Skip navigation
Currently Being Moderated

Unsupported GPU for CS5

May 5, 2010 8:59 AM

With all the interest of the "unsupported" GPUs, I thought it was time to start a specific thread.

 

Please post your questions and experiences.

 

Hacking is not advised and the unsupported cards are not ready for production use.

 

You've been warned! 

Branched to a new discussion.
 
Replies 1 2 3 ... 6 Previous Next
  • Currently Being Moderated
    May 5, 2010 12:24 PM   in reply to cts51911

    I've only seen the GPU go to about 20% at most with the few things I've played with so far, but the GTX 275 did smooth certain things out even if it didn't drop the CPU usage too much for some of the effects (for sharpen it cut CPU usage in half though and even where CPU usage appeared to remain the same, the video sometimes went from a bit choppy to smooth).

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 12:35 PM   in reply to cts51911

    The Boss is out of the office for a bit and I had a chance to test his i7 975 Sager Notebook (3 internal hard drives because we hate it when Harm gets irritable).  His GTX285M is not officially supported.  The hack has got it crankin.  I placed two layers of Panasonic P2 1080i with a title layer on top.  On the second layer of video, I cut the clip's opacity by 46% added gaussian blur maxed out and the Fast Color Corrector dialed in to various levels.  Yellow Render line.  The playback is awesome.  Smooth as can be.  I exported a finished video in various formats.  No render errors or problems to report.  CS5 is amazing.

     
    |
    Mark as:
  • Admiral Cook
    169 posts
    Jul 3, 2009
    Currently Being Moderated
    May 5, 2010 12:43 PM   in reply to chupacabracobra

    I would like to use it with my 8800GT. Where is the patch?

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 12:47 PM   in reply to Admiral Cook

    Well, it is not a "patch," but an unsupported "hack." Reply #2 has the steps to do the hack.

     

    Disclaimer - I have never used this, and know nothing about it, so cannot vouch for the effectiveness, or the dangers of doing this.

     

    Good luck,

     

    Hunt

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 1:11 PM   in reply to Admiral Cook
    function(){return A.apply(null,[this].concat($A(arguments)))}

    function(){return A.apply(null,[this].concat($A(arguments)))}

    Admiral Cook wrote:

     

    I would like to use it with my 8800GT. Where is the patch?

    not a patch, just follow the instruction above (basically you just type the name of your graphics card into the approved cards text file in the PPCS5 directory and it's good to go

     

    one thing to keep in mind is that the 8800GT doesn't support double-precision, i have no idea if they used double-precision for any of the effects, but keep that in mind (i believe, but am not sure, that all of the 200 series, aside from the 240/250 do support double). I think I saw someone using a card that didn't support double and they said it worked, so maybe it would be ok, at least for most effects. I would tend to think they might not have used double at all, but who knows.

     

    actually more importantly I think the 8800GT, at the very least most of them, would fail though since I think most had only 512MB of memory.

    Did they ever make an expanded memory edition of that card??

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 2:04 PM   in reply to Admiral Cook

    The 8800GT will not work even with a hack.  The minimum required GPU memory for the hack to work is 756MB.

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 2:19 PM   in reply to cts51911

    I was taking a wild stab at another notebook we have that has two GeForce 8800M GTX's in SLI.  I remember from the pre-release MPE reports, that SLI won't be supported.  The cards memory of 495MB is what GPUSniffer.exe lists as excluding it from use within Premiere.    

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 2:49 PM   in reply to Chuck A. McIntyre

    Updated test now with testing done with a GTX 480 card.  I should preface that the Cineforms files were actually color corrected in FirstLight and really didn't need to do the color correction, but wanted to keep it consistent.  Also, all the filters below are accelerated effects.  Not sure if the accleration works in software mode or not.  Let me know if you have any questions or want me to test something else.  I may drop all the acclerated effects and put on non-accelerated ones for testing.

     

    Cineform Files - Software
    Total Time- 3min 30sec
    No Filter CPU Playback Quality - Yellow/Smooth
    Export Time No Filters - 4min 17sec
    Filter CPU Playback Quality - Red/Few Stutters
    Filter List - Gaussian Blur(47), Sharpen(43), Fast Color Corrector, Tint(100), Noise(16)
    Export Time Filters - 7min 17sec

     

    Cineform Files - MPE
    Total Time- 3min 30sec
    No Filter CPU Playback Quality - Yellow/Smooth
    Export Time No Filters - 4min 22sec
    Filter CPU Playback Quality - Yellow/Smooth
    Filter List - Gaussian Blur(47), Sharpen(43), Fast Color Corrector, Tint(100), Noise(16)
    Export Time Filters - 4min 49sec

     

    5D Mark II Files - Software
    Total Time- 3min 30sec
    No Filter CPU Playback Quality - Yellow/Smooth
    Export Time No Filters - 3min 42sec
    Filter CPU Playback Quality - Red/Few Stutters
    Filter List - Gaussian Blur(47), Sharpen(43), Fast Color Corrector, Tint(100), Noise(16)
    Export Time Filters - 6min 40sec

     

    5D Mark II Files - MPE
    Total Time- 3min 30sec
    No Filter CPU Playback Quality - Yellow/Smooth
    Export Time No Filters - 3min 40sec
    Filter CPU Playback Quality - Yellow/Smooth
    Filter List - Gaussian Blur(47), Sharpen(43), Fast Color Corrector, Tint(100), Noise(16)
    Export Time Filters - 4min 17sec

     


    Cineform Conclusions
    No Filters - No Difference
    Filters 34% less time using MPE (437sec vs 289sec)

     

    5D Mark II Conclusions
    No Filters - No Difference
    Filters 36% less time using MPE (400sec vs 257sec)

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 3:37 PM   in reply to cts51911

    The big thing hurting me right now is that I'm not running a RAID setup.

     

    Doing an export now with same filters on the Cineform files via CS4 AME.  Lets see what the time differences are.  Also, the playback is pretty choppy.

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 3:55 PM   in reply to metalsaber

    Just for the fun of it, I ran the same thing through CS4.  Prepare for  the horrors!!!!

    Cineform Files – CS4
    Total Time- 3min 30sec
    Filter CPU Playback Quality – RedStutters
    Filter List – Gaussian Blur(47), Sharpen(43), Fast Color Corrector,  Tint(100), Noise(16)
    Export Time Filters – 15min 32sec

    Cineform Conclusions
    69% less time using CS5 MPE vs CS4 (289sec vs 932sec)

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 4:14 PM   in reply to cts51911

    So far, pretty amazing.  Importing 1080p 29.97 clips.  Exporting to 720p 29.97.

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 6:13 PM   in reply to metalsaber

    @metalsaber:

     

    How many video layers wil the hacked card run before it tips up or needs to render?

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 6:45 PM   in reply to cts51911

    I couldnt determine exactly how many layers he actually used in his "test"

     

    He seemed to be having ssues at 4  with both cards.(AVCHD admittedly) and also had scrubbing problems

     

    I have gone past 8 layers and kept throwing fx at them all .  Still havent broken it or seen a RED render line and scrubbing is perfect.

     

    BTW - I have my source and Program monitors at Hi Res setting...  and I get same result playing out to the full screen. (tilde key)

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 7:17 PM   in reply to cts51911

    mxf (DVCPROHD), avi uncompressed, mov uncompressed, Prores

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 7:50 PM   in reply to shooternz

    shooternz wrote:

     

    @metalsaber:

     

    How many video layers wil the hacked card run before it tips up or needs to render?

     

    Will do a test tomorrow afternoon/evening.

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 8:09 PM   in reply to metalsaber

    That will be interesting. Thanx for doing that and lok forward to the result (even though its kind of academic  to me now)

     

    I think people really want to compare hacked card performance against the Approved cards.

     

    I dont know how that is quantifed all else being equal but the Layer Limits maybe  one way.

     

    For me... I regularly use 6 to 9 Video layers so I am well pleased with my results.

     

    For the Record I am using a Quadro FX3800.

     
    |
    Mark as:
  • Currently Being Moderated
    May 5, 2010 8:49 PM   in reply to cts51911

    I dont think he has any other files apart from what he has already used.  He should stick with those for relativity.

     

    Some one else will fill in the gaps with experiences in other formats CODECs

     

    I will state now - what I use works perfectly with FX3800.

    If what I use works perfectly for someone with a hacked card - thats their good luck as well.

     

    Do not forget this is not a singularity  process.

     

    It involves:  OS, Hardware, Setup, Graphics Card, User input before you even get to CODECS / Formats etc...

     
    |
    Mark as:
  • Currently Being Moderated
    May 6, 2010 6:29 AM   in reply to shooternz

    shooternz wrote:

     

    That will be interesting. Thanx for doing that and lok forward to the result (even though its kind of academic  to me now)

     

    I think people really want to compare hacked card performance against the Approved cards.

     

    I dont know how that is quantifed all else being equal but the Layer Limits maybe  one way.

     

    For me... I regularly use 6 to 9 Video layers so I am well pleased with my results.

     

    For the Record I am using a Quadro FX3800.

     

    Ok, just a few tests.

     

    1. I put 6 layers of 5D Mark II video into the timeline with 0 filters.  The bar was still yellow and played back smoothly.
    2. I then removed the top 3 layers and started with the bottom 3.  I added 3 filters to each (Sharpen, Gaussian Blur, and Brightness/Contrast).  Yellow Bar and Smooth Playback.
    3. I then added a clips 4-6 back in with no filters.  Still yellow time line and smooth playback
    4. I then removed clips 5-6.  Added on the Brightness/Contrast filter to the 4th video layer.  Clip turns red/pretty much smooth playback with a few little blips.  Nothing major.  Interesting that the bar does not cover the entire 4th video layer length.  This has to do with the bottom 3 layers all being at various lengths with the 3rd layer being the shortest.  So the red bar on layer 4 only extends to the length of layer 3.  Probably since the remaining length of layer 4 is now the 3rd layer of video allowed to have filters.
    5. I then added 2 more filters to layer 4 (3 effects total) and the playback is about as smooth as in #4.  Pretty usable imo.
    6. If I add a 5th layer of video with 0 effects, the timeline actually goes back to yellow for the length of the 5th layer.  Even with a 50% opacity on layer 5, it still plays back rather smooth.
    7. If I add a single effect to layer 5, timeline turns red again.  Playback is a bit more choppy at this point.

     

    Seems like 3 layers of video with effects will keep the timeline yellow.  Adding a 4th video layer with effects turns the timeline red, but still useable.  Odd part was adding a 5th layer with 0 effects on top turned the timeline back to yellow again.

     

    So my guess is that Adobe is limiting the GTX series to 3 layers of video with effects.  Probably has to do with not ticking off those that purchased the Quadro series cards.

     
    |
    Mark as:
  • Currently Being Moderated
    May 6, 2010 7:43 AM   in reply to cts51911

    Great to see all the testing with the "unsupported" cards and the hack!  Love it, althoguh I already have a GTX 285.  Thought I'd post some experiences I've had with my system and GTX 285.  Would be great to see some system specs apart from the various video cards being tested so we know where we could improve apart from the video cards.

     

    My system:

     

    GTX 285

    8GB DDR3 1600

    Core i7 920 (perfectly stable and cool at 3.8Ghz)

    Two (2) Raptors in Raid 0 for OS

    Three (3) Raptors in Raid 0 for Working-Project Disk

     

    Video:  Canon 7D, AVCHD Full 1080 24p

     

    MPE definitely improves playback and render, but will not run seemlessly through a timeline with longer non-GPU accelerated effects such as Additive Dissolves or Non-Additive Dissolves. Adding a second layer over an Additive Dissolve and I'm back to CS4 poor performance.  I've tried Full resolution playback, 1/2 and 1/4 and it just won't do it.

     

    Does anyone have this same experience with the unsupported hack or supported cards?

     

    I hope Adobe is looking to add/replace the non-GPU accelerated effects with GPU/MPE accelerated ones during CS5 updates.  They picked some common effects, but so what - expand these!  Adobe is saying "we have this great program, but it will only benefit you sometimes - oh, and if we make all of our effects GPU accelerated we will put all of our third party effects people like Boris out of business, so we'll wait for them to create some instead"

     
    |
    Mark as:
  • Currently Being Moderated
    May 6, 2010 8:38 AM   in reply to mylifevisual

    We should be getting our two supported 285 GTX cards either today or tomorrow.  Just for the fun of it, I applied the hack to one of our systems that's going to get a supported card.  Right now it has a GeForce 9500 GT.  Everything works without errors etc. with this hack applied.  No render line and a yellow render line in some cases with 3 layers of video with MPE supported effects applied.  One problem though, and it's a big one, I don't see any playback benefit and exports take twice as long with MPE enabled.  So... It MAY be that Adobe has focused upon supporting cards based upon their actual benefit to the editing workflow (speed) as opposed to stability/errors etc.

     

    * I haven't seen smoke yet, so I don't think this minor hack can do any hardware or even software damage.  I don't spend much time behind Command Prompt windows, but you can save some time if you decide to try this hack out by dragging and dropping the GPUSniffer.exe into the Command Prompt window and hitting Enter.  The whole thing only takes a couple of minutes.

     
    |
    Mark as:
  • Currently Being Moderated
    May 6, 2010 8:45 AM   in reply to mylifevisual

    I'm hoping Adobe expands their lineup of acclerated effects.   As you said, they got quite a bit of them and a lot of the popular ones.

     

    I'm just running a i7 920 proc oc'd to 3.2GHz with 12GB of 1333 RAM.  2 - 500GB and 2 - 1TB hard drives not in RAID.  I put the machine together for photo editing/video games before I even considered using it for video editing, so I'm stuck with the way it is.  I have too many 3rd party programs on it that I don't want to go through the grief of re-activating.

     

    Next machine I build will be in RAID from the beginning.

     
    |
    Mark as:
  • Currently Being Moderated
    May 6, 2010 11:09 AM   in reply to cts51911
    function(){return A.apply(null,[this].concat($A(arguments)))}

    cts51911 wrote:

     

    I'd like to reiterate that this is a

    HACK!

    Use at your own peril.

     

    Isn't saying stuff like "use at your own PERIL" going a bit extreme? It is not like it will melt your computer down or anything (and if it did, then the next time you played a video game it would've melted down anyway and likely a lot faster at that). Even if it were to mess up it's not like it will do anything more perilous to your hardware than a hundred other programs (with no warnings) could. I guess it depends how you interp use at your own peril.

    '

     
    |
    Mark as:
  • Currently Being Moderated
    May 6, 2010 11:36 AM   in reply to cts51911
    function(){return A.apply(null,[this].concat($A(arguments)))}

    cts51911 wrote:

     

    chupacabracobra wrote:

     

    function(){return A.apply(null,[this].concat($A(arguments)))}

    cts51911 wrote:

     

    I'd like to reiterate that this is a

    HACK!

    Use at your own peril.

     

    Isn't saying stuff like "use at your own PERIL" going a bit extreme? It is not like it will melt your computer down or anything (and if it did, then the next time you played a video game it would've melted down anyway and likely a lot faster at that). Even if it were to mess up it's not like it will do anything more perilous to your hardware than a hundred other programs (with no warnings) could. I guess it depends how you interp use at your own peril.

    '

    It is supposed to be funny.....Most people get it...Some don't (obviously)


    OK, OK I missed that, but well hey it is kinda of hard to tell considering all the kool-aid drinkers who write the exactly the same (only even more dramatically at times hah) and even some of the proponents had been saying things somewhat along those lines, so again hard to tell what jokes are actually meant to be jokes hah.

     
    |
    Mark as:
  • Currently Being Moderated
    May 6, 2010 2:00 PM   in reply to metalsaber

    Thanx for doing the test.

     

     

    Seems to show the hacks dont make Quadros out of that series of cards.

     
    |
    Mark as:
  • Currently Being Moderated
    May 6, 2010 2:15 PM   in reply to shooternz
    function(){return A.apply(null,[this].concat($A(arguments)))}

    shooternz wrote:

     

    Thanx for doing the test.

     

     

    Seems to show the hacks dont make Quadros out of that series of cards.

    if you don't use tons of layers and keep it to three with effects, then I would bet a Fermi would do even better than the Quadros out there (or at least any non G100-type Quadros). G100 CUDA blows away G80 CUDA. Granted some of the increase might need a different programming style, but even same code, as is, should run faster on G100 than ANY G80 chip quadro or not (and let us be honest a Quadro is just a consumer card with, for the most part, just some driver features enabled, such as more more streams allowed; they hardly go to the expense to design a different architecure for Quadro, they need to huge sales of gaming cards to subsidize it all) and I would think a consumer card of same level as a particular QUadro would also do every bit as well so long as you don't use too many layers.

     
    |
    Mark as:
  • Currently Being Moderated
    May 6, 2010 2:18 PM   in reply to shooternz

    Just received and installed the supported BFG 285 GTX from Newegg.com in one of our systems.  I have 7 layers of P2 1080i with different effects on the different clips:

    Gaussian blur, color correction, spinning picture in pictures, different blending modes all playing back at full resolution perfectly smooth with a yellow render line.  I thought the 285's were only supposed to support 3 layers without the red render line.  Amazing.

     
    |
    Mark as:
1 2 3 ... 6 Previous Next
Actions

More Like This

  • Retrieving data ...

Bookmarked By (0)