Resolved Geekbench 5 killed my cMP 5.1

[apt80]

I just read that the new Geekbench 5 was out. I thought lets have a try, but I think it killed my cMP 5.1.

I have a 3,33 GHz 6 core cMP 2009 updated to 5.1. 48 GB memory. Sapphire RX 580 Pulse GPU. I think the gpu just killed the mainboard. The GPU has a single 8p power connection and it is connected with a mini 6 pin to 6 pin cable. I read in the Sapphire manual that it can run on a 6p port, and it has been running OK for a few weeks. I ran the GeekBench 4 benchmark and the Unigine benchmark a few times without problem.

When I started the Geekbench5 / compute / metal benchmark the machine went down and won't boot anymore. There is zero response from the power button.

Are there any fuses that may protect the motherboard?
What can I check?
Any advice?

I'm a bit worried...

Of course GB5 didn't kill my MP, I probably messed up the installation of the GPU.

 

[apt80]

it's not dead after all. I took the mainboard, the memory and the GPU out and put it back together again. When I plugged the power cord it immediately started to boot.

Maybe I started the thread too soon, but with no reaction from the power button at all I was a bit worried.

Still: does anyone know whether this was caused by some kind of protection?
Is it allright to run the RX 580 on a single 6pin cable? I'm assuming that I just get less performance. I have a 2x 6 pin (mini) to 8 pin ordered.

 

[JedNZ]

Northbridge rivets?

 

[apt80]

Northbridge rivets?

Could be, but I just cleaned it 2 weeks ago. I didn't take it completely apart but vacuumed and dedusted it.
The Northbridge T diode was 56 C at the time of the crash. I have the single processor version, it doesn't run very hot.

I guess that's pretty cool.

 

[mattspace]

The GPU has a single 8p power connection and it is connected with a mini 6 pin to 6 pin cable. I read in the Sapphire manual that it can run on a 6p port, and it has been running OK for a few weeks.

traditional practice here is to run it on both 6 pin feeds - wouldn't surprise me if you triggered an overload protection shutdown, and required (for whatever reason) an SMC / NVRAM reset , ie full disconnect from power. The disassembly was probably not necessary.

 

[apt80]

traditional practice here is to run it on both 6 pin feeds - wouldn't surprise me if you triggered an overload protection shutdown, and required (for whatever reason) an SMC / NVRAM reset , ie full disconnect from power. The disassembly was probably not necessary.

OK, so there is a protection, and it was reset by the disconnect. Good to know, thanks. I'm going to install the double cable asap.

 

[h9826790]

You didn't power your graphic card correctly.

You should use DUAL mini 6pin to single 8pin.

The PULSE RX580 can draw more than 120W via the 8pin when under high stress (which only connected to ONE mini 6pin in your case). Once go beyond this point, it may trigger the cMP's shutdown protection.

You can't boot up again initially, because it require SMC reset to boot again (also, a kind of protection). And most likely when reseat the GPU / RAM, you did the SMC reset unintentionally. That's why your Mac can boot again.

 

[Kai.Def]

You didn't power your graphic card correctly.

You should use DUAL mini 6pin to single 8pin.

The PULSE RX580 can draw more than 120W via the 8pin when under high stress (which only connected to ONE mini 6pin in your case). Once go beyond this point, it may trigger the cMP's shutdown protection.

You can't boot up again initially, because it require SMC reset to boot again (also, a kind of protection). And most likely when reseat the GPU / RAM, you did the SMC reset unintentionally. That's why your Mac can boot again.

My cMP 2010 12 core isn’t starting up - nothing when I press the power button nor when I try to jump start. I suspected to have done damage by not having had sufficient power to the gpu. Still trying to fix it. The other day I plugged in the power cable in the back & it just started (without having pressed the power button). Any ideas?

 

[bsbeamer]

My cMP 2010 12 core isn’t starting up - nothing when I press the power button nor when I try to jump start. I suspected to have done damage by not having had sufficient power to the gpu. Still trying to fix it. The other day I plugged in the power cable in the back & it just started (without having pressed the power button). Any ideas?

Using a UPS at all? Change any configurations internally?

I've had reboot/restarts happen after power outages, but that's it. Usually triggered by a UPS shutdown and setup in power preferences to do so. It's not always 100% reliable and occasionally does require intervention to manually restart.

 

[MIKX]

Yes, definitely ! You MUST get a two X Mini-6pin to a single 8 pin cable.

EG : from Amazon : https://www.amazon.com/COMeap-Express-Video-Adapter-15-inch/dp/B07J336WY4

 

[apt80]

It had been ordered already, now I know why it is needed.
Delivery in two weeks.

 

[tpivette89]

traditional practice here is to run it on both 6 pin feeds - wouldn't surprise me if you triggered an overload protection shutdown, and required (for whatever reason) an SMC / NVRAM reset , ie full disconnect from power. The disassembly was probably not necessary.

This is what would happen to my secondary 5,1 that my kids use to play Roblox. After a few minutes of gameplay the computer would automatically power off and I would have to unplug the power cord, then plug it back in, for the power button to become responsive again.

Although I have an RX 480 with only a single 6 pin connector, the solution was to get a dual mini 6 pin to 6 pin cable. This resolved the issue and I haven't had a problem since. I firmly believe if the OP gets the correct cable (that he said he has ordered) he will not have any future problems.