I finally retired my two HD 5850's and jumped onto the HD 7000 series bandwagon .
I bought myself a Sapphire Radeon HD Dual-X 7970. It's the non-Boost, non factory overclocked version. It uses a custom blue PCB (non-reference board). It goes by the model / revsion number 11197-11-40G.
From what I've found out, this card has a lower stock voltage than most; it uses 1.075V versus the more commong 1.175V or 1.2V. GPU-Z indicates my particular chip has a ASIC Quality of 75% (still kinda gimmick IMO).
I am having trouble running the card much past stock (yes I am overclocking it). It has no problems running at stock 925 / 1375 speeds.
With the overclock, the drivers seems to constantly crash and recover..over and over again while in-game, as a result, black screen. I will also run into a DirectX Error in BF3. To be more clear
DirectX functions "GetDeviceRemovedReason" failed with DXGI_ERROR_DEVICE_HUNG: Device hung due to badly formed commands...
From what I've seem to have discovered the card requires more than the stock 1.075V to run beyond 1000MHz core. The problem seems to be cuased by the voltage resetting back to the stock value even though I set it in the overclocking software.
I know the voltage for my card is not locked; GPU-Z, AIDA64, and HWMonitor detects the voltage if I change it in the software.
The logs from MSi Afterburner and Sapphire TriXX shows that the voltage resets...which makes me assume causes the overclock to become unstable...and the driver "dies."
I've disables ULPS in registry and in TriXX, I've disabled PowerPlay in MSi Afterburner, Enabled unofficial overclocking in Afterburner, and Force constant voltage in both MSi Afterburner and TriXX. Non of this seems to fix the issue.
I've even tried the 13.3 Beta driver from AMD (the notes say it resolves hanging issues). I've also used the non-beta 13.1 drivers. I've installed the latest CAP (12.10?) from the driver page.
I wouldn't have both Afterburner and TriXX running at the same time (duh), but some say that even though they are not running, they can interfer with each other. So, I had only one of each installed at a time - no help.
I had Graphics OverDrive enabled; maybe that was causing issues so I disabled it - again, no help.
I know the drivers are constantly crashing and recovering...because it says in the Event Viewer Logs AND if I Alt-Tab out (if the PC doens't fully lock up) I can see the driver crash icon in the bottom-right Icon Tray.
" Display Driver amdkmdap stopped responding and has successfully recovered...Event ID 5101."
Examining the GPU-Z and MSi Afterburner logs (if exiting the game is sucessful or rebooting my PC), the Core & Memory (overclocked) frequencies were very consistent. I had the voltage...say set to 1.186V, during in-game the voltage would vary from 1.15V to 1.18V (1.174V, 1.162V, 1.158V, etc). I understnad voltage droop plays a role so...
Near the end of the log, the Core & memory remains at the set overclocked value (i.e. 1150 / 1385) BUT the voltage has dropped down to ~1.049V for a few seconds.
Shortly after, the speeds drop down 500MHz Core and 150MHz Memory (2D clocks?).
The card then increases its clocks back up to 1150 / 1385 BUT the voltage remains down at ~1.049V!
I believe where the voltage drops down to 1.049V and the card's frequencies dips and what-not is where the drivers crash, DirectX Hang issues, and everything goes to hell...
I might just flash another card's BIOS that runs at stock 1.175V / 1.2V and see if that helps )since my card does have the Dual BIOS and BIOS switch).
EDIT: System specs is in the sig...but I should update that (did a few upgrades to it). Done!
Phenom II X6 1090T @ 4.1GHz // ASUS Crosshair V Formula 990FX // Sapphire Radeon HD Dual-X 7970 // Thermaltake Frio w/ push-pull (using Antec Formula 7 Nano Diamond paste!!) // 8GB (2 X 4GB) G.Skill RipJaws DDR3 1600MHz @ 8-9-8-34-21-1T // Corsair TX850 850W // Corsair Force GT 120GB SSD (OS) // Kingston 200 V+ 120GB SSD // WD Caviar Black 1TB // Windows 7 Ult x64 // Logitech X-540
02:29 AM by