Jump to content

blowntaha

Registered User
  • Posts

    213
  • Joined

  • Last visited

Everything posted by blowntaha

  1. Wondering if someone could help me out with this error 12. I've got an AIMB-272 mini-ITX motherboard with a QM67 chipset. It's not exactly a noteboook, but it uses laptop-sized RAM and CPU (2720QM) and has a x16 PCIe 3.0 slot. I've got a GTX 970 on here and was working just fine until I installed a new Windows 10 build and started to get error 12 on my iGPU (Intel HD 3000). I can run some games on here but it crashes whenever I run GPU-intensive programs; I can't even run the generic GPU-Z render test in fullscreen for more than 2-3 minutes before it force restarts. I tried rolling back my Windows 10 build to what it was including a fresh install but same problem. I also get a black screen whenever I try to install any NVIDIA driver and have to blind restart my computer (by pressing Win+D -> Alt+F4 -> Down -> Enter). Whenever I try doing a DSDT override, it will successfully compile and apply, but error 12 persists on my iGPU upon restarting. I am aware that DSDT override is for eGPU setups and might not even fix my GTX 970 from crashing but I am just trying everything to get things back to normal. @Tech Inferno Fan
  2. It's been a while since I used that laptop, but I think what I did to get it working was to disable the iGPU in device manager in windows, then reboot and go into BIOS and set Primary Display to auto, and Always Enable PEG to enable/on. I don't think I even did a DSDT override for that model, might be wrong. You might get a yellow bang in device manager under iGPU once your eGPU starts working, but nothing to be alarmed over.
  3. I wrote that guide so that you would not run into any trouble under any circumstance (that I could think of). IME it was pretty much plug and play, aside from trying different mPCIe ports. and if I remember correctly, I think I maybe had to disable the iGPU in device manage
  4. "worked" as in you successfully dumped it or successfully OC'd it?
  5. I'm using a mini-ITX mobo with a sandy bridge 2720QM CPU, w/ +3.4 MHz I got a +200-300 higher 3dmark11 bench. If BCLK control is what you're aiming for then you have to change the clock range definition record in addition to the hex bit you mentioned above. FYI, on my mobo when I flash a modded ME FW matching the clock range definition record from ocme.bin in OP, I also have issues with booting (stuck at splash logo screen). Khenglish suggested I change 0x0180 to 0x0800 in DIV2 for sandy bridge, after doing that change I was able to successfully boot; although, I have to turn off and turn on several times before I can get it to POST. This is probably the value that needs tinkering with for an ideal FW.
  6. An easy way is to look at the surrounding bits to see if they are the same when modifying me.bin For example, when you find the bit differences when comparing the two images you built in FITC, your me.bin should have similar looking values surrounding the change. Searching for 3 bits is not specific enough. Try searching for a series of 5 consecutive bits in the line above or below the value you want to change. Also, you don't need to worry about coverting into little endian because that can get confusing and make u wanna pull ur hair out lol
  7. I discovered a workaround for those trying this on a 6-series chipset: Download HxD Open the ocme.bin from the OP in FITC Go here and change Number of Flash Components to zero After doing that, tap f5 to build it and leave FITC opened. Locate outimage.bin which will either be in the root directory of FITC or in a folder called ocme or Build. Once found, rename it and copy it to another location, I'll call this OC.bin Open another instance of FITC with your ME FW dump, let's call this me.bin Go back to the previous FITC instance (ocme.bin) and change all of its settings to match me.bin Build the image, locate the new outimage.bin (same location as before), verify it is the one you just changed by looking at the timestamp and rename it to original.bin Open original.bin and OC.bin in HxD, tap CTRL + K and then enter, this will compare the two images and highlight each different bit which coorresponds to the changes made in FITC (converted into little endian but you don't need to know that). You can use F6 to cycle to the next difference or Shift + F6 for the previous difference Open your me.bin in HxD. Proceed to carefully change ONLY the bits which are different between original.bin and OC.bin. DO NOT change it according to the line number/bit location because they will never be aligned. Instead, locate a series of 3-4 bits around the target difference and use CTRL + F to find the same series of bits in me.bin, then change the necessary bit accordingly to match the value in OC.bin After you're done, save the file and flash using FPT.
  8. I came into three "industrial" motherboards, AIMB-272 QM67 chipset, 6-series ME7.x. I remember reading somewhere the FITC for ME7 is bugged? Is this also true for the 5mb variant as well? I have three of these so no biggie if it bricks, just wanted to first confirm that 6-series FITC produces corrupted FW images before I invest time into manually modding it in hex. Odd how it's not enabled by default, almost all of the necessary values in FITC are already set except for two values.
  9. That's only for dx9 though, right? I don't typically play games that are dx9, especially more so since dota 2 is now fully dx11. But thanks for the tip.
  10. I'd rather not go for anything pre 6-series, but thanks for the info. If by CPU swapping you mean changing CPUs, most if not all laptops, especially 1st gen i-CPUs do 'support' CPU swapping. But the best 1st gen i7 just doesn't cut it with me :\
  11. I don't know much about PCI 1.0/2.0/3.0 because it has never really been an issue for me, I've used my PE4C on about ~ 10 laptops. I may go for the E6430, but first I'm trying to see if I can do a x2 link on my E5420. I don't have its expresscard attachment and only the mobo though, but would much rather buy that part than a whole new laptop.
  12. I was looking around for laptops to buy that are capable of x2 link on my PE4C v2.1 w/ my GTX 970. I have lots of ivy bridge CPUs so preferred to be 3rd gen intel with HM77 chipset (or any other chipset with an "Extreme" SKU). I had a hard time finding a laptop that is both capable of x2 link and have extreme SKU. The eGPU candidates list is not as complete as I'd like it to be, so here I am asking anyone who knows of such a combination. ^.^
  13. You should really try searching the thread to get your answer because this is something that has been answered many times. Modify platform.ini in notepad by changing RESSDIN to RESSEN. I think it's one of the first couple lines.
  14. @cpzz i think what you're seeing is ur gpu changing power states. It shouldnt be a problem when under load, at least i dont think. You could run a bench and log gpuz stats to file to see the voltage fluctuations
  15. Instead of making temporary reference vars, etc you could use the DSDT parser and follow the guide here by angerthosenear. Even if you aren't trying to do simultaneous iGPU + dGPU + eGPU, it makes everything much easier. The only extra required step not included in the linked guide is to copy/paste or extract everything from the attached .zip file, along with the necessary iasl & asl files to your C:\program files\java\jre1.8.xx\bin directory, open admin cmd from that directory and do all the commands from there.
  16. First, specs: Model: HP Pavilion dm4-1009tx CPU: i5-430m GPU: gtx 970 via PE4C-mPCI eGPU: Have to boot with a wifi card, put windows to sleep, and then hotplug egpu due to whitelist. RAM: 2 GB TRYING to get 8 GB since that's the max this old cpu can support.. After (several) successful DSDT overrides, I still get error 12. When I say successful, that means I added the QWordMemory entry in the correct location, compiled with few errors and then doing the override after some minor debugging. Tried different versions of iasl, with/without egpu connected as well as both x86 and x64 java versions. I did this about 5 times and still get the error.. Could not find sufficient solutions searching, or maybe im just bad at searching =d does anyone have an idea or a viable solution? thanks
  17. Too noob a question to be answered... ? Reason I ask, other than to find out if I need to get another CPU, is because I'm finding mixed answers about whether or not you can use Ivy Bridge CPUs on mobos intended for Sandy Bridge...
  18. Hi, My laptop bugged out on me and I'm trying to use a mobo I had lying around as a replacement, waiting on its DC cable to arrive at the moment. In the meantime, looking for a suitable CPU, could I use my 3920XM on it? The mobo belonged to a Dell Latitude E5420 and the CPUs that usually came with it were with 2nd gen i5s/i7s (example 2529m), but has PPGA988 which is the same socket as 3920XM... will this work? Mobo model: from a Dell Latitude E5420 Chipset: HM65
  19. Disclaimer: I don't know of any instances where this has caused a brick, but that doesn't mean it won't happen. I'm highly doubtful of this method's potential for bricking a laptop. In the event it does, just do a regular BIOS recovery using stock bios. dont blame me for anything ! Here's a short how-to for unlocking the flash descriptor on most laptops where the bios that is not encrypted, regardless of manufacturer. All you will need is a hex editor, your FULL bios image, and know how to perform BIOS recovery. Obtain/download a copy of your stock bios. Doesn't have to be stock, but just to be safe. Extract it to a folder (if it's not already), and open the bios image in a hex editor, Search for these bits, 5A A5 F0 0F 03 00 04 If you're in the right spot, 5 lines below that should be where your flash descriptor's lock is set. The block of hex bits for the desc region should look something like this, Save it somewhere. Perform BIOS recovery using your newly modified bios, afterwards you should have ME r/w access. (if required) re-install using the latest IMEI drivers. The reason you can't just flash this as you normally would when updating BIOS is because ME region does not get updated during typical BIOS updates, only the BIOS region does. When you do recovery, it flashes all available regions, including ME.
  20. So... I have an SSD which I'd been using for about 3 years now, but due to its small size I switched over to my higher-capacity HDD and started using a 14GB RAMdisk. Over the years I frequently switched the SSD back and forth between my laptops for various reasons, to the point where the plastic pieces (protectors?) that are attached to the pins have fallen off (see spoiler if you don't know what I mean). This wasn't a major concern for me since I could still plug it in and get it working, but worth mentioning. When I decided to switch back to my SSD, my laptop wouldn't turn on. If I remove the SSD, it will turn on; also if I power on my laptop with no HDD attached, then hotplug the SSD, the laptop immediately (+/- 1-2s) shuts off. HDD still works fine using the same SATA HDD slot. ~Although I doubt it matters, here are my specs: Model: HP 15-r210dx Notebook CPU: 5200U GPU: GTX 970 eGPU RAM: 16 GB SSD in question: Kingston HyperX 90GB Anyone have ANY idea WTF is going on? -Thanks
  21. For ME8, you can issue the cct sl command. This has to be done before "EOP." In my experience, unless you have a programmer and you're using it in conjunction with the Clock Commander Tool, enabling all ICC registers this way is a dead end because issuing that command before EOP is impossible or unknown. Alternatively, you can enable all/most ICC registers via FITC.
×
×
  • 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.