Water Flow Rate v's CPU/GPU Temperatues... my own findings...

pugheaven · 6 Jul 2020 at 17:55

So, basically just spent a couple of hours testing what happens when I crank up my pump from minimum to maximum (I don't know what minimum and maximum actual rates are as I don't have a flow meter)... however maybe someone can give me the specs of my pump, I never bothered finding out when I bought it lol.

So, this testing was performed in a customer loop with the following core elements:

Water Pump: Barrow Water-cooling DDC Pump, PWM Combination Pumps, LRC 2.0 RGB (not connected to PWN, manually controlled)
CPU Waterblock: Barrow CPU Water Block use for Intel X99 X299 Socket LGA2011 2066 Acrylic AURA RGB Light Copper Radiator Block (LTYK3X-04-V2)
GPU Waterblock: Eisblock Aurora Plexi GPX-A AMD Radeon RX 5700/5700XT Water Block
Waterloop Coolant: 1.5 litres approx of EK-Cryo Fuel Clear Premix Water-cooling Fluid
Radiators: 1x360mm and 2x240mm Radiators
Manual Water Temp: (although my inline was usually 1C higher than I recorded with the manual temp gauge physically stuck in the water at the filler cap)

So the room is pretty much set to 24.1C during all testing and the water is around 27-28 during all testing and I left each run for about 10 minutes to keep skewing the results with water temps etc... so, it's not an exact science, but it's as accurate as I could do in the time allowed and I ensured that the water temp was down to 27.5C before I started the next test! Also, I've picked the most strenuous testing simply to MAX out both CPU and GPU which then tests the efficiency of my loop inclusive of the radiator maximums just icnase they effect things.

So, obviously when in idle, whether the pump is on maximum or minimum, the temps really don't differ at all!!! So, here's the testing! I recorded all temperatures through HWiNFO64 which gives you current, maximum and average temps on all compenents.

For tests 1 and 2 for the GPU, Heaven Benchmark was ran for 5 complete loops then the maximums were recorded.

For tests 3 and 4 for the CPU, AIDA64 with only the FPU testing was ran for exactly 5 minutes for the CPU (I chose this rather than other tests as FPU calculations absolutely hammer ANY system rather than just CPU normal stress testing that don't use FPU. The temps are around 15C higher using FPU calcs than a mixture or non FPU so, I wanted absolute maximum to test against

For tests 5 and 6 which was a combination running both CPU and GPU at the same time, AIDA64 was ran as long as the 5 heaven cycles took which I didn't time and just left sitting there, but they ran for the same time so this is absolute worst case scenario which tbh would never happen in real life where you're maxing a CPU with FPU testing maxing 100% and maxing a GPU at 97-99%... however suppose it's nice to see if the loop can handle the huge heat generated with everything on maximum:

1. GPU ONLY Minimum Water Flow: GPU Core Max = 72.0C, GPU Hot Spot Max = 87.0C

2. GPU ONLY Maximum Water Flow: GPU Core Max = 69.0C, GPU Hot Spot Max = 82.0C3

3. CPU ONLY Minimum Water Flow: CPU Package Max = 93C

4. CPU ONLY Maximum Water Flow: CPU Package Max = 88C

5. GPU & CPU with Minimum Water Flow: GPU Core Max = 79.0C, GPU Hot Spot Max = 89.0C,

CPU Core 0 = 95C
CPU Core 1 = 97C
CPU Core 2 = 98C
CPU Core 3 = 85C
CPU Core 4 = 86C
CPU Core 5 = 97C
CPU Core 6 = 96C
CPU Core 7 = 93C
Overall Average = 93.38C

6. GPU & CPU with Maximum Water Flow: GPU Core Max = 72.0C, GPU Hot Spot Max = 86.0C,

CPU Core 0 = 93C
CPU Core 1 = 94C
CPU Core 2 = 91C
CPU Core 3 = 82C
CPU Core 4 = 84C
CPU Core 5 = 94C
CPU Core 6 = 92C
CPU Core 7 = 90C
Overall Average = 90.00C

So can we draw a conclusion from this? Well with my system and my config yes I think we can conclude that during all tests increasing the water flow decreases both GPU, CPU with both CPU/GPU running to their absolute maximum. Is this a surprise? Probably not, as I'd expect that the minimum on this pump is so low that either the GPU or CPU could in theory saturate the water greatly enough around the blocks because the pump isn't pulling that hot water away quite quickly enough so therefore it starts to build and once that water temp near the CPU/GPU heats up the actually elements heat as well??? It's kind of what I'm seeing possibly?

Also, I know these temps on the CPU look high but bear in mind, the CPU is an 8 core Xeon overclocked from 3.0ghz to 4.6ghz and the GPU is running maximum allowed values of 2200/1900 on core and memory with 1.200v maxed so... this is gonna generate a lot of heat and the testing i've done here is absolute mental and would never ever occur in real life. Anyone don't believe me, then test your kit using AIDA64 with the FPU only selected on a stress test, it's a system killer but great for stability testing lol. You'll see way higher temps recorded than you will ever get in real life. In gaming my CPU's are sitting in the 50-60's and the GPU is around the same as above actually in gaming so not much difference there.

So, what's everyones thoughts on this, don't get me wrong this is NOT an exact science BUT I've been as controlled as I can! Was good to just actually see for myself how it effects as I've heard a lot of differing opinions... and everyones systems will be different.

However guys, if you ran the same tests as me and published your results, would be interesting to compare?

EDIT:

Ok, as an amendment and looking at temps, the one thing that occured to me (and from a comment to this thread so cheers it's all good), I've not, since building this loop 4 weeks plus, re-tightened everything up i.e. the block on the GPU or the CPU and it does reccomend this for obvious reasons i.e. the thermal paste is now working properly and everything settles down and I could tighten the screws on both getting a better contact... so, I just re-ran the flow tests again with new results in RED: NOTE, the water delta is exactly 3C cooler today than yesterday so you can't really compare yesterdays original results with the updated results today, but you CAN compare everything in red as that was all done today and based on same water temps...

... and this show the same thing, in my system water flow DOES make a difference to CPU and GPU temps... by around 3C on the maximums and averages... I can also conclude that I do believe even though it's hard to compare now as my water delta was lower today that tightening up the blocks HAS made a better connection and is conducting heat a little better as well but that's up for debate.

I'd also conclude that running both the CPU and the GPU at the SAME time, obvously inceases temps on both which seems an obvious statement, however is true lol. And yes the GPU may seem high temps, but it's an 5700XT which top at 110C on the hot spot as per AMD spec, hence why people undervolt them to keep temps down, however, they dont' run like nvidia cards or AMD 570's and 580's etc they run HOT.

1. GPU ONLY Minimum Water Flow: GPU Core Max = 72.0C, GPU Hot Spot Max = 87.0C GPU Core Max = 65.0C, GPU Hot Spot Max = 79.0C (post tightening of block)

2. GPU ONLY Maximum Water Flow: GPU Core Max = 69.0C, GPU Hot Spot Max = 82.0C GPU Core Max = 63.0C, GPU Hot Spot Max = 76.0C (post tightening of block)

3. CPU ONLY Minimum Water Flow: CPU Package Max = 93C, (CPU Package Max = 91C post tightening of block)

CPU Core 0 = 90C (90C post tightening of block)
CPU Core 1 = 92C (89C post tightening of block)
CPU Core 2 = 93C (91C post tightening of block)
CPU Core 3 = 82C (79C post tightening of block)
CPU Core 4 = 83C (80C post tightening of block)
CPU Core 5 = 93C (90C post tightening of block)
CPU Core 6 = 93C (89C post tightening of block)
CPU Core 7 = 92C (89C post tightening of block)
Overall Average = 89.75C (87.12C post tightening of block)

4. CPU ONLY Maximum Water Flow: CPU Package Max = 88C (CPU Package Max = 87C post tightening of block)

CPU Core 0 = 88C (86C post tightening of block)
CPU Core 1 = 87C (86C post tightening of block)
CPU Core 2 = 88C (88C post tightening of block)
CPU Core 3 = 83C (76C post tightening of block)
CPU Core 4 = 84C (78C post tightening of block)
CPU Core 5 = 88C (87C post tightening of block)
CPU Core 6 = 88C (87C post tightening of block)
CPU Core 7 = 87C (84C post tightening of block)
Overall Average = 86.62C (84.00C post tightening of block)

5. GPU & CPU with Minimum Water Flow: GPU Core Max = 79.0C, GPU Hot Spot Max = 89.0C, (GPU Core Max = 72.0C, GPU Hot Spot Max = 85.0C, post tightening of the blocks)

CPU Core 0 = 95C (92C post tightening of block)
CPU Core 1 = 97C (92C post tightening of block)
CPU Core 2 = 98C (95C post tightening of block)
CPU Core 3 = 85C (82C post tightening of block)
CPU Core 4 = 86C (84C post tightening of block)
CPU Core 5 = 97C (94C post tightening of block)
CPU Core 6 = 96C (93C post tightening of block)
CPU Core 7 = 93C (91C post tightening of block)
Overall Average = 93.38C (90.38C post tightening of block)

6. GPU & CPU with Maximum Water Flow: GPU Core Max = 72.0C, GPU Hot Spot Max = 86.0C, (GPU Core Max = 68.0C, GPU Hot Spot Max = 77.0C, post tightening of the blocks)

CPU Core 0 = 93C (89C post tightening of block)
CPU Core 1 = 94C (90C post tightening of block)
CPU Core 2 = 91C (92C post tightening of block)
CPU Core 3 = 82C (80C post tightening of block)
CPU Core 4 = 84C (82C post tightening of block)
CPU Core 5 = 94C (91C post tightening of block)
CPU Core 6 = 92C (90C post tightening of block)
CPU Core 7 = 90C (88C post tightening of block)
Overall Average = 90.00C (87.75C post tightening of block)

tyler_jrb · 6 Jul 2020 at 18:08

Nice testing :cool:

P95 26.6 on max stress is worse than aida. Aida pulls about 160w on my 9900k, p95 pulls about 175w.

Custom torture test. Min + Max FFT size in KB = 12, and run FFT's in place ticked.

GPU temps are a waay high so i'm hoping its on air!

With regards to pump speed, never really found it makes much difference. Perhaps a degree or so from low-maximum speed on both GPU and CPU for me. But i'd rather my pumps run a little slower to increase lifespan, decrease noise etc. Even when i had a blockage in my loop at one point, a literal trickle of coolant flow the temps only increased by 5c or so.

pugheaven · 6 Jul 2020 at 18:15

tyler_jrb said:
Nice testing

P95 26.6 on max stress is worse than aida. Aida pulls about 160w on my 9900k, p95 pulls about 175w.

Custom torture test. Min + Max FFT size in KB = 12, and run FFT's in place ticked.

GPU temps are a waay high so i'm hoping its on air!

With regards to pump speed, never really found it makes much difference. Perhaps a degree or so from low-maximum speed on both GPU and CPU for me. But i'd rather my pumps run a little slower to increase lifespan, decrease noise etc. Even when i had a blockage in my loop at one point, a literal trickle of coolant flow the temps only increased by 5c or so.

Yeah I was gonna use P95 myself as I do also use it, but I know a lot of people don't like it so went AIDA but yeah, OCCT, AIDA64 and Prime95 all major system killers, if you system can run them for hours, it's stable lol. As for my GPU temps, 5700XT run hot as hell... when this was on AIR it wouldn't run and woudl throttle as the hot spot temp are allowed to hit 110C and my core was in the high 80s... as soon as I went to water it's 20-25C down compared to on air maxed out... whcih I could never do on air.

Yeah I'll be honest I've always said, in general use, I set my pump to minimum myself mate for exactly the same reason as yourself, less wear and tear and less noise and in the real world running games there's barely any difference, HENCE why I maxed everything out to try and eak out any differences when you're putting the system under massive strain and there is a difference clearly in my results between minimum and maximum... but I bet if I was just gaming you'd barely notice the difference.

Since I complete this testing, I literally set my pump back to minimum and unless benching or something I just don't bother turning it up!

Anyway was good fun tbh and interesting so all opinions are wlecome here, interesting to see other peoples findings on this subject as I can see so many conflicting opinions but not that many people actually doing it on different spec gear.

tyler_jrb · 6 Jul 2020 at 18:26

pugheaven said:
Yeah I was gonna use P95 myself as I do also use it, but I know a lot of people don't like it so went AIDA but yeah, OCCT, AIDA64 and Prime95 all major system killers, if you system can run them for hours, it's stable lol. As for my GPU temps, 5700XT run hot as hell... when this was on AIR it wouldn't run and woudl throttle as the hot spot temp are allowed to hit 110C and my core was in the high 80s... as soon as I went to water it's 20-25C down compared to on air maxed out... whcih I could never do on air.

Yeah I'll be honest I've always said in general use I set this to minimum myself mate for exactly the same reason as yourself, less wear and tear and less noise and in the real world running games there's barely any difference, HENCE why I maxed everything out to try and get out any differences and there is a difference clearly in my results between minimum and maximum... but I bet if I was just gaming you'd barely notice the difference.

Since I complete, I literally set my pump back to minimum and unless benching or something I just don't bother turning it up!

Should be a lot cooler than that I’d have thought though. Unless it’s pulling something silly like 600w I’d definitely investigate.

My 2080Ti rarely goes over 50c even in a locked cabinet with maxed voltage, power, core and memory speeds where the ambient gets to probably 30c in there.

Even when I had 980Ti SLI that drew 1100w from my hx1000i once on furmark they hit around 60c. Could be 5700XT nature but core temp should be IMO around 50c unless that’s after a while of testing and the loop is perhaps getting overloaded.

ttaskmaster · 6 Jul 2020 at 18:32

I think that, like everything else, there will be a number of variables that with give very different results depending on build. Different components will affect things differently, so while a very restrictive CPU block will hamper flow rates, the water will remain in there longer and pick up more heat. Same for a restrictive rad, the water will be in there and shedding heat for longer than if it just blasted through.
So as with everything else cooling-loop related, it's partly just a case of trying things out with the build you've got.

pugheaven · 6 Jul 2020 at 22:51

ttaskmaster said:
I think that, like everything else, there will be a number of variables that with give very different results depending on build. Different components will affect things differently, so while a very restrictive CPU block will hamper flow rates, the water will remain in there longer and pick up more heat. Same for a restrictive rad, the water will be in there and shedding heat for longer than if it just blasted through.
So as with everything else cooling-loop related, it's partly just a case of trying things out with the build you've got.

Yep completely agree mate, what works for one, will not work for another. The only reason I did this is to just see for myself, and more really for information that sometimes flow rate does matter when other say it doesnt. The person doing rthe build should be the one deciding what speed they want and by doing some simple testing like I've done you can then decide on what speed you run at.

pugheaven · 6 Jul 2020 at 22:59

tyler_jrb said:
Should be a lot cooler than that I’d have thought though. Unless it’s pulling something silly like 600w I’d definitely investigate.

My 2080Ti rarely goes over 50c even in a locked cabinet with maxed voltage, power, core and memory speeds where the ambient gets to probably 30c in there.

Even when I had 980Ti SLI that drew 1100w from my hx1000i once on furmark they hit around 60c. Could be 5700XT nature but core temp should be IMO around 50c unless that’s after a while of testing and the loop is perhaps getting overloaded.

Compauired to what these run, it'll be the voltage which AMD stadnard pusj 1,200mv I'll try and reduce the voltage, and I knwo that'll bring temps down mental, howver these 5700XT's out the box run into triple figures on the hot tspots and AMD knows this, however you can undervolt, wghich I used to do on my other 5700XT, not tried it on this one tbh simple because I ramped the clock to max... might have a tinker actually and see if I can reduce the voltage a little as that does make a massive difference to temps...

pugheaven · 7 Jul 2020 at 08:22

Ok, as an amendment and looking at temps, the one thing that occured to me (and from a comment to this thread so cheers it's all good), I've not, since building this loop 4 weeks plus, re-tightened everything up i.e. the block on the GPU or the CPU and it does reccomend this for obvious reasons i.e. the thermal paste is now working properly and everything settles down and I could tighten the screws on both getting a better contact... so, I just re-ran the flow tests again with new results in RED: NOTE, the water delta is exactly 3C cooler today than yesterday so you can't really compare yesterdays original results with the updated results today, but you CAN compare everything in red as that was all done today and based on same water temps...

... and this show the same thing, in my system water flow DOES make a difference to CPU and GPU temps... by around 3C on the maximums and averages... I can also conclude that I do believe even though it's hard to compare now as my water delta was lower today that tightening up the blocks HAS made a better connection and is conducting heat a little better as well but that's up for debate.

I'd also conclude that running both the CPU and the GPU at the SAME time, obvously inceases temps on both which seems an obvious statement, however is true lol. And yes the GPU may seem high temps, but it's an 5700XT which top at 110C on the hot spot as per AMD spec, hence why people undervolt them to keep temps down, however, they dont' run like nvidia cards or AMD 570's and 580's etc they run HOT.

1. GPU ONLY Minimum Water Flow: GPU Core Max = 72.0C, GPU Hot Spot Max = 87.0C GPU Core Max = 65.0C, GPU Hot Spot Max = 79.0C (post tightening of block)

2. GPU ONLY Maximum Water Flow: GPU Core Max = 69.0C, GPU Hot Spot Max = 82.0C GPU Core Max = 63.0C, GPU Hot Spot Max = 76.0C (post tightening of block)

3. CPU ONLY Minimum Water Flow: CPU Package Max = 93C, (CPU Package Max = 91C post tightening of block)

CPU Core 0 = 90C (90C post tightening of block)
CPU Core 1 = 92C (89C post tightening of block)
CPU Core 2 = 93C (91C post tightening of block)
CPU Core 3 = 82C (79C post tightening of block)
CPU Core 4 = 83C (80C post tightening of block)
CPU Core 5 = 93C (90C post tightening of block)
CPU Core 6 = 93C (89C post tightening of block)
CPU Core 7 = 92C (89C post tightening of block)
Overall Average = 89.75C (87.12C post tightening of block)

4. CPU ONLY Maximum Water Flow: CPU Package Max = 88C (CPU Package Max = 87C post tightening of block)

CPU Core 0 = 88C (86C post tightening of block)
CPU Core 1 = 87C (86C post tightening of block)
CPU Core 2 = 88C (88C post tightening of block)
CPU Core 3 = 83C (76C post tightening of block)
CPU Core 4 = 84C (78C post tightening of block)
CPU Core 5 = 88C (87C post tightening of block)
CPU Core 6 = 88C (87C post tightening of block)
CPU Core 7 = 87C (84C post tightening of block)
Overall Average = 86.62C (84.00C post tightening of block)

5. GPU & CPU with Minimum Water Flow: GPU Core Max = 79.0C, GPU Hot Spot Max = 89.0C, (GPU Core Max = 72.0C, GPU Hot Spot Max = 85.0C, post tightening of the blocks)

CPU Core 0 = 95C (92C post tightening of block)
CPU Core 1 = 97C (92C post tightening of block)
CPU Core 2 = 98C (95C post tightening of block)
CPU Core 3 = 85C (82C post tightening of block)
CPU Core 4 = 86C (84C post tightening of block)
CPU Core 5 = 97C (94C post tightening of block)
CPU Core 6 = 96C (93C post tightening of block)
CPU Core 7 = 93C (91C post tightening of block)
Overall Average = 93.38C (90.38C post tightening of block)

6. GPU & CPU with Maximum Water Flow: GPU Core Max = 72.0C, GPU Hot Spot Max = 86.0C, (GPU Core Max = 68.0C, GPU Hot Spot Max = 77.0C, post tightening of the blocks)

CPU Core 0 = 93C (89C post tightening of block)
CPU Core 1 = 94C (90C post tightening of block)
CPU Core 2 = 91C (92C post tightening of block)
CPU Core 3 = 82C (80C post tightening of block)
CPU Core 4 = 84C (82C post tightening of block)
CPU Core 5 = 94C (91C post tightening of block)
CPU Core 6 = 92C (90C post tightening of block)
CPU Core 7 = 90C (88C post tightening of block)
Overall Average = 90.00C (87.75C post tightening of block)

George Hudson · 12 Jul 2020 at 12:19

I trust you had your fan speeds locked while doing that testing, i.e. they were not ramping up and down. You need to have as many constants as possible so that it is only the water pump speed that can be attributable to the temp changes. I tested my system for a couple of months to achieve optimal setup, and it was the fan speeds on the radiators that made the biggest changes to temps for me.

pugheaven · 12 Jul 2020 at 21:18

George Hudson said:
I trust you had your fan speeds locked while doing that testing, i.e. they were not ramping up and down. You need to have as many constants as possible so that it is only the water pump speed that can be attributable to the temp changes. I tested my system for a couple of months to achieve optimal setup, and it was the fan speeds on the radiators that made the biggest changes to temps for me.

Hi mate, yeah they're not PWM fans well they are but not set to the mobo, theyre all controlled from a controller so all 7 fans are set at a constant speed. Agreed it's very hard to keep all th variables the same, however I believe I've got as close to that as possible... each run was at same water temp etc, all fans set to constant speed as they are everyday... yeah I believe I did, but there is always that error.

However, I'm folding now, and my GPU temps are currently 57C GPU/65V Hot Spot... and simply by upping the pump the temps will drop within a minute so there is deffo a precedent for fluid flow. However agreed, I'm doing minimum v's maxium, but I had to, as that will show the diffference. I generally have my pump set to 33% ish on the dial (taking minimum as 0 and max as 100 obviously) as day to day, why strain the pump when the day to day temps are perfect, waste of time as who cars if a CPU is 60C or 57C... it makes no odds until maxing everything out.

enjoyed doing this though just for my own benefit, and always good to share results./

Bubo · 13 Jul 2020 at 08:40

With that distro plate, are your gpu and cpu running in parallel? This may explain why your gpu temps look so high. There was another guy on here recently with a distro plate set up who was having temperature issues I think, was that you, can't remember. Anyway, this may be evidence that if you have unbalanced gpu/cpu flow due to the differing restrictions, then the only way to lower temps in the hotter side is up pump speed. Though gpu temps do still look high. But as you say probably not a problem in normal gaming.

pugheaven · 13 Jul 2020 at 09:42

I'm not convinced the GPU temps are high, on air this thing was hitting over 100C on the hot spot and now it's 25C lower and running at higher boost speeds maxing the card physically out, I dont' believe the temps for a 5700XT are high at all, if anything they're low. However, using something like FOLDING@HOME 24/7 on the GPU, it's currently on 59C GPU/65C Hot Spot. Don't forget, that I'm running 1,200mv which is the max to get the 1200/1900 boosts and reducing this voltage and boost massivly brings the card temps down, but to run what I want to run at I need that big voltage. If I run the GPU at default settings I can run the voltage at 1,110mv and it reduces temps upto by another 10C. Also in anything other than these tests, the CPU never gets past the low 70s i.e. gaming etc.

https://www.tomshardware.com/uk/news/amd-rx-5700-graphics-card-thermal-management,40144.html

AMD explained:

"Instead of setting a conservative, ‘worst case’ throttling temperature for the entire die, the Radeon RX 5700 series GPUs will continue to opportunistically and aggressively ramp clocks until any one of the many available sensors hits the ‘hotspot’ or ‘Junction’ temperature of 110 degrees Celsius. Operating at up to 110C Junction Temperature during typical gaming usage is expected and within spec. This enables the Radeon RX 5700 series GPUs to offer much higher performance and clocks out of the box, while maintaining acoustic and reliability targets."

... the order for the cooling is: Distro Plate > Pump > 240mm Rad > GPU > 360mm Rad > 240mm Rad > CPU > Distro Plate.

Jay85 · 13 Jul 2020 at 12:58

When i looked at your results i thought the gpu temps were extreme for watercooling! Then realised it was a 5700xt so kinda explains that. I think in comparison to my gpu temps on a 2080, running overclocked i hit around 44-45c on the gpu. I reckon i could be even lower if my cpu temps were a little lower still from a lower core voltage.

pugheaven · 13 Jul 2020 at 13:50

Jay85 said:
When i looked at your results i thought the gpu temps were extreme for watercooling! Then realised it was a 5700xt so kinda explains that. I think in comparison to my gpu temps on a 2080, running overclocked i hit around 44-45c on the gpu. I reckon i could be even lower if my cpu temps were a little lower still from a lower core voltage.

Yeah nVidia deffo have their temps under control, AMD just seems to have not cared this time around, however we all know that voltage and temps kill things... so, they may allow 110C but would any of us want that? Nope no chance... 45C on GPU is mental, mine idles around 39C

Jay85 · 13 Jul 2020 at 14:30

pugheaven said:
Yeah nVidia deffo have their temps under control, AMD just seems to have not cared this time around, however we all know that voltage and temps kill things... so, they may allow 110C but would any of us want that? Nope no chance... 45C on GPU is mental, mine idles around 39C

No way in hell would i want to run anything near those temps long term, its insane! I still remember when 80-90c on fermi cards was considered hot but these really do get hot especially when amd expects 110c to be fine still?!

Have you considered swapping the 2 x 240's for 2 x 360's to see if it helps with dissipating the heat? I feel like you have a lot of headroom still to drop those temps down.

pugheaven · 13 Jul 2020 at 14:50

Jay85 said:
No way in hell would i want to run anything near those temps long term, its insane! I still remember when 80-90c on fermi cards was considered hot but these really do get hot especially when amd expects 110c to be fine still?!

Have you considered swapping the 2 x 240's for 2 x 360's to see if it helps with dissipating the heat? I feel like you have a lot of headroom still to drop those temps down.

Yeah the mad thing is, the CPU side of it, when I run default the temps barely above 55C on absolute OCCT full load, it's mental the temps, so I know the water cooling is working like a charm... but who runs things at stock hey hahaha. I've found a nice point with the Xeon though where it's maxing out but temps and voltage are emaning it'll last along time. I can actually go 1.435v on that chip and run 4.8GHZ, but why the hell pump that much through, it's just gonna degrade and I'd like to keep this rig running for more years just upgrading GPU... it's more than powerful enough for my needs. TBH, I could run this 5700XT at default which is also fast enough and knock quite a bit of heat out of the system, but where's the fun in that lol