AMD Catalyst 14.12 Omega WHQL (14.501.1003.0 November 20)

Discussion in 'Videocards - AMD Radeon Drivers Section' started by shadow_craft, Dec 9, 2014.

  1. Spathi

    Spathi Master Guru

    Messages:
    446
    Likes Received:
    0
    GPU:
    Galax GTX980 V2
    Another random? Some of the people posting do not have UEFI. Actually most of them have legacy and some complaining on the net have weird server boards not designed for graphics. They could try to slow down the cards BIOS to see if the... well I already explained.

    The cpu caches from mapped main memory, kernel panic, lol.

    Easier to swap the excessively large PSU's, ahaha, seriously they have a hard time in standby or even running at 100W which is what they need to do. To use your card as an example it will only need 200w+100w+a bit extra, so a 400w PSU would be fine. Ignore the candy, FSP is a good brand if you need a new one.

    I did start building and programming computers in 1981, but I have rambled enough.
     
    Last edited: Feb 10, 2015
  2. The Postman

    The Postman Ancient Guru

    Messages:
    1,773
    Likes Received:
    0
    GPU:
    MSI 980 TI Gaming
    Wow I just got a headache from reading so much nonsense...
     
  3. xodius80

    xodius80 Master Guru

    Messages:
    789
    Likes Received:
    11
    GPU:
    GTX780Ti
    hi, can i join here and say somthing about HPET?!?!?!?



    LOOOOL jk, anyways im here to report that these drivers have proven stable in all my games including mwo!

    k thanks! bye.
     
  4. Spathi

    Spathi Master Guru

    Messages:
    446
    Likes Received:
    0
    GPU:
    Galax GTX980 V2
    They work for me as well now, but underneath they do something strange.

    HPET can be used by windows to sync Invariant TSC the first time, and to do something else for Invariant TSC, I forget. So leave it on.
     

  5. The Mac

    The Mac Ancient Guru

    Messages:
    4,404
    Likes Received:
    0
    GPU:
    Sapphire R9-290 Vapor-X
    you clearly missed the sarcasm there.

    im going to guess English isnt your primary language.

    steel case...wtf?

    anyone on a 64bit OS turning on shadow vbios is asking for trouble.

    and if you are on a 32bit OS, you dont belong on this forum.


    lol
     
    Last edited: Feb 10, 2015
  6. Spathi

    Spathi Master Guru

    Messages:
    446
    Likes Received:
    0
    GPU:
    Galax GTX980 V2
    It was sarcasm back, which you missed.

    So do you all have the "AMD sleep bug"?

    I suggested that if it is bothering you then try a different PSU. Pretty simple test.

    When questioned, I explained how the PSU is important. Do you guys all have voodoo dolls with pins in them, you can shadow the video bios to test PSU latency, it has nothing to do with 64-bit and does not cause the world to end. If you shadow the BIOS then it uses the PCIe bus and system memory to operate which takes a bit longer. Worth a shot... if sleep then works get a new PSU.

    A steel case ensures that the motherboard earth and the video card earth lead to the outside of the PSU, rather than traveling up into the monitor earth.
     
    Last edited: Feb 10, 2015
  7. The Mac

    The Mac Ancient Guru

    Messages:
    4,404
    Likes Received:
    0
    GPU:
    Sapphire R9-290 Vapor-X
    wow, you havent a clue what you are talking about.

    shadowing vbios copies it from bios eprom to system memory and moves the pointer to the new location, it has absolutely nothing to do with the PCIe buss.

    and it has zero effect on any PSU latency (whatever the hell that is, and i have a EE degree and am fully versed in power systems designs)

    If you are referring to either PFC (Power Factor Coefficient), or ripple, you have the wrong concepts.

    and no, a steel case makes no difference as "gound" traces exsist on the MOBO itself and is sent to the Powersupply therough the 3rd conductor on the cable thats goes to the recepticle, as does the monitor. Any additional ground paths through a case are purely redundant, as they have to back to the 3rd conductor on the power cable, same as the mobo (the black wires on the ATX connector)

    Your understanding of electricity is error.
     
    Last edited: Feb 10, 2015
  8. Spathi

    Spathi Master Guru

    Messages:
    446
    Likes Received:
    0
    GPU:
    Galax GTX980 V2
    The video card is connected to the pcie bus, which is connected to the cpu, which is then connected to system memory. Anyway I saw it work, so I have some empirical evidence. I am not saying it will work, the only real test is to swap.

    PSU latency... if different rails are coming on at different times. More likely to happen on a big PSU.

    Video cards earth to the case as well as the MB, that is why the MB has an earthing screw.
     
    Last edited: Feb 10, 2015
  9. The Mac

    The Mac Ancient Guru

    Messages:
    4,404
    Likes Received:
    0
    GPU:
    Sapphire R9-290 Vapor-X
    no such animal, that not how power supplies work.

    All rails are synchonised and a "power good" signal is sent through a special wire.

    With out that signal the PS wont engergize.
     
  10. Spathi

    Spathi Master Guru

    Messages:
    446
    Likes Received:
    0
    GPU:
    Galax GTX980 V2
    Maybe that was what was faulty with mine then.
     

  11. The Mac

    The Mac Ancient Guru

    Messages:
    4,404
    Likes Received:
    0
    GPU:
    Sapphire R9-290 Vapor-X
    now THATS certainly possible,

    ive seen "power good" signals go faulty, but usually the mobo wont engergise the PS when that happens.

    youd need an o-scope to figure that one out.

    To the best of my knowledge, the recover from sleep bug was casued by poor mobo design that didnt supply enough current to the +5v pins on the Pcie bus in combination with certain PSs.

    It was partially mitigated through bios updates, however some PSs were still problematic.

    Several Corsair models were suspect among others.
     
    Last edited: Feb 10, 2015
  12. SlackerITGuy

    SlackerITGuy Master Guru

    Messages:
    393
    Likes Received:
    1
    GPU:
    SAPPHIRE Radeon R9 Fury
    Gonna give it one more try...

    Evolve is launching tomorrow
    Tomorrow is Tuesday.

    15.2 right?
     
    Last edited: Feb 11, 2015
  13. Spathi

    Spathi Master Guru

    Messages:
    446
    Likes Received:
    0
    GPU:
    Galax GTX980 V2
    I just assumes the 12V because I thought the 5V would be OK, I did wonder.

    The PS worked for 9 years, then I noticed the other day the lights on the MB were flicking and the PC was going into sleep and not waking, so I quickly swapped it. It has always had a problem starting video cards though and I found various workarounds as the PS had a watt meter I like, lol..
     
  14. The Mac

    The Mac Ancient Guru

    Messages:
    4,404
    Likes Received:
    0
    GPU:
    Sapphire R9-290 Vapor-X
    its the mobos power regulation that's the root cause, not the Power Supply. VRMs are mostly programmable these days, which is how they managed it mitigate a lot of through bios updates.

    However, PSs are cheaper to replace than a $300 high-end mobo, so that's the route people went.

    12v is generally deactivated in S states, the 5v is necessary to "wake' the device

    Its pretty rare these days, it was a continuous issue on the P67 (cougar point) chip-set.
     
    Last edited: Feb 10, 2015
  15. Spathi

    Spathi Master Guru

    Messages:
    446
    Likes Received:
    0
    GPU:
    Galax GTX980 V2
    It was the PS for me. I assumed the 12V was "ready" a few nano seconds before it could provide power for many years (the MB can not measure the 12V that goes to the Video Card), then the 5V started to die last week. Maybe the 5V was dodgy the whole time? Anyway, swapping the PS fixed it, no problems.

    My current MB has 250A VRM to the CPU. I thought that was a gimmick? If it was the 5V then that does not explain it because the MB would wait or not turn on (like it did, when the 5V died). If it was the 2x12V direct to the Graphics all those years then that explains why the GPU/Monitor sometimes stayed off after sleep on different MB's and cards.

    Are you saying the GPU can start doing things using the 12V before the MB PCIe BUS is woken up? I had assumed it needs the BUS and CPU to start doing things, rofl.
     
    Last edited: Feb 10, 2015

  16. Espionage724

    Espionage724 Guest

    A different PSU didn't do anything for the issue for me, but I guess it could be a possible cause for some people.

    The only power-saving-related things I had enabled at the time (that I can account for anyway; Windows likes to do more-hidden stuff like dynamic ticks (which I also disabled) and stuff) was the monitor suspend, and ULPS, and ULPS being enabled or disabled also didn't have an effect on the monitor suspend bug.

    Pretty sure Legacy vs UEFI boot didn't matter either. I've ran Windows 8/8.1/10 in UEFI exclusively for the most part, and ran 7 in Legacy, and have seen it happen in all those scenarios.

    Not entirely sure what Linux is doing differently though to not have it happen, but it'd probably help to isolate the cause if I knew what was causing it in Windows lol.
     
  17. The Mac

    The Mac Ancient Guru

    Messages:
    4,404
    Likes Received:
    0
    GPU:
    Sapphire R9-290 Vapor-X
    Its a combination of the mobo VRM and the PS.

    A different mobo would have solved the issues just as readily as a new power supply.

    current on the 12v rail is irrelvent for S states (sleep)..

    unless you have 3000 watt power supply, i seriously doubt there are 250a on the 12v rail.

    on the 5v rail however, 250a would be 750 watts.

    again, the VRM on the mobo limits this. If its faulty, its not delivering to its destination what it should. Again, because its somewhat programmable, any issues can be mostly mitigated.

    for the most part, when power design is involved, voltage is always constant. Power is controlled by current control.

    Ohms law: E = I x R, and P=I^r

    solving for P, using E and I, you get P = I x E

    Voltage being a constant, power is controlled by current.

    If you need a certain power envelope to enable a function, current control is vital.

    hense the purpose of VRMs.

    (for the non-mathie people, I=current, R=resistance, E=voltage, P=Power.)
     
    Last edited: Feb 10, 2015
  18. Spathi

    Spathi Master Guru

    Messages:
    446
    Likes Received:
    0
    GPU:
    Galax GTX980 V2
    This is how I thought about it (as a defective PS design)...

    Would the VRM's slow down the powering up, that would explain why the z68 worked better than my old board. 5V+12V to MB comes on, PSU says 5V ready, delay from VRMs while 12V to GPU comes on, PC starts OK (old PC needed to copy VBIOS to come on). Out of sleep... 5V is on, 12V to MB comes on, 5V ready comes on, VRMs are already charged, 12V to video comes on too late and monitor stays off.
     
  19. The Mac

    The Mac Ancient Guru

    Messages:
    4,404
    Likes Received:
    0
    GPU:
    Sapphire R9-290 Vapor-X
    Generally, i would say no. However signaling is a complicated thing.


    From S1 (sleep), memory is never de-energized, so its never gone

    From S2 (hybernate), bios is reintialized, all functions are overridden from disk cache. Bios state would be the same from S mode initialization.

    From S3 (hybrid), depending on what mode is used, see the other 2.

    Again, its less about the power supply, and more about how the VRMs are handling current control.
     
    Last edited: Feb 10, 2015
  20. Spathi

    Spathi Master Guru

    Messages:
    446
    Likes Received:
    0
    GPU:
    Galax GTX980 V2
    Any errors is in your event log at the times you go into and come out of sleep?

    btw, when I google you CPU and "sleep" there are 10 pages of hits, one guys solution was to try another model graphics card, and it worked.
     

Share This Page