NVIDIA Driver 353.06 with Kepler fix is OUT

Discussion in 'Videocards - NVIDIA GeForce Drivers Section' started by moab600, Jun 1, 2015.

  1. CK the Greek

    CK the Greek Maha Guru

    Messages:
    1,291
    Likes Received:
    15
    GPU:
    GB 2060S,Gsync,3DV2
    Again...what? Even for 980 you are wrong too. First that link you posted doesn't work and second check this out
    http://www.techpowerup.com/reviews/Gigabyte/GTX_980_Ti_G1_Gaming/28.html
    or even this to see hardcore measured for the most power hungry 980 something that you won't never see with any game,only with Furmark BURNER.
     
    Last edited: Jun 16, 2015
  2. CrazyGenio

    CrazyGenio Master Guru

    Messages:
    446
    Likes Received:
    32
    GPU:
    Msi rtx 2080ti trio
    idk why my link was censored, but i know i'm telling the truth because i'm using one, i have the gtx 980 gaming g1, i know what i'm talking about, the link you posted was for the 980ti gaming g1 not the 980, so your link is irrelevant because it's show TDP usage of a reference gtx 980 in the comparison, again i'm talking about gtx 980 gigabyte gaming g1, not the reference or any other.

    [​IMG]

    the gtx 980 gaming g1 uses a lot of power, more than a reference or any other non reference card.
     
  3. EdKiefer

    EdKiefer Ancient Guru

    Messages:
    2,670
    Likes Received:
    279
    GPU:
    MSI 970 Gaming 4G
    That's total system power, not just the card .
    A stock G1 (980) won't pull more than 250w , the bios won't let it .

    And yes I know they cheat by setting power factor value (366w) higher than TDP (250w).
    What this does is now when you hit power limit it will be like 85% .

    But yeh if you mode bios and run like 1312v and crank up power/TDP it will probably draw 400w or way higher that stock, thats because the relationship with voltage to power .
    I wouldn't go by furmark either , use real world BM or games , that just not real load .
     
    Last edited: Jun 16, 2015
  4. CrazyGenio

    CrazyGenio Master Guru

    Messages:
    446
    Likes Received:
    32
    GPU:
    Msi rtx 2080ti trio
    even with real load the gigabyte consume more power BTW why you people are so defensive about nvidia power efficiency, even you changed the topic and the first point was some new cards are so energy efficient that they become unstable i agree with that and that's why i put the gigabyte thing

    i was thinking that these drivers are having crashes just browsing and being idle maybe because a new P-state switching or whatever, i was thinking about this because user with reference cards aren't having issues and i'm using this gigabyte card i only have TDR in chrome while other people are having TDR in a lot of games and all browsers, and because gigabyte card draws more power it's the reason i'm thinking i wasn't having the issues other are having. Just Chrome for me, nothing more, even i can play witcher 3 6 hour straight without any problem.
     
    Last edited: Jun 16, 2015

  5. EdKiefer

    EdKiefer Ancient Guru

    Messages:
    2,670
    Likes Received:
    279
    GPU:
    MSI 970 Gaming 4G
    Sure , over reference card , all 3rd party performance/OC 9xx draw more than reference cause they can OC them so much higher than other versions and on some voltages are higher to than reference . MSI , Gigabyte , and even Asus strix all run higher .
    G1 are one of highest clocked cards ,and the way they have bios setup allows higher peaks .
     
  6. CK the Greek

    CK the Greek Maha Guru

    Messages:
    1,291
    Likes Received:
    15
    GPU:
    GB 2060S,Gsync,3DV2
    As already mentioned you were speaking for TOTAL SYSTEM and not only gpu.
    Here is the G1 (stock speeds)
    [​IMG]

    and here is the most hungry 980

    [​IMG]

    sorry but I have to say again..read more carefully next time so we clearly understand each other in here.
     
  7. -Tj-

    -Tj- Ancient Guru

    Messages:
    17,182
    Likes Received:
    1,928
    GPU:
    Zotac GTX980Ti OC
    Is this still a driver thread? :p



    OT
    Its ok for me now. No browser crashes anymore, idle @desktop was always fine.. Weird stuff.
     
  8. BadDriver

    BadDriver Master Guru

    Messages:
    285
    Likes Received:
    2
    GPU:
    STRIX-GTX1070-8G
    Is it normal that by only enabling shadowplay the gpu and memory clock go up to max????

    Made a movie from it.

    https://www.youtube.com/watch?v=87iP-CBIDYs


    Enabling shadowplay= gpu and memory clocks on max
    Dissabling shadowplay= gpu an memory clocks back to low.
    (no gaming no recording)

    GFE version 2.4.5.44
     
    Last edited: Jun 18, 2015
  9. fluidz

    fluidz Master Guru

    Messages:
    250
    Likes Received:
    32
    GPU:
    Msi trio x 2080ti
    Update :

    Just in case anybody runs into this same issue. All fixed. I uninstalled Chrome and reinstalled it. Simple. Artifact no longer shows.

    Still getting periodic crashes in Chrome. (Driver stopped responding messages) Usually happens after I close down a game, then open Chrome straight afterwards. Happens quite a lot after shutting down Fc4.
     
    Last edited: Jun 19, 2015
  10. Kishkumen32

    Kishkumen32 New Member

    Messages:
    7
    Likes Received:
    0
    GPU:
    GTX 980 Ti SC 6GB
    This thread has me thinking now. If the driver crashes are related to changing power states, clocks etc. Could you enable KBoost in evga precision to disable that?

    I wonder if the reason why people aren't crashing in Witcher 3 (using 353.06) for instance, is because that game will max out any gpu at 99%. Thus, it is the same effect as having KBoost enabled.

    If you were to enable KBoost, then enable gpu acceleration in chrome, and the crashes go away, I think that definitely points the finger towards some kind of (overly aggressive) power savings features that have been introduced in the drivers since 347.88.
     

  11. FDisk

    FDisk Master Guru

    Messages:
    766
    Likes Received:
    0
    GPU:
    ASUS STRIX GTX970 OC 4GB
    KBoost is not going to max your GPU power. It will just lock the BOOST clock so it wont drop. Power will still go up and down.
     
  12. vandinz

    vandinz Member

    Messages:
    39
    Likes Received:
    0
    GPU:
    Nvidia GTX 970
    I've been trying to narrow the TDR's down on my system and it seems to be ZOTAC Firestorm for me. It's their version of Afterburner and every time it's running I get random crashes. Not all the time, but I do get them. When I do start getting them, I turn off FS and they instantly stop. I've had a full day with FS running and had a couple of crashes and a full day with it off and nothing.

    It may just be coincidence. Going to keep testing.

    EDIT - FS is fine with older drivers (two drivers back). So I'm not saying it's am FS issue, it's the driver not working with it this time around for some reason.
     
  13. fluidz

    fluidz Master Guru

    Messages:
    250
    Likes Received:
    32
    GPU:
    Msi trio x 2080ti
    I've found events that relate closely to each "Display driver nvlddmkm stopped responding and has successfully recovered." message.

    Events 4624 + 4672

    These events usually happen one or two seconds before the dreaded "Display driver nvlddmkm stopped responding and has successfully recovered" popup appears. I've found this to happen both whilst using Chrome during gaming, or after exiting a gaming session then opening Chrome (Chrome being the only app running at the time)

    If you look at the 4624 log below, it says : "An account was successfully logged on." This happens whilst Chrome is being used.

    Blue highlight = nvlddmkm error (system log)
    Green highlight = 4672 Auditing event (security log)
    Red highlight = 4624 Auditing event (security log)

    [​IMG]

    Larger version here - http://s7.postimg.org/7vhx6p2jd/353_06_crash_related.png

    Also I noticed a large fluctuation of activity in Afterburner at the time of the crash, maybe this is due to the driver reloading?

    [​IMG]

    Larger version here - http://s28.postimg.org/bc0va5thn/afterburner_graph.png

    Related thread with people mentioning it causes their system to freeze, no mention if its related to graphics cards though - http://answers.microsoft.com/en-us/...g/8d8761ef-7937-4fda-9b8c-b864578ddec1?page=2

    Aside - These crashes are :bang:
     
    Last edited: Jun 19, 2015
  14. evgenim

    evgenim Master Guru

    Messages:
    236
    Likes Received:
    17
    GPU:
    RTX 2060 OC
    That's right. But kboost will some heat up the gpu. Near +8...10c more warmer in idle mode.

    But lock P-State is the best solution (first of all for overclockers).
    http://forums.guru3d.com/showpost.php?p=5100475&postcount=11
     
    Last edited: Jun 19, 2015
  15. -Tj-

    -Tj- Ancient Guru

    Messages:
    17,182
    Likes Received:
    1,928
    GPU:
    Zotac GTX980Ti OC
    I OC'ed my system to the "max" again - slightly different voltage and so far nothing unusual yet..

    Last time I blamed it on bad system OC, but now I see its mostly nvapi polling rate interaction issue.
    Like others said, something changed in r352/r353 and that affects all monitoring tools, including RealTemp if using nvapi access for extra gpu °C monitoring.
     
    Last edited: Jun 19, 2015

  16. fluidz

    fluidz Master Guru

    Messages:
    250
    Likes Received:
    32
    GPU:
    Msi trio x 2080ti
    Applied this, I'll test and report back if drivers stop crashing. Idle vddc was 0.82 before, now at 1.01v. Is there any way to bring that voltage down a little? My Gpu has jumped up by 20c, sitting at 51c. :3eyes:
     
  17. -Tj-

    -Tj- Ancient Guru

    Messages:
    17,182
    Likes Received:
    1,928
    GPU:
    Zotac GTX980Ti OC
    If you set to P0 state it will force gpu to run @ base 3d frequency all the time and that needs such voltage + heats extra.


    Where do you have TDR in idle or web browsers? And do you use some monitoring tool at windows startup?



    You can also try nvinspector's multimonitor tool, this controls these p-states too, e.g. force gpu to use max low 3d freq. instead of full 3d in idle.

    [​IMG]
     
    Last edited: Jun 19, 2015
  18. RdJokr

    RdJokr Member Guru

    Messages:
    144
    Likes Received:
    0
    GPU:
    MSI GTX 750 Ti TF / 2GB
    So, I'm guessing next Tuesday we'll have a new Game Ready driver for Batman Arkham Knight?
     
  19. gozunz

    gozunz Member

    Messages:
    12
    Likes Received:
    1
    GPU:
    Nvidia GTX 1080
    337.88 vs 353.06

    Just to compare as i have seen people complain about performance...

    System specs:
    i7 4770
    16GB Ram
    EVGA GeForce 770 GTX 2GB
    Windows 8.1 x64

    337.88:

    Unigine Heaven 4.0 - All extreme/ultra 8xaa 1080p
    38.5fps, Score 970

    GTA5 - 2560x1400, All High, MSSA, 16xAF
    Tesellation Off:
    Frames Per Second (Higher is better) Min, Max, Avg
    Pass 0, 12.032136, 113.986618, 77.764023
    Pass 1, 41.231308, 130.050369, 82.287903
    Pass 2, 38.498646, 113.511200, 78.791458
    Pass 3, 36.701550, 140.115463, 81.541756
    Pass 4, 16.110609, 153.939255, 79.099716
    Tesellation High:
    Pass 0, 13.440018, 132.341507, 75.438927
    Pass 1, 28.039656, 126.918549, 79.825813
    Pass 2, 34.715363, 129.009445, 76.707344
    Pass 3, 35.351467, 141.567154, 80.818939
    Pass 4, 19.020514, 164.804489, 80.559067

    353.06:

    Unigine Heaven 4.0 - All extreme/ultra 8xaa 1080p
    39.2fps, Score 988

    GTA5 - 2560x1400, All High, MSSA, 16xAF
    Tesellation Off:
    Frames Per Second (Higher is better) Min, Max, Avg
    Pass 0, 24.405258, 105.863762, 80.427902
    Pass 1, 43.819183, 125.964973, 82.798706
    Pass 2, 38.445545, 136.408157, 81.184570
    Pass 3, 41.649136, 140.352158, 82.248634
    Pass 4, 32.843800, 142.539322, 83.603355
    Tesellation High:
    Pass 0, 40.245216, 143.500351, 78.632988
    Pass 1, 40.166348, 127.361374, 80.442291
    Pass 2, 54.702148, 143.376404, 77.075211
    Pass 3, 38.446880, 141.223724, 81.105392
    Pass 4, 22.701738, 138.386658, 82.527885
     
  20. Darren Hodgson

    Darren Hodgson Ancient Guru

    Messages:
    16,031
    Likes Received:
    553
    GPU:
    EVGA GTX 1080 Ti SC2
    I was thinking the very same thing.

    I have experienced TDRs in a few specific games but most of them are stable...thankfully (or least for as long as I have been playing them anyway). Granted I have over 600 games installed and have only be4en playing perhaps 5-7% of them at the moment.

    However, The Witcher 3 has been almost completely stable for the 65 hours I have been playing it, except for one unexpected exit to the desktop (I am loathe to call it a crash because their was no error window nor is there any entry for it in the event log). My GPU usage in that game has almost been 99% constantly with a core clock of 1,291 MHz for almost the entire session except for drops during loading screen (when adaptive performance kicks in and lowers the clocks).

    I had two TDRs while playing Batman: Arkham Origins over the weekend, the only game out of the two dozen or so I have been playing to suffer from that issue. I do have PhysX enabled and set to High plus 4xTXAA and maxed out 1920x1200 settings, the game seems to be less demanding on the GPU overall than The Witcher 3.

    During the 9 months that I owned my EVGA GTX 980 SC, I only experienced one TDR issue and that was because I tried overclocking the core and memory clocks. Once I set them to their factory defaults (already overclocked) then the card was 100% stable. I have experienced more TDRs in the last month than I have in the last 5 years prior to owning a GTX 980 Ti SC. Could the overclock set by EVGA itself be unstable?
     
    Last edited: Jun 22, 2015

Share This Page