Failed/Bad Asrock Phantom Radeon 6800?

Discussion in 'Videocards - AMD Radeon' started by MrSleezeBag1964, Dec 29, 2020.

  1. rflair

    rflair Don Coleus Staff Member

    Messages:
    4,854
    Likes Received:
    1,725
    GPU:
    5700XT
    Did you ever end up testing the card in your girlfriends PC?
     
  2. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    Unfortunately I haven't yet. Needed to do some research on the Thermaltake power supply. She's been running a Vega 56 fine for over a year, but when were installing her NVME, I found that the Vega is connected with a single PCIe power cable with 2X 8-Pin connectors. The plastic on the connectors is red though.... not sure if that means they are possibly rated for higher power draw?

    Will respond to some of the other posts a bit later. Going to take some pictures of my BIOS info and run another OCCT test. I was able to get another hard shutdown running a 5 minute OCCT Power test.
     
  3. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    Quick thought.... earlier this morning I ran a quick OCCT power test and saw the CPU hit 102°. Also noticed that my CPU fan is vibrating at higher RPM's. Likley due to dust on blades. I'm wondering if these vibrations could be playing a role. The C: drive is mounted behind the mobo.

    A very interesting observation too... back when I was helping my younger cousin build his first PC last month, his "VGA" error light would come on after he moved his PC. (We are both using the Asus X570 Tuf Gaming Plus.)

    We thought we had a bad GPU, but come to find out, the NVME SSD wasn't mounted all the way. I have no idea why the VGA error light was coming on, but once the NVME was properly secured, no more VGA error light.
     
  4. rflair

    rflair Don Coleus Staff Member

    Messages:
    4,854
    Likes Received:
    1,725
    GPU:
    5700XT
    Have you tried moving the card to the other PCI-E slot?

    And your CPU should never hit 102c, not even running OCCT.
     

  5. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    OMG.

    Major facepalm. Got in to BIOS and took some pictures of settings and started researching.

    For some reason the CPU core ratio was set to 4.2 Ghz. There were no other adjustments to support/optimize a 300Mhz base frequency increase.

    I'm not sure how it happened... perhaps I accidently loaded a BIOS profile by mistake. The OC Profile created in Clock Tuner for Ryzen isn't capable of changing that in the BIOS is it?
     
  6. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    I haven't yet. I have two fans below I need to remove first. I will try it if I can't reach stability with current settings.
     
  7. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    A glimmer of hope?

    Decided to take a step back and rethinking my assumptions.

    It appears other users have had success turning off Freesync. First round of testing in Apex Legends Firing Range... About 35 minutes of stable gameplay, more than I've achieved with anything else.

    Fingers crossed. Will report back after more testing.
     
  8. JonasBeckman

    JonasBeckman Ancient Guru

    Messages:
    17,564
    Likes Received:
    2,961
    GPU:
    XFX 7900XTX M'310
    Interesting, Enhanced Sync is the one that is kinda buggy and not recommended to have enabled, FreeSync (With E-Sync still off.) I would assume should be mostly functional but it's a mix here of the monitor, monitors firmware and AMD's drivers so it's not the easiest to say what is acting up where if this mode causes problems and the display driver option I believe can show either VRR or FreeSync so while they should be the same the driver might handle it differently when hooked up to a proper FreeSync 2 Premium display. :)

    Might also be worth testing HDMI and Display Port if it happens with one or the other, connectivity issues should be better too compared to the Navi10 cards but HDMI 2.1 in particular seems to still be a bit hit & miss.
    Would hope that's mainly from the firmware though and not a hardware error though a few displays have had a lot of issues here on both NVIDIA and AMD GPU's with NVIDIA doing some driver workarounds even to lessen these issues.
    (From my understanding at least until newer monitor firmware code can fix it up as it should be working.)

    Alternatively just disable it and wait for a newer driver if that confirms it working as intended instead of messing with different cables and if that's a problem (Again more a Navi10 problem.) or if it's something with DP but not HDMI or the reverse.


    EDIT:
    https://www.amd.com/en/support/kb/release-notes/rn-rad-win-20-12-2-rgb
    https://www.amd.com/en/support/kb/release-notes/rn-rad-win-20-2-2

    Been going for a while.
     
    Last edited: Jan 10, 2021
  9. cliffgamerz

    cliffgamerz Master Guru

    Messages:
    426
    Likes Received:
    105
    GPU:
    Inno3D RTX 3070 8GB
    Your GPU Core clocks and Memory clocks are low in Furmark and the results for 6800 is definitely less than normal i think because my 3070 results are much better than yours.
     

    Attached Files:

  10. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    Thanks everybody for providing some feedback! I really appreciate it!!

    I'm going to have to make a decision to get a Newegg refund, or RMA with ASrock soon. Problem now is that if I refund and try to buy the same card to today, MSRP has gone from $680 to $835 due to the tariff exemption expiring on Jan 1st. I can afford it, sure, but also refuse to pay that much for a 6800. Lol... that's just sad!!

    Good news seems to be that Freesync disabled in Radeon and in monitor resulted in perfectly stable gaming session for 3 hours in Apex Legends last night.

    Aside from of course some screen tearing - I was able to get pretty acceptable results by setting an FPS range in Radeon Chill, and using in-game V-sync. Not great, but it works at least. Zero crashes!

    In response to previous questions and to help provide more context, here's additional hardware info:

    Displays & Settings:
    Primary monitor is an LG32GK650F-B connected with included Displayport cable. My secondary monitor is an old Asus 1080P model VS278 connected with HDMI.

    On the "Enhanced Sync" - I've never run it since I saw that in the patch notes...almost a year ago? Truth be told, I never really saw any difference running it on or off.

    Motherboard/BIOS/RAM:
    I'm running on the Asus X570 Tuf Gaming Plus:
    • BIOS Version 3001 x64 (12-4-20 Build Date
    • AGESA: "ComboV2_1180"
    [​IMG]

    Detailed RAM Info from Thaiphoon:

    [​IMG]

    GPU Frequency:
    GPU Clock Behavior in OCCT & Furmark after a few more tests today seems to result in the same ~1500 Mhz clock with near 100% utilization.

    Here's example from Radeon Monitoring while running OCCT. Temps seem great - this card runs COOL. I've never seen junctions exceed 65°. Yet, can't seem to hit anywhere near the base frequency.

    [​IMG]

    However, it appears something is limiting performance in Time Spy. I can barely seem to break above the 12,000 Mark, while the average for a 3800X and Radeon 6800 seems to be closer to 14,000. The GPU frequency is only peaking at ~1,883 Mhz.

    [​IMG]
     

  11. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    Quick note - GPU Does seem to boost over 2,000 Mhz during gaming. Did a quick run on Doom Eternal and saw almost 2,100 Mhz:

    [​IMG]
     
    JonasBeckman likes this.
  12. JonasBeckman

    JonasBeckman Ancient Guru

    Messages:
    17,564
    Likes Received:
    2,961
    GPU:
    XFX 7900XTX M'310
    That's about where it should be, 2105 being the default boost speed and then up and down a bit from how AMD is scaling these.
    Was wondering if there was something in regards to the GPU and power draw and going up to higher clock speeds which would then cause instability.

    But it seems it might just be the display firmware or a driver bug in regards to FreeSync I'll have to check but if toggling that off has resulted in stability then yeah that might have resolved it.


    I'm on the 850-GL from LG myself first thing I did was plug that USB-A cable in and go through the slow and annoying process of their back then earlier OSD 4.90 something software to get the firmware updated.
    Display Port 1.4 here not that it should matter too much HDMI is really robust and solid up to 2.0b at least.

    This is a VRR display though variable rate refresh it doesn't do G or Free sync it had some quirks but not on the level some of the newer Samsung G Odyssey or other HDMI 2.1 displays had going by NVIDIA Ampere reports and some few AMD Navi20 ones.

    Some of the bundled cables are not that great but should still be serviceable without needing the full fee for these certified rated cables of way too expensive pricing aside from some of the potential issues in regards to length but others are downright broken or can even be somewhat poorly configured if not dangerous to the hardware display and GPU both but again that's the cable and display in general not something that adaptive sync would be affected by.

    Digital after all so it either sends or receives or it does not and if it loses signal it can often be restored it shouldn't be some bigger crash from it. :)


    So overall, nice!
    Everything seems like it's working after all just that something with FreeSync / adaptive-sync looks like it's causing some problems but the hardware appears to be just fine without that enabled.


    EDIT: Oops missed a little bit.
    Oh and yeah if you're doing multiple displays that can be buggy.

    But I am now having problems finding recent posts about that so perhaps it got fixed.
    Still finding plenty of reports and comments in regards to flickering but that's mostly in regards to VA panels and it's not a crash it's just annoying display flickering.

    There's also the higher memory clock speeds on idle related to VESA and timings I believe is how it's related to why the GPU VRAM can run at full speed or also when using a high refresh rate single display.

    CRU or Custom Resolution Utility can tweak that if desired.
     
    Last edited: Jan 10, 2021
  13. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    Interesting... I just checked and I don't believe one can update this monitor's firmware. No USB port. Only 2X HDMI and 1X Display port.

    Much to my dismay, this monitor cannot do over 1440p/60Hz via HDMI. Found that out the hard way when I plugged in my Xbox Series X.

    In regards to higher memory clock speeds, it appears VRAM is showing ~1980 MHz at 5% utilization under normal use. (Internet browser open).

    upload_2021-1-10_15-29-14.png

    I'm beginning to think there could be a deeper problem afoot... right after making my previous post, the computer locked up and restarted. No VGA Error light on Mobo, but I did find this in Windows Event Viewer:

    upload_2021-1-10_15-30-48.png
     
  14. BuildeR2

    BuildeR2 Ancient Guru

    Messages:
    3,208
    Likes Received:
    437
    GPU:
    ASUS 4090 TUF OG OC
    After catching up with all the new information, you aren't still trying to run with 1800FCLK/3600MHz memory speeds are you? Especially after seeing WHEA errors on the CPU, I would make absolutely sure things are at stock speeds.

    CPU and/or FCLK instability can make it seem like a GPU is on the fritz. I was testing at 1900FCLK back when Mechwarrior 5 released and thought my 2080Ti was dying, but it was just the game CTD'ing and PC rebooting due to massive WHEA errors since my 3600X can't handle Infinity Fabric over 1833.
     
  15. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    Initial testing with Freesync Off that achieved stability was done at 3200 RAM/1600 Mhz Infinity Frabric.

    BUT.... I decided to take a somewhat different route and haven't done much more testing. (Have had to work 11 hours on Monday & Tuesday.) I had an unused Samsung 2TB NVME I needed to put to work. Did a fresh install of Windows and working to restore all the files & get organized.

    Over the years I've created a bit of a mess...
    upload_2021-1-12_22-58-55.png

    My previous version of Windows was actually installed on an Intel Core i5 6600 system. Figured I'd take this opportunity to get everything cleaned up, fresh installs of games, and remove some of my old SATA drives before moving forward with additional testing.
     
    BuildeR2 likes this.

  16. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    Curious though... I had been running my RAM and CPU overclock with absolute perfect stability for at least 6 months. I'm confused on why introducing a new GPU to the mix could start causing instability?
     
  17. MrSleezeBag1964

    MrSleezeBag1964 Member

    Messages:
    36
    Likes Received:
    4
    GPU:
    Radeon 6800 16 GB
    Quick update: With default speed settings now applied to RAM and Infinity... And running 3DMark Timespy from a fresh NVME C: drive with fresh Windows install... Out of 5 runs, 3 of them hard crashed, GPU fans kick on the afterburners, and I get the Mobo error light.

    I called up ASRock yesterday to ask about the RMA process. They offered to let me speak to their technicians.

    I fear that RMA times could be 90+ days. But with this model now selling for $840 on Newegg, seems that exchanging/refunding to Newegg isn't the best idea since I only paid $680.

    ASRock also responded in email saying the RMA can take ~5 business days, however I'm pretty sure they are wrong.
     
  18. BuildeR2

    BuildeR2 Ancient Guru

    Messages:
    3,208
    Likes Received:
    437
    GPU:
    ASUS 4090 TUF OG OC
    Dang, seems like you may have to return or RMA that graphics card. If you aren't able to re-paste it without consequences then you are reliant on Newegg or ASRock at this point.

    That is not good to get hard crashes with a fresh OS and default hardware settings. Regarding my last post, GPU's from different vendors or in totally different classes (like 480 to 6800) will use your hardware in different ways and push things harder likely exposing instability.
     
  19. ninja750

    ninja750 Master Guru

    Messages:
    366
    Likes Received:
    89
    GPU:
    RX 6800 REF
    I remember reading about whea errors with first bioses for ryzen 3000, then gone with the latest
    specially with 570 tuf
     
  20. Astyanax

    Astyanax Ancient Guru

    Messages:
    17,011
    Likes Received:
    7,352
    GPU:
    GTX 1080ti
    they were never entirely fixed, AMD turned off the AER reporter for the PCIE controller.

    there are people literally with machines that can't be stabilised because of it.
     
    JonasBeckman likes this.

Share This Page