Tuesday, January 30th 2018

An EPYC Threadripper: Der8auer Gets EPYC CPU Working on X399 Motherboard

So, maybe it isn't really working - but at least the system boots up all the way to the BIOS memory checks, where it then stops emitting life signs. Der8auer went through a sort of blind process to discover that there is a particular ID pin on EPYC that when covered, allows the CPU to be booted up by a X399 motherboard (in this case, an ASUS X399 Zenith Extreme). ID pins are nothing new, and basically tell sockets whether or not they should be powering up a particular CPU.

So what exactly does this mean? Nothing much - only that the sockets and pinouts are the same. The approach towards detecting the ID pin was a crude, brute force one, appending a piece of electrical tape to different parts of the CPU, narrowing down the search for a single pin. When this particular pin was covered, standby power finally kept on, and the motherboard ran through some initial boot steps until stopping at the D0 memory boot code. Der8auer thinks that a "simple" BIOS switch on this TR4 motherboard to an EPYC motherboard's BIOS would suffice to get the EPYC CPU running on this Threadripper motherboard. Check out the full video after the break.
Source: Der8auer's YouTube
Add your own comment

37 Comments on An EPYC Threadripper: Der8auer Gets EPYC CPU Working on X399 Motherboard

#26
RejZoR
I remember back in the day literally everyone modded R9500 to R9700. It was free unlock of texture units and you got a high end card from a mid end one. If money can be saved, people will do it.
Posted on Reply
#27
Punx223
Not so new member
I work with both TR and EPYC based platforms and I can tell you there is definitely more to this.

First off, there are specific pin differences that reference communication between different parts of the platform tropology..

by blocking the pin as he did, he simply stopped a hard fault condition. you could likely force a EPYC to perform a full bootup if you ousted specific pins but stability would likely be compromised as therte really is quite a difference between the two from a functional level especially in IMC.
Posted on Reply
#28
eidairaman1
The Exiled Airman
Punx223I work with both TR and EPYC based platforms and I can tell you there is definitely more to this.

First off, there are specific pin differences that reference communication between different parts of the platform tropology..

by blocking the pin as he did, he simply stopped a hard fault condition. you could likely force a EPYC to perform a full bootup if you ousted specific pins but stability would likely be compromised as therte really is quite a difference between the two from a functional level especially in IMC.
Thats where a eprom flash would need to occur for stability reasons
Posted on Reply
#29
lexluthermiester
RejZoRI remember back in the day literally everyone modded R9500 to R9700. It was free unlock of texture units and you got a high end card from a mid end one. If money can be saved, people will do it.
Ah ok, gotcha. Heck I did that. Had one of the good ones that literally became a 9700Pro with just the FW flash. Needed a new HS/fan but otherwise run great for years. That was still an elitist/hobbyist thing. And what this article describes is a serious rigmarole.

EDIT; Fricken auto-correct. Always getting it wrong..
Posted on Reply
#30
eidairaman1
The Exiled Airman
lexluthermiesterAh ok, gotcha. Heck I did that. Had one of the good ones that literally because a 9700Pro with just the FW flash. Needed a new HS/fan but otherwise run great for years. That was still an elitist/hobbyist thing. And what this article describes is a serious rigmarole.
He is testing the possibility of getting an epyc stable in a TR platform, sort of like 771 xeons being used in 775 boards.

If he succeeds it opens the platform up to further upgrades, now if he figures out a multiplier hack there would be some sweet rigs out.
Posted on Reply
#31
lexluthermiester
eidairaman1He is testing the possibility of getting an epyc stable in a TR platform, sort of like 771 xeons being used in 775 boards. If he succeeds it opens the platform up to further upgrades, now if he figures out a multiplier hack there would be some sweet rigs out.
Oh, totally get what he's doing. It's interesting for sure. Just a heck of a process. The 771 CPU to 775 socket mod was a great example of just what can be done. Did it once with a C2Q 9775. It worked and is still running(AFAIK), but was a serious effort. But that was an extreme edition CPU the owner bought for a great price thinking he could just slap it in. So that mod was handy and saved his day. However, it was a very rare exception to the rule and happened after Intel had abandoned the socket/platform.

My original point was that AMD has nothing to worry about as this will likely never become mainstream.
Posted on Reply
#32
eidairaman1
The Exiled Airman
lexluthermiesterOh, totally get what he's doing. It's interesting for sure. Just a heck of a process. The 771 CPU to 775 socket mod was a great example of just what can be done. Did it once with a C2Q 9775. It worked and is still running(AFAIK), but was a serious effort. But that was an extreme edition CPU the owner bought for a great price thinking he could just slap it in. So that mod was handy and saved his day. However, it was a very rare exception to the rule and happened after Intel had abandoned the socket/platform.

My original point was that AMD has nothing to worry about as this will likely never become mainstream.
Yeah only likes of myself would do such a mod lol
Posted on Reply
#33
lexluthermiester
eidairaman1Yeah only likes of myself would do such a mod lol
Exactly. People like us are hobbyist/elitists/tinkerers willing to give such an endevor a go because such things do not scare or intimidate. Like the guy in this article, we are the in the very small minority. Which is why folks come to people like us to solve problems most mainstream techs would scoff at. While I can't speak for you, it seems likely this applies.
Posted on Reply
#34
Punx223
Not so new member
eidairaman1Thats where a eprom flash would need to occur for stability reasons
The ePROM flash is not the end all tell all.... there are physical topology differences in the platform which have to be accounted for.

you can possibly "Make it work" but it would not be ideal by any means and likely from just a cursory look on my part of both platforms, there is a high likelihood you will lose core functionality if you get it to work at all.
Posted on Reply
#35
lexluthermiester
Punx223there are physical topology differences in the platform which have to be accounted for.
Correct, and the same problem exists for the 771>775 mod. It will be interesting to see if this works.
Posted on Reply
#36
Punx223
Not so new member
oh I definitely agree and Roman is a very good dude and a friend.... and he's quite intelligent with this...

I would love to see how it sorts out... I kinda wish I had time to allocate as I could potentially double the research effort on this... but I digress.... things to do.... (adulting sucks)
Posted on Reply
#37
eidairaman1
The Exiled Airman
Punx223The ePROM flash is not the end all tell all.... there are physical topology differences in the platform which have to be accounted for.

you can possibly "Make it work" but it would not be ideal by any means and likely from just a cursory look on my part of both platforms, there is a high likelihood you will lose core functionality if you get it to work at all.
I know about bridging, it would need contact trace tape like 771 xeons did
lexluthermiesterExactly. People like us are hobbyist/elitists/tinkerers willing to give such an endevor a go because such things do not scare or intimidate. Like the guy in this article, we are the in the very small minority. Which is why folks come to people like us to solve problems most mainstream techs would scoff at. While I can't speak for you, it seems likely this applies.
It does apply, this kind of mod I would do, course there are times I just don't bother. @buildzoid does very insane franken-mods though.
Posted on Reply
Add your own comment
Jul 13th, 2025 02:23 CDT change timezone

New Forum Posts

Popular Reviews

TPU on YouTube

Controversial News Posts