Tarot
-
Posts
1 -
Joined
-
Last visited
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.
[GUIDE] DSDT override to fix error 12
in DIY e-GPU Projects
Posted
I have a toshiba satellite laptop with Windows 10 Pro 64, Intel HM55 motherboard, core i7 cpu, 16g ram, with nvidia 310m dgpu. I am setting up a new external gpu using the PE4C-EC v3.0 interface, with 730 watt psu from an older system. The new gpu is MSI geforce 1060 gtx (gaming) with 6g RAM. The large memory is present and has no problems.
My question is: The graphics card has 6g of ram so does that mean that i need to set aside that much for the large memory?
I am using EGPU 1.30, RW Everything and DSDT Editor. I have compiled the DSDT per the instructions at the beginning of this thread several times. I used the same space (2g right?) for my large memory as the op used in the demonstration . The compile resulted in a 0 error, 0 warning dsdt.dat. I uploaded the resulting aml file with no problem. I can run the compaction with success using the 56g end point in 36 bit space. However, my status window in egpu 1.3 shows:
tolud: n/a
dsdt: no
using devcon_txt: yes
pciw: yes
all-alloc: yes [128 mb]
this is the relevant section of the pcidmpw.txt i ran from the egpu 1.3 diagnostics.
PCI\VEN_10DE&DEV_1C03&SUBSYS_32811462&REV_A1\4&20119B3C&0&00E2
Name: NVIDIA GeForce GTX 1060 6GB
Device has no resources reserved.
but this is the pci map... which clearly shows resources being used.
PCI bios found, version 03.00, last bus # 1f
codes:
n\p: (not) prefetchable, r\R: ROM ((dis\EN)abled), i: IO space
start end c bus device fun bar venid devid classcode
0000000c 0000000d i 0c 00 0 4 10de 1c03 030000 display
00002000 0000207f i 0c 00 0 5 10de 1c03 030000 display
00000000 0fffffff p 0c 00 0 1 10de 1c03 030000 display
00000000 0000000f p 0c 00 0 2 10de 1c03 030000 display
00000000 0007ffff r 0c 00 0 8 10de 1c03 030000 display
fd000000 fdffffff n 0c 00 0 0 10de 1c03 030000 display
fe000000 ffffffff p 0c 00 0 3 10de 1c03 030000 display
when i chainload i get a black screen and have to force a shut down. I have tried every possible combination of compaction (both dgpu/egpu, just egpu, ignoring dgpu, etc.) When i do not chainload, the 310m works, but the 1060 still has the code 43. When i chainload and do get past the black screen both of the nvidia graphics cards are on the bus/in device manager, but they both show an error code 43. I even tried uninstalling the 310m and installing only the 1060 but even alone the egpu has code 43. I am trying to figure out why it shows up in device manager but will not get past this error. I have tried at least a dozen driver releases. is there something that i'm missing? I have been trying different things for 2 months now. I give up.
I hope someone can tell where i have gone wrong.
I should add that i was trying to use the dsdt replace method, but the new table is the tiniest bit larger than the previous table since i added the declaration for the large memory. Should i use it anyway?