimimi085181]

I'd stil like some performance regression testing on this. But i don't know what i was thinking when i mentioned RS2 and Zelda TP, those don't use partial updates.

[JMC47]

New Super Mario Bros. Wii - Master - 257 - 260 fps
New Super Mario Bros. Wii - PR3752 - 256 FPS
Silent Hill Shattered Memories - Master - 13 FPS
Silent Hill Shattered Memories - PR3752 - 13 FPS

Note 3x IR

[mimimi085181]

Ok, i'm not getting a performance regression either, 35 vs 35 fps at 8xIR in New Super Mario Bros.

@degasus or @phire can you take a look at this? I think this pr can't do any damage, but i'd like to get a 2nd opinion.

[phire]

So the downside of this is that same partial texture might be created multiple times?

[mimimi085181]

Yes, it would do the same thing possibly several times. This could be optimised of course, but thinking about the 5.0 release, i wanted to make something that is less likely to miss something.

[phire]

Yeah, this is fine for now and doesn't show major performance impacts. Optimization can be done later.

LGTM

[phire]

Reviewed 1 of 1 files at r1.
Review status: all files reviewed at latest revision, all discussions resolved.

---

Comments from the review on Reviewable.io

[JMC47]

LGTM

[degasus]

Reviewed 1 of 1 files at r1.
Review status: all files reviewed at latest revision, 2 unresolved discussions, all commit checks successful.

---

Source/Core/VideoCommon/TextureCacheBase.cpp, line 311 [r1] (raw file):
Is there still a check to not update this texture per use? How did you profile it? Also with a very high IR?

---

Source/Core/VideoCommon/TextureCacheBase.cpp, line 319 [r1] (raw file):
unrelated, but this case is always true because of :
TexCache::iterator iter = textures_by_address.lower_bound(entry_to_update->addr);

---

Comments from the review on Reviewable.io

[mimimi085181]

Review status: all files reviewed at latest revision, 2 unresolved discussions.

---

Source/Core/VideoCommon/TextureCacheBase.cpp, line 311 [r1] (raw file):
I was thinking of different ways to optimise this, but maybe it's better to wait for the 5.0 before starting some new optimisation. I tested New Super Mario Bros at 8xIR and got 35 vs 35 fps.

---

Comments from the review on Reviewable.io

[JMC47]

All checks have passed...

I look at ye with the merge rights with accusing eyes.