XXL_Folding
- Thread starter Stan_Lite
- Start date
- Joined
- 27 Apr 2004
- Posts
- 107,330
- Location
- In bed with your sister
This is turning into a real problem child 
Got home from work on Friday and had a think about things over the weekend and decided to change a couple of minor things. I ordered a couple of bits and set about making the required changes to what was already done. One of the things I had to do was change one of the hoses on the mosfet block so I set about removing said hose. Unfortunately, in my haste, I yanked too hard and the whole block came away, stripping the threads. I had to take the whole thing apart to put the original heatsink back on. Disappointing but not a show stopper.
So, today I was expecting the new bits to come so I put it all together again and when the stuff came this afternoon, I got the loop set up just the way I wanted it. Plugged in the power and set about filling the loop. That was where I encountered the next problem. The loop wasn't filling properly. It was getting as far as the main motherboard block and not getting any further. "great", I thought, "a blockage". I switched off and took the whole thing apart again and dismantled the block but there was no sign of any obstruction
I tried the block on it's own but nothing. It would pump as far as the first port and stop. I tried switching the inlet and outlet but got the same result. It refuses to pump through, even though there appears to be no reason why it shouldn't. I decided at that point to abandon the motherboard block and I've re-fitted the original.
I'll have to leave it for now since, not only can I feel the rage coming on but, when I switched the pump off originally, the back pressure caused water to shoot out of the open fill port of the res, splashing all over the motherboard. I'll have to let that dry out.
For now, I'm having a cup of strong tea and some ginger snaps, after which, I shall tidy up, have tea, have a bath and then get blind drunk. I shall resume building tomorrow if my hangover isn't too bad. If it is, it'll have to wait another day
Got home from work on Friday and had a think about things over the weekend and decided to change a couple of minor things. I ordered a couple of bits and set about making the required changes to what was already done. One of the things I had to do was change one of the hoses on the mosfet block so I set about removing said hose. Unfortunately, in my haste, I yanked too hard and the whole block came away, stripping the threads. I had to take the whole thing apart to put the original heatsink back on. Disappointing but not a show stopper.
So, today I was expecting the new bits to come so I put it all together again and when the stuff came this afternoon, I got the loop set up just the way I wanted it. Plugged in the power and set about filling the loop. That was where I encountered the next problem. The loop wasn't filling properly. It was getting as far as the main motherboard block and not getting any further. "great", I thought, "a blockage". I switched off and took the whole thing apart again and dismantled the block but there was no sign of any obstruction
I tried the block on it's own but nothing. It would pump as far as the first port and stop. I tried switching the inlet and outlet but got the same result. It refuses to pump through, even though there appears to be no reason why it shouldn't. I decided at that point to abandon the motherboard block and I've re-fitted the original.I'll have to leave it for now since, not only can I feel the rage coming on but, when I switched the pump off originally, the back pressure caused water to shoot out of the open fill port of the res, splashing all over the motherboard. I'll have to let that dry out.
For now, I'm having a cup of strong tea and some ginger snaps, after which, I shall tidy up, have tea, have a bath and then get blind drunk. I shall resume building tomorrow if my hangover isn't too bad. If it is, it'll have to wait another day
Soldato
- Joined
- 7 Mar 2013
- Posts
- 3,321
- Location
- Caistor
This is turning into a real problem child
Got home from work on Friday and had a think about things over the weekend and decided to change a couple of minor things. I ordered a couple of bits and set about making the required changes to what was already done. One of the things I had to do was change one of the hoses on the mosfet block so I set about removing said hose. Unfortunately, in my haste, I yanked too hard and the whole block came away, stripping the threads. I had to take the whole thing apart to put the original heatsink back on. Disappointing but not a show stopper.
So, today I was expecting the new bits to come so I put it all together again and when the stuff came this afternoon, I got the loop set up just the way I wanted it. Plugged in the power and set about filling the loop. That was where I encountered the next problem. The loop wasn't filling properly. It was getting as far as the main motherboard block and not getting any further. "great", I thought, "a blockage". I switched off and took the whole thing apart again and dismantled the block but there was no sign of any obstruction
I'll have to leave it for now since, not only can I feel the rage coming on but, when I switched the pump off originally, the back pressure caused water to shoot out of the open fill port of the res, splashing all over the motherboard. I'll have to let that dry out.
For now, I'm having a cup of strong tea and some ginger snaps, after which, I shall tidy up, have tea, have a bath and then get blind drunk. I shall resume building tomorrow if my hangover isn't too bad. If it is, it'll have to wait another day
Sorry to hear that
RMA the block maybe?
PSU input current at maximum rating with 80% efficiency equals approx 7.8A....hardly massive
- Joined
- 27 Apr 2004
- Posts
- 107,330
- Location
- In bed with your sister
She lives!!!!
Bought off MM. To be honest, I got such a bargain on the bundle I bought, I wouldn't have the neck to complain - the guy only charged me a fraction of what the stuff was worth. Anyway, I haven't put it back together and tested it again yet, it might still work but, the fact that it has failed once puts me off using it - just in case.
Tried that. Nearly ruptured myself blowing so hard.
Anyway, enough of that and on with the update.
This is the rascally mosfet block - "it came off in me hand, honest guv"
Got the loop put together this morning and ran it for a few hours to test - all good.
Then came the moment of truth. I was dreading switching it on after all the trials I've been through getting the damn thing built. I had visions of switching it on and nothing happening or loads of beeping. Anyway, I took the plunge and on she went. After what seemed an eternity of flashing and churning - SHE LIVES!!!
Had a bit of an issue installing Windows. The SSD showed up fine in the BIOS but, when I tried to install Windows, it couldn't find it. Did a bit of Googling and came across something that said Windows installation sometimes doesn't recognise a drive if it has a dodgy MBR so I followed the directions and opened a command prompt, started diskpart and did a clean of the drive. Hey presto, when I tried the installation again, there it was
Temps are pretty good. The lowest core at idle is 17C and the highest is 24C. Running prime it's between 34C and 40C should give me a bit of headroom for overclocking.
Now Windows is installed and monitoring programs and updates installed, it's time to start overclocking the CPUs and RAM. I'm just about to have dinner then I'll get cracking with that. I'll find the max OC and run Prime 95 overnight to test thoroughly and tomorrow, I'll install the GPUs, OC them and benchmark. After that I'll tackle the cabling. maybe then I can get my dining table back.
Sorry to hear that
RMA the block maybe?
Bought off MM. To be honest, I got such a bargain on the bundle I bought, I wouldn't have the neck to complain - the guy only charged me a fraction of what the stuff was worth. Anyway, I haven't put it back together and tested it again yet, it might still work but, the fact that it has failed once puts me off using it - just in case.
Tried that. Nearly ruptured myself blowing so hard.
Anyway, enough of that and on with the update.
This is the rascally mosfet block - "it came off in me hand, honest guv"
Got the loop put together this morning and ran it for a few hours to test - all good.
Then came the moment of truth. I was dreading switching it on after all the trials I've been through getting the damn thing built. I had visions of switching it on and nothing happening or loads of beeping. Anyway, I took the plunge and on she went. After what seemed an eternity of flashing and churning - SHE LIVES!!!
Had a bit of an issue installing Windows. The SSD showed up fine in the BIOS but, when I tried to install Windows, it couldn't find it. Did a bit of Googling and came across something that said Windows installation sometimes doesn't recognise a drive if it has a dodgy MBR so I followed the directions and opened a command prompt, started diskpart and did a clean of the drive. Hey presto, when I tried the installation again, there it was
Temps are pretty good. The lowest core at idle is 17C and the highest is 24C. Running prime it's between 34C and 40C should give me a bit of headroom for overclocking.
Now Windows is installed and monitoring programs and updates installed, it's time to start overclocking the CPUs and RAM. I'm just about to have dinner then I'll get cracking with that. I'll find the max OC and run Prime 95 overnight to test thoroughly and tomorrow, I'll install the GPUs, OC them and benchmark. After that I'll tackle the cabling. maybe then I can get my dining table back.
- Joined
- 27 Apr 2004
- Posts
- 107,330
- Location
- In bed with your sister
Seems to be going well so far. Currently testing on 3.7GHz - max temp so far 65C on small FFT. Should hopefully be able to squeeze a bit more out of it.
- Joined
- 27 Apr 2004
- Posts
- 107,330
- Location
- In bed with your sister
Spoke too soon. 3.7GHz was as fast as I could get it to run stable. I've got 3.7 rock solid on 1.36V. Can't get above that even with Vcore at 1.45V. I haven't really touched anything else as I'm not sure what is likely to be effective. Do you think fiddling with the VTT would help any? I'm perfectly happy with 3.7 but, if I can get higher, I'd be even happier.
Never thought about it until you mentioned it. I think Coretemp may have been a bit out there. My room temp is about 21C through the day and about 17-18C in the evening. Would need to be a very special watercooling loop to cool it to below ambient
Nope. Won't be doing anything with that until I've finished with the build.
- Joined
- 27 Apr 2004
- Posts
- 107,330
- Location
- In bed with your sister
OK. Got her up and running for now but I'm not totally happy. For one thing, I can only get 3 GPUs to work in it. If I put the 4th 670 in it, the whole machine slows down and becomes jittery with erratic CPU usage. I've tested all the 670s individually and all 4 slots in use individually with no issues at all. With 3 cards in, it runs beautifully; as soon as I put the 4th card in, it all goes wonky I've had to put the 4th 670 in with the 690 in the gaming rig to fold on.
Secondly, the CPU performance is a huge disappointment. I thought with 16 threads at 3.4GHz I'd be getting awesome PPD. Instead, it's not even close to the performance from the 3770K (4.5GHz). I'm not sure whether it's a problem with my set-up or whether these E5620 Xeons are just pants. Either way, I'm not impressed
Having posted in the DC section, I've got a few things to try when I get home in 3 weeks. If anyone else has any suggestions, I'd be happy to hear them.
Anyway, here's a photo as she stands just now.
Plan for the future is to hopefully get it to run with 4 GPUs and to watercool them. I'm not sure if I'm going to upgrade anything now as I'm a bit fed up with it and really don't want to spend any more cash on it. I might just finish up the watercooling and call it a day. Thinking about re-doing the gaming rig. I quite fancy the Corsair Carbide 540 cube case. I might get a better motherboard and move everything over to that and get another 690 with block and make a watercooled quad sli gaming rig.
I've had to put the 4th 670 in with the 690 in the gaming rig to fold on.Secondly, the CPU performance is a huge disappointment. I thought with 16 threads at 3.4GHz I'd be getting awesome PPD. Instead, it's not even close to the performance from the 3770K (4.5GHz). I'm not sure whether it's a problem with my set-up or whether these E5620 Xeons are just pants. Either way, I'm not impressed
Having posted in the DC section, I've got a few things to try when I get home in 3 weeks. If anyone else has any suggestions, I'd be happy to hear them.
Anyway, here's a photo as she stands just now.
Plan for the future is to hopefully get it to run with 4 GPUs and to watercool them. I'm not sure if I'm going to upgrade anything now as I'm a bit fed up with it and really don't want to spend any more cash on it. I might just finish up the watercooling and call it a day. Thinking about re-doing the gaming rig. I quite fancy the Corsair Carbide 540 cube case. I might get a better motherboard and move everything over to that and get another 690 with block and make a watercooled quad sli gaming rig.
Well it certainly looks good. These CPUs gice as much as an overclocked 3930k if I am correct. What clocks are you running them on?
If you want bigger PPD I guess you have to run them in Linux overclocked to try and do big adv. WUs.
If you want bigger PPD I guess you have to run them in Linux overclocked to try and do big adv. WUs.
- Joined
- 27 Apr 2004
- Posts
- 107,330
- Location
- In bed with your sister
I'm running at 3.4GHz. I did get them Prime stable at 3.8GHz but, as I've found many times in the past, Prime stable ≠ Folding stable. I'll have to have a good tinker with things when I get back as I never really had time to fiddle before I left for work.
As for Bigadv, I think I'd probably be struggling to make deadlines just CPU Folding, never mind with GPUs as well. From what I've read hex cores would keep me comfortably within the deadlines but I'd think GPU Folding alongside would be a no-no. Apparently, adding one hex core to give 10 cores (20 threads) would be fine as well. Not sure If I want to go down that route though. The whole point of building this was for a kickass, multi-GPU system and I think I'd like to keep it that way. I'll look into other ways of getting into Bigadv mega points.
Shame about the quad gpu performance, its definitely possible to do, I didn't try it in the SR-2 I had, as in those days it was all about linux bigadv performance. But as you know I've been running a quad gpu's in a P6T7WS 1366 for ages without issue. Hmm.. wonder if it is the CPU's?
- Joined
- 27 Apr 2004
- Posts
- 107,330
- Location
- In bed with your sister
I know it's been a while but there have been a few developments. Due to hardware accidents (2 burned out 670s and a knackered motherboard), the new gaming rig has been cancelled. I had been running bigadv on Ubuntu on the SR-2 and was going to put the 690s in the new gaming rig. I bent some of the pins on the motherboard for the new gaming rig and rendered it useless so I've shelved it for now and I've decided to run the SR-2 with the 690s as my gaming rig (also Folding). I also got hold of a couple of hex core X5650 Xeons for a good price which has beefed things up a bit.
To that end, I have been redoing things and have made a few alterations.
Got her built back up and now setting up for gaming and Folding. Finished overclocking and now testing for stability - so far so good, all stable and temps at perfectly acceptable maximums. running Prime95 small FFTs and Unigine heaven at the same time. CPU temp maxing out at about 46C and CPUs at about 67C. Got a pretty decent OC on the CPUs - running at 4.0GHz (50% OC). 690s are a huge disappointment though. Best stable OC is +45 on GPU, that's 957MHz with a boost of 1084MHz
Never mind, here are some pics:
Side on - check out the neat GPU cabling Tubing is XSPC 13/19mm and fittings are all monsoon. The monsoon fittings look really good but are a bit of a pain; you need to tighten as many as you can before installing as it's really difficult getting into tight spots with the tightening tool.
Top rad - Alphacool NexXxoS UT60 Full Copper 360mm with 1450 rpm Gentle Typhoons in push pull configuration. You can just about get a fag paper between the bottom fans and the RAM - I've tried it
Front rad from the side - EK Coolstream XTX240 with the same fan configuration as the top rad.
CPUs - 2X X5650 Xeons with Koolance waterblocks. Oc'd to 4.0GHz
GPUs - 2X GTX 690s in quad SLI with EK waterblocks. Paltry OC of 957 (1084) gpu and 6480 mem.
Cabling at the back - about as neat as I can make it.
Front on. 2X Akasa fan controllers, cheapo DVD-RW and XSPC bay-res/pump combo with a D5 pump. Also showing front view of 240 rad. Leaving the front cover off for now so I can top up the res. Experience has shown me that the water level takes a few days to drop to its final level and the front panel is a pig to get off.
With the side panel on. I hate the side panel, it's fugly. Might consider modding it in the future.
That's all for now. As I said, I'm currently setting up and benching, will update again if there are any further significant developments.
I also need to solder the power switch. I managed to pull one of the wires off when I was building it (clumsy sod). I've currently got the reset switch wired up as the power switch until I can find someone with a soldering iron I can borrow (not buying one to fix one switch as it will probably never get used again).
To that end, I have been redoing things and have made a few alterations.
Got her built back up and now setting up for gaming and Folding. Finished overclocking and now testing for stability - so far so good, all stable and temps at perfectly acceptable maximums. running Prime95 small FFTs and Unigine heaven at the same time. CPU temp maxing out at about 46C and CPUs at about 67C. Got a pretty decent OC on the CPUs - running at 4.0GHz (50% OC). 690s are a huge disappointment though. Best stable OC is +45 on GPU, that's 957MHz with a boost of 1084MHz
Never mind, here are some pics:
Side on - check out the neat GPU cabling
Tubing is XSPC 13/19mm and fittings are all monsoon. The monsoon fittings look really good but are a bit of a pain; you need to tighten as many as you can before installing as it's really difficult getting into tight spots with the tightening tool.Top rad - Alphacool NexXxoS UT60 Full Copper 360mm with 1450 rpm Gentle Typhoons in push pull configuration. You can just about get a fag paper between the bottom fans and the RAM - I've tried it
Front rad from the side - EK Coolstream XTX240 with the same fan configuration as the top rad.
CPUs - 2X X5650 Xeons with Koolance waterblocks. Oc'd to 4.0GHz
GPUs - 2X GTX 690s in quad SLI with EK waterblocks. Paltry OC of 957 (1084) gpu and 6480 mem.
Cabling at the back - about as neat as I can make it.
Front on. 2X Akasa fan controllers, cheapo DVD-RW and XSPC bay-res/pump combo with a D5 pump. Also showing front view of 240 rad. Leaving the front cover off for now so I can top up the res. Experience has shown me that the water level takes a few days to drop to its final level and the front panel is a pig to get off.
With the side panel on. I hate the side panel, it's fugly. Might consider modding it in the future.
That's all for now. As I said, I'm currently setting up and benching, will update again if there are any further significant developments.
I also need to solder the power switch. I managed to pull one of the wires off when I was building it (clumsy sod). I've currently got the reset switch wired up as the power switch until I can find someone with a soldering iron I can borrow (not buying one to fix one switch as it will probably never get used again).
Associate
- Joined
- 3 Nov 2013
- Posts
- 320
- Location
- Surrey
Dat motherboard... Woah