Dolphin, the GameCube and Wii emulator - Forums

@ georaldc nope not even delided i just got a really sweet chip (only need 1.2v for 4.5ghz)
i peak 94*C @ 1.33v for 4.8stable though. id probably hit thermal limit at 1.35-1.38ish.

and i need roughly 1.4v to make 5ghz game stable.

i run with a hyper 212+ /w Scythe Ultrakaze fan in a Push config that blows air directly into my exhausti fan. i use antec nano diamond forumla 7. its nice paste

First:
FPS: 107 / VPS: 213 / Speed 355%





Second:
FPS: 81 / VPS: 162 / Speed 270%




Third:
FPS: 128 / VPS: 256 / Speed 426%




Specs:

i5 2500k @ 5.1Ghz 1.355v
Corsair 8GB 1600mhz (stock) 9-9-9-24-1T
GTX 570 @ 1000/2000
Windows 7 x64

Well this is just for you mr. screenshots and benchmarks.


amd turion II x2 P540 @2.4ghz
4GB of ddr3 1066
amd 512MB GDDR3 5470m

vs

core 2 duo E7500 (underclocked via Pstate to 2.4ghz)
4GB of ddr2 800
amd 1GB GDDR3 4670 (@stock settings)



Using this dolphin test
set to 3x IR, 4x MSAA, 16x AF

4670:


5470M


now for just 3x IR, 0x AA/AF

4670:

5470M:


Now hmmmm....
5470m vs 4670, there is only 2-3 fps difference in benchmarks??? and its a higher demand???? less delta at a higher demand??? blasphemy!!!!!!!!

You want to know why it is? because the demand is so far above both cards, there will be little difference (something you cant understand).
When 1 card is within ability to perform at the settings, while one isnt (at high or lower settings/resolutions), there will be a much larger gap between the 2 cards.
When both cards are able to perform at the settings, there is usually less delta between them, but it really goes on a case by case basis, as this sub category breaks into many others of like whether one card is hitting so high of an fps its giving diminishing returns and etc.
The reason a 5470m and a 4670 (or any other better graphics card) will perform identical in something like this dolphin test, is becuase emulation will always be held back by your cpu (well assuming you meet the actual requirements for the IR/settings you are using). Since both can play at 1x resolution, the cpu will always hold back the graphics. As at say 1x IR something like a 4670 could probably get like 200fps, but getting that with any cpu is almost impracticable.


When the graphics cards arnt being pushed extremely past their limits, you can get a better idea of the differences between them.
So i choose 3x IR which would be probably just enough to cause less then 100% speed with the 4670. This way the 4670 is being pushed to its limit w/o major slow down, while the 5470m is being pushed way past its means to play at 100% speed. (giving the largest delta between the 2 cards)

Even at this test, the difference were noticeable between the 4670 and 5470m, but its not a super huge margin. and this is a 4670 vs a 5470m we are talking about. the 4670 is 2x better then a 3650.... so take the difference between the 4670 vs 5470m and divide by 2 to give a some what guess at the difference between a 5470m vs 3650.

1) 24-17= 7fps difference, divide by 2, 3.5fps difference in 5470m vs 3650
2) 30-22= 8 fps difference, divide by 2, 4 fps difference between 5470m vs 3650
3) 25-17 = 8 fps difference, again 4fps difference between 5470m vs 3650.

This will be probably about the largest delta between a 5470m and a 3650 in dolphin. like 3-4fps difference. wouldnt you say that the 5470m would perform "PROBABLY ABOUT" the same?
and that the 5470m can easily play at 1x IR w/o any slow downs, as to not hold back any cpu from its potential?
(which that is the point you were disputing with me????)

All my statement was, was an educated guess based on performance of cards i have experience with. and it turns out im not too far off from reality (like you are).
This is the difference between real testing vs JUST basing your statistics on NOTHING but other peoples benchmarks.
You earlier asked whats the difference? well the fact that you are physically present for the test and can 100% see everything going on. Which ofcourse you cant know for 100% certainty my testing (just like any other benchmarks you find online), but thats why you do your own testing, and base your knowledge on something more tangible. (liek ive seen the 4670 and 5470m in real play, and thats how i base this off of, rather then using just benchmarks which time and time again even when perfectly recreated, never turn out the exact same way)

but anyway, im moving on now (i think i know why squall doesnt really like dolphin community much)
kthxbye

(04-05-2013, 07:20 AM)lor Wrote: [ -> ]First:
FPS: 107 / VPS: 213 / Speed 355%





Second:
FPS: 81 / VPS: 162 / Speed 270%




Third:
FPS: 128 / VPS: 256 / Speed 426%




Specs:

i5 2500k @ 5.1Ghz 1.355v
Corsair 8GB 1600mhz (stock) 9-9-9-24-1T
GTX 570 @ 1000/2000
Windows 7 x64

nice scores for 5.1ghz. its very close to my 5ghz scores with ivy.

@jbone1337

why are you argueing about gpus when we are in a cpu benchmark where Gpu Bottleneck isnt Present... at all. . you wont really see the difference in power in gpus till you crank up the Resolution. which in this benchmark... we dont.


Also Guys i just tried cranking my thing to stability for benchmarking at 5.1ghz. it wont budge. tried 1.49v then called it a day. lol. dont wanna prematurely kill off my chip maybe i can tweak another setting.

(04-05-2013, 12:04 PM)Venomx1 Wrote: [ -> ]nice scores for 5.1ghz. its very close to my 5ghz scores with ivy.

@jbone1337

why are you argueing about gpus when we are in a cpu benchmark where Gpu Bottleneck isnt Present... at all. . you wont really see the difference in power in gpus till you crank up the Resolution. which in this benchmark... we dont.


Also Guys i just tried cranking my thing to stability for benchmarking at 5.1ghz. it wont budge. tried 1.49v then called it a day. lol. dont wanna prematurely kill off my chip maybe i can tweak another setting.

idk why i am lol
whole thing started over the fact that someone stated the 5470m was under the min requirements for 1x resolution, and was throwing off the cpu benchmark of the 2.4ghz turion.
And then i had just replied that it def should be above the min requirements, and then just stated the 5470m is "probably about the same about of performance as a 3650" (since in the dolphin faq for 1x resolution, its recommended to have a 3650 or better).

Even though i know you can do it with less, but i just was saying 5470m is def capable of 1x res on this benchmark, and probably most every gamecube games requirements.


But apparently for no reason what so ever, natural decided to go crazy on the subject and say theres no way a 5470m is anywhere close to the performance of a 3650..... (ya i also dont get why that would be relevant in this thread either.....)
and he figured it would be better to also dispute this with me in this thread, completely derailing it, rather then pming me about it or something....

Since he was SO AMENDMENT about me showing him benchmarks to for some reason defend my claim that a 5470m is more then enough to play dolphin, due to the fact its close to the performance of a 3650....
i decided to compare my brothers 4670 to the 5470m in a few different tests using this dolphin benchmark. (and the 4670 is almost 2x the performance of a 3650, and its the closest thing i have to a 3650 around me to be able to do the benchmark with).



@other comment
a 3570k @ 5.1ghz O.O
thats insane lmao

the 2500k overclocks further then the 3570k, so i wouldnt expect a 3570k to be able to hit 5.0ghz+ mark. (and if you did hit 5ghz you would have to have a super lucky chip in which it like was perfect in the silicon bonding process and is able to overclock way more then normal ones at lower voltages and have like crazy liquid cooling going)

yeah i only need 1.4v to make it stable @ 5ghz. and 1.2v at 4.5. like i said in previous posts i got decent air cooling setup with a really solid chip.. hell mine isnt even delided lol. i might throw intel 25 dollars for OC protection so i might attempt 1.5+

FOR SCIENCE!

yes 2500ks in general were able clock higher because they didnt use crappy TIM and it was sodered so temps were easy to control so even the lowly silicon could do 4.5 with enough volts. but 5.1ghz sandy is pretty rare too. most top out at 4.8-5ghz gap range

jbone1337 Wrote:and that the 5470m can easily play at 1x IR w/o any slow downs, as to not hold back any cpu from its potential?

(which that is the point you were disputing with me????)

I will repeat this a third time:

NaturalViolence Wrote:

NaturalViolence Wrote:If this were another game I might suspect a gpu bottleneck. However given the slow cpu and the fact the WW is extremely light on the GPU (which is one of the reasons that it was chosen) I can safely say that the gpu is not a bottleneck in this instance.

I said from the very beginning that this would not impact dolphin. I never tried to argue against this. When he said that the performance levels of the two cards would be the same (not performance in dolphin, just performance). I pointed out the specs and that prompted him to reply with this:

jbone1337 Wrote:But in actual gaming performance at a 720P resolution (1280x720/1366x768), they will score about the same (3650 would get maybe 1-2 fps more lol), which that level of difference is negligible.
If you were playing at like 640x480 or something, there would be a larger difference in performance between the 2. But really thats not a real world benchmark when benchmarking a gpu, since no one is ever going to be really playing at that resolution....

He was the one that originally brought up gaming performance being nearly identical. He even specifically said gaming performance. In every post since up until the last one he has not tried to take this back but instead kept coming up with examples and analogies to try and hammer his support of this claim further. All he had to do was respond with "yes games will run significantly faster on the 3650 but this won't effect WW" and that would have been the end of it. But he chose to continue arguing against me as if he disagreed.

Please take the time to read my posts thoroughly before you respond to them. For the last time I have never been arguing about WW performance this entire time. I don't know where you got that idea from. I stated at the very beginning what you were right about (that this won't impact WW performance) and what you were wrong about (that the 3650 performs about the same as a 5470m in PC games). I even started repeating the claims being made by both sides in list form at the end of each post I made in order to prevent something like this from occurring. When I asked you to back up your claim about the 3650 having about the same performance as a 5470m in game benchmarks with evidence you continued to argue against me. You immediately lost your shit and started calling me names. You then made the situation worse by coming up with increasingly more ridiculous claims to add to your original claim for me to dispute including a claim that all peer reviewed benchmarks were unreliable. That is why this went on for so long. I correct people on things like this all the time and I have never gotten a response as uncivil as yours was. And as such I didn't feel the need to show you the courtesy of using a PM. I felt that if you were responding to me in a public place you would be less likely to resort to things like namecalling with your image at stake and more likely to come up with the data that I was looking for.

Now while this isn't proper PC game benchmarking you have delivered some real data here that we can use. And for that I thank you. Your latest post is a lot cleaner and more readable than your earlier posts mainly due to the lack of incoherent rage. Though it still needs some work in my opinion. But I appreciate the effort you made this time. Since you were civil this time I can respond without fear that it will escalate into name calling.

Anyways the data doesn't include any 3650 results so a lot of guess work still needs to be done but there are some instances above where the 4670 is beating the 5470m by more than 50%. By your own logic a 3650 should be beating a 5470m in those same scenarios by at least 25%. While that is lower than I would have predicted based on the specs that's still pretty significant if you ask me. I'm not sure I would group the 3650 together with the 5470m and call them almost equal. Without the communication overhead that dolphin imposes the difference between them should be substantially larger in PC games as we have seen with past dolphin GPU benchmarks.

jbone1337 Wrote:but anyway, im moving on now (i think i know why squall doesnt really like dolphin community much)

Squall hates this community because starscream bans him whenever he starts acting like an asshole to everyone (including insulting the admins directly). Quite frankly it's amazing how much starscream let him get away with before he finally brought down the banhammer.

(04-05-2013, 12:58 PM)jbone1337 Wrote: [ -> ]...

(04-05-2013, 01:59 PM)NaturalViolence Wrote: [ -> ]...

Anymore off-topic will be deleted. We're not here to watch you two debate whatever it is you're debating, we're here for the CPU benchmark. Feel free to create a new thread to discuss other things.

(04-05-2013, 01:59 PM)NaturalViolence Wrote: [ -> ]

jbone1337 Wrote:but anyway, im moving on now (i think i know why squall doesnt really like dolphin community much)

Squall hates this community because starscream bans him whenever he starts acting like an asshole to everyone (including insulting the admins directly). Quite frankly it's amazing how much starscream let him get away with before he finally brought down the banhammer.

jbone1337: You don't like this community because NV is asking you questions that you can't or don't want to answer. NV doesn't represent the "Dolphin community" as a whole. No one is forcing you to respond to any particular person on this forum. If someone is bothering you, ignore them or PM someone on the forum staff so we can discuss any legitimate problem you may have.

NaturalViolence: Don't talk about things you don't understand please. Squall acts certain ways, yes, but he wasn't banned by me alone, and he is currently not banned.

(04-03-2013, 01:36 PM)Venomx1 Wrote: [ -> ]@ georaldc nope not even delided i just got a really sweet chip (only need 1.2v for 4.5ghz)
i peak 94*C @ 1.33v for 4.8stable though. id probably hit thermal limit at 1.35-1.38ish.

and i need roughly 1.4v to make 5ghz game stable.

i run with a hyper 212+ /w Scythe Ultrakaze fan in a Push config that blows air directly into my exhausti fan. i use antec nano diamond forumla 7. its nice paste

90+ C is certainly pushing it. I'm at 4.6 now and hitting 80C with prime, will try benching again in a bit. I wonder though, does your asrock board exhibit the same problem as mine where voltages reported by programs like cpuz is off by 0.1? Like cpuz says I'm at 1.25 at full load when in reality I'm actually close to 1.35

(04-08-2013, 06:37 AM)georaldc Wrote: [ -> ]

(04-03-2013, 01:36 PM)Venomx1 Wrote: [ -> ]@ georaldc nope not even delided i just got a really sweet chip (only need 1.2v for 4.5ghz)
i peak 94*C @ 1.33v for 4.8stable though. id probably hit thermal limit at 1.35-1.38ish.

and i need roughly 1.4v to make 5ghz game stable.

i run with a hyper 212+ /w Scythe Ultrakaze fan in a Push config that blows air directly into my exhausti fan. i use antec nano diamond forumla 7. its nice paste

90+ C is certainly pushing it. I'm at 4.6 now and hitting 80C with prime, will try benching again in a bit. I wonder though, does your asrock board exhibit the same problem as mine where voltages reported by programs like cpuz is off by 0.1? Like cpuz says I'm at 1.25 at full load when in reality I'm actually close to 1.35

no. if it was i probably wouldve fried my chip by now. thats a pretty big error. and yeah 90+C is while in prime95. realistically doesnt go higher than 70*C if im at like 5ghz @ 1.4v

what voltage do you need for 4.6?