Jump to content

timohour

Registered User
  • Posts

    295
  • Joined

  • Last visited

  • Days Won

    12

Everything posted by timohour

  1. I looked on many hackintosh sites but I can't find the reason those glitches occur. It should be a BIOS setting not a hardware problem since it works with a BIOS revision but not with another. Do you have any other info that can help us narrow it down? you can check other variables related to the iGPU... I can't tell which one could do the trick. Setting: DVMT Total Gfx Mem, Variable: 0x1C8 Option: 128M, Value: 0x1 Option: 256M, Value: 0x2 Option: MAX, Value: 0x3 End of Options Setting: Aperture Size, Variable: 0x17D Option: 128MB, Value: 0x1 Option: 256MB, Value: 0x2 Option: 512MB, Value: 0x3 End of Options Setting: GTT Size, Variable: 0x17C Option: 1MB, Value: 0x1 Option: 2MB, Value: 0x2 End of Options Setting: Gfx Low Power Mode, Variable: 0x178 Option: Enabled, Value: 0x1 Option: Disabled, Value: 0x0 End of Options Setting: Graphics Performance Analyzers, Variable: 0x179 Option: Enabled, Value: 0x1 Option: Disabled, Value: 0x0 End of Options keep a note which was the default one before changing it. I have never touched any of the above settings so I don't know how your laptop will react.
  2. Great news! I am happy that you finally did it. I think that I might understood what caused the problem. The OC_ME region supplied by khenglish is based on an A13 (if memory serves) FTP dump. This is probably why it didn't down/upgrade the BIOS when you tried. It seems like there was a mismatch between BIOS version and ME FW version. So, from now on I recommend that everyone who wants to patch his ME FW in order to enable BCLK OC, to mod his own ME FW version instead of using the one supplied. This way we probably won't have such problem again. What's the default value of 0x1c7 for the A07 version? You can try all values with the A17 to see if anything changes... Option: 32M, Value: 0x1 Option: 64M, Value: 0x2 Option: 96M, Value: 0x3 Option: 128M, Value: 0x4 Option: 160M, Value: 0x5 Option: 192M, Value: 0x6 Option: 224M, Value: 0x7 Option: 256M, Value: 0x8 Option: 288M, Value: 0x9 Option: 320M, Value: 0xA Option: 352M, Value: 0xB Option: 384M, Value: 0xC Option: 416M, Value: 0xD Option: 448M, Value: 0xE Option: 480M, Value: 0xF Option: 512M, Value: 0x10 Option: 1024M, Value: 0x11 It most probably won't have any difference but wouldn't hurt if you tried. (I know it's a dirty job, but somebody's gotta do it)
  3. offcourse it does.. use the dump of your original BIOS to revert to «stock ME FW». If I understand your question correctly, you will flash the old ME FW using the command you used to flash the OCed one.
  4. I have no other ideas... Last resort would be to revert your ME FW from the original dump you made and see if it makes any difference (I doubt it)... Let us know how it goes. Did you get this laptop with A16 or did you just updated?
  5. And you tried as you said with Windows and Freedos? and the result is the same? What is the result when you try with Freedos? It is maybe useless but would you like to test the other older than your A16 BIOS available? Is the result the same? A15,A15 24 Jun 20144:09:16 PM A14,A14 28 Jan 20143:22:08 PM A13,A13 17 Oct 20137:51:00 AM A12,A12 18 Jul 20131:31:34 PM A11,A11 03 Apr 20139:51:00 AM A09,A09 08 Jan 20132:16:51 AM
  6. Default is 0x0 Did you try setting your DVMT allocation memory with your current BIOS version? Did it have any difference?
  7. It's been a while for the E6530... I was advised from Local Dell Pro Support to update to Α17 for better W10 support. It was time to introduce that for the E6430... Without changing anything (regarding the Flash Descryptor) try to remove the CMOS battery for 10-20 seconds powerdrain the machine and try again to flash the BIOS version you want. Also try setting 0x1c7 to 0x1 (32MB) or 0x3 (96MB) and check if you still have glitches in your OSX installation running your current BIOS version.
  8. I think that the best he has to do first is use UEFI variables to change the DVMT allocated memory to less than 96MB the UEFI variable for DVMT Pre-Allocated memory is 0x1c7. So, following the information above, check what is the default allocated memory.
  9. A user reported a while back that flashing the BIOS region from one laptop to another (actually one motherboard to another) bricked his E6430. So, I wouldn't recommend it unless you have a hardware SPI programmer handy.
  10. If you unlocked your descryptor when you first got your dump you opened a hex editor, and looked for this HEX string 00 00 0B 0A 00 00 0D 0C 18 01 08 and change it to 00 00 FF FF 00 00 FF FF 18 01 08 the first FF FF where to unlock your Flash Descryptor region while the latter where to unlock your ME FW region. If you revert the second FF FF to 0D 0C you will lock ME FW region and probably fool the BIOS to continue with the flash. If you revert your Flash Descryptor region (first FF FF) you will have to reapply the pinmod to unclock. I assume that locking the ME FW region could probably solve this non-updating-BIOS issue. Try it and report back. I have never tried it so I am not taking any responsibility if you brick your machine. I am positive it will work though. EDIT: Also, try what kondilac suggested above and change the memory allocation of your iGPU to anything less than 96MB...
  11. There was another user that reported that he couldn't apply a bios update after applying the ME FW hack. I didn't have any issue in my E6430 to upgrade from A11 to the latest a few months back. It can probably be a problem because your BIOS try to unlock the ME FW region which is already unlocked. Try to lock again the ME FW (not the Flash Descryptor region) and check if the same problem occurs. I never tried OSX on my E6430. Does everything work? (I currently use OSX only on my desktop)
  12. Ok thanks. Have you set new limits with UEFI variables and it still downclocks to x37 or is it the default limits?
  13. Check then with the approved adapter @ x38 what's your consumption @ 3.8GHz? Will it work? (use throttlestop to lower the multi)
  14. Format a USB 2.0 flash drive to FAT32 place the file under EFI\Boot\bootx64.efi Start up your Dell press F12. Select the UEFI option of your flash drive. Check the NEWBIE GUIDE for the rest. You can place it in your ssd's recovery partition too in the same directory..
  15. I am sorry . You are right this is the note but I failed to explain it on the first port that this is for Optimus compression (it does say so that on the post headline though ) But again, not clear so my fault. Connecting eGPU after post though would have the same effect. This is the hack for a more permanent solution.
  16. Correct info. I was talking about the latest result... There is no secondary card there If he boots with eGPU disconnected though and hotplug after post, iGPU will be set on and primary. He will need to set the EFI variables for a more permanent solution. Just keep it informative...
  17. OK last favor. When you go home, try to startup the laptop as it is now (PCI Bridge disabled) confirm that it boots with HD4000 enabled and then reboot and hotplug the GTX670 AFTER post. It is reported from other users that with the E6x30 series when you connect an eGPU before boot, it will disable internal screen and iGPU making impossible for Optimus compression to kick in. This is maybe why in your x230 score it states HD4000 as secondary GPU while in your E6430 score there is no secondary GPU as noticed by @Tech Inferno Fan above. Also check your x1.2 bandwidth using CUDA-z performance on both your x230 and your E6430 cause some other users reported that they had trouble establishing a stable x1.2 link. Such test will maybe prove that the E6430's ExpressCard link is problematic. I never had such problem connecting my R9-280x to my E6430 and my scores where on par with the HP Probook 6470b I owned. Keep in mind though that my E6430 is one of the latest (Aug 2013) and there was maybe a silent revision.
  18. This result is with only HD4000 and GTΧ670 shown in Device Manager? and pls when/if you have time do the CUDA-z performance test to see if it gives lower values than with your x230.
  19. in order to easily disable the NVS5200M within Windows go to Device Manager -> System devices and disable NV2000 PCI bridge (Xeon® processor E3 - 1200 v2/3rd Gen Core processor PCI Express Root Port - 0151). In order to do this you should set Optimus Enabled in BIOS and make sure that Intel HD4000 is enabled in Device Manager. This way your dGPU will be disabled and Optimus compression will be enabled. Try that, reboot and report back.
  20. Do you have Optimus enabled in BIOS? I assume yes. I also assume you checked with GPU-z that your card runs @ x1.2 speeds. Could you try to disable your NVS5200 inside Windows? You can do it easily if you go to Device Manager -> System devices and disable the PCI Express Root Port that your dGPU is attached. Then try again to OC your GTX670 and check if the results change. On the second link you provided the results on the x230 are with an OCed GPU while with the E6430 it seems like it is not OCed (check under the graphics) that's what caused the huge difference. There is still a big difference though in the non OCed results. If you still have the X230 could you please run CUDA-z and post a screenshot of the performance tab for each laptop? (use the heavy load mode test) It will be helpful to see if there is a big difference in the host to device or device to host benchmarks.
×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use. We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.