![]() | ||
|
|
General Liquid/Water Cooling Discussion For discussion about Full Cooling System kits, or general cooling topics. Keep specific cooling items like pumps, radiators, etc... in their specific forums. |
![]() |
Thread Tools |
![]() |
#51 | |
Thermophile
Join Date: Sep 2002
Location: Melbourne, Australia
Posts: 2,538
|
![]() Quote:
Yesterday I installed my old Hydra prototype onto my X800XT-PE. With stock air-cooling the GPU would climb to 65C under full load and was only capable of around 560MHz core before it started locking up in 3DMark. It didn't actually artifact until around 590MHz core, but it was not stable. I went out and picked up one of the newest and latest Zalman flower GPU sinks. The GPU core under load climbed to 59C, and I picked up pretty much no extra overclock. The Zalman was MUCH quieter than the stock heatsink though, so this was a good thing in itself. I finally put on the waterblock, and now the GPU peaks at a reported 29C under full load. Artifact free overclock rose to 625MHz, and lock-up free overclock rose to 595MHz, or around a 6% gain. I was somewhat surprised at the ultra low temperatures, but then we need to factor in that the GPU die is ~240mm^2 in size, and therefore has a MUCH lower heat density than most any CPU die. The BIG difference though was the noise level. No fans, the GPU is now silent, and is cooled FAR better than anything air-based that I could ever stick on the card. So to me it was worth it on the basis of having no noise, and enjoying some extra overclock, even if it was only 5%, and perhaps extending the life-span of the GPU somewhat as it now runs much cooler. |
|
![]() |
![]() |
![]() |
#52 | |
Cooling Neophyte
Join Date: Jan 2005
Location: UK
Posts: 96
|
![]() Quote:
![]() What's the point in overclocking if you don't take it all the way? ![]() I know it's a case of diminishing returns in a monetary sense, but I don't think anyone overclocks to save money. Hobbies cost money! |
|
![]() |
![]() |
![]() |
#53 | |
Cooling Savant
Join Date: Dec 2003
Location: France
Posts: 291
|
![]() Quote:
|
|
![]() |
![]() |
![]() |
#54 |
Cooling Savant
Join Date: May 2003
Location: New Smyrna, FL
Posts: 258
|
![]()
Once you've got the watercooling rig, adding a gpu block just a bit more expensive than a decent gpu hsf.
Hose routing, installing/removing a 'blocked' vid card, and pump and radiator capacity become minor issues as well. That zalman flower is pretty dinky, btw. I was thinking more along the lines of this: http://www.ocfaq.com/reviews/ArcticC...ilencer_Rev3/2 My 9800 r35 core is watercooled, but I'll certainly be looking for a card that has adjustable Vcore on the next go round. |
![]() |
![]() |
![]() |
#55 | |
Thermophile
Join Date: May 2001
Location: UK
Posts: 1,064
|
![]() Quote:
__________________
Once upon a time, in a land far far away... |
|
![]() |
![]() |
![]() |
#56 |
Cooling Savant
Join Date: May 2003
Location: New Smyrna, FL
Posts: 258
|
![]()
What's default?
|
![]() |
![]() |
![]() |
#57 |
Thermophile
Join Date: May 2001
Location: UK
Posts: 1,064
|
![]()
1.3V on GTs, 1.4V on ultras. Ultra bioses may be able to go to 1.6V, but I'm not sure on that (I have GTs).
Also people have reported that 6800s have a reasonably high mortality rate at 1.5V+
__________________
Once upon a time, in a land far far away... |
![]() |
![]() |
![]() |
#58 | |
Cooling Savant
Join Date: Nov 2002
Location: winnipeg
Posts: 129
|
![]() Quote:
|
|
![]() |
![]() |
![]() |
#59 |
Thermophile
Join Date: May 2001
Location: UK
Posts: 1,064
|
![]()
No only 1.4V and I got about 10% iirc. 6800 series cards can have issues with running at over voltage as the regulators tend not to be up to the job, the ultras will tend to do better as they have a higher power regulator setup installed.
I don't normally run my cards over 1.3 as the increased load on my cooling loop means I have to run the rad fan annoyingly loud.
__________________
Once upon a time, in a land far far away... |
![]() |
![]() |
![]() |
#60 | ||
Cooling Savant
Join Date: May 2004
Location: Northern VA
Posts: 383
|
![]() Quote:
Quote:
|
||
![]() |
![]() |
![]() |
#61 |
Cooling Savant
Join Date: May 2003
Location: New Smyrna, FL
Posts: 258
|
![]()
Can you softmod and get the extra pipes too? Best of both worlds.
|
![]() |
![]() |
![]() |
#62 |
Cooling Savant
Join Date: May 2003
Location: Okotoks, A.B. Canada
Posts: 726
|
![]()
was wondering if anyone knows the TDP for the 6600GT?
can't seem to find it on the net anywhere
__________________
"Great spirits have always encountered violent opposition from mediocre minds" - (Einstein) |
![]() |
![]() |
![]() |
#63 |
Thermophile
Join Date: May 2001
Location: UK
Posts: 1,064
|
![]()
GPU TDPs and power specs in general are notoriously hard to find.
![]()
__________________
Once upon a time, in a land far far away... |
![]() |
![]() |
![]() |
#64 | |
Cooling Savant
Join Date: May 2003
Location: Okotoks, A.B. Canada
Posts: 726
|
![]() Quote:
or should i change the peltier for a higher rated one... 120 watt or 172 watt? 80 Watt is the highest 40x40mm 12volt based peltier i can find... after that they are all 24 volt... a 120 watt 12 volt would be great cause i proboly could still plug that into my power supply... but i havn't seen any...
__________________
"Great spirits have always encountered violent opposition from mediocre minds" - (Einstein) |
|
![]() |
![]() |
![]() |
#65 |
Thermophile
Join Date: Oct 2002
Location: U.S.A = Michigan
Posts: 1,243
|
![]()
Reading the minor debate regarding O/Cing of vid cards made me curious, so I went and took a average of 25 members of Xtreme.
Average O/C of these 25 guys is 643/600 on X800 Pro - X850XT PE. Lowest I noted was a 519/510 & the highest was a whopping 770/661. A fairly high number, (most?), of these cards are volt modded. Overclocks like these would jack up the heat disapated a good deal I would think. And that would make the GPU block a much larger secondary heat supplier. Last edited by Blackeagle; 02-20-2005 at 03:51 AM. |
![]() |
![]() |
![]() |
#66 | ||
Cooling Savant
Join Date: May 2003
Location: Okotoks, A.B. Canada
Posts: 726
|
![]() Quote:
Quote:
__________________
"Great spirits have always encountered violent opposition from mediocre minds" - (Einstein) |
||
![]() |
![]() |
![]() |
#67 | |
Thermophile
Join Date: May 2001
Location: UK
Posts: 1,064
|
![]() Quote:
__________________
Once upon a time, in a land far far away... |
|
![]() |
![]() |
![]() |
#68 | |
Thermophile
Join Date: Oct 2002
Location: U.S.A = Michigan
Posts: 1,243
|
![]() Quote:
Which gives a LOT more weight to the power consumption figures Butcher gave earlier in this thread IMO. And if a highly O/C'd & overvolted GPU is dissipating 80w+ then we are very soon going to require better designed water blocks closer to advanced CPU blocks in ability to remove heat quickly. Worked on the Hydra lately Cathar? ![]() EDIT: And from posts I've seen at Xtreme Resources & Xtreme Systems for some time the 80 watt TEC set ups DON'T handle a highly overclocked & voltmoded card at all well any more. |
|
![]() |
![]() |
![]() |
#69 |
Cooling Savant
Join Date: Sep 2004
Location: Los Angeles, Ca
Posts: 310
|
![]()
eagle, by ur statement I assume that you are suggesting that we start using impingement blocks on our gpus. blocks that work like cathars, the antartica, TDX, RBX, etc.
If this is what you are suggesting there is one slight problem with this. with two impingement style blocks, you would need to start with an iwaki class pump to get decent flow. a mcp650 or DDC would be brought to its knees with two impingement style blocks attached to it. Over at XS we have discussed this at some lenght and thought it might be a novel idea to try but realisticly its not feasable for most users. and on a further note, i really havnt noticed hte diff on my x800's oc between it running with in a 5C range of temperatures. 5C is the difference between a low end free flowing block like a maze4 and a storm block. and if im not going to see much of an increae in OC for the flow killing, why would i bother. I dont know if this is on a more wide scale, but i wish some people would test it. to test it i would just put ur fans up on full (rad fans) and see how high the oc is stable, and then turn them down and let the water get a bit warmer and see ur oc results. I know that this is far from scientific, but its better than nothing.
__________________
Excuse me, I believe you have my stapler. |
![]() |
![]() |
![]() |
#70 |
Thermophile
Join Date: Oct 2002
Location: U.S.A = Michigan
Posts: 1,243
|
![]()
How high are you runing your voltage & O/C Maxxx?
Point of that question is this. With higher voltage & overclocks the better cooling of the more advanced block would be of increased value. Will extend your cards life while giving higher performance as well. And while you saw no gain, your card may be pretty well maxed out regardless of a cooling increase 10X as great. If I recall right you're not thrilled with the O/Cing of your card, which tends to support my point regarding it being your limit, not the cooling. While increasing the temps by increasing water temps would offer some indications of the cards tolerance of temp variation at THAT setting, it's not the same as increasing the overclock and temps generated within the GPU. I'm sorry, I wouldn't trust that as really proof of anything other than how the card reacts to temps being increased from outside it. As to the Iwaki pump, I can arrange that, as can you. My interest in this is two fold, first and of greatest importance is what I can learn regarding how to best tweak my own rig. Secondary is the improtance to general usage. My comment towards Cathar was in jest, I know he's far to busy right now to be spending time on the Hydra. Though if Cathar does develop a GPU design in time I will most surely be interested in it. Cathar has never brought out a new block design that didn't push back the limits on cooling/performance. Nor do I think he'd disregard a GPU blocks head loss. Were he to design a GPU block with jets of the style of the Cascade/Storm I'm sure he'd draw from the SS Cascade's jet size/pattern. Look at the flow rate the SS had, hardly to restrictive & if he tweaked it a bit for more flow as the goal with nearly the same temps, even better. I'm not suggesting a silver GPU here, just use of the SS's impingement pattern. As heat output of next gen cards comes about this will be a increasing need. Give vid cards just 2 more generations and it's going to get interesting. I know some think TEC is the answer to this, but I'm not the only one who won't use them. There is going to be a large market segment for new and advanced GPU blocks. Blocks that offer better cooling, and without prohibative pressure drop. EDIT: The Storm's increase in performance vs the Cascade, as far as O/Cing goes is most likly limited. But it does cool a bit better. Last edited by Blackeagle; 03-06-2005 at 05:11 PM. |
![]() |
![]() |
![]() |
#71 |
Cooling Savant
Join Date: Sep 2004
Location: Los Angeles, Ca
Posts: 310
|
![]()
I see your point egale.. striving for that last bit of performance. but that ha been kept to the cpu more than anything. hence the lack of high restriction gpu blocks.
the gpu was/is a x800xt at 655mhz core (mem doesnt matter much in this situation) and the voltages was in the range of 1.8... not sure on the exact voltage as i have since removed the volt mod (has my WC setup in pieces and air cooling cant handle 1.8 volts), but it was in that range.. suffice to say this sucker would dump alot of heat. it was actually dumping more heat than my winchester.
__________________
Excuse me, I believe you have my stapler. |
![]() |
![]() |
![]() |
#72 |
Thermophile
Join Date: May 2001
Location: UK
Posts: 1,064
|
![]()
Don't forget the lower heat density of GPUs compared to CPUs is going to mean that maze vs impengement makes a bit less difference.
__________________
Once upon a time, in a land far far away... |
![]() |
![]() |
![]() |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
|
|