Opteron 175 OC result, please comment

C

C-Otto

Weaksauce
Joined
Oct 26, 2004
Messages
95
Hi there.

Opteron 175 (OSA175DAA6CD, CCBWE, 0550UPMW, 1387356C50140)
Abit AV8
4*1 GB DDR400 (Infineon and Kingston)
copper water cooling getting water from two 10L buckets (picture)

As my board does not allow DDR400 with 4GB of RAM, I normally run with DDR333. Enforcing DDR400 results in crashes.

With the RAM set to DDR266 (for the devider) and HTT multiplier to 4x (to be safe) I climbed my way up to 2.6 GHz with voltage increases up to 1.575V. After some tests I think that 2.5 GHz with 1.55V is stable. I even could switch to DDR333 (giving DDR378 effectively) without a problem.

The temperature of the power thingies on my mainboard is at about 80°C. They are rated at 100°C and more, so that should be OK (I hope so). I think having no fan inside the PC outside of the power supply is not that good :D

Turning the power thingies' alarm off made my sleep more enjoyable. The CPU is at about 60°C when I look in the BIOS. Unfortunately ABIT decided to use crap, so I cannot read the sensors from within the OS.

Comments and congratulations welcome :)

Bye,
Carsten
 
Thanks. I don't really thing 60° is much, do you? The thing is rated at something more I think.
 
You are on water, so you should be ok with higher votls, BUT looking at your temps, if at all accurate from BIOS. It's too high for my taste, I really wouldnt let it past 55C @ full load.

just my .02
 
Ok 60c is WAY too high for these chips too, also u need to keep in mind that they're made to run on less voltage than the regular A64 chips, stock is 1.35, if you're running temps that high on both your mobo AND you're CPU, id have to agree, it's only a matter of time before you fry that CPU. Your decision but if your planning on keeping it running for a while i'd tone it down abit. The extra gain from those crazy temps is not worth the the time you're taking off the lifespan of your chip, but thats just my opinion. ;)
 
SoCalDJ said:
Ok 60c is WAY too high for these chips too, also u need to keep in mind that they're made to run on less voltage than the regular A64 chips, stock is 1.35, if you're running temps that high on both your mobo AND you're CPU, id have to agree, it's only a matter of time before you fry that CPU. Your decision but if your planning on keeping it running for a while i'd tone it down abit. The extra gain from those crazy temps is not worth the the time you're taking off the lifespan of your chip, but thats just my opinion. ;)
Click to expand...

The chip is rated at 65°C, isn't it? I think 60° is fine, then.
 
I think in modern times water cooling means having a big radiator and built-in systems. I am "oldschool" and still have my buckets holding the water and a pump inside :)
I am considering using a radiator, though.
 
something is odd. No way in hell you should be hitting 60C with water at such a low voltage. Also your PWMIC is very hot...

Whats your watercooling setup and tell me you have fans running inside the case just for airflow..
 
60c on water? I've running water too and my temps never go over 39-40c under full load. also my PWMIC temps needed a fan over them to control those temps too.
 
Wait did you say that you "DO NOT" have a radiator? That would be your problem my friend. I am running a 175 also at 2.82 with two GTX's in the loop at 630 and I never go above 40C even after hours of looping 3DMark and Prime.

Something is wrong.

Oh and I would never run a CPU past 50C-55C. I do not care if its rated for 65C, running it at 60C is just nuts.
 
Yeah the hottest i have EVER seen the CPU core temp on my water setup was 45C, and that was full load on my hot 3200+ and it was doing 2.95 Ghz with 1.6 volts.

My X2 3800+ @ 2.85 Ghz with 1.6 volts only loaded (dual prime!) at 38C, the IHS was removed, but still, 60C is quite insane, and 80C on the PWMIC is also crazy. Readings are incorrect, or something is quite wrong.
 
I woudlnt trust the AV8's temps. I had one, and it said my 3500 was running @ 80C, I moved it into another board, and it reads @ 35C. I just thinks abits temp probe on the av8 sucks
 
I have no clue whether the sensor is correct. Because I can only check inside the BIOS I have no real data from the CPU before...

My water setup:

- huge copper block on the CPU (others have great results with it)
- pump inside a bucket
- two buckets connected, one with the pump, one receiving the water from the CPU
- total 20 liters

The only moving parts in my tower, besides the water, are the disk drives and the fan inside the power supply (BeQuiet 400W or so). I guess the air inside the tower is quite hot, so that would explain the temperature of the power thingies.

I will add some silent fans later and the radiator will be bought soon.

Thanks for your comments!
 
C-Otto said:
I have no clue whether the sensor is correct. Because I can only check inside the BIOS I have no real data from the CPU before...

My water setup:

- huge copper block on the CPU (others have great results with it)
- pump inside a bucket
- two buckets connected, one with the pump, one receiving the water from the CPU
- total 20 liters

The only moving parts in my tower, besides the water, are the disk drives and the fan inside the power supply (BeQuiet 400W or so). I guess the air inside the tower is quite hot, so that would explain the temperature of the power thingies.

I will add some silent fans later and the radiator will be bought soon.

Thanks for your comments!
Click to expand...

If you dont' have a radiator your setup is not going to work. Your not removing the heat from the water, in fact you're KEEPING the heat in the water. So 60C cpu + 60C water = 60C temp.
 
The water is not going to get to 60C... there's 20 liters of it.

trying some basic chemistry (which i hope is right)

60C - 20C assumed ambient = 40C delta, heat capacity of water = 4.186 j/ml, 20 L = 20,000 ml.

so, 40*4.186*20,000 = 3.3 million joules. Now if you run the computer a while you're eventually going to get there, but i'm assuming that wherever he's storing the 20 L, it's going to dissipate the heat before the water even begins to get near 60C. I'd be surprised if the water was more than 30C... but then again, maybe he's storing the water in some insulated container lol.

My advice is to open up the case, get some ventilation in there, see if that helps out at all. If you're still getting those crazy temps, just assume its the AV8 messing up.
 
Nice calculation!
The two buckets of normal size are just standing around and based on the look of my windows they lose quite a lot of water over time. I guess not enough, though...
I installed some fans and will buy a radiator.
 
OK.
1) I ordered the radiator, the Airplex Evo 360
see image
I will equip this with three Papst fans running at 5V (perhaps more)
2) I used some (more or less) ugly program to read out the µguru sensor. At first, with two buckets of warm water, I got 74°C. With fresh water I can lower the temperature to about 55°C.
tempdetail500.png
 
Okay, now I have my radiator.
Temperature is constant at about 60°C (for now?).
I overclocked to 2,7 GHz, but that is not stable. Now I am at 2640 MHz, I think somewhere around here it's stable.
 
C-Otto said:
Okay, now I have my radiator.
Temperature is constant at about 60°C (for now?).
I overclocked to 2,7 GHz, but that is not stable. Now I am at 2640 MHz, I think somewhere around here it's stable.
Click to expand...

What do you mean "constant?" Constant at load?
 
Are you still using the submerged pump in the water pails?
I am thinking you have a lot of tubing, which is no doubt effecting your flow rate, which in turn is hurting cooling, even with the rad.
 
Pump in bucket - Radiator - CPU - Bucket

The flow is quite good I think. I just measured 0.9 liters per minute.
 
Thats rather low.
It's only 54 LPH.

Why don't you just do a standard loop
Pump >>> Chip >>> RAD >>> Res >>> Back to Pump

It will clean up all that mess of pails and tubing you have all over the place.

Or get a second pump in that line, say right before the chip and the sub pump right after the rad.
 
60C at load is nuts man. Are you sure you are reading this correctly?

What is the reading at idle?
 
I don't think that 60° is that much. I have contact to a few German overclockers with similar processors, there 50° is quite normal with good cooling (and great OC results). Because they use a lower voltage than me, the 10° increase is OK I think.

At idle the temperature drops to about 35° (almost instantaneously). I have to use a strange program to read out the value from the microguru sensor, because ABIT does not provide information to the developers of lmsensors, but I think the values are what the chip tells me (the voltages are correct for sure). But I don't know how far I can trust the onboard sensor ABIT uses. I don't think it is mounted on the CPU's die.

Borgschulze:
You are using a single core CPU and have only 84% of my voltage. Now guess why your temperature is lower.

PS: I noticed video encoding (two threads) does not produce as much heat as distributed.net (OGR) does.
 
I'm also using very very bad cooling.

There is something wrong with your cooling if you reach 60c.
 
25C under load differential means the something is wrong dude. A decent W/C set up should be no more than 12-15C above ambient. A great W/C set-up might be more along the lines of 7-12C above ambient.

You are trying to argue that people you see are running 50c and 10C difference is "not that much". I would say 20% increase is a ton man.

I dunno, do what yo want, but you are asking for trouble IMO.
 
im still on air and my 148 is at 3+Ghz, and I a not close to 60c.....ouch.
 
You must log in or register to reply here.
Back
Top