Limits of a 4P OC with 6180SEs

Joined
Aug 18, 2012
Messages
18
Greetings! I'm a n00b when it comes to AMD & 4P rigs so I hope someone can answer a few questions. I've put together a 4P rig with a SM H8QGL-IF+, 4 6180SE's, and 32GB of Ballistix ddr3 1600 memory (1.5V 8-8-8-24). I've flashed the [H] Bios and have been testing overclock settings and stability for the last few weeks.

I'd been using a saved copy of a 8101 WU and trying to complete the unit at different settings. I found that a refclock of 225 was no problem but 230, and 228 results in a segmentation fault (although 228 made it 95% through the WU). I've now been running at 227 for a week and completed 10 regular SMP units and 4 8101 bigadv units. CPU temps are 40-44C and the system is drawing ~820 watts.

This morning, I noted that a new 8101 had aborted at 35% with a seg-fault and core status 8B (139). So am I still too close to the edge? Does anyone have any suggestions other than backing off? I don't see much mention of 6180SE's when building these rigs--was getting them a mistake?

Thanks!
 
If 225 is stable I'd just run it there and leave it. The 6180SE are clocked quite high at stock speeds and have the highest TDP rating in the 6100 series so that doesn't really leave much headroom for OC'ing.
 
I think most people are going in the range of 220 to 230. It really is depenant on the weakest chip of the group. Kepp in mind the windows, as they can make the difference:

Post
 
If 225 is stable I'd just run it there and leave it. The 6180SE are clocked quite high at stock speeds and have the highest TDP rating in the 6100 series so that doesn't really leave much headroom for OC'ing.

Yea, I realize that I am flirting with the top end of the 6100 family so my OC would be starting from a higher CPU clock. I was just a little surprised that 13%OC was all I could get. Still, I guess 11:30 TPF on a 8101 isn't too bad ;).

I think most people are going in the range of 220 to 230. It really is depenant on the weakest chip of the group. Kepp in mind the windows, as they can make the difference:

Post

Yep, I'd seen tear's post, though I'm not completely sure I understand it. From what it says, I am currently at the 'right edge' of the (215-230) range. I suppose I could try going up to 231+ and see how that goes. FWIW, I do not see any HT retries.
 
My three 6180SE systems run 225, 225, and 218. I have heard of very few sets of these chips that could do over 225.
 
My three 6180SE systems run 225, 225, and 218. I have heard of very few sets of these chips that could do over 225.

The two sets I had were running 232 and 228 :D

Lowering temperatures will help stabilize at higher refclocks. I ended up moving CPUs around in the sockets so CPUs with a lower freq. limit were in sockets that delivered more power. It takes way too much time to do the testing, so I'd suggest just keeping with 225.
 
Last edited:
The two sets I had were running 232 and 228 :D

Lowering temperatures will help stabilize at higher refclocks. I ended up moving CPUs around in the sockets so CPUs with a lower freq. limit were in sockets that delivered more power. It takes way too much time to do the testing, so I'd suggest just keeping with 225.

Wow ... I wish! You say 'had'--you got rid of them? I've got my four on water (wanted to do something different) and was debating a little extra cooling. Maybe I will give that a try...
 
If you have them on water look at adjusting the flow and/or getting a stronger pump or two in series.
 
If you have them on water look at adjusting the flow and/or getting a stronger pump or two in series.

With 4 blocks and 2 or more radiators, I would definitely do two pumps in a push/pull config.
 
If you have them on water look at adjusting the flow and/or getting a stronger pump or two in series.

With 4 blocks and 2 or more radiators, I would definitely do two pumps in a push/pull config.

I currently have MCP35X -> MCR420 w/ 8x GT AP15 push/pull -> 4x Koolance 370SA and temps max out @ 38-42C with ambient @ 27C now that I've backed off to a 226 refclock. Flowrate is around 1 gal/min.

Is it worth adding more pump / res given that most of procs max out around 225-226? If I do this, does it make more sense to put the extra pump / rad in the middle of the loop like:

res->pump1->4x120rad->cpu1->cpu2->pump2->2x120rad->cpu3->cpu4->res

or would one of the pump pairing blocks like this be just as effective? I think having the second rad after the second cpu block makes sense regardless, yes?
 
I have one rig with dual DDC's in a double XSPC top similar to the above, and the other with a D5
 
I have one rig with dual DDC's in a double XSPC top similar to the above, and the other with a D5

Thanks, Biffa. How do you like that XSPC top? Anyone reading this use the Swiftech top?

I'll do a little more digging on the dual pump tops as an option. It would make my plumbing alot easier if I didn't have to add a second pump in the middle of the loop.
 
I don't think another pump will give u much for better temps. I'd think adding another radiator or some sort MAY help some, though temps in the mid 40's seems pretty good IMO tho I've not run amd cpu's in this type of config. But from a WC perspective, you can never have to much radiator :) Tho i think diminishing returns will quickly become apparent if you add more gear.

Putting a pump mid loop vs 2 pumps back to back really won't make much difference. The water doesn't really care where the pumps are as long as the water is moving. You'd probly add more tubing to plumb the pump mid loop that would counteract any supposed advantage there I'd think as well.
 
I don't think another pump will give u much for better temps. I'd think adding another radiator or some sort MAY help some, though temps in the mid 40's seems pretty good IMO tho I've not run amd cpu's in this type of config. But from a WC perspective, you can never have to much radiator :) Tho i think diminishing returns will quickly become apparent if you add more gear.

Putting a pump mid loop vs 2 pumps back to back really won't make much difference. The water doesn't really care where the pumps are as long as the water is moving. You'd probly add more tubing to plumb the pump mid loop that would counteract any supposed advantage there I'd think as well.

I agree completely on "no such thing as too much radiator." I'm thinking about a second pump because my flow rate is down around 1gpm right now which is ok but not great. I'm sure that any additional radiator will drop that below 1 gpm which is less than ideal from what I've read.

And, the only reason I would consider any of this is the possibility of pushing these cpus to a higher refclock and I'm still not sure this would be worth it.

Thanks!
 
Back
Top