I suspect that the limitation on ASRock’s website for the Taichi stating eGPU is not possible is only valid when using ASRock’s own Thunderbolt AIC that only has an Alpine Ridge chipset.
I don’t have an eGPU enclosure so I cannot test this personally. Maybe @wendell can check on this; in the video for the ASrock X570 ITX/TB3 motherboard there was an eGPU enclosure visible, if I remember correctly (the X570 ITX/TB3 and Creator both use Titan Ridge chipsets onboard however in a configuration with limited power).
Unfortunately I don’t know if Level1techs is planning a video with the Taichi and Gigabyte GC-Titan Ridge combo.
The GC-Titan Ridge also supports 6 TB devices per port instead of 3 (ASRock’s solution) so I would be cautiously optimistic that eGPU might also work with this combination.
Since there would a nice symbolism if you reviewed every* current plug-‘n’-play AMD Thunderbolt solution could you maybe do a follow-up to the upcoming X570 Taichi review where you’re comparing the Taichi’s Thunderbolt performance with ASRock’s own TB Rev 2 AIC and the Gigabyte GC-Titan Ridge?
(*well, there only are the few ASRock motherboards)
As far as I know no reviewer has done this yet… so there would be something extra unique to your content maybe drawing in new people (A/V-Music production hoomans).
As detailed in this thread, I (with the replacement Taichi) and @carsonxyz are running the X570 Taichi/GC-Titan Ridge combo without any issues. However I don’t have any testing experience with eGPU enclosures.
Your safest bet would probably be the X570 Creator however its Thunderbolt is - as described in this thread - less “powerful” compared to the Taichi+GC-Titan Ridge IF everything really works as it should (so far it seems to be the case but again, not enough feedback reports to be absolutely positive).
And I don’t think that the Creator is worth the hefty price difference compared to the Taichi. My thoughts:
I have an Intel NUC 6i7kyk that has thunderbolt through an Alpine Ridge controer and it supports eGPU. So if AsRock don’t that is their own deal as there is nothing technically stopping it in the hardware.
It could just be, as stated above, because it has not been validated by them. But that is a pretty lame excuse if so.
This thread convinced me the time was right to add Thunderbolt 3 to my Ryzen 3600 for a audio interface upgrade I’ve been planning.
I bought a x570 Taichi and the Gigabyte Titan Ridge and crossed my fingers it would be a trouble free install. Pleasantly surprised to find the revised Taichi MB had relocated the USB 3.2 Gen 2 (type C) so it no longer is blocked by longer graphics cards.
Plugged in all the headers and PCIe power cables, booted into bios and turned TB3 on and Security off. Dont know if its necessary to turn security off?
Checked in Win 10 device manager and found it was unrecognised as was to be expected. Tried installing the TB3 driver I downloaded from Gigabyte only to find it would not run. I had to update the driver from the device manager. Just browsed to the Gigabyte folder I had downloaded the driver to and it installed perfectly.
Next I ran the firmware update tool I downloaded from Gigabyte and updated to firmware version 43.0. The update utility was a simply click and let run and without any hassle like I experienced with the driver.
All worked perfectly and correctly identified in the device manger. Win 10 also installed the Thunderbolt Control Centre App, so all I have to do now is work out the best TB3 audio interface for my needs.
This thread made it possible as I wanted to use my existing Ryzen cpu and not spend a fortune on a MB with unwanted features. Recommend to anyone looking to do the same!
Happy to hear another story of a successful implementation!
Yeah, if you’re used to drivers being installed by setup routines it takes a short while to realize that you got to manual install the new Thunderbolt drivers and get that shitty Thunderbolt service App from the MS Store now…
It’s a bit bittersweet since when looking at upcoming TRX40 boards by ASRock it seems that they are now also leaving the Thunderbolt header off (the intended location is still visible).
Maybe with some firmware/driver update you can get Thunderbolt to function just like any other additional Interface PCIe add-in card without the need for a physical header or BIOS support (and yeah, being able to hot-plug devices is important to me).
Yeah the sooner TB3 or USB4 starts being included on all brands of Ryzen suitable MBs the better!
I take it the Display Port cable does not have to be attached from the graphics card to the Titan AIC unless you want to use a TB3 display attached to one of the two TB3 ports on the card?
The documentation does not make it clear but it seems reasonable to assume if you dont want to drive a TB3 display then you dont need a display input to the card. Can you confirm TB3 ports work without needing the display input?
Might have called the Titan Ridge AIC install a success a little early. The PC now wont boot into win 10 without a reset, it stops at Dr Debug error code 88. The manual does not list a code 88 but from checking online it seems it has something to do with chipset initialization?
So it stops in initial boot up and after pressing reset finishes boot sequence and starts win 10 reliably.
My PC is running only 2 sata drives and not using any NVMe (yet) with a MSI 1070 DUKE graphics card. I have 32 GB of 3600 G.Skill Trident Z Neo, (2x16) and it runs at full speed without any issues. The PC performs well and once started shows TB3 in Device Manager and in Windows Thunderbolt Manager.
Have a HX850i power supply and a H115i Pro 280 AIO cooler both connected to internal USB running Corsair’s Link app.
I have tried changing RAM to other brand modules and running the PC without the power supply and AIO cooler connected to internal USB and using Link. Made no difference to the start up routine and still required a reset to start.
Removed the Gigabyte Titan Ridge card and disconnected all its associated headers from PC and problem is fixed. The PC without the AIC card installed starts nice and quick without requiring a reset.
Any ideas from users who have successfully installed and run this card without these issues would be much appreciated!
I’ve had such a behavior with an add-in ethernet card, not the Titan Ridge - what helped me was:
Disconnect the PSU from AC outlet
Press the Power On button for about 10 sec
Remove the CMOS battery
Wait 5-10 min
Reseat the battery
Download latest UEFI (currently 2.50), decompress it to a USB thumb drive and make a copy of the BIOS file and rename that to CREATIVE.ROM
Follow manual instructions for USB flashback procedure
After that has completed, start the system again with the USB thumb drive still attached
Go into UEFI, select load UEFI defaults, save & restart system
Go into UEFI and reflash the latest BIOS file with its original file name from the USB thumb drive - reboot
Repeat Step 9
Now make and save the changes to the UEFI for your final configuration.
I’d re-install the Titan-Ridge Add-In card after step 7 - be sure to disconnect the PSU again and press the Power On button for a few seconds before installing the add-in card.
Interesting because I had already tried reloading bios and setting CSM to disabled. I had not tried clearing cmos, so I went through the process you outlined.
Long story short, it works great and boots up normally without a reset being needed. Glad you suggested a complete clearing out and reloading of the bios.
I have tried it for a while at default settings and its stable. I have now resumed running the memory at full speed and the cpu is running precision boost overdrive without any problems. Bios is back to CSM enabled and did not make any difference enabled or disabled.
One other thing I noticed and wondered if it might cause a problem on boot up was my PC was set to keep USB power onto the ports when shut down. Noticed this morning the USB light on my audio interface was on and the PC off. This must have happened when upgrading to this MB as I normally have the USB ports off.
As the Titan AIC requires a single USB 2 internal port to operate along with the TB3 header and PCIe power, the USB port not powering down with the PC may cause a boot issue ? Reasoning is that the AIC is only partly powered on from the USB port. Anyway its back to USB ports being powered off now when PC is powered off.
Since I had to pull the PC out to remove the battery I also decided to put the two USB Corsair Link cables together in one header. What a simple solution and now only one internal USB header used for the two Link cables. The Titan AIC is plugged into one and the two Link cables combined into the other header. No longer have to leave the HX850i unplugged and can monitor it and the 280 AIO cooler together again.
Without your suggestion of completely clearing out the cmos and reloading bios I would not have considered it necessary as I had already reloaded the latest bios.
I’ll give it a week or so before I go out and buy a TB3 audio interface just to make sure its completely stable and booting normally. Thanks again for your help as the card would have been still sitting in its cardboard box for a quite a while before I tried to install it again.
One last thing! I take it the DisplayPort cable from the graphics card to the Titan AIC only needs to be connected for when using a TB3 port to drive a display. My AIC seems to be working fine without the DisplayPort cable connected!
Happy to hear that your issues seem to have been resolved!
Hope that this thread remains a shining beacon visible via Google showing that there is in fact demand for the latest Thunderbolt revision on AMD platforma
I’m curious how Intel’s going to sabotage Thunderbolt functionality within the upcoming USB 4.
I think from what I have read on USB4 implementation, that its 2 years away from becoming mainstream. Not only is there the time lag there is also the fact that TB3 is an optional inclusion in the USB4 spec. Its up to the manufacturer’s to implement the TB3 protocol into USB4 for their devices.
Device makers who opt to support Thunderbolt have to support Thunderbolt 3-compatible rates of 10.3125Gbps (for Gen 2) and 20.625 Gbps (for Gen 3) devices.
Big advantage is that the USB Type C connection will become the standard for all USB connections. Problem there is that depending on what type of USB4 implementation is used on the device will dictate if the cable needs to be TB compliant.
USB4 will be backwardly compliant to USB2 and above, but will obviously require adapters for the Type C connection. Hopefully TB compliant cables will become cheaper once USB4 becomes popular.
All of the above is why I bought a X570 Taichi MB and the Gigabyte Titan Ridge AIC. Combining the MB with the AIC gives me every combination that will come with a full implementation of USB4. I have the connectivity now and without the premium I bet gets added to USB4 with TB included.
Asrock is smart enough to get the jump on the other AMD, MB manufacturer’s by providing TB headers on some MBs and the TB chipset on others. The thing I dont get is why Asrock have not updated their TB3 AIC, maybe waiting to introduce a USB4 AIC?
Gigabyte Titan Ridge seems to be the best value TB3 AIC (Intel compliant) out there at the moment. It seems combined with Asrock MB a pretty good combination for future proof USB connectivity.
Forgot to mention: If you like to have no power at the USB devices when the system is shut down, enable “Deep Sleep in S5” (or S4/S5) in the UEFI.
Note:
Don’t physically disconnect AC to your PSU when this setting is enabled regularily - I don’t know why exactly but I’ve read that this might damage hardware.
S5 is the system properly shut down via the operating system, S4 is the hibernate mode where the system memory content is copied to the disk and then power is turned off.
Of course, if you enable this option you cannot wake the system from S4 or S5 with an USB keyboard et al., you have to use the Power Button.
Sleep Mode (S3) is not affected by this setting, although I wish there was such an option to cut power to specified USB ports in Sleep Mode (I use it daily since booting up and launching various services takes a while) to reduce wasted power (even if it is very little).
I disabled “Fast StartUp” in Win 10 to allow the PC to turn off power to the USB ports. Disabling Fast Start Up allows the PC to truly power off when shut down.
I didn’t need to change any bios setting to power down the ports. Fast StartUp is not needed by me for my use and really looks troublesome as it enables a hibernation mode.
Hi 1anX,
i’m using a X570 Taichi(2.50) with a GC Titan Ridge and I encountered your same behaviour with the 88 error in Dr. Debug Led.
I solved this issue downgrading to 2.10.
In my case, i found on the internet that it could be related to my iGPU on my 3400G.
Infact, in 2.50 i couldn’t disable the iGPU(BIOS automatically reset the “disable iGPU” to “enable”, probably a BIOS bug)
For the 2.10, my iGPU is automatically disabled and in around 15 boots i’ve had zero issues, so probably BIOS related.
Glad i found someone with my same error!
Could just be bugs in AMD’s AGESA (and nothing to do with the Gigabyte GC-Titan Ridge AIC):
Currently Zen 2 gets many optimizations/fixes with AGESA updates and there was even a note that only users with Zen 2 CPUs should use the latest BIOS versions.
By product name you might suggest that the 3400G is a Zen 2 CPU, but it is in fact a Zen+ CPU like the Ryzen 2000 series.
Thanks Normie for your replay.
Infact I still cannot reach 4.2 Ghz in CinebenchR15 even with 2.50.
2.50 slighlty Increased my Performance but since is a momentary CPU(waiting for the 3950X) I could wait at 2.10 for the Moment.
