• Login
  • Register
  • Dolphin Forums
  • Home
  • FAQ
  • Download
  • Wiki
  • Code


Dolphin, the GameCube and Wii emulator - Forums › Dolphin Emulator Discussion and Support › Hardware v
« Previous 1 ... 77 78 79 80 81 ... 189 Next »

I want to build a 4k Dolphin machine.
View New Posts | View Today's Posts

Pages (7): « Previous 1 ... 3 4 5 6 7 Next »
Jump to page 
Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Thread Modes
I want to build a 4k Dolphin machine.
02-07-2015, 01:44 PM
#51
admin89 Offline
Overclocker™ ✓ᵛᵉʳᶦᶠᶦᵉᵈ
*******
Posts: 6,889
Threads: 127
Joined: Nov 2009
GTX 970 outperforms GTX 780 in most benchmark . My point still stands
Laptop: (Show Spoiler)
Clevo W230SS : 3200x1800 IPS | i7 4700MQ @ 3.6GHz (Intel XTU + Triple fan mod) | GTX 860M GDDR5 | 128GB Toshiba CFD SSD | 16GB DDR3L 1600MHz
Aspire 715 43G : 1080p 144Hz |  R5 5625U @ 4.3GHz | Nvidia RTX 3050 4GB | 500GB WD SSD  | 16GB DDR4 3200MHz 
Mini PC :: (Show Spoiler)
G3258 @ 4.6GHz | ELSA GTX 750 | Asrock Z87E ITX | 600W SFX 80+ Gold Silverstone + SG06-LITE | Corsair Vengeance 8GB 2000MHz | Scythe Kozuti + Ao Kaze | 45TB 2.5" Ex HDD (in total) , Zelda Gold Wiimote , LE Wii Classic Controller , Gold LE PS3 DualShock , BlackWidow Chroma ,
Now Playing : Xenoblade Definitive Edition on Yuzu - Switch Emu 

 
Find
Reply
02-07-2015, 02:10 PM (This post was last modified: 02-07-2015, 02:26 PM by Soupa Face.)
#52
Soupa Face Offline
Tetra's Bitch
**
Posts: 13
Threads: 0
Joined: Feb 2015
(02-07-2015, 01:04 PM)admin89 Wrote:
Quote:Windows Task Manager
GPU usage
CPU usage
Doesn't matter
https://dolphin-emu.org/docs/faq/#which-cpu-should-i-use
Keep in mind Dolphin is a dual core app . No matter how fast your CPU is , it will never be enough
If you're not ready for CPU overclocking then just underclock your CPU to 3.2GHz via thottlestop and benchmark Dolphin again .
Dolphin is completely different to PC games . If you really have GPU bottleneck in PC games , CPU overclocking or underclocking will not affect the overall FPS
CPU overclocking will increase your FPS in Dolphin
Btw GTX 260 was tested with the most graphic demanding SMG and it can handle 4xIR . GTX 750 Ti is superior to GTX 260...
If I want to play any Gamecube / Wii game through Dolphin at 4K I'm now certain that my current 750 ti can almost do just that having tested my whole library recently at 6x IR.

However, to play on this 4K TV with an HDMI 2.0 input, I'll need a minimum of a gtx 970 (or a gtx 960 Ti if one of those ever comes out). The video in the spoiler made that clear to me:
Spoiler: (Show Spoiler)
https://www.youtube.com/watch?v=_KlvHabHxZk

So yes, I am aware of what you've pointed out. In fact under normal circumstances there's not a single game I can't play (of those I've tested so far) @ 5x IR except when I'm trying to play a game with the frame-rate limited to 120. As I mentioned in my post at the top of the previous page:
Quote:[...]The way I see it, more and more people are going to be coming out with these 60fps patches which *surprise surprise* really taxes both processors.[...]
I see this Super Mario Sunshine 60fps thing as getting better in time with fewer and fewer issues. I can also see more and more people trying to make 60fps patches for any and all games that normally run at 30fps as Dolphin's performance on both the CPU and GPU increases.

This relates to 4K in that my ultimate goal is to play Super Mario Sunshine @ 60fps with the patch at at least x5 IR, preferably higher. So the test I've done in the last post seems to show that it's my GPU that's the problem, not the CPU except maaaaaaaybe in the case of Super Mario Sunshine which doesn't appear to be stressing out either processor at 4x IR according to GPU-z and Core Temp (and Windows Task Manager / Resource Manager) but since the CPU usage is a bit higher I kinda-sorta think it may be the culprit somehow.

I'd honestly just like to know how my CPU is throttling me at 60fps if both Core Temp and Windows Task Manager show the CPU usage peaking at 70% for both cores being used?

From the few PC games I've tested this with on my emulation pc , even reaching around 85% usage in a couple of the 4 cores being used (again referring to PC games that use 4 cores and not Dolphin) hasn't brought the frame-rate below 60fps, it's always been my GPU (formerly an Radeon HD 7770). So hopefully someone now sees my confusion as to how mid-high CPU usage on both CPU cores being used in Dolphin instead the expected high (99%-85%) usage could contribute to dropped frames??

EDIT: Also why would it be my CPU that's the issue if I'm loosing frames from just changing the IR up and down? I thought that was purely a GPU thing.
I'll gladly play pretty much any video game -+- Dolphin version 4.0-r6251
~=~---------------------------------------------------------------------------------~=~
Windows 7 x64
Intel i5-4670k (stock turbo @ 3.8ghz)
Nvidia gtx 750 Ti Superclocked
G.Skill RipJaws 8GB 1600mhz
Find
Reply
02-07-2015, 03:53 PM
#53
BONKERS Offline
Member
***
Posts: 82
Threads: 0
Joined: Jul 2012
(02-07-2015, 01:36 PM)admin89 Wrote:
(02-07-2015, 05:25 AM)shreduhsoreus Wrote:
(02-06-2015, 03:18 PM)admin89 Wrote: When Nvidia announce GTX 970 , it's faster than GTX 780 . GTX 960 failed to deliver higher performance than GTX 770 (Bah , GTX 770 totally crush GTX 960)

That's kind of irrelevant considering that the 770 is 1.5-2 times the price of the 960. If it were designed to outperform the 770 it wouldn't be $200.
GTX 780 cost twice as much as GTX 970 !!!

The GTX 780 is also effectively discontinued from manufacturing. This is why the price is high. Some are still in the sub 400$ range. Same thing with the 780Ti, some are still sub 500$ most are priced high simply due to discontinuation.
Find
Reply
02-07-2015, 04:59 PM
#54
admin89 Offline
Overclocker™ ✓ᵛᵉʳᶦᶠᶦᵉᵈ
*******
Posts: 6,889
Threads: 127
Joined: Nov 2009
It's actually getting cheaper than before
Release Price : $649 for non-TI version , $699 for Ti version , $999 for GTX Titan
http://en.wikipedia.org/wiki/GeForce_700_series
Do some searching before posting please
Laptop: (Show Spoiler)
Clevo W230SS : 3200x1800 IPS | i7 4700MQ @ 3.6GHz (Intel XTU + Triple fan mod) | GTX 860M GDDR5 | 128GB Toshiba CFD SSD | 16GB DDR3L 1600MHz
Aspire 715 43G : 1080p 144Hz |  R5 5625U @ 4.3GHz | Nvidia RTX 3050 4GB | 500GB WD SSD  | 16GB DDR4 3200MHz 
Mini PC :: (Show Spoiler)
G3258 @ 4.6GHz | ELSA GTX 750 | Asrock Z87E ITX | 600W SFX 80+ Gold Silverstone + SG06-LITE | Corsair Vengeance 8GB 2000MHz | Scythe Kozuti + Ao Kaze | 45TB 2.5" Ex HDD (in total) , Zelda Gold Wiimote , LE Wii Classic Controller , Gold LE PS3 DualShock , BlackWidow Chroma ,
Now Playing : Xenoblade Definitive Edition on Yuzu - Switch Emu 

 
Find
Reply
02-07-2015, 09:19 PM (This post was last modified: 02-07-2015, 09:23 PM by Garteal.)
#55
Garteal Offline
「Lab Mem. 004」
********
Global Moderators
Posts: 2,095
Threads: 24
Joined: Aug 2011
(02-07-2015, 11:47 AM)Soupa Face Wrote: I'm aware that increasing the internal resolution of any pc game uses a little more of the CPU but from what I understand we aren't talking about the game using 20% more of the CPU going from 1080p to 1440p or anything crazy like that.... right?

This is not to disregard what you said, in fact, this clearly shows that it's not the GPU that's bottle-necking me after all, yet I still would like to know what the heck is throttling my CPU?? ...because it doesn't seem like it's something Core Temp can detect.
Sunshine is forced to run with EFB Copies to RAM in the game ini which is (a lot) slower than EFB to Texture due to stalling and waiting for the EFB copy.

Set your IR to 4x and open up the graphics window ingame and make sure EFB Copies is set to Texture; see if you can retain that 60FPS or not. Don't forget to add Dolphin to the NVIDIA CP and set the power management "Prefer maximum performance".
Also you might want to use OpenGL instead of Direct3D.
Find
Reply
02-09-2015, 11:36 AM (This post was last modified: 02-09-2015, 01:54 PM by Soupa Face.)
#56
Soupa Face Offline
Tetra's Bitch
**
Posts: 13
Threads: 0
Joined: Feb 2015
(02-07-2015, 09:19 PM)Garteal Wrote:
(02-07-2015, 11:47 AM)Soupa Face Wrote: I'm aware that increasing the internal resolution of any pc game uses a little more of the CPU but from what I understand we aren't talking about the game using 20% more of the CPU going from 1080p to 1440p or anything crazy like that.... right?

This is not to disregard what you said, in fact, this clearly shows that it's not the GPU that's bottle-necking me after all, yet I still would like to know what the heck is throttling my CPU?? ...because it doesn't seem like it's something Core Temp can detect.
Sunshine is forced to run with EFB Copies to RAM in the game ini which is (a lot) slower than EFB to Texture due to stalling and waiting for the EFB copy.

Set your IR to 4x and open up the graphics window ingame and make sure EFB Copies is set to Texture; see if you can retain that 60FPS or not. Don't forget to add Dolphin to the NVIDIA CP and set the power management "Prefer maximum performance".
Also you might want to use OpenGL instead of Direct3D.

Those 2 screenshots I took and posted were using EFB copies to Texture which I thought I didn't have a problem with back when I first started my original Sunshine save (on Dolphin v4.0.2) but clearly I imagined that because going back and trying to play a level involving that goop didn't work properly just as it states in the wiki.


While I never did test setting EFB Copies to Ram using the original 60fps patch, I did try it with the "slightly slower" 60fps patch that allows level portals to function as they should among other things, both of which can be found here. With the new "slightly slower" patch all my processor sensor readings finally make sense where as it looks like my GPU is clearly the thing slowing me down.

In the Spoiler below are photos of slowdown at 5x IR, no slowdown at 4x IR and crazy bad slowdown at 5x IR with OpenGL as the backend:
Spoiler: (Show Spoiler)
Slowdown at x5 IR (Direct 3D Backend):
[Image: Zz5N6Y0.png]

No slowdown what-so-ever at 4x IR even when really trying to use up my GPU's resources (Direct 3D Backend):
[Image: sqlCZuq.png]

5x IR using the other graphics api with some less-than-desirable results... (OpenGL Backend):
[Image: WCJhC0S.png]

So basically my Nvidia gtx 750 ti is some-what capable of handling 5x IR even with the 60fps patch going on but runs smooth as butter with 4x IR which with my current 1080p setup is fine but later on when I do get a fancy 4K TV with HDMI 2.0 input, this GPU will not cut it with the 60fps patch. Also thanks to this test I remembered why I've always used Direct3D as the backend.

EDIT: Again, I just want to confirm that the main reason on a technical level the gtx 960 can't achieve 6x IR in everything (with a consistent frame-rate) is due to its low memory bus, right?

The only reason I still think the gtx 960 may be capable of 6x IR is due to other forums / sites claiming it to not just be a class above the gtx 750 ti but that it's basically the equivalent of 2 gtx 750 ti's in terms of performance. With my gtx 750 ti doing so well already at 5x (not with several Wii games as it turns out) how can the gtx 960 not put out 6x if not for some technical limitation?
I'll gladly play pretty much any video game -+- Dolphin version 4.0-r6251
~=~---------------------------------------------------------------------------------~=~
Windows 7 x64
Intel i5-4670k (stock turbo @ 3.8ghz)
Nvidia gtx 750 Ti Superclocked
G.Skill RipJaws 8GB 1600mhz
Find
Reply
02-09-2015, 04:49 PM
#57
admin89 Offline
Overclocker™ ✓ᵛᵉʳᶦᶠᶦᵉᵈ
*******
Posts: 6,889
Threads: 127
Joined: Nov 2009
GTX 960 may perform better than R9 285 at 1080p
but at 4k benchmark ? GTX 960 bite the dust . All i can think of is low memory bandwidth
Laptop: (Show Spoiler)
Clevo W230SS : 3200x1800 IPS | i7 4700MQ @ 3.6GHz (Intel XTU + Triple fan mod) | GTX 860M GDDR5 | 128GB Toshiba CFD SSD | 16GB DDR3L 1600MHz
Aspire 715 43G : 1080p 144Hz |  R5 5625U @ 4.3GHz | Nvidia RTX 3050 4GB | 500GB WD SSD  | 16GB DDR4 3200MHz 
Mini PC :: (Show Spoiler)
G3258 @ 4.6GHz | ELSA GTX 750 | Asrock Z87E ITX | 600W SFX 80+ Gold Silverstone + SG06-LITE | Corsair Vengeance 8GB 2000MHz | Scythe Kozuti + Ao Kaze | 45TB 2.5" Ex HDD (in total) , Zelda Gold Wiimote , LE Wii Classic Controller , Gold LE PS3 DualShock , BlackWidow Chroma ,
Now Playing : Xenoblade Definitive Edition on Yuzu - Switch Emu 

 
Find
Reply
02-23-2015, 10:42 AM
#58
NaturalViolence Offline
It's not that I hate people, I just hate stupid people
*******
Posts: 9,013
Threads: 24
Joined: Oct 2009
Incoming wall of text.  I've had this thread on my backlog for quite some time since some of the claims involve me personally and most of soupa face's questions have not been answered yet.

DatKid20 Wrote:I guess this is what people call clearly better.

You're cherry picking benchmarks again and I'm sure that you know it at this point.  My previous post explains in great detail why the 980 is the better card and backs it up with broad data.

DatKid20 Wrote:At 4K The GTX 980 isn't clearly better at all.

It clearly is.  In 1/3 of the 4K tests they tie, in 1/3 the 980 wins very slightly (1-3 fps), and in 1/3 it wins by a greater margin (3-9 fps).  You don't think that after looking at that data the 980 is clearly the better card?  I don't know how the data can be any clearer on this.

DatKid20 Wrote:(Intel actually has the best OpenCL support for whatever reason.)

Intels recent IGP architectures are very GPGPU focused despite them not making a big deal about it in marketing material (unlike AMD).  They use a much larger SP with much more complex pipelines that is more flexible with non-graphics workloads.

Sweezy Wrote:If I did buy one of their CPU's, If I used a different motherboard, would I get a different overclock?

Not really.  Despite common belief there is almost no difference in overclocking potential from using different motherboards for a modern Intel cpu.  Haswell has the VRM and clock generator integrated now so for the most part nothing involved in overclocking even touches the motherboard.

@People talking about silicon lottery

Quite frankly I don't think the extra cost is justified since the additional performance gains are so minimal.

Soupa Face Wrote:Apparently a gtx 960 Ti is in the works maybe, perhaps, we'll see

There is no real evidence to support this at the moment.

admin89 Wrote:Not gonna happen . The maximum IR it can handle is 4xIR
Each GPC is accompanied by a 64-bit memory controller bus, the reduction would mean the GeForce GTX 960 will only have a 128-bit wide graphics memory bus.
admin89 Wrote:Would you really buy GTX 960 if you were him ? GTX 960 can't handle 6xIR without a doubt

You have no data to back this up with as no one has tested it with dolphin yet.

admin89 Wrote:128 bit for $200 price range ? You got to be kidding Nvidia . R9 280 has 384 bit wide graphics memory bus
Not to mention the memory bandwidth , 112GB/sec is a joke . Just take a look at GTX 760 or 660's specs (almost twice as much)
Base on linus's review , GTX 960 is a let down .

And yet the 960 outperforms them in almost every test I've seen (there are a few 4K tests where the 760 ties or wins slightly, and all of those were super recent games where they haven't had time to optimize the drivers for them yet).  Which again reinforces that notion that memory bandwidth (or any spec for that matter) doesn't matter!  Performance matters, the big picture matters, not the individual components that make up that picture.

Nvidia has made a number of complex caching and compression improvements that have dramatically reduced the need for memory bandwidth.  This allows them to dedicate more transistors and power to the SPs which improves the architectures performance and efficiency.  This is a good thing.  No engineer wants to use super wide memory buses, it creates all sorts of problems.  They do it out of necessity.  That's why so many companies are working so hard to try and reduce the need for wide memory buses.  

admin89 Wrote:When Nvidia announce GTX 970 , it's faster than GTX 780 . GTX 960 failed to deliver higher performance than GTX 770 (Bah , GTX 770 totally crush GTX 960)

The 960 delivered on its expectation of beating the 760, just like all its predecessors before it.  The 970 beating the 780 was a fortunate surprise, not an expectation.  Without a die shrink there is not much they can do to improve performance per dollar in this price range.  But they did manage to significantly reduce the size and power consumption while adding new features and higher performance compared to the 760.  I would say that they did a good job with the hand that they were delt.  The big performance gains that we're all expecting won't come from either nvidia or AMD until the 20nm TSMC shrink happens.  Which won't be until late 2015 at the absolute earliest.

admin89 Wrote:What is so wrong about it ?

The data is 5 years outdated (dolphins video backends have changed A LOT since then) and was limited to begin with (which forced me to make assumptions and predictions based on what I had).

admin89 Wrote:Natural Violence is no god . No matter what he said , 112GB/sec and 128 bit bus are epic fail

I may not be a god but you still don't have the data you need to back up your claims.  And it's counterintuitive to assume that it's going to be slower than a 660 when it beats it in everything that has been tested.

DatKid20 Wrote:He was the one who started the Memory Bandwidth rumor.

It's wasn't a rumor.  It's based on real data (which is now outdated) that is still availible in the thread.  It was an accurate conclusion to draw from the data at the time that it was written.

admin89 Wrote:Btw GTX 260 was tested with the most graphic demanding SMG and it can handle 4xIR

SMG is not the most graphically demanding game.

admin89 Wrote:GTX 750 Ti is superior to GTX 260...

And yet the 260 has more memory bandwidth.

Soupa Face Wrote:I'm aware that increasing the internal resolution of any pc game uses a little more of the CPU but from what I understand we aren't talking about the game using 20% more of the CPU going from 1080p to 1440p or anything crazy like that.... right?

No.  CPU load will not substantially increase.

Soupa Face Wrote:I've also checked my CPU's status with Windows Task Manager and that too showed that there was next to no difference in CPU usage from x2.5 IR to x4 IR, but I've been told Core Temp and GPU-z were pretty darn accurate anyway.

CPU load and task manager have nothing to do with cpu bottlenecks.  If for example you were to change the games settings to increase the cpu requirements and reduce performance thread 1 would still be at 100% load in either scenario.  And as admin pointed out task manager is no longer an accurate way to observe thread/core load.

Soupa Face Wrote:This is not to disregard what you said, in fact, this clearly shows that it's not the GPU that's bottle-necking me after all, yet I still would like to know what the heck is throttling my CPU?? ...because it doesn't seem like it's something Core Temp can detect.

Nothing is throttling your cpu.

Soupa Face Wrote:UPDATE: I've just checked to see if my CPU would bottleneck me in 3 other games if I forced the frame-rate to 120: Luigi's Mansion, Pokemon Collsuseum, and Wind Waker, using the same programs I used above, and sure enough, in all 3 cases it turned out to be VERY clear that my GPU was throttling things this time around; at least at x4 IR. Although in all cases just setting the IR to x3 (even with FXAA) turned out to bring my GPU usage from 95-88% to 85-68%ish (I'm honestly only remembering Wind Wakers usage but the others were pretty close to this).

Don't use gpu load to gauge gpu bottlenecks.  It's notoriously innaccurate due to long idle periods.  Use framerates to check for bottlenecks instead.  It does however sound like your GPU is a bottleneck at 4x IR in some of these games, although they're still getting above 60 fps so I don't think it matters too much.

Soupa Face Wrote:If the rumors of a gtx 960 Ti are to be believed I'm still going to hold onto my gtx 750 Ti due to me needing a card that supports HDMI 2.0 for 4k 60hz on the HDTV I've got my eye on

The 960 supports HDMI 2.0

Soupa Face Wrote:However, to play on this 4K TV with an HDMI 2.0 input, I'll need a minimum of a gtx 970 (or a gtx 960 Ti if one of those ever comes out). The video in the spoiler made that clear to me:

That video was likely made before the 960 existed.  The 960 supports HDMI 2.0

Soupa Face Wrote:This relates to 4K in that my ultimate goal is to play Super Mario Sunshine @ 60fps with the patch at at least x5 IR, preferably higher. So the test I've done in the last post seems to show that it's my GPU that's the problem, not the CPU except maaaaaaaybe in the case of Super Mario Sunshine which doesn't appear to be stressing out either processor at 4x IR according to GPU-z and Core Temp (and Windows Task Manager / Resource Manager) but since the CPU usage is a bit higher I kinda-sorta think it may be the culprit somehow.

Again your testing method is flawed.  See above and below.

Soupa Face Wrote:I'd honestly just like to know how my CPU is throttling me at 60fps if both Core Temp and Windows Task Manager show the CPU usage peaking at 70% for both cores being used?

So far what you've shown us hasn't shown that.

Soupa Face Wrote:From the few PC games I've tested this with on my emulation pc , even reaching around 85% usage in a couple of the 4 cores being used (again referring to PC games that use 4 cores and not Dolphin) hasn't brought the frame-rate below 60fps, it's always been my GPU (formerly an Radeon HD 7770). So hopefully someone now sees my confusion as to how mid-high CPU usage on both CPU cores being used in Dolphin instead the expected high (99%-85%) usage could contribute to dropped frames??

Again you're reading it wrong.

Dolphin has two main threads.  The OS switches which core they're running on every few microseconds to keep the heat more evenly distributed across the cpu.  So 4 cores are being used but only 2 at a time.  Task manager reports average use of each core over a 1-4 second interval depending on the settings.  In your example I see about 70, 50, 30, 30 load respectively.  Sum this together and you get a real load of around 100, 80, 0, 0.  There is your bottleneck.  Make sense?

Soupa Face Wrote:EDIT: Also why would it be my CPU that's the issue if I'm loosing frames from just changing the IR up and down? I thought that was purely a GPU thing.

It's not.  It's a gpu bottleneck.

admin89 Wrote:It's actually getting cheaper than before
Release Price : $649 for non-TI version , $699 for Ti version , $999 for GTX Titan
http://en.wikipedia.org/wiki/GeForce_700_series
Do some searching before posting please

That doesn't disprove his assertion.  Or even relate to it.  He's right in that the card is effectively discontinued and that the price hasn't reflected its performance because of that.  People shouldn't be buying these anymore.

Soupa Face Wrote:EDIT: Again, I just want to confirm that the main reason on a technical level the gtx 960 can't achieve 6x IR in everything (with a consistent frame-rate) is due to its low memory bus, right?

We don't know anything about this yet and anyone who says otherwise is lying unless they can provide data to back it up or have one themselves.  Nobody has tested the 960 with dolphin yet so we just don't know one way or the other.

Soupa Face Wrote:The only reason I still think the gtx 960 may be capable of 6x IR is due to other forums / sites claiming it to not just be a class above the gtx 750 ti but that it's basically the equivalent of 2 gtx 750 ti's in terms of performance. With my gtx 750 ti doing so well already at 5x (not with several Wii games as it turns out) how can the gtx 960 not put out 6x if not for some technical limitation?
Soupa Face Wrote:...that said, if what I've read is true, therefore the gtx 960 really is the equivalent of 2 gtx 750 ti's; I should just stick with that right?

In PC games it is but as for dolphin we don't know yet.

Soupa Face Wrote:Is the gtx 960 worth buying for x6 - x5 internal resolution on everything*, 60fps patch or otherwise? Or should I just wait a while for a gtx 960 Ti to come about?

There is no evidence tht the 960 TI is going to be released so I wouldn't factor that into your considerations at all.  Either get the 960 or the 970.  The 960 will probably work well for what you're doing but it might not, it's more of a risk than a 970.

Soupa Face Wrote:My one fear with the gtx 960 is that no graphics card manufacturer makes one with 3gb of vram or more which may be useful when it comes to Dolphin, but it also may not which is of course why waiting for a more advanced 960 with (I can only imagine) more than 2gb of vram would be a smart idea.

Dolphin will not use over 2GB+ of vram even at 4K.  I doubt it will even get close to that.

The rest of your post I have already answered in previous sections.

Good lord that was long.
"Normally if given a choice between doing something and nothing, I’d choose to do nothing. But I would do something if it helps someone else do nothing. I’d work all night if it meant nothing got done."  
-Ron Swanson

"I shall be a good politician, even if it kills me. Or if it kills anyone else for that matter. "
-Mark Antony
Website Find
Reply
02-25-2015, 11:12 PM (This post was last modified: 02-25-2015, 11:16 PM by Link_to_the_past.)
#59
Link_to_the_past Offline
Link on steroids really
*******
Posts: 1,767
Threads: 17
Joined: Feb 2010
I just tried my 960 with dolphin, used Super Mario Galaxy 2 for testing, latest dev with d3d and efb to texture. Up to 4x scale and therefore QHD with efb to texture i didn't get a single fps drop, specifically i got 172 fps at the mario spaceship from 1x up to 4x. Framerate takes a hit from 5x and up, specifically 108 fps at 5x and 90 fps at 6x (4K resolution). So the game is playable even at 4K if you pair such a card with a decent processor (i have a 2500k at 4,1 ghz), and the card can handle up to 4x without a sweat (you won't get any gpu bottleneck with this card up to QHD res with dolphin, only cpu).
Find
Reply
02-26-2015, 07:31 AM
#60
NaturalViolence Offline
It's not that I hate people, I just hate stupid people
*******
Posts: 9,013
Threads: 24
Joined: Oct 2009
Please redo the test with efb to ram and LLE audio. Which build was this on?
"Normally if given a choice between doing something and nothing, I’d choose to do nothing. But I would do something if it helps someone else do nothing. I’d work all night if it meant nothing got done."  
-Ron Swanson

"I shall be a good politician, even if it kills me. Or if it kills anyone else for that matter. "
-Mark Antony
Website Find
Reply
« Next Oldest | Next Newest »
Pages (7): « Previous 1 ... 3 4 5 6 7 Next »
Jump to page 


  • View a Printable Version
  • Subscribe to this thread
Forum Jump:


Users browsing this thread: 1 Guest(s)



Powered By MyBB | Theme by Fragma

Linear Mode
Threaded Mode