Jump to content
Slice

RadeonMonitor problems

129 posts in this topic

Recommended Posts

I think RadeonDump can onyl dump (read) not write.

PS: I found that newest Darwindumper has perhaps newer Version (finder show conent, Tools..) of RadenDump.

With that version (same dump values) the range end works better in case shows only the range wanted, not always a complete line - more bytes shown as given range.

RadeonDump -r 6b4,6b4

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/GFX0@0/RadeonPCI
it matched on name "ATY_GPU"
0x06B4: 0x0001E083
GA-EP35:~ andreasm$ 
 

https://bitbucket.org/blackosx/darwindumper/downloads/

 

Do we need that DirectHW.kext (also within DarwinDumper) to write in some addresses?

I dont have installed  - i think also Darwindumper may kextload that at runtime? And how could we write in that adress space you told us?

Share this post


Link to post
Share on other sites
Advertisement

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/PEGP@0/RadeonPCI

it matched on name "ATY_GPU"

0x0000:  FFFF7E5C FF404138 00000000 00000000 00000000 00000000 00000000 00000000

0x0020:  00000000 00000000 00000000 00000000 100100A2 00D816E6 100100A2 00D816E6

0x0040:  00000000 00000000 00010000 00010000 00000000 00000000 00000000 00000000

0x0060:  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000

0x0080:  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000

0x00A0:  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000

0x00C0:  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000

0x00E0:  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000

0x0100:  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000

 

my 290x


Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/PEGP@0/RadeonPCI

it matched on name "ATY_GPU"

0xFFFF7E5C:  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000

 

Share this post


Link to post
Share on other sites

Hi, Slice.

I think the working RadeonPCI.kext (for use with that RadeonDump) is a few posts later. http://www.insanelymac.com/forum/topic/317915-clover-test-and-patches-for-polaris-gpu/page-8?do=findComment&comment=2340126 (otherwise all results are zeros)

How do you mean your range: in range 6A0,40 ... If crash then 6B0,10

In the examples the command looks:

RadeonDump2 -r 0,200 , how are the parameters for in range 6A0,40 (you mean start=6A0 and 40 length, so 6A0, 6EF would be OK ?

RX 460 0x67ef , Sierra working (for this test: no RadeonMonitor in use, only kextutil  RadeonPCI.kext, in Clover Lilu + whatevergreen )

RadeonDump2 -r 6A0,6FF

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/GFX0@0/RadeonPCI
it matched on name "ATY_GPU"
0x06A0:  00000000 AAAA5555 00000000 AAAA5555 0003F010 0001E083 00000000 AAAA5555
0x06C0:  00000000 AAAA5555 00000000 AAAA5555 00000000 AAAA5555 00000000 00000064
0x06E0:  00000002 00000000 00000000 00000000 00000000 00000000 00000000 00000000
 
/RadeonDump2 -r 1700,1750  (found that you looked for that range in the past : ) )
Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/GFX0@0/RadeonPCI
it matched on name "ATY_GPU"
0x1700:  00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
0x1720:  00000000 00000000 00000000 00000000 00000000 00010000 00000000 00000000
0x1740:  00000200 00000000 00000000 00000000 00000000 00000000 00000000 000
 
​RadeonDump2 -r 5400,5450 (found that you looked for that range in the past : ) )
Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/GFX0@0/RadeonPCI
it matched on name "ATY_GPU"
0x5400:  00000000 00000001 00000000 00000000 00000000 0000001A 00000002 00000002
0x5420:  00000000 00000000 00000800 80000000 00000100 00040000 00000000 00000000
0x5440:  00000000 00000000 00000000 00000000 00018003 00000000 00040040 4010E110
 
PS: attached the used two tools.
Usage:
sudo chown -R root:wheel ~/Downloads/RadeonPCI.kext

sudo chmod -R 755 ~/Downloads/RadeonPCI.kext

sudo kextutil -v ~/Downloads/RadeonPCI.kext

..

~/Downloads/RadeonDump2 -r start,end 

I seems have no those sources. Did you have it?

Share this post


Link to post
Share on other sites

Corrected RadeonPCI.kext. Hope it will work on most Radeon cards.

Utility I'll remake later.

attachicon.gifRadeonPCI.kext.zip

not worked at my machine

Picos-Mac-Pro:radeon pico$ ./RadeonDump1 -r 0,200

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI

it matched on name "ATY_GPU"

Assertion failed: (KERN_SUCCESS == kr), function main, file ./RadeonDump1.c, line 301.

Abort trap: 6

 

with old RadeonPci.kext and newesr RadeonDump1

Picos-Mac-Pro:radeon pico$ ./RadeonDump1 -w 6b0,c0300014

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI

it matched on name "ATY_GPU"

Picos-Mac-Pro:radeon pico$ ./RadeonDump1 -r 6b4,6b4

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI

it matched on name "ATY_GPU"

0x06B4: 0x00000000

Share this post


Link to post
Share on other sites

Same for me, newer RadeonPCI.kext has that Abort Trap: 6. (AMD RX 460)

 

Using older RadeonPCI.kext together witj new RadeonDump1

 

RadeonDump1 -n 6b0,c0300014
Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/GFX0@0/RadeonPCI
it matched on name "ATY_GPU"
0xc0300014: 0x    5a2d
bash-3.2# 
 
RadeonDump1 -r 6B0,6BF
Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/GFX0@0/RadeonPCI
it matched on name "ATY_GPU"
0x06B0: 0xC0300014
0x06B4: 0x00005028
0x06B8: 0x00000000
0x06BC: 0xAAAA5555
 

 

Share this post


Link to post
Share on other sites

not worked at my machine

Picos-Mac-Pro:radeon pico$ ./RadeonDump1 -r 0,200

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI

it matched on name "ATY_GPU"

Assertion failed: (KERN_SUCCESS == kr), function main, file ./RadeonDump1.c, line 301.

Abort trap: 6

 

with old RadeonPci.kext and newesr RadeonDump1

Picos-Mac-Pro:radeon pico$ ./RadeonDump1 -w 6b0,c0300014

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI

it matched on name "ATY_GPU"

Picos-Mac-Pro:radeon pico$ ./RadeonDump1 -r 6b4,6b4

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI

it matched on name "ATY_GPU"

0x06B4: 0x00000000

Should be

./RadeonDump1 -n 6b0,c0300014

 

Same for me, newer RadeonPCI.kext has that Abort Trap: 6. (AMD RX 460)

 

Using older RadeonPCI.kext together witj new RadeonDump1

 

RadeonDump1 -n 6b0,c0300014
Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/GFX0@0/RadeonPCI
it matched on name "ATY_GPU"
0xc0300014: 0x    5a2d
bash-3.2# 
 
RadeonDump1 -r 6B0,6BF
Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEG0@1/IOPP/GFX0@0/RadeonPCI
it matched on name "ATY_GPU"
0x06B0: 0xC0300014
0x06B4: 0x00005028
0x06B8: 0x00000000
0x06BC: 0xAAAA5555

 

Good!

Now see the behavior of output

5a2d -> 0x2d = 45 celsium degree

What if heavy load?

Share this post


Link to post
Share on other sites

HD 7790

Picos-Mac-Pro:radeon pico$ ./RadeonDump1 -n 6b0,c0300014
Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI
it matched on name "ATY_GPU"
0xc0300014: 0x       0

Share this post


Link to post
Share on other sites

Picos-Mac-Pro:radeon pico$ ./RadeonDump1 -r 714,7f4

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI
it matched on name "ATY_GPU"
0x0714: 0x00000000
0x0718: 0x00000000
0x071C: 0x00000000
0x0720: 0x00000000
0x0724: 0x00000000
0x0728: 0x00000000
0x072C: 0x00000000
0x0730: 0x00000000
0x0734: 0x00000000
0x0738: 0x00000000
0x073C: 0x00000000
0x0740: 0x00000000
0x0744: 0x00000000
0x0748: 0x00000000
0x074C: 0x00000000
0x0750: 0x00000000
0x0754: 0x00000000
0x0758: 0x00000000
0x075C: 0x00000000
0x0760: 0x00000000
0x0764: 0x00000000
0x0768: 0x00000000
0x076C: 0x00000000
0x0770: 0x00000000
0x0774: 0x00000000
0x0778: 0x00000000
0x077C: 0x00000000
0x0780: 0x00000000
0x0784: 0x00000000
0x0788: 0x00000000
0x078C: 0x00000000
0x0790: 0x00000000
0x0794: 0x00000000
0x0798: 0x00000000
0x079C: 0x00000000
0x07A0: 0x00000000
0x07A4: 0x00000000
0x07A8: 0x00000000
0x07AC: 0x00000000
0x07B0: 0x00000000
0x07B4: 0x00000000
0x07B8: 0x00000000
0x07BC: 0x00000000
0x07C0: 0x00000000
0x07C4: 0x00000000
0x07C8: 0x00000000
0x07CC: 0x00000000
0x07D0: 0x00000000
0x07D4: 0x00000000
0x07D8: 0x00000000
0x07DC: 0x00000000
0x07E0: 0x00000000
0x07E4: 0x00000000
0x07E8: 0x00000000
0x07EC: 0x00000000
0x07F0: 0x00000000
0x07F4: 0x00000000


Picos-Mac-Pro:radeon pico$  ./RadeonDump1 -n 6b0,c0300e0c
Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI
it matched on name "ATY_GPU"
0xc0300e0c: 0x       0

Share this post


Link to post
Share on other sites

Should be

./RadeonDump1 -n 6b0,c0300014

Good!

Now see the behavior of output

5a2d -> 0x2d = 45 celsium degree

What if heavy load?

 

Sir, you want heavy GPU  load?!, Sir :)

Then you get it :)

 

running Geeks3D Tessmark X64 -

 

post-110586-0-92068100-1511953512_thumb.jpg

 

Before (idle, Desktop):

RadeonDump1 -n 6b0,c0300014 ... 0xc0300014: 0x    562b (43 Celsius?)

Start GPU stress:
0xc0300014: 0x    6a3a > 0xc0300014: 0x    703e (62 C)
Near Eend of GPU Test:   0xc0300014: 0x    7641 (65 C)

Very short after gpu load ends :

0xc0300014: 0x    6a36 (54 C) > 0xc0300014: 0x    6633 (51 C)

Now, after about 3 Min idle: 

0xc0300014: 0x    5028 (40 C)

 

PS:

I have question about FakeSMC (yours) vs Rehabman ( https://github.com/RehabMan/OS-X-FakeSMC-kozlek: also heavy load, but in this case of GPU FAN full speed :)

Both work as they should - i use rehabmans.

As i tried yours, for radeonMonitor tests there was some ??? about the FAN of the RX 460:

With your FakeSMC the GPU FAN runs full speed (for about 5 Seconds) at startup (Desktop appears) and at every wakeup (for me main problem).

With rehabmans the gpu fan stay quite in same situations (as it should).

Both with Sierra, no other change.

Can it be that some diff in SMC keys (exist or not or diff in value) can make that different FAN handling?

 

EDIT: I uploaded SMC -l of your & rehab Fakesmc

What can be they keys, values belong to the annoying full speed fan at start & wake?

SMC Slice: GPU fan runs full speed for 5 sec at startup  & wakeup
Last login: Wed Nov 29 12:49:54 on ttys000
GA-EP35:~ andreasm$ /Users/andreasm/Desktop/smc 
Apple System Management Control (SMC) tool 0.01
Usage:
/Users/andreasm/Desktop/smc [options]
    -f         : fan info decoded
    -h         : help
    -k <key>   : key to manipulate
    -l         : list all keys and values
    -r         : read the value of a key
    -w <value> : write the specified value to a key
    -v         : version

GA-EP35:~ andreasm$ /Users/andreasm/Desktop/smc -l
  #KEY  [ui32]  40 (bytes 00 00 00 28)
  $Num  [ui8 ]  1 (bytes 01)
  ACID  [ch8*]  (bytes ba be 3c 45 c0 03 10 43)
  ALI0  [{ali]  (bytes 06 01 02 00)
  ALRV  [ui16]  1 (bytes 00 01)
  ALV0  [{alv]  (bytes 01 01 03 b7 00 a3 00 15 6a f4)
  CLKH  [{clh]  (bytes 00 00 70 80 00 01 19 40)
  CLKT  [ui32]  162 (bytes 00 00 b3 a2)
  EPCI  [flag]  (bytes 06 90 07 00)
  FNum  [ui8 ]  (bytes 00)
  LSOF  [flag]  (bytes 01)
  LSSB  [{lso]  (bytes 01 01)
  LsNM  [ui8 ]  (bytes 01)
  LsbV  [{rev]  (bytes 01 04 0a 00 06)
  MOST  [ui16]  3 (bytes 80 03)
  MSDW  [ui8 ]  (bytes 01)
  MSPS  [{msp]  (bytes 00)
  MSSD  [si8 ]  (bytes 03)
  NATJ  [ui8 ]  (bytes 02)
  NTOK  [ui8 ]  (bytes 01)
  NVPR  [ui8 ]  (bytes 00)
  OSK0  [ch8*]  (bytes 6f 75 72 68 61 72 64 77 6f 72 6b 62 79 74 68 65 73 65 77 6f 72 64 73 67 75 61 72 64 65 64 70 6c)
  OSK1  [ch8*]  (bytes 65 61 73 65 64 6f 6e 74 73 74 65 61 6c 28 63 29 41 70 70 6c 65 43 6f 6d 70 75 74 65 72 49 6e 63)
  REV   [{rev]  (bytes 02 11 0f 00 00 16)
  RMde  [char]  (bytes 41)
  RPlt  [ch8*]  (bytes 64 38 00 00 00 00 00 00)
  RBr   [ch8*]  (bytes 64 38 00 00 00 00 00 00)
  BEMB  [flag]  (bytes 00)
  TC0D  [sp78]  (bytes 1f 00)
  FRC0  [freq]  (bytes 06 40)
  VC0C  [fp2e]  (bytes 3b a5)
  TC1D  [sp78]  (bytes 20 00)
  FRC1  [freq]  (bytes 06 40)
  TC2D  [sp78]  (bytes 1f 00)
  FRC2  [freq]  (bytes 06 40)
  TC3D  [sp78]  (bytes 20 00)
  FRC3  [freq]  (bytes 06 40)
  MPkC  [ui16]  0 (bytes a0 00)
  NATi  [ui16]  30 (bytes 00 1e)
  HE0N  [ui8 ]  (bytes 10)
GA-EP35:~ andreasm$ 

SMC rehabman https://github.com/RehabMan/OS-X-FakeSMC-kozlek
GPU FAN sosnt run full speed at startup / wakeup

GA-EP35:~ andreasm$ /Users/andreasm/Desktop/smc -l
  #KEY  [ui32]  46 (bytes 00 00 00 2e)
  FNum  [ui8 ]  0 (bytes 00)
  HWS0  [ch8*]  (bytes 47 69 67 61 62 79 74 65)
  HWS1  [ch8*]  (bytes 5a 37 37 2d 44 53 33 48)
  RPlt  [ch8*]  (bytes 64 38 00 00 00 00 00 00)
  RBr   [ch8*]  (bytes 64 38 00 00 00 00 00 00)
  TC0D  [sp78]  (bytes 27 00)
  TC1D  [sp78]  (bytes 29 00)
  TC2D  [sp78]  (bytes 27 00)
  TC3D  [sp78]  (bytes 28 00)
  TCAD  [sp78]  (bytes 2a 00)
  MlC0  [fp88]  (bytes 20 00)
  CC0C  [ui32]  128 (bytes 00 00 0c 80)
  MlC1  [fp88]  (bytes 20 00)
  CC1C  [ui32]  128 (bytes 00 00 0c 80)
  MlC2  [fp88]  (bytes 1f 00)
  CC2C  [ui32]  28 (bytes 00 00 0c 1c)
  MlC3  [fp88]  (bytes 1f 00)
  CC3C  [ui32]  28 (bytes 00 00 0c 1c)
  CCPA  [ui32]  21 (bytes 00 00 0c 15)
  PCTR  [sp78]  (bytes 17 92)
  PCPC  [sp78]  (bytes 12 4e)
  PCPG  [sp78]  (bytes 00 00)
  $Adr  [ui32]  0 (bytes 00 00 03 00)
  $Num  [ui8 ]  1 (bytes 01)
  ACID  [ch8*]  (bytes 85 fc 04 4d d8 12 7a f4)
  ALI0  [{ali]  (bytes 07 01 02 00)
  ALRV  [ui16]  1 (bytes 00 01)
  ALV0  [{alv]  (bytes 01 10 06 75 01 20 00 b6 0e e0)
  LsNM  [ui8 ]  1 (bytes 01)
  LsbV  [{rev]  (bytes 01 04 0a 00 06)
  MSLD  [ui8 ]  0 (bytes 00)
  MSSD  [si8 ]  (bytes 05)
  MSSP  [si8 ]  (bytes 05)
  NATJ  [ui8 ]  2 (bytes 02)
  OSK0  [ch8*]  (bytes 6f 75 72 68 61 72 64 77 6f 72 6b 62 79 74 68 65 73 65 77 6f 72 64 73 67 75 61 72 64 65 64 70 6c)
  OSK1  [ch8*]  (bytes 65 61 73 65 64 6f 6e 74 73 74 65 61 6c 28 63 29 41 70 70 6c 65 43 6f 6d 70 75 74 65 72 49 6e 63)
  REV   [{rev]  (bytes 02 11 0f 00 00 16)
  RMde  [char]  (bytes 41)
  RVBF  [{rev]  (bytes 01 30 0f 00 00 03)
  RVUF  [{rev]  (bytes 01 30 0f 00 00 03)
  BEMB  [flag]  (bytes 00)
  EPCI  [ui32]  0 (bytes 06 90 07 00)
  NATi  [ui16]  30 (bytes 00 1e)
  NTOK  [ui8 ]  1 (bytes 01)
  HE0N  [ui8 ]  16 (bytes 10)
GA-EP35:~ andreasm$ 

Share this post


Link to post
Share on other sites

Someone with Vega 10 test, please,

./RadeonDump -r 16658,16660

or

./RadeonDump -r 59968,59970

 

@mitch_de

Great!

So we know how to get temperature from Baffin family cards. And so my RadeonMonitor latest version should work for you? It already contains this method.

May be you test it with other FakeSMC?

About GPU FAN I don't know. It looks like GPU initializing but I nowhere do this. Only in Clover RadeonDeInit what I think is common for both your cases.

 

 

Ok that sounds like a lot of work. Is it something you are able to do ?

 

Thanks !

It is easy but requires a time. May be today evening I'll make it.

Share this post


Link to post
Share on other sites

Hope this kext will not crash and will show Radeon temperature.

doesn't crash anymore, but GPU 0 Core Temp -127° C ?!? (and NO, it is NOT cooled by liquid oxygen, just a one mile diameter fan with a million RpS)

 

Bildschirmfoto2017-11-29um18.45.38.png

but anyway... nice job.

Share this post


Link to post
Share on other sites

doesn't crash anymore, but GPU 0 Core Temp -127° C ?!? (and NO, it is NOT cooled by liquid oxygen, just a one mile diameter fan with a million RpS)

 

 

but anyway... nice job.

Can you, please, look into system.log or dmesg to find messages from RadeonMonitor?

Share this post


Link to post
Share on other sites

Hi, also for RX 460 - static (no changes under gpu load, idle) at -127 Celsius.

Nothing shown in dmesg & any log - Sierra.

Share this post


Link to post
Share on other sites

HD 7790  GPU 0

 

17    0 0xffffff7f8397a000 0x5000     0x5000     org.slice.RadeonMonitor (1.3.1) A388BA0F-9378-333A-8061-C9FE496C5B40 <16 12 7 4 3>

 

post-1113740-0-13405200-1512020964_thumb.png

Share this post


Link to post
Share on other sites

HD 7790  GPU 0

 

17    0 0xffffff7f8397a000 0x5000     0x5000     org.slice.RadeonMonitor (1.3.1) A388BA0F-9378-333A-8061-C9FE496C5B40

 

attachicon.gifScreen Shot 2017-11-30 at 12.47.46.png

 

Picos-Mac-Pro:radeon pico$ ./RadeonDump1 -r 714,7f4

 

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI

it matched on name "ATY_GPU"

0x0714: 0x00000000

0x0718: 0x00000000

0x071C: 0x00000000

0x0720: 0x00000000

0x0724: 0x00000000

0x0728: 0x00000000

0x072C: 0x00000000

0x0730: 0x00000000

0x0734: 0x00000000

0x0738: 0x00000000

0x073C: 0x00000000

0x0740: 0x00000000

0x0744: 0x00000000

0x0748: 0x00000000

0x074C: 0x00000000

0x0750: 0x00000000

0x0754: 0x00000000

0x0758: 0x00000000

0x075C: 0x00000000

0x0760: 0x00000000

0x0764: 0x00000000

0x0768: 0x00000000

0x076C: 0x00000000

0x0770: 0x00000000

0x0774: 0x00000000

0x0778: 0x00000000

0x077C: 0x00000000

0x0780: 0x00000000

0x0784: 0x00000000

0x0788: 0x00000000

0x078C: 0x00000000

0x0790: 0x00000000

0x0794: 0x00000000

0x0798: 0x00000000

0x079C: 0x00000000

0x07A0: 0x00000000

0x07A4: 0x00000000

0x07A8: 0x00000000

0x07AC: 0x00000000

0x07B0: 0x00000000

0x07B4: 0x00000000

0x07B8: 0x00000000

0x07BC: 0x00000000

0x07C0: 0x00000000

0x07C4: 0x00000000

0x07C8: 0x00000000

0x07CC: 0x00000000

0x07D0: 0x00000000

0x07D4: 0x00000000

0x07D8: 0x00000000

0x07DC: 0x00000000

0x07E0: 0x00000000

0x07E4: 0x00000000

0x07E8: 0x00000000

0x07EC: 0x00000000

0x07F0: 0x00000000

0x07F4: 0x00000000

 

Picos-Mac-Pro:radeon pico$  ./RadeonDump1 -n 6b0,c0300e0c

Found a device of class RadeonPCI: IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/PEGP@1/IOPP/GFX1@0/RadeonPCI

it matched on name "ATY_GPU"

0xc0300e0c: 0x       0

It looks like HD7790 is other family. I see no a good method to get temperature for it.

 

Waiting for tests with other cards.

Share this post


Link to post
Share on other sites

It looks like HD7790 is other family. I see no a good method to get temperature for it.

 

Waiting for tests with other cards.

 

wake after sleep shown GPU Core -68

 

post-1113740-0-10317600-1512025892_thumb.png

Share this post


Link to post
Share on other sites

I found the mistake
Here
https://sourceforge.net/p/hwsensors/hwsensors3/code3/HEAD/tree/trunk/plugins/GPUSensors/RadeonMonitor/ATICard.cpp#l21
line:
IOPhysicalAddress bar = (IOPhysicalAddress)((VCard->configRead32(0x20)) & ~0x3f);
Must be 0x24 instead of 0x20

I can compile only in evening. If someone do it now it will be great!

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.

  • Similar Content

    • By macinsane
      Mojave (and maybe High Sierra 10.13.6) on Asrock H370M-ITX/ac and getting RX560 to work with Intel graphics for full hardware acceleration
       
      Components used:
       
      Core i7 8700, UHD 630
      Gigabyte Radeon RX560
      Wifi/BT BCM943602BAED (DW1830) M.2
      Samsung PM961 NVMe SSD
       
      *This guide requires you to know the basics, meaning how to create a macOS install drive, installing and using Clover. There are great beginners guides out there! As always you need at least FakeSMC to boot, every other kext I’ve used is mentioned in the guide. For Mojave beta you need to be enlisted with your Apple-ID in Apple’s beta programme. What is described here may work for High Sierra 10.13.6, but I doubt it to be reliable. Even in early beta status (PB3) Mojave runs much more stable and smoother than High Sierra ever did for me.* 
       
      UEFI-BIOS settings:
       
      The board comes with BIOS version 1.20, there is an update 1.50 on the Asrock website. Usually it is good to stay updated but in this case I’ll recommend to stay on 1.20, because the 1.50 version seems to be unstable and I had restarts upon wake and things like that I couldn’t get fixed even with dsdt patches. So I just downgraded to 1.20 which runs fine.
      Settings needed: Just the usual. Disable fast boot and security boot (everything Windows-related), disable vt-d and enable XHCI hand-off. Activate multi monitor support and set VRAM to 64MB to activate the Intel IGPU. From what I’ve read most people recommend 128MB but I find the 64MB setting to be more reliable, getting artifacts with settings below or above 64MB. 128MB may be needed for 4k support, though. As for nVidia users: You may have to disable the Intel graphics. I’m using an AMD card and that is the brand to go these days since Apple decided to work with AMD
       
      Clover and power management:
       
      I recommend upgrading to 4586. It seems that Aptiomemoryfix is now working with H370 chipsets. I stayed on OsxAptiomemoryfix2 together with Emuvariable and it works. Feel free to experiment.  
      For power management with your CoffeeLake cpu you can just tick plugin type within Clover or compile a ssdt with Pikeralphas script. Mojave reports my cpu correctly as Core i7 since PB3, before that it is just “Unknown” but that is just cosmetic. 
       
      Graphics:
       
      This is the worst part. Getting the Radeon to run correctly alongside the Intel graphics took me weeks to figure out. 
       
      While RX570 and 580 run out of the box in most cases, the RX560 models often do not or at least not correctly. Apple even banned them from eGPU use. For my approach you cannot use “Whatevergreen.kext”, it simply didn’t work for me the way I needed it. Instead I added some fixes to my DSDT to set the correct number of ports used (DP, HDMI, DVI) and tell macOS to use the correct frambuffer. For those who are not into dsdt editing I added my DSDT, look at the “PEGP” section how it is done. This works for the RX560 using the “Acre” framebuffer, you can replace it with “Radeon”, that works too. Other cards need a different framebuffer and a different number of ports. You can do it!   BTW: Do not buy this Gigabytecard!!! It has some weird VBIOS on it, which caused me nightmares before I got the card to work! Buy a Sapphire card instead, they are used by Apple too.   
       
      To get full hardware acceleration for encoding videos you have to have the Intel graphics enabled alongside your AMD card. Problem is, you cannot have it running simply as a second graphics card but you’ll need to “hide” it, means telling macOS that it doesn’t have any display connectors. Otherwise you’ll have screen artifacts or in the worst case just constant crashes. This can be done by adding a certain ig-platform-id in Clover. If you use 59120003 the Intel graphics will not show up in system profiler but it is there. You can check it with iStats or HWMonitor or look if the Kaby Lake graphics kexts are loaded.  This ID makes the CoffeeLake UHD 630 to be recognized as a Kaby Lake gpu that has no display ports, is just used as a helper card.
       
      You’ll need it as a helper card for video encoding. That is what Intel graphics are used for in Macs, to support video encoding on the hardware side. You can check if hardware encoding is supposedly enabled by using MacX Video Converter Pro and VDADecoderChecker. If both say “yes” and “fully supported” this is a good start. But despite the widespread belief this doesn’t mean everything is set correctly! You now have to find out, how to get both cards working together. This is the moment where you have to choose between work and fun: Are you using FinalCut Pro for video editing or are you planning to use iTunes for entertainment? Up until this day (Mojave PB3, High Sierra 10.13.6) you cannot have both working correctly! I’ll show you how it is done both ways. 
       
      The best settings for Final Cut:
       
      Use the DSDT I’ve added with the respective AMD settings for the RX560 or modify them for your graphics card.
       
      Use SMBIOS for iMac17,1 (14,2 does work but picture over DP port is ugly, iMac 18,1/2/3 do not work) in Clover.
       
      Have AppleIntelKBLGraphicsFramebufferInjector_3e9x.kextin your Clover/kexts/Other folder.
       
      Until Mojave PB2 (may work on High Sierra 10.13.6):
       
      Fake-ID 3E928086 plus ig-platform-id 59120003
       
      These are the correct setting when installing Mojave (usually you will start with PB1 as the full installer).
       
      For Mojave PB3 and higher:
       
      Fake-ID 59128086 plus ig-platform-id 59120003
       
      After upgrading to PB3 you will get a black screen with 3E92 fake id because PB3 introduces Coffee Lake kexts and macOS will use these kexts instead but that doesn’t work with our ig-platform-id. So you can download the update to PB3 but before installing it you’ll need to change the fake-id to a usual Kaby Lake id. While this didn’t work in High Sierra and the first two Mojave betas (meaning you didn’t get full hardware encoding) it is now magically working with PB3. It is safe to remove AppleIntelKBLGraphicsFramebufferInjector_3e9x.kext after upgrading to PB3.
       
      The best setting for iTunes and why it is the worstfor Final Cut:
      Just do the same thing as for FinalCut, put Shiki.kext into your Clover/kexts/other folder and add -shikigva boot argument. You can use -shikigva=1 or 2 or 12 it will work either way. Now you can watch movies with iTunes, yay! But it has some sideeffects. You cannot use this setting when seriously editing with Final Cut.
       
      While this works by tricking macOS to use the IntelGraphics to play protected video content, it doesn’t seem to be the way real Macs are doing this. As soon as you open FinalCut you will notice (check this with iStats or HWMonitor) that the Intel graphics is not or only barely utilized, instead cpu usage skyrockets, indicating that slow software rendering is used instead of hardware encoding through Intel graphics. To make it worse the Radeon still “thinks” there is an Intel graphics helping so it does not run at highest potential. You may never notice it when using HD material with 24 or 25p because it still fast enough for that. But 50p or 4k material will most likely cause stuttering. I was astonished when I had stutters while previewing my 50p material, impossible to watch it. Thinking: What is wrong? I have a freaking i7 8700 plus RX560! Even my old hackbook with HD4000 is doing better!     
       
      Wifi and Bluetooth
       
      I swapped the Intel Wifi card that was coming with the mainboard plugged into the onboard M.2 port with a compatible Dell card. The DW1830 or BCM943602BAED. The wifi part runs out of the box, but if you want the card to appear as Apple AirportExtreme and set the country code according to your country you need to put FakePCIID_Broadcom_WiFi.kext into Clover/kexts/other and add some Clover kext-to-patches by PMHeart. 
       
      Name: com.apple.driver.AirPort.BrcmNIC
      Find: 75166644 3BA3581A 0000750C 4183FE04 7406
      Replace: 66906644 3BA3581A 00006690 4183FE04 6690
      Comment: AirPortBrcmNIC - PCIe in 10.13.x
       
      Name: com.apple.driver.AirPort.BrcmNIC
      Find: 4183FCFF 7435488D 55D0
      Replace: 66C70644 45EB348D 55D0    (*this is for DE country code, change for your need)
      Comment: BCM4352-Country Code 
       
      Bluetooth will work thanks to Rehabman’s kexts. You’ll need BrcmPatchRAM2.kext and BrcmFirmwareData.kext OR BrcmFirmwareRepo.kext. The Data.kext can be injected through Clover, just throw it into the kexts/other folder, the Repo.kext has to be installed in /Library/Extensions on your mac drive, it won’t work with Clover. I had issues with the Data.kext, Bluetooth was gone on every second or third reboot. The Repo.kext is more reliable, haven’t had any issues yet. I really don’t like the kexts to be installed in my system but I guess I have to live with it for the time being.
       
      Oh, handoff and continuity don’t work, doesn’t matter that system profiler shows them as enabled. At least it works in one way: the mac can handoff Safari pages to my iPhone, but not the other way around. Documents opened on my phone show up in the dock, but they are not loaded. 
      This issue is around since at least High Sierra and from what I’ve read Rehabman is not going to fix this, because he founds this features not worth it. I have to disagree, these are very convenient and valuable features if you work with desktop, laptop and tablet in different situations, I miss this capability very much!    
       
      Audio
       
      Audio works with Lilu.kext plus AppleALC.kext and you need to inject Layout ID=1 in Clover. For Mojave use the boot setting -lilubetaall to force the kexts to load. If you like to have DP or HDMI audio add a HDAU device to your dsdt (see mine for reference).
       
      Credits:
      Credits go to this great community! I hope I mentioned everyone in this guide whose insights made this hack possible. If I have forgotten someone please let me know!
      DSDT.aml.zip
    • By vector sigma
      HWSensor Features
      Control temperature of Intel and AMD CPU Control temperature of AMD, Intel and NVidia cards Control temperature of motherboard Control FANS Control Voltages Control HDD/SSD state (SATA & NVMe) with S.M.A.R.T. monitoring  Control frequencies Laptop Battery Monitoring Different applications support High Sierra compatible Installation to /S/L/E or in the ESP (Clover only)  
      Supported languages:
      English (base), Russian, Italian, Korean and semplified Chinese
       
      Source code: at https://sourceforge.net/p/hwsensors/hwsensors3/code3/HEAD/tree/
      Bugs report at https://sourceforge.net/p/hwsensors/hwsensors3/tickets/
      Topic for discussion at: https://www.insanelymac.com/forum/topic/299861-hwsensors3/
       
      Chief Developer @Slice, new HWMonitorSMC2.app and package by @vector sigma
    • By Slice
      Laptop battery monitoring

       
       
       
      Using this driver you have to exclude other battery drivers: ACPIBatteryManager, AppleSmartBattery and so on.
      This driver assumes to using HWSensors3 with FakeSMC 3.5 as being part of it.
       
      19.01.2018
      Updated version of VoodooBatterySMC will get extended battery information and so it will be more exact.
      VoodooBatterySMC.kext.zip
       
      Updated HWMonitorSMC by vector-sigma
      HWMonitorSMC.app.zip
       
      HWMonitor2 by vector-sigma
      https://sourceforge.net/projects/hwsensors3.hwsensors.p/files/latest/download
    • By zeg66
      Please help me with this 10.9.5 system. Motherboard: Gigabyte 965p-s3 rev3.3, gpu: Ati Radeon HD6450 (dvi/hdmi/vga). 
      Using any of the 3 ports on the 6450 (dvi, hdmi, vga), I see all the white letters on black background during boot but when they complete and it's time for the graphics to start I get a black screen and sometimes a restart after that. Using Clover 4411. Is there any way to get past this?
×