Overclock.net banner
3041 - 3060 of 3091 Posts

·
Registered
Joined
·
363 Posts
Yes, at 10K they are screaming.

I think you would achieve a better result with spot cooling than putting it in case.

You would not run them at 10k speed permanently. You would / could set a PWM curve in BIOS monitoring your thermal probe. At their lowest speed they are no more noisy than a normal case fan. I'm pretty sure that most 'boards support at least one header that does zero RPM too.
 

·
Registered
Joined
·
60 Posts
Yes, at 10K they are screaming.

I think you would achieve a better result with spot cooling than putting it in case.

You would not run them at 10k speed permanently. You would / could set a PWM curve in BIOS monitoring your thermal probe. At their lowest speed they are no more noisy than a normal case fan. I'm pretty sure that most 'boards support at least one header that does zero RPM too.
Ok I’ll take a look at it. Thanks BTW. It is sad that high end board needs me to do this. BTW would you able to touch the caps between the vrm and cpu socket on your TUF Mk 2? I saw the picture and it has similar caps. See if they are hot after running stress test for sometime.
 

·
Registered
Joined
·
363 Posts
The TUF died a few weeks ago, Amazon have it now.

I think with the frequencies, voltages and core counts we are all using it is to be expected that we need to pay more attention to how and what we cool. We are all using pretty high end stuff these days - just look how they cool similar kit it in the "real" world to maintain stability. If we are fully loading our CPU's we should fully load the cooling too. If this means we have to spot cool certain areas to maintain our overclock, so be it.

Next time i'm stress testing I will try and put my hand around the areas you ask about, see if i can give you details. My stress runs are usually with all fans on 100%.
 

·
Registered
Joined
·
802 Posts
I think adjusting ODT is necessary to run high frequency with encore board. Can’t leave it on auto or it won’t even boot. But with simple tweak it can run 4000c15 at 1.54v or 4000c16 at 1.46v. I found out the most impact is PARK value.
I'll have to mess with it after this heat wave we are getting, Weird thing is it would go a week without a single issue and then act up. Then sometimes it would go a few days without issues and then act up.
 

·
Registered
Joined
·
60 Posts
I'll have to mess with it after this heat wave we are getting, Weird thing is it would go a week without a single issue and then act up. Then sometimes it would go a few days without issues and then act up.
Yes exactly my problem before I manually set the ODT/RTT. It sometimes train the wrong value.

By the way you have encore board care to check the temp of the electrolyte capacitors above the cpu socket during stress test? Say after running for 10 mins of cb23.

2515346
 

·
Old and Decrepit
Joined
·
2,314 Posts
I think that EK velocity is killing you in a bad way. They are OK at best on smaller chips but just cant handle the load of HEDT much less OC 10980XE. Best Block EK has going for HEDT is the Magnitude. Another alternative is the Optimus Signature V2. I tested both and they are comparable as is out of the box so it comes down to your specific needs. What EK has over optimus is less flow restrictions so better performance at low flow rates and a lot of customization variabilities to change the look, inlet side (without flipping the block upside down) of a "goofy" mount, and jet plates to adjust the bow of the cold plate. With optimus you get what you get and any modifications are up to you. I found with the Optimus it cools better and has better contact with the inner oring pulled (no jet plates like EK) I bought an optimus block with whats stated as a flat with the same cold plate but the construction of the block itself is supposed to give less deflection of the cold plate so less bow. I haven't had a chance to run it yet, that's coming soon. I have however found that the flat cold plate (not quite flat) for the magnitude works better than the standard one on my 10980XE sample and to date has given me the best results. They vary a bit, no two are the same.....go figure...Takes some time, patience and testing to see what works best. The next block that would outperform a velocity is a HealtKiller IV pro. I was running that before the Optimus and magnitude came out and it beat the pants off of anyting EK had to offer at the time and still holds true with the exception of the magnitude but it does come at a premium price.
 

·
Registered
Joined
·
60 Posts
I think that EK velocity is killing you in a bad way. They are OK at best on smaller chips but just cant handle the load of HEDT much less OC 10980XE. Best Block EK has going for HEDT is the Magnitude. Another alternative is the Optimus Signature V2. I tested both and they are comparable as is out of the box so it comes down to your specific needs. What EK has over optimus is less flow restrictions so better performance at low flow rates and a lot of customization variabilities to change the look, inlet side (without flipping the block upside down) of a "goofy" mount, and jet plates to adjust the bow of the cold plate. With optimus you get what you get and any modifications are up to you. I found with the Optimus it cools better and has better contact with the inner oring pulled (no jet plates like EK) I bought an optimus block with whats stated as a flat with the same cold plate but the construction of the block itself is supposed to give less deflection of the cold plate so less bow. I haven't had a chance to run it yet, that's coming soon. I have however found that the flat cold plate (not quite flat) for the magnitude works better than the standard one on my 10980XE sample and to date has given me the best results. They vary a bit, no two are the same.....go figure...Takes some time, patience and testing to see what works best. The next block that would outperform a velocity is a HealtKiller IV pro. I was running that before the Optimus and magnitude came out and it beat the pants off of anyting EK had to offer at the time and still holds true with the exception of the magnitude but it does come at a premium price.
Thanks. In fact I’m waiting for EK new waterblock. The one didn’t use thumb screw. Because my retention bracket start to get rusty already. The drawbak of this new waterblock is maybe if oneday I wanted to do direct die it may not be possible. Is magnitude possible to be used for direct die?

I heard optimus is good and it does looks good but I’m not sure if I like it because it doesn’t use spring so I think it is good because the pressure is too high? I’m not sure how it works without spring.
 

·
Old and Decrepit
Joined
·
2,314 Posts
I haven’t tried direct die so can’t say.
What I can say is with the IHS tighter is better. I tighten the hardware as tight as I could without going crazy. I saw at least one member saying he cut slots in the Optimus springless mount for more torque. I didn’t go that far. Just as tight as I could by hand and with the magnitude I went until the springs bottomed out with tool, you can feel it. Manufacturer rep for EK said it’s supposed to be to where top of mounting screws are supposed to be flush with the bracket recess where the screw goes. I’d start there, do a nice 30 minute Heat run then tighten it on down while TIM is still warm. Direct die changes everything and I’d probably follow EKs recommendation. You wouldn’t want to use the flat cold plate either.

You might want to reach out to EKJake either here or he frequents performance PCs FB group to see if he can better answer direct die questions.
 

·
OC...D
Joined
·
28,239 Posts
Always a good read (this thread)... albeit a drive-by reading. I hope to get back to regular OCN time soon!

hires_110301-F-RR679-717.jpg
 
  • Rep+
Reactions: MrTOOSHORT

·
Registered
Joined
·
60 Posts
So I found proof that the power plane from the vrm to cpu has not enough surface area.

I attached temperature sensor which came with the board to the back side of the socket exactly where the burnt mark on my previous board. And connect it to water in temp header (because I use the other header already but anyway it doesn’t matter).
Then I ran prime95 small no avx (need heavy synthetic load which I can easily replicate consistently).
Core at stock so it ran 38 all cores, cache at 33 and memory 4000c16.

This is with vccin 1.8 v (default).
2515561

Temp of the substrate reaching 77C.


And this is with vccin 2v.
2515562


The temp of the substrate is only 73C!
The lower substrate temp despite 200mv higher vccin is because at the same power draw (same oc same load), higher voltage needs lower current thus smaller power plane doesn’t run as hot.

I also posted this on ROG forum as I need Asus rep to chime in on this matter since I’m really afraid to oc with this board now. Notice that the substrate temp is much higher than the cpu!

I’m not sure how high board substrate can handle temperature and granted this simulation is using prime95 which is overkill but I imagine running cinebench at 47 will draw much much more power than running prime95 at 38.

I wonder if Asus made larger power plane will the bottle neck move to the socket pins and may damage the cpu... not sure how much amps lga 2066 power pins (don’t know how many are they) can take.

Any thought will be appreciated.
 

·
Old and Decrepit
Joined
·
2,314 Posts
Im not sure 77C is enough to cook the the substrate like what you experienced. That one is still baffling me. Mine is barely getting warm with an extended AVX workload with no offset at 4.8 VCCIN at 1.85, at least from what I can tell. I'm not talking a quickie either but Im not taping temp probes either. I ran the snot out of it almost 24 hours a day for a couple of months. I reached my fingers in there at about the 30 minute mark on a blender animation of the classroom render (again 4.8 with no offset) and it was warm but not scorching.
 

·
Registered
Joined
·
60 Posts
Im not sure 77C is enough to cook the the substrate like what you experienced. That one is still baffling me. Mine is barely getting warm with an extended AVX workload with no offset at 4.8 VCCIN at 1.85, at least from what I can tell. I'm not talking a quickie either but Im not taping temp probes either. I ran the snot out of it almost 24 hours a day for a couple of months. I reached my fingers in there at about the 30 minute mark on a blender animation of the classroom render (again 4.8 with no offset) and it was warm but not scorching.
Since I went thru multiple boards and they are behave similarly, does it means my cpu is power hogging?

77C is at stock cores (38 all cores with non AVX load) and I just tried 42 and it went straight to 86 within 9 mins. This is the substrate temp near the cpu socket where the power plane having the smallest surface (before it went into the pins).

I imagine with my previous 47 oc when my board burnt it went above 100C. Not sure if that is healthy temp for PCB substrate.

By the way thanks for the water vlock recommendation. I’m ordering magnitude see how it goes. If I don’t like it for some reason then I can return it and get optimus instead. I can’t try optimus first since from where I come from there is no optimus dealer so I can’t return it easily if I don’t like it.
 

·
Old and Decrepit
Joined
·
2,314 Posts
Optimus blocks are thin on the supply side. More of a boutique shop at this point. They were doing OK with the supply until they branched into GPU blocks, reservoirs and fittings. I’ve had private conversations with them and they said they were running several CNC machines with robotics but I’m having a hard time swallowing that when they post up they have to change the set up after running a product and it takes a couple of days. The CPU blocks roll out fairly regularly but no retail stores that carry them, that’s for sure. Only place I’ve seen them other than their own website is performance PCs. All the GPU blocks are pay me now and it will ship in a few months. GPU Blocks look nice and seem to perform well from the statements but I’ve yet to see any controlled comparisons. I’m just not patient enough to wait that long.
 

·
Premium Member
Joined
·
139 Posts
GPU Blocks look nice and seem to perform well from the statements but I’ve yet to see any controlled comparisons. I’m just not patient enough to wait that long.
I haven't seen any controlled comparisons either, but I have one on my FTW3 3090 and I am happy with it. In Hardware Info I see max GPU Temps of 33C, Max Hot Spot Temps of 46C and Mem Junction Temps of 54C after stress testing with Real Bench 2.56. Not the most rigorous workload, but will give one point of comparison. My loop is overkill to keep things quiet and runs a max delta T (air to water) of about 3C.
 

·
OC...D
Joined
·
28,239 Posts
I haven't seen any controlled comparisons either, but I have one on my FTW3 3090 and I am happy with it. In Hardware Info I see max GPU Temps of 33C, Max Hot Spot Temps of 46C and Mem Junction Temps of 54C after stress testing with Real Bench 2.56. Not the most rigorous workload, but will give one point of comparison. My loop is overkill to keep things quiet and runs a max delta T (air to water) of about 3C.
Yeah, I have an Optimus block on my 3090FTW3 Ultra (with a 450W bios loaded) and it works really well. It's the only block I have for the 3090 so no comparos... hot spot temps (memory junction node) stay well within the AOR and will reach 80C only during mining. Folding or gaming is high 50s low 60s.
I have compared the CPU blocks to Ek and Koolance. The Foundation block is the sweetspot for cost-performance. The V2 is a serious chuck of well machined metal, but cost $$$. The foundation is within a degree of the Koolance 380 and 390, both are better than the EK supremacy (no velocity here). Only down side to the Op and EK is the presence of a "venturi" jet design that forces the flow thru only a narrow part of the fin bed, and will tend to accumulate "gunk" if there is any in the loop. This is a design flaw in both, not a defect. Koolance never suffer this problem and performs the same in my experience.
That said, the Optimus products are very well made and look amazing!
 

·
Registered
Joined
·
46 Posts
Hey gang -

I'm a year or two late to this thread it looks. I just picked up a 10980XE from Microcenter last week and have been working on OC'ing it. I have a great deal of ignorance about all of the various bits that can be tweaked. Primarily I'm just used to pushing the core ratio, any AVX offsets needed, and Vcore values. I'm replacing a Silicon Lottery binned and delidded 7900X which could comfortably hit 4.7Ghz with a 3/5 offset and 1.237 Vcore. The unofficial goal of the 10980XE was to see if I could hit 5GHz all-core and keep it cool enough.

Uh. Sorta. Initial findings see a 1.367(!) Vcore will keep the chip from crashing CPU intensive applications at 5GHz. Anything lower, and it's crash city. After 10-20 minutes of gaming, the temps hit in the low-mid 80s*C, with a dedicated water cooling loop. Now I know the EK R6E monoblock I have isn't perfectly ideal, and I'm looking to replace it. But I'll bet that's only good for a few degrees C.

Dropping the speed to 4.8GHz sees a required Vcore of 1.250. That's pretty substantial, and easily drops the gaming temps by 10 or so degrees. Ultimately I'm not sure that 200MHz is going to make a huge difference when fragging, but after all of that, here's my question:

Is a 1.367 Vcore actually bad for the CPU? Is it just a case of struggling with cooling? Or are there other things I'd have to keep aware of?

Specifics:
  • Cooling path: Pump --> EK R6E monoblock --> AlphaCool 480x60mm radiator in a push/pull setup --> reservoir/pump.
  • Idle temps drop right down to 40*C and stay there.
  • Gaming temps hit low 80s*C at 5Ghz. 70-75*C at 4.8GHz.
Any thoughts? Thanks!
 

·
Registered
Joined
·
802 Posts
Hey gang -

I'm a year or two late to this thread it looks. I just picked up a 10980XE from Microcenter last week and have been working on OC'ing it. I have a great deal of ignorance about all of the various bits that can be tweaked. Primarily I'm just used to pushing the core ratio, any AVX offsets needed, and Vcore values. I'm replacing a Silicon Lottery binned and delidded 7900X which could comfortably hit 4.7Ghz with a 3/5 offset and 1.237 Vcore. The unofficial goal of the 10980XE was to see if I could hit 5GHz all-core and keep it cool enough.

Uh. Sorta. Initial findings see a 1.367(!) Vcore will keep the chip from crashing CPU intensive applications at 5GHz. Anything lower, and it's crash city. After 10-20 minutes of gaming, the temps hit in the low-mid 80s*C, with a dedicated water cooling loop. Now I know the EK R6E monoblock I have isn't perfectly ideal, and I'm looking to replace it. But I'll bet that's only good for a few degrees C.

Dropping the speed to 4.8GHz sees a required Vcore of 1.250. That's pretty substantial, and easily drops the gaming temps by 10 or so degrees. Ultimately I'm not sure that 200MHz is going to make a huge difference when fragging, but after all of that, here's my question:

Is a 1.367 Vcore actually bad for the CPU? Is it just a case of struggling with cooling? Or are there other things I'd have to keep aware of?

Specifics:
  • Cooling path: Pump --> EK R6E monoblock --> AlphaCool 480x60mm radiator in a push/pull setup --> reservoir/pump.
  • Idle temps drop right down to 40*C and stay there.
  • Gaming temps hit low 80s*C at 5Ghz. 70-75*C at 4.8GHz.
Any thoughts? Thanks!
Most people running a 10980XE at high clocks 24/7(4.8-5ghz) are using more radiator surface area than what you are.
A 10980XE at all core 5ghz and 1.367 VID is going to be well over 500w and that 480 isn't going to cope.
You say your idle temp is 40c, which to me seems higher than normal, My 10940x at 5ghz 2 core are 4.9ghz the rest of the cores idles at 30-33c.

At the end of the day you will more than likely never be able to cool all 18 cores at 1.367 VID. I personally run adaptive voltage as I have figured out how to get it to work properly but I personally would not push much past 1.3 VID.

What games are you playing that push the core to 80c?
 

·
Registered
Joined
·
46 Posts
Most people running a 10980XE at high clocks 24/7(4.8-5ghz) are using more radiator surface area than what you are.
A 10980XE at all core 5ghz and 1.367 VID is going to be well over 500w and that 480 isn't going to cope.
Even a fat 480x60? Fair enough. Adding more will be very challenging given my case is no longer made and finding parts for it is exceptionally difficult. I may be able to add another 240x45mm to the loop in push/pull. Uncertain if that will make enough of a difference to bother.

but I personally would not push much past 1.3 VID.
Any reason why other than the cooling issue?

What games are you playing that push the core to 80c?
The two I played are Rainbow Six Siege and CoD Warzone. Any FPS game is going to do the same thing: work one core enough to push it to its max speed, which will warm it up. The same thing happened with my OC'd 7900X, just not to the same temp levels (high 60s).
 

·
Registered
Joined
·
60 Posts
Hey gang -

I'm a year or two late to this thread it looks. I just picked up a 10980XE from Microcenter last week and have been working on OC'ing it. I have a great deal of ignorance about all of the various bits that can be tweaked. Primarily I'm just used to pushing the core ratio, any AVX offsets needed, and Vcore values. I'm replacing a Silicon Lottery binned and delidded 7900X which could comfortably hit 4.7Ghz with a 3/5 offset and 1.237 Vcore. The unofficial goal of the 10980XE was to see if I could hit 5GHz all-core and keep it cool enough.

Uh. Sorta. Initial findings see a 1.367(!) Vcore will keep the chip from crashing CPU intensive applications at 5GHz. Anything lower, and it's crash city. After 10-20 minutes of gaming, the temps hit in the low-mid 80s*C, with a dedicated water cooling loop. Now I know the EK R6E monoblock I have isn't perfectly ideal, and I'm looking to replace it. But I'll bet that's only good for a few degrees C.

Dropping the speed to 4.8GHz sees a required Vcore of 1.250. That's pretty substantial, and easily drops the gaming temps by 10 or so degrees. Ultimately I'm not sure that 200MHz is going to make a huge difference when fragging, but after all of that, here's my question:

Is a 1.367 Vcore actually bad for the CPU? Is it just a case of struggling with cooling? Or are there other things I'd have to keep aware of?

Specifics:
  • Cooling path: Pump --> EK R6E monoblock --> AlphaCool 480x60mm radiator in a push/pull setup --> reservoir/pump.
  • Idle temps drop right down to 40*C and stay there.
  • Gaming temps hit low 80s*C at 5Ghz. 70-75*C at 4.8GHz.
Any thoughts? Thanks!
The voltage is fine but the current draw isn’t. Low 80 does seems a bit high for gaming but i is not dangerous IMO.
As above poster said, if you’re looking at performance then I would suggest to push mesh instead of the core. Set the core to whatever temp you are comfortable.
 
3041 - 3060 of 3091 Posts
Top