Xeon Owners Club

[Toothless]

Ouch time to start looking on the board to see what gone wrong it could be from anything like burnt/missing pins, blown caps/mosfets, bad bios and the list goes on from there


That why i always check what way the pin header is facing before plugging in the fan as it only goes one way in. I managed to kill the fan header on the asus p5q premium by probing with the dmm and i had a sneeze and the probes touches something else and kaboom magic smoke came out of somewhere of the board that was hard to swallow to be honest

I'm waiting to be able to afford two coolers for it. We just moved and first thing GF got was a couch. Woooo.

 

[freeagent]

The one on my R3F literally exploded lol. Same with the Z77 except that shut the board down, and to be honest it wasn't the same after that. My Rampage Formula X48 let out a spark but it kept on going like its bigger brother. It still works fine too, minus the header. Nice thick pcb, the Z77 is flimsy. But it was still quick

 

[SlamHammer]

I'm waiting to be able to afford two coolers for it. We just moved and first thing GF got was a couch. Woooo.

Sounds like you also need a new GF

 

[Caring1]

No boot.

Leave the Primary CPU in, and remove the other from the socket, then try.

 

[Toothless]

Sounds like you also need a new GF

6 and a half years with her and we finally got own place. Of she's happy I'm happy.

Leave the Primary CPU in, and remove the other from the socket, then try.

Just need coolers first.

 

[8BitZ80]

I'm waiting to be able to afford two coolers for it. We just moved and first thing GF got was a couch. Woooo.

Hmm, I guess she bought a couch because that's where you're gonna be sleeping when you buy two LGA1366 coolers without asking first. I'm joking lol

 

[lexluthermiester]

oooo yeah I am definitely waiting to snag one for a steal. 200 dollars or more for a 6.7 yr old processor? NO THANK YOU lol

No need to be unpleasant. This thread is a friendly one...

 

[Dinnercore]

I'm waiting to be able to afford two coolers for it.

Why do you have to be across the pond, I just recently sold my SR-2 and got the air-coolers left over since I sold with waterblocks. But shipping to the states is more expensive then buying these new

 

[Regeneration]

Two of my Westmere Xeons IMC degraded. WHEA errors started appearing in the eventlog. Had to reduce uncore and voltage. Guess even 1.3v is too high.

 

[Zyll Goliat]

Two of my Westmere Xeons IMC degraded. WHEA errors started appearing in the eventlog. Had to reduce uncore and voltage. Guess even 1.3v is too high.

Are you sure that CPU's the one that degraded?Maybe your mobo is the one that is become more "sensitive"

 

[Dinnercore]

Two of my Westmere Xeons IMC degraded. WHEA errors started appearing in the eventlog. Had to reduce uncore and voltage. Guess even 1.3v is too high.

What life did they live? OC, temps, usage?

I ran my X5690s for over half a year at 1.4V+ on Vcore and 1.375V VTT with up to 1.7V VDIMM to the memory and saw no change. I reverted them back to stock recently and stress tested them at defaults to see if I can still sell them, they passed every test fine and I had no trouble running them at 4,6GHz OC.

I would assume that if these can hold 1.4V for half a year of daily usage, they should be fine at 1.3V for years.

 

[Regeneration]

2 light load servers, air cooled, HT on, moderate 4.2 GHz OC, 3x 4GB at 2300 1T 1.6v, 1.3v for both vcore and IMC, temps in the 50-60s.

Been running for 2 years for 24/7. Now both IMC slightly degraded. Need to drop uncore by 50 MHz or to increase QPI/VTT to 1.3125v.

 

[Dinnercore]

2 light load servers, air cooled, moderate 4.2 GHz OC, 3x 4GB at 2300 1T 1.6v, 1.3v for both vcore and IMC.

Been running for 2 years for 24/7. Now both IMC slightly degraded. Need to drop uncore by 50 MHz or to increase QPI/VTT to 1.3125v.

Might also be the caps on the board degrading, just a thought.

 

[lexluthermiester]

3x 4GB at 2300 1T 1.6v

That is what did it. I hate to say this, but you overvolted them to much. There is amplification effect at play with IMC voltage. 1.6v on the RAM itself and 1.3v for the controller equals fried CPU after prolonged use. If you had kept the RAM voltage at or near stock and kept the IMC voltage at or near stock, things would still be running strong.

If you want to get them to run stable, you'll need to return all of the RAM related settings to stock or lower.

 

[Dinnercore]

That is what did it. I hate to say this, but you overvolted them to much. There is amplification effect at play with IMC voltage. 1.6v on the RAM itself and 1.3v for the controller equals fried CPU after prolonged use. If you had kept the RAM voltage at or near stock and kept the IMC voltage at or near stock, things would still be running strong.

If you want to get them to run stable, you'll need to return all of the RAM related settings to stock or lower.

I slightly doubt that, as up to 1.65V DIMM is still stock for DDR3. The controller just wants to be within a margin of the VDIMM, for some termination reasons. You should never set them apart from each other too much, as in if you want to raise your VDIMM over 1.6V you also HAVE to increase the VTT over stock or else it gets troublesome. But when you run a safe 1.3V VTT you can go over 1.7V VDIMM just fine, I ran it like that without problems.

I still support the degraded caps, as you probably OC'd right to the edge of stability 2 years ago and these old server components have a busy life in warm conditions. Add your 2 years of 24/7 with air-cooled CPUs which probably did not help to keep the components close to the VRM super cold and you start to push your caps rated lifetime with these 10 year old components.
This is no big deal, just your power filtering getting slightly worse. That leads to more fluctuations for voltage during certain loads -> dips below the point of stability.

 

[Edwired]

I thought the vtt and vdimm have to be within 0.5v between 1.30 to 1.70v that would be under the safely margin of that?

In my case the vtt is 1,33750 and vdimm is 1.64v = 0.3025 it below the safely margin of 0,5v correct me if im wrong

 

[Dinnercore]

I thought the vtt and vdimm have to be within 0.5v between 1.30 to 1.70v that would be under the safely margin of that?

In my case the vtt is 1,33750 and vdimm is 1.64v = 0.3025 it below the safely margin of 0,5v correct me if im wrong

That is the number I had in mind as well. And from my limited experience with three Xeon platforms with Westmere that I used so far, these are safe settings.

 

[Edwired]

Then again it depend on the board overvolting some of the voltages. I may have to find all the voltages read points by dmm to confirm if they are overvolting or undervolting according to the bios set voltage as I dont feel comfortable setting the voltages blind as i dont know what the board is doing at the set voltages

 

[Toothless]

Why do you have to be across the pond, I just recently sold my SR-2 and got the air-coolers left over since I sold with waterblocks. But shipping to the states is more expensive then buying these new

Cheapest I can find half decent coolers is about 30-50 each.

 

[Dinnercore]

Cheapest I can find half decent coolers is about 30-50 each.

I have a pair of these, don't even know what brand or model those are. Got the mounting hardware, but given the weight and size shipping will be upwards of 50$.

 

[Toothless]

I have a pair of these, don't even know what brand or model those are. Got the mounting hardware, but given the weight and size shipping will be upwards of 50$.
View attachment 158081

PM incoming

 

[Regeneration]

These are 2 separate systems and 2 separate motherboards. Need to be very careful with DDR and IMC voltages on Westmere for a long term use.

 

[SlamHammer]

No need to be unpleasant. This thread is a friendly one...

Apologies!

Wasn't so much a Xeon bash but rather the pumped-up price for the 1660/80V2 Xeons.

The lowest I've seen reputable people sell them for is 170/180! If I'm not looking in the right places please tell me! lol

Keeping my fingers crossed though come fall/Christmas time for the killer sales!

 

[Edwired]

These are 2 separate systems and 2 separate motherboards. Need to be very careful with DDR and IMC voltages on Westmere for a long term use.

Every components will react differently depending on voltages that it can be misleading for overclockers stated safe and maximum allowed when the motherboard can play havoc on over or under voltages depending on set clocks and etc. Just wishing theres guide or diagram of read out points on motherboards but it limited since then

 

[lexluthermiester]

Apologies!

No worries!

Wasn't so much a Xeon bash but rather the pumped-up price for the 1660/80V2 Xeons.

A lot of prices in the used market have gone wonky. I think everyone can feel you on that one.