Jump to content

Unigine Valley (Highend OpenGL) Benchmark


  • Please log in to reply
77 replies to this topic

#21
eep357

eep357

    Triple Platinum

  • Retired
  • 2,527 posts
  • Gender:Male
  • Location:Dark Side of The Wall
  • Interests:things and stuff
It would be min of 31.1, but the 8FPS happens at the same spot each time, it freezes for half a second right when the first lightning strike of the benchmark happens no matter what settings are used. IDK?

#22
mitch_de

mitch_de

    InsanelyMacaholic

  • Retired
  • 2,902 posts
  • Gender:Male
  • Location:Stuttgart / Germany
yep, your massive framedrop at that bench position (when first rain drops came up) can have many reasons:
- bench (+ OS X gpu driver) cant handle VRAM memory usage correct
- bench (+ OS X gpu driver) cant handle the used shader programm correct
....
Also my gpu stalls a very short time at thos bench position drops from 17 fps very short down to 9 fps.

Maybe the bench (and the specific os x driver) has some prob using vram memory correct oder with the used rain drop shader programm.
VRAM usage problems could be, because also AA (antialiasing) makes probs on all AMD gpus. AA has an major effect for VRAM usage, so perhaps some driver specific VRAM usage probs?

You could use atMonitor (attached, don't update because newer versions removed gpu part) to look (in realtime) after the VRAM / gpu usage beside running valley bench in windowed mode and setup atMonitor gpu % and gpu VRAM usage for menue or floating window values.
I am unsure if the AMD drivers also support VRAM usage "looking" as Nvidia. Give a try. rob from barefeats tested it with GTX 580 and there valley used about 1 GB VRAM (higer res + AA + ultra)

Looks like ( if used as menue setting) for my 512 MB card - even without AA and BASIC preset near all VRAM is used.
VRAM usage never gets over around 95% - if more VRAM needed, the driver swaps VRAM<>RAM for new content over the PCIe Bus. That swappings takes very much time compared to already loaded textures into VRAM.

Attached File  Bildschirmfoto 2013-02-28 um 08.02.23.jpg   32.4KB   1 downloads
Attached File  Bildschirmfoto 2013-02-28 um 08.01.53.jpg   30.78KB   1 downloads

PS: I also tested windowed vs fullscreen basic. Very minimal diff between both. Only 1 FPS diff average.

Attached Files



#23
eep357

eep357

    Triple Platinum

  • Retired
  • 2,527 posts
  • Gender:Male
  • Location:Dark Side of The Wall
  • Interests:things and stuff
Osx broke VRAM monitoring on 5xxx series and newer cards since around 10.7.3, it can still be monitored with openGL driver monitor to get figures, problem was/is that total available VRAM became a negative number making all monitoring apps show 100% usage. With openGL driver monitor you can expand it's range to allow you to see negative and positive numbers so it is still a way to get comparitive figures even though the exact numbers reported might not be accurate. I'll give it a run in a little while and post it's output during that scene

#24
eep357

eep357

    Triple Platinum

  • Retired
  • 2,527 posts
  • Gender:Male
  • Location:Dark Side of The Wall
  • Interests:things and stuff
OK here is the results, highlighted the relevant time frame (it's pretty obvious), only displaying the data sets that have info, most the other monitoring options are just 0. This is not when the rain starts, but in the following scene almost exactly when the first lightning flash happens and the screen is focused on a large boulder.
Attached File  AMDRadeonX4000GLDriver_.jpg   124.59KB   12 downloads
What's also interesting is in windowed mode, while same thing does happen, if a portion of the window is covered up by another open window (monitor tool or a finder window) then it does not have the same problem at that spot in the bench and there's nothing unusual in the monitor readings. Moving half or most of the window off the desktop it will still stutter. Maybe a problem with an object or a buffer being brought forward?

#25
mitch_de

mitch_de

    InsanelyMacaholic

  • Retired
  • 2,902 posts
  • Gender:Male
  • Location:Stuttgart / Germany
I also used OpenDL Driver Monitor (Dev tools).
Running Valley in small window (640x360) to show Monitor beside.
I used low quality up to ultra + 8*AA (AA uses much more VRAM.
Beside that i run an other VRAM usage OpenGL/CL App OceanWave (needs about 70 MB VRAM).
I can see, that using ultra+8*AA (or for sure ultra + much higher res) the 5120 MB VRAM gets "out". Means cpu must swap VRAM.

That can be seen with the parameters : CPU Wait for VRAM Heap Allocation - which never gets more than 0 if VRAM is enough.
Also Texture Page Off Data gets > 0. Current Mapped DMA Memory in this case gets near 800 MB.
All those 3 parameters never got > 0 ( or in case of Mapped DMA much lower max value) if i didnt run Oceanwave beside Ultra+ 8*AA.

Your 1GB++ card should not show any non zero results in Texture page off or CPU Wait for VRAM Heap Allocation i think.

Pic shows running Valley in LOW and not running OceanWave OPenCL bench beside. If i run in Ultra+8*AA beside OceanWave i get big positive CPU Wait for VRAM Heap alloc and also Texture Page Off Data > 0.

Attached Files



#26
eep357

eep357

    Triple Platinum

  • Retired
  • 2,527 posts
  • Gender:Male
  • Location:Dark Side of The Wall
  • Interests:things and stuff
I included all options; then after running it, any that never went over a 0 value I unchecked so they wouldn't clutter the table view, so any that aren't seen in that screen shot were all zero. 3gb of VRAM on the card, based on the figures it doesn't seem to be getting anywhere near maxing that out. Tried rolling back to earlier 10.8.3 kexts, no luck. Low quality actually doesn't do it and runs fine. Going to try different smbios out.

#27
k3nny

k3nny

    InsanelyMac Legend

  • Members
  • PipPipPipPipPipPipPip
  • 560 posts
  • Gender:Male
Here my results. Presets Basic w/o AA and Extreme HD w/o AA:
Attached File  valley_basic_wo_aa.jpg   84.69KB   3 downloadsAttached File  valley_extremeHd_wo_aa.jpg   85.49KB   2 downloads

#28
mitch_de

mitch_de

    InsanelyMacaholic

  • Retired
  • 2,902 posts
  • Gender:Male
  • Location:Stuttgart / Germany
Interesting - that AMD 7870 (min FPS 33) hasn't that big FPS drop as the other AMD 7970 (min FPS 8) gpu.
CPU speed diff cant be the reason: i5 (no fps drop) vs i7 (fps drop). Other fps values of 7870 / 7970 - average fps & max fps are very close.

#29
eep357

eep357

    Triple Platinum

  • Retired
  • 2,527 posts
  • Gender:Male
  • Location:Dark Side of The Wall
  • Interests:things and stuff
Basic with AA enabled, although with usual AA color problems is more inline with what it should get:
Attached File  Unigine Valley Benchmark AAon.jpg   52.92KB   16 downloads
And Extreme HD with AA off looked gorgeous and smooth with only a little hiccup at that same spot, and it always exactly 8.4fps exactly when it happens making me think it's with the openGL functions in Valley and not hardware or OSX problem:
Attached File  Screen Shot 2013-02-28 at 7.35.34 AM.png   907.66KB   21 downloads

Edit: So after some additional testing, running bench in Win 8 using openGL api gives lower overall FPS vs OSX, but it also freezes at exact same spot and drops min FPS to exact same amount at 8.4. DX11 give higher average/peak FPS but also freeze at same spot, lowering min FPS to 16.4. Tests ran using both Ghz edition VBIOS and orig release 7970 VBIOS with same results. So it's either the applications handling of 7970 or issue with the card its self.

#30
Maniac10

Maniac10

    InsanelyMac Legend

  • Members
  • PipPipPipPipPipPipPipPip
  • 1,258 posts
  • Gender:Not Telling

BASIC benchmark without AA. The first time I had the issue eep357 mentions but the second time it was gone.

 

 

Attached Files



#31
gothic860

gothic860

    InsanelyMac Protégé

  • Members
  • PipPip
  • 84 posts
  • Gender:Male
  • Location:Germany, Bavaria
BASIC with AA + fullscreen:
Posted Image

BASIC with AA no fullscreen:
Posted Image

#32
blackosx

blackosx

    InsanelyMacaholic

  • Coders
  • 3,170 posts
  • Gender:Male
  • Location:UK
Hi Mitch

Nice Benchmarking app!
Here's my result using Preset Basic without AA for Radeon 5770.

Attached Files



#33
fau7i

fau7i

    InsanelyMac Geek

  • Members
  • PipPipPip
  • 111 posts
  • Gender:Male
  • Location:Malaysia
Sapphire HD 7870 XT with preset basic (w/o AA) in windowed and fullscreen.

Posted Image

#34
Slice

Slice

    InsanelyMacaholic

  • Local Moderators
  • 3,192 posts
  • Gender:Male
  • Location:Moscow
I hear it is emulated OpenGL and it is like to be true.

#35
theconnactic

theconnactic

    Stubborn AMD user

  • Local Moderators
  • 2,973 posts
  • Gender:Male

I hear it is emulated OpenGL and it is like to be true.


Hi, Slice!

Could you please elaborate?

All the best!

#36
RobertX

RobertX

    Yosemite Sam

  • Members
  • PipPipPipPipPipPipPip
  • 572 posts
  • Gender:Not Telling
...slightly better on 10.8.3


Attached File  Untitled2.png   1.16MB   12 downloads

#37
Slice

Slice

    InsanelyMacaholic

  • Local Moderators
  • 3,192 posts
  • Gender:Male
  • Location:Moscow

Hi Mitch

Nice Benchmarking app!
Here's my result using Preset Basic without AA for Radeon 5770.

Good result with poor card? It is because of good CPU i7 8 cores.

#38
mitch_de

mitch_de

    InsanelyMacaholic

  • Retired
  • 2,902 posts
  • Gender:Male
  • Location:Stuttgart / Germany

I hear it is emulated OpenGL and it is like to be true.


No, its real OpenGL. But as every complex OpenGL Bench (as DirectX too) which uses real game content/tasks as bench, like Unique does, CPU does matter of course.
That happens also to real Win DirectX games with complex usage of gpu. Same GPU , highend AMD 7970 / GTX 680 running C2D up vs i7 OCed will give up to 30% FPS diff!!
Depends on the stuff benched but as more the bench uses real game content, like Unique does and not only syntetic bench tasks, the more the CPU does matter.
Also, number of cpu kernels isnt the only parameter! More diff happens by different L1/l2/L3 cpu cache sizes & their different Lx optimisations - because the opengl/directX driver "stays" often in those L1/l2/L3 cpu caches running benches. I guess that not more than 3 real cpu kernels will get "work" to do (means getting > 40% kernel load) running that bench. Perhaps someone checks that with activity cpu shown beside running an windowed Valley bench.
So >4 (real) cpu kernels will not make diff - only the speed of each used(>= 40%) kernel+ speed of L1/2/3 Cache + speed of systembus/RAM will have an effect.

But !!!!
Never will a lowend GPU come close or get same FPS as an highend gpu. C2D with AMD 7970 stay much faster than i7 with AMD 5770!!!
So gpu speed can be benched more real game like = highend and lowend gpus getting closer together, or more synthetic (unreal ;) which shows more speed diffs.
Also if you take a look on the game fps benches in Win, you see that the ranking of the gpus often changes between different games (= usage of directx/opengl code different, different driver speeds, different optimizations....)
For example OpenGLExtensionsViewer Bench results will be sure have less CPU decency BUT doesnt show GPU speed in REAL LIFE (= gaming).

To get an more overall overview of an gpu speed, its useful to compare many benchmark types - more real game benches and the more synthetic ones.

#39
Slice

Slice

    InsanelyMacaholic

  • Local Moderators
  • 3,192 posts
  • Gender:Male
  • Location:Moscow
OK, but how you explain that this test in not working with AA on Radeons? Radeon do support AA in real OpenGL.
I have very poor results on my comp#1. Ten times worst then blackosx with HD5770.
But my system works fine with 40fps in real game 4x4Evolution ported from Windows game by the mean of Wineskin. As far as I understand it emulates DirectX by system OpenGL that is hardware OpenGL in my case.

#40
mitch_de

mitch_de

    InsanelyMacaholic

  • Retired
  • 2,902 posts
  • Gender:Male
  • Location:Stuttgart / Germany
1. The AA bug shows that the benches (both, Valley & Heaven) from them have some probs with AMD AA setting and perhaps other bugs too for AMD. Its OS X related - Win OpenGL doesnt have that AMD AA prob.
I think they know already about that and will fix that - perhaps also Apple must do some work on the AMD drivers too (10.8.3 fixed AA prob not i think).
Some bad results with some benches doesnt matter if other benches show your AMD 6670 working good compared to such 5770 (which performs much better you said )
here)

Q: If your 6670 is really 10 times slower than

blackosx with HD5770, you would have only around 5 FPS avr?! (They have 45-50 FPS, windowed, BASIC, AA OFF)



Perhaps others with 6770 & 6670 post their results, because until now no other 667o is listed here.







0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

© 2014 InsanelyMac  |   News  |   Forum  |   Downloads  |   OSx86 Wiki  |   Mac Netbook  |   PHP hosting by CatN  |   Designed by Ed Gain  |   Logo by irfan  |   Privacy Policy