MP 1,1-5,1 testing TB3 AIC with MP 5,1

[syncom]

Kingston DataTraveler 100 G3-DT100G3/64GB​

Kingston DataTraveler 100 G3-DT100G3/64GB USB 3.0, PenDrive, 64 GB, 1 Piece, Black

Kingston DataTraveler 100 G3 (DT100G3) USB Flash Drive is compliant with USB 3.0 specification and is able to take full advantage of technologies installed on desktops, notebooks and digital devices. With the DT100G3 drive, storing and transferring documents, presentations, music, videos and othe...

www.amazon.it

Lexar LJDS47-64GABBK JUMPDRIVE S47 64GB​

Lexar LJDS47-64GABBK JUMPDRIVE S47 64GB USB 3.1 Black

Yes, we've heard the buzz, too, but the fact is that our roots with Lexar are rooted in memory solutions, so we're still standing, stronger than ever.

www.amazon.it

EasyULT Case Hard Disk 2.5 Pollici, USB 3.0 Enclosure (with WD Blue 320GB)​

EasyULT Case Hard Disk 2.5 Pollici, USB 3.0 Enclosure Hard Disk Esterno Box Caso SATA I II III HDD SSD 9.5mm e 7mm [Supporto UASP] Tool Free, con OTG Adattatore(Nero)

Supporta tutti gli SSD HDD SATA I / II / III da 2,5 pollici, compatibili con dischi rigidi con spessore da 7 mm a 9,5 mm. Questo box esterno SATA III da 2.5 "è fino al 70% più veloce di una normale connessione USB 3.0 quando è associato a un controller host dotato di UASP. Uscita: USB 3.0 Mic...

www.amazon.it

With Inateck Hub all drives are mounted at boot while with USB-C to USB-adapters I must disconnect and reconnect them After Mac Pro starts.

These are the adapters I used:

EasyULT 3 Pieces USB C to USB 3.0 Adapter, Male to Female, 5Gbps OTG Type C Cable for MacBook PRO 2017/2016, Huawei P30 Lite P20 Lite Mate 20 (Grey)

Maximum transfer rate 5 Gbps 10 times faster than USB 2.0, just a few seconds to transfer the large file. Durable and stable. Plug and Play No need to install hardware or software, easy to use. Cable length: 20 cm, easy to connect Supports OTG Transfer videos, photos, music and other files...

www.amazon.it

nonda USB C Adapter USB 3.1 Gen1,USB Type C to USB OTG Adapter Thunderbolt 3 to USB Female for MacBook Pro2019, MacBook Air 2020, iPad Pro 2020, Other Devices with USB Type C (Silver)

Introducing the world's smallest USB-C to USB-A adapter! Use it to charge your phone and other standard USB devices or transfer data to USB 3.0 super fast (up to 5Gbps). Made of durable aluminum alloy, this mini adapter is available in 3 colors to match the new 12-inch MacBook. It is also compati...

www.amazon.it

 

[Quadra950Fan]

I believe your Syba card should be plugged in to Slot 2 as it requires the x16 PCIE lane for "full" speed.

I am looking to get one of these cards for my cMP 4,1/5,1. I have an RX580 in Slot 1. The Syba card (Syba Dual M.2 M-Key NVMe Ports to PCIe 3.0 x16 Bifurcation Riser Controller - Support Non-Bifurcation Motherboard (SI-PEX40129)) would be quite close to it and I wonder about having enough of a gap for the fans to cool effectively.

Does anyone have a similar setup and is there enough of a gap between the two cards for the fans to draw air in?

It looks to be shorter than the RX580, but is that enough cooling capacity? I do not do much heavy GPU work a majority of the time. Attached are some images (Yes, I need to dust it!). TIA

 

[mikas]

If all your PCIe slots are not needed, you could install RX580 in slot 2(x16) and Syba in slot 1(x16). Like here my Syba+Vega56..

It fits the other way too, but like you said, it's not ideal for cooling. Like in here my RX580+Syba..

But what the heck, some just want/need the power..

😉
edit. we seem to discussing this in wrong thread here. (testing TB3 AIC)

 

[Quadra950Fan]

Didn't even think about moving the GPU, great idea. The only other thing I want to add is USB 3 but that's much smaller...

Alas, you are right, this is the wrong thread for this topic. If I have any more questions I'll start a new thread for it. Thanks!
/End

 

[KevinClark]

Does someone Know if it's possible to recreate a SSDT file For Titan Ridge with theses drivers from Monterey 12.5.1

I have Flashed titan with some of them and it works but maybe we need to copy the Plist in the open core package to be the nearest true of apple thunderbolt 2

Thanks

 

[cdf]

Just about done with the design. In about the same footprint as the original Bluetooth module, the device provides a USB hub and basic GPIO functionality. I hope to eventually document the device in a dedicated thread.

Teaser:

 

[KevinClark]

Hello

I wonder if we could flash a titan ridge V2 with those firmware extract from apple Monterey 12.5.1 installer

Some in the package have one firmware with info

some them have two maybe It could be possible to flash one firmware in the blue chip and the other in green chip ?

Then we could get working the two thunderbolt 2 of titan ridge in our Mac Pro 5.1 ?

What do you think about ?

let me know

Thanks in advance

 

[Gustav Holdoff]

I found one very unpleasant effect of the presence of a thunderbolt card in my cMP
after working in macos reboot in windows ok,
and after working in windows, rebooting into macos fails - macpro gets a panic
to boot in macOS, I need to reset SMC and NVRAM (if I tried to reset only SMC - the "no boot disk" screen appears, but the macos eventually loads too- without reseting smc - bootloop and panic)
It seems, that windows in smc or nvram writes something that is not compatible with macos

I had something similar when I tried to experiment with special SSDTs for my previous W5700 - then also working on Macos, followed by rebooting into Windows and back, caused failures
but i read on hackintosh forums that ssdt in windows can cause crashes,

the only difference for me with ssdt <true> is that the computer recognizes the card and controllers correctly- as Alpine
But in Mac OS, working with Thunderbolt absolutely does not depend on which argument is <true> or <false>
With the argument <true>, the sleep function does not work
I decided to go through the hardware step by step and turn it off one by one
in order to find out the cause of the panic
And at the first step, I managed to find out that this is an Alpine Ridge Thunderbolt card
without thunderbolt card i tested boot macos-windows-macos-windows and so on
the panic is gone
I put my inateck USB-3 card instead of a thunderbolt, everything works
So the Thunderbolt card after working in Windows has some residual information that causes panic in MacOS or has the wrong power management that Windows gave it
because if I work in a macos with a thunderbolt card, then I can reboot the system many times and there is no panic,
the charge of the capacitors cannot be the cause, because the power management of this card in macos gives the system and nvram the correct information
- it means that Windows writes something in the nvram about this card, which then the macos is not perceived
Alpine ridge i bought on ebay as a flashed
but I can't be sure if the information was rewritten on the chips, or if the connectors were just connected correctly
(the resistor that is responsible for sleeping in place - it is not removed)
My CMP 4.1/5.1 dual cpu xeon 5680, 96gb RAM, radeonpro W6800, nvme crucial 2tb, evo plus 1tb, TB3 alpine ridge, Inateck USB-3, oc 0.8.4, mac os 12.6

 

[cdf]

I found one very unpleasant effect of the presence of a thunderbolt card in my cMP
after working in macos reboot in windows ok,
and after working in windows, rebooting into macos fails - macpro gets a panic
to boot in macOS, I need to reset SMC and NVRAM (if I tried to reset only SMC - the "no boot disk" screen appears, but the macos eventually loads too- without reseting smc - bootloop and panic)
It seems, that windows in smc or nvram writes something that is not compatible with macos

This is very interesting. A flashed Thunderbolt add-in card works in ECM mode, but Windows expects ICM mode, so some undesirable things can happen. Usually a cold boot into macOS is enough to return things to normal. Are you using the Thunderbolt Boot Camp driver (TbtHostControllerA.inf) or the one that Windows automatically installs (TbtHostController.inf)?

the only difference for me with ssdt <true> is that the computer recognizes the card and controllers correctly- as Alpine

If you're referring to what is shown in System Information > PCI, then that information is really just cosmetic.

With the argument <true>, the sleep function does not work

(the resistor that is responsible for sleeping in place - it is not removed)

Thunderbolt sleep issues are serious: @tsialex has seen cases of corrupt BootROMs filled with Thunderbolt sleep related panic logs. A poorly written SSDT can cause such issues. Is sleep a problem only if a device is connected to your card? Does your Mac Pro try to sleep and is then immediately waken up? (If you've answered yes to these questions, then removing the resistor could help.) Can you post your SSDT file?

So the Thunderbolt card after working in Windows has some residual information that causes panic in MacOS or has the wrong power management that Windows gave it
because if I work in a macos with a thunderbolt card, then I can reboot the system many times and there is no panic,
the charge of the capacitors cannot be the cause, because the power management of this card in macos gives the system and nvram the correct information
- it means that Windows writes something in the nvram about this card, which then the macos is not perceived
but I can't be sure if the information was rewritten on the chips, or if the connectors were just connected correctly

I wonder if this information consists of logs or configuration data. It's also possible that data gets written to the flash chip of the Thunderbolt card itself. This data (which appears in the scratch region above the active region of the memory contents) depends on what devices you have connected and can change from one boot to the next. It would be interesting to see this data in your case. Note that because your card is flashed for Mac, you don't need a flash programmer to dump the data. You could simply use the tbpatch utility:

GitHub - osy/ThunderboltPatcher: Patch Thunderbolt firmware through TI PD interface

Patch Thunderbolt firmware through TI PD interface - osy/ThunderboltPatcher

github.com

 

[Gustav Holdoff]

This is very interesting. A flashed Thunderbolt add-in card works in ECM mode, but Windows expects ICM mode, so some undesirable things can happen. Usually a cold boot into macOS is enough to return things to normal. Are you using the Thunderbolt Boot Camp driver (TbtHostControllerA.inf) or the one that Windows automatically installs (TbtHostController.inf)?
If you're referring to what is shown in System Information > PCI, then that information is really just cosmetic.

I'm not a big windows expert, my last computer was probably in 2004,
and I've already forgotten a lot about how to really set up the system
in this case i let windows test the hardware and install the drivers by itself.
However, I downloaded the alpine gigabyte package from official site and tried to install the driver,
but the installer gave an error (at the very end of the installation)
about systeminfo- yes, maybe cosmetic
with ssdt <false> there is unknown controller
with true- there is Alpine controller
About sleep:
with SSDT <true> cmp doesn't sleep at all (whether something is connected to the thunderbolt or not)
with SSDT <false> i see usb-3 WD MyBookDuo (but thunderbolt WD MyBook Thunderbolt only after warm reboot)
in windows with <false> all devices (and USB-3 and Thunderbolt) are visible only after warm reboot from MacOS
about sleep crashes - I made post earlier
behavior with W5700 was different than with W6800
it is obvious that the presence of the patched W6800 introduces some new obstacle in the work of the macPro

https://forums.macrumors.com/threads/testing-tb3-aic-with-mp-5-1.2143042/post-31110390

EDETED
I have already tested many times and I can confirm that after macos reboot in windows and from windows shutdown I can't load macos again without smc reset.
although I did not notice this behavior before with the W5700 GPU.
probably some other nuance from the patched W6800 intervenes - there Windows can also write something to nvram

As for the patcher, I will study this topic next weekend, since I need a stable job during working hours, and I will continue the experiments next weekend
SSDT.aml is in attachment
My CMP 4.1/5.1 dual cpu xeon 5680, 96gb RAM, radeonpro W6800, nvme crucial 2tb, evo plus 1tb, TB3 alpine ridge, Inateck USB-3, oc 0.8.4, mac os 12.6

 

[cdf]

SSDT.aml is in attachment

One thing that sticks out with this SSDT is that it doesn't include a ThunderboltDROM...

 

[Gustav Holdoff]

One thing that sticks out with this SSDT is that it doesn't include a ThunderboltDROM...

and this one? I downloaded it from the site zephyr (please check -AR)

 

[cdf]

and this one? I downloaded it from the site zephyr (please check -AR)

This one is more standard. Does it also break sleep?

 

[AlexMaximus]

I am looking to get one of these cards for my cMP 4,1/5,1. I have an RX580 in Slot 1. The Syba card (Syba Dual M.2 M-Key NVMe Ports to PCIe 3.0 x16 Bifurcation Riser Controller - Support Non-Bifurcation Motherboard (SI-PEX40129)) would be quite close to it and I wonder about having enough of a gap for the fans to cool effectively.

Does anyone have a similar setup and is there enough of a gap between the two cards for the fans to draw air in?

It looks to be shorter than the RX580, but is that enough cooling capacity? I do not do much heavy GPU work a majority of the time. Attached are some images (Yes, I need to dust it!). TIA

I have the same Syba card and I used to have the same problem. The cooling was not good enough with the Saba card in slot #2. In the end I got a single PCIe card (aqua computer KryoM2) card for a single 2TB NVME SSD. This card is extremely short and works really great for my silent RX580 "Seahawk" solution.

 

[Gustav Holdoff]

This one is more standard. Does it also break sleep?

As far as I remember, when I bought the card, I tested with different ssdt, and there seemed to be no difference.
I can't check right now because I disconnected the alpine, I'll try next weekend
My CMP 4.1/5.1 dual cpu xeon 5680, 96gb RAM, radeonpro W6800, nvme crucial 2tb, evo plus 1tb, TB3 alpine ridge, Inateck USB-3, oc 0.8.4, mac os 12.6

 

[cdf]

Update: GPIO and USB hub module installed and working.

The module provides four internal USB data connections: one for the BT module, two for the GC-TITAN RIDGE, and a spare (pictured above, only the BT connection is used). It also provides GPIO for the GC-TITAN RIDGE (see above on the right). To the best of my knowledge, this is the first active add-on module for interacting with a Thunderbolt card through the THB_C header. The module not only takes care of FORCE_PWR but also detects SLP_S3 to reset one of the PD controllers for proper wake-from-sleep behavior. As a result, my LG UltraFine 5K display no longer stays black when my Mac Pro wakes from sleep!

 

[AlexMaximus]

Update: GPIO and USB hub module installed and working.

View attachment 2079053

View attachment 2079052

The module provides four internal USB data connections: one for the BT module, two for the GC-TITAN RIDGE, and a spare (pictured above, only the BT connection is used). It also provides GPIO for the GC-TITAN RIDGE (see above on the right). To the best of my knowledge, this is the first active add-on module for interacting with a Thunderbolt card through the THB_C header. The module not only takes care of FORCE_PWR but also detects SLP_S3 to reset one of the PD controllers for proper wake-from-sleep behavior. As a result, my LG UltraFine 5K display no longer stays black when my Mac Pro wakes from sleep!

Wow, that is really convenient for those of us with a Titan Ridge card. Could you provide the source where to get that hub card from?
Does this make the little jumper cable obsolete on that TR card?

 

[cdf]

Could you provide the source where to get that hub card from?

This one is actually my first prototype. If I make the module available, it'll be in the form of an open-source project, maybe with some units made-to-order.

Does this make the little jumper cable obsolete on that TR card?

Yes. It provides that functionality along with a basic Thunderbolt header. The firmware currently addresses a wake-from-sleep issue, but it could also address hotplug issues, which would be interesting to investigate.

In case anyone is curious, here's a shot of the back of the module (the silkscreen really came out nice and crisp):

 

[superparati]

This is a great job!

 

[AlexMaximus]

Perfect, you got my order, count me in. For your development efforts, I‘ll fund you 50 bucks for one of those baby’s With cables.

 

[KevinClark]

Idem intéressé

 

[AlexBurdov]

This one is actually my first prototype. If I make the module available, it'll be in the form of an open-source project, maybe with some units made-to-order.


Yes. It provides that functionality along with a basic Thunderbolt header. The firmware currently addresses a wake-from-sleep issue, but it could also address hotplug issues, which would be interesting to investigate.

In case anyone is curious, here's a shot of the back of the module (the silkscreen really came out nice and crisp):

View attachment 2080094

Nice work! Count me in for 2 sets...

 

[Anglezark]

Hi folks,

Has anyone tried a Maple Ridge yet? They seem readily available and about the same price as a Titan Ridge, but with better spec's..? What would the down sides be?

 

[joevt]

Has anyone tried a Maple Ridge yet? They seem readily available and about the same price as a Titan Ridge, but with better spec's..? What would the down sides be?

I use Alpine Ridge, Titan Ridge, and Maple Ridge for USB 3.1 gen 2. There's probably no upside to using Maple Ridge compared to Titan Ridge. You should probably consult a hackintosh website for Maple Ridge with macOS info. The difference with Hackintosh is that it might have BIOS support for Maple Ridge which might be needed for whatever methods they use to get it to work with macOS.

 

[KevinClark]

I use Alpine Ridge, Titan Ridge, and Maple Ridge for USB 3.1 gen 2. There's probably no upside to using Maple Ridge compared to Titan Ridge. You should probably consult a hackintosh website for Maple Ridge with macOS info. The difference with Hackintosh is that it might have BIOS support for Maple Ridge which might be needed for whatever methods they use to get it to work with macOS.

It’s possible to flash titan ridge with usbc thunderbolt driver from osx package installer like latest monterey the card works in the Macpro maybe the maple ridge could work