Author |
Topic Search Topic Options
|
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-06-2020 at 4:23pm |
Fly happy wrote:
I can just move it to M.2_3 for peace of mind, and down there I won't have any problems getting a fan just over that heatsink. I don't think I will ever use more than 4 SATA drives either. |
Well what I am saying is.. if that huge cover plate is the heatsink for everything (top and bottom) then moving the drive isn't the solution, chilling the entire plate with focused air probably is, even if a flat fan is below the video card and blowing strait onto the plate.
As I said, the only slot I saw in the manual (quick look) that is PCIe x4 is the one your drive is currently located in. I did not dig any further.
So if you are sure the same speed will be present in the bottom slot it may be easier to cool.
One other consideration..
Benchmarks place the highest load on a drive. if the max you saw was 60c it should be OK but there is no video card in the system as of yet which will cover that location.
Its a bit of a conundrum right now. But if that huge cover plate is the actual drive heatsink it may be possible to focus air on that plate and bring the drive temp down.
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-06-2020 at 4:50pm |
In HWinfo64 this is basically all you need to roll
unless you wish to look at other items.
And let me be clear.. VCCIO/VCCSA will fluctuate from the BIOS preset so ignore those unless outlined
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-07-2020 at 12:26am |
Thanks, I'll hide a bit more in HWInfo64.
There are actually two drive heatsinks on that board.
One larger combined heatsink for M.2_1 and M.2_2. And a separate heatsink for the M.2_3.
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-07-2020 at 12:43am |
I think the deciding factor will be how it reacts to a video card covering the area.
They say those drives are ok up to 70c but my personal deal with any solid state device is if is rolling over 60c it needs a little help. If it is not going any higher than 60 then I can see letting that go but check it after a card is installed and see where it sits.
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-07-2020 at 1:43am |
A while back I had an idea of maybe getting a 2nd larger NVMe drive for Flight Simulator and games and so on. Another reason why I used M.2_1 was that the 2nd drive would be easier to install in M.2_3.
But now I'm leaning towards a 2.5" SATA SSD instead because of the price difference. An RTX3090 if I get one, will dig a big hole in my wallet. I was expecting a 3080Ti too, sort of filling the gap between the 3090 and 3080 but it seems that's not gonna happen at this stage.
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-07-2020 at 2:07am |
Yea they are playing the Vram game this round. The tech is awesome but they neuter the area the tech can work best with.
I honestly think that is disgusting.
They could have knowing how much is leaning on the graphics card today easily made a 10-11GB, 14-16GB and then the 24GB card but instead they played it based on what folks knew in the past.
I do have to say the price of the 24GB card in compare to the original Titan is far more in line but still.. I mean, they know what is coming and they know there will be the point where folks will need little extra Vram and then they will ring the cash register again on it. I am sure they will come out with another card after the initial 6-9 month grab that has more Vram specifically for that purpose.
I run Windows and the FS installs from the same NVMe on 2 partitions which allows me to back the entire drive up for a single recovery of the OS partition or the 2nd, or both.
I also have a old 2011 OCZ REVODRIVE3 X2 which was the predecessor to these NVMe drives on a PCIe slot which I modified with a little fan on the card heatsink and it still runs like a rocket today probably because I applied that little fan on the heatsink 9 years ago.
Everything else is mechanical SATA for backup and storage
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-07-2020 at 3:54am |
I might just go ahead and partition the NVMe drive as you did and use that partition for FS. And I need to create that overprovisioning partition. How large is your Windows partition and how large would you make that overprovitioning partition?
And for games I will probably get a 2.5" 2TB or larger SATA SSD.
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-07-2020 at 4:25am |
Partitions sizes are based on what you need to install.
I left 120GB for Windows over time and its 70GB used
That can be changed with partition software if needed later.
Overprovision is typically 9-10% of the entire drive itself and the Samsung software should define that for you and a simple click to set it.
remember my old mechanical drive recommendations? Don't fill a mechanical drive past 60% (that was for head scan and disk spin performance) and drive failure can occur if at least 10% is not free-space.
Duhhh same-same except you don't have to worry about the 60%
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-07-2020 at 12:36pm |
I read 930 GB total capacity of my drive in Local Disk (C:) Properties. I assume the 70 GB lost is the factory provisioning? 48 GB out of the 930 GB is used at the moment.
If I from the 930 GB capacity subtract 120GB for a Windows partition and 100GB for overprovisioning, it leaves me with 710 GB for Flight Simulator.
Heck, there is gonna be space for more software there. How much of that partition would you say can be used?
Is there a partition limit as well as a drive limit?
From the Bible: NEVER FILL A SSD MORE THAN 85%
Is that valid for NVMe drives too and if applicable, can the over provisioning partition be included in the 15% of space not to be filled?
And oh, since engineers love to test things, I removed the heatsink over my drive and ran the Benchmark in Magician, with everything else being the same. After 45 seconds the drive temperature reached 71C so I aborted the Benchmark. Enough discovery was made 
Next test will be to mount the drive in M.2_3, put the heatsink on and compare the temperature with the M.2_1 results. The heatsink for the M.2_3 is smaller than the other heatsink though. But we'll see what the results will show.
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-07-2020 at 1:18pm |
Ok there are some crossed wires here
I thought you said in the thread responses that the large plate you pictured was the actual heatsink for the drives.. you cant take the heatsink off the drive.
I only suggested removing it if it was simply a beauty plate for the motherboard and the real heatsink was underneath it.
NEVER FILL A SSD MORE THAN 85% a good engineer always leaves room for f**kups
You will naturally lose some drive space in formatting however it is the Samsung software that sets the over provisioning area, nothing else. Look for: Over Provisioning option in the Samsung Magician software and let it do the work
It may have done that automatically when the software was installed.. that I cant tell you but the Over Provision area in the software will show you if that is the case with the GUI
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-07-2020 at 1:29pm |
Heatsinks and they are going back in.
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-07-2020 at 1:38pm |
The heatsink for the chipset in the red circle is covered by the plastic cover so wouldn't you say that cover is best left removed?
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-07-2020 at 3:43pm |
I may have taken the wrong drive temp reading before
The drive is now in slot M.2_3 and the temperature in Magician was 48C at the end of Benchmark run 3. I didn't notice before that you could go back to Drive Information and see the temperature during the benchmark so I took the reading from HWInfo.
In HWInfo you can see there is a sensor that reads 63C and that sensor read 3C degrees higher than during the Benchmark Run with the drive in slot in M.2_1.
Anyways, I've been hard at work and I have a fan over the M.2_3 heatsink now, which I will connect and test with tomorrow.
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
Avidean
Senior Member
Joined: February-04-2012
Points: 1349
|
Posted: September-07-2020 at 6:38pm |
If I may interject. For what its worth, I working on the same hardware and I have a 2080ti above that spot. I have 3 970 EVO's in there and they are all low to mid 30's. I have 2 x 120mm fans at an angle from the edge of the board to the top of the GPU. They are just balanced there right now. I'll make a bracket to hold them roughly in that position:
|
 |
Avidean
Senior Member
Joined: February-04-2012
Points: 1349
|
Posted: September-07-2020 at 7:34pm |
And again if I may. We have another problem. I don't think that OCCT is going to be safe to use because it seems for the 10 series in the setting it does not show individual core temps sensors only the CPU temp. 
If you guys want me not to post in this thread just say so. I'll be very offended but I know Nick doesn't care if he hurts my feeling! 
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-07-2020 at 11:37pm |
Interesting fan solution. I see you got your RAM sticks sorted too. So you got all three M.2 slots occupied? And are those drive temperatures from when running benchmarks in the Magician software using the magician temperature? As you may have seen, I was taking the reading out of HWInfo and there the max temp was 63C and in Magician 48C.
Let's what Nick has to say about OCCT, but maybe we can still use it but monitor the core temps manually in HWINFO?
It's totally fine and appreciated if you share your findings and knowledge here.
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 12:32am |
Nick is seriously spent.. been hacking BIOS code and other stuff
@ Hans.. never take drive temps from anything other than the manufacture software that installed the driver and reads it correctly. If that drive is reading 48c after 3 back to back benchmarks from the same software, its ducky, and don't remove any plates from the motherboard unless you are sure of what you are doing.
HW monitor and HWInfo are for clocking and doing such jobs.. don't use them for making changes unless you know the areas in them you are looking at are reading the outputs correctly.
@ Avidean.. I don't understand this statement at all: I don't think that OCCT is going to be safe to use because it seems for the 10 series in the setting it does not show individual core temps sensors only the CPU temp
I have testers that are using OCCT 4.5.1 that have not brought up any issues with what I posted and how to set the core temp limit in the software settings I outlined. If there is something you are seeing that I have not been informed of I need to see a image of this difference you are referring to.
And never us Asus or any other Windows software to make voltage BIOS changes. We make changes directly in the BIOS and that means lazy folks have to deal with reboots for simple changes. You use Windows software to do that, even if it is made by Asus.... you are asking for real trouble.
and of course I don't mind the shares here.. 
|
 |
Avidean
Senior Member
Joined: February-04-2012
Points: 1349
|
Posted: September-08-2020 at 12:55am |
As can be seen from the screen shot, the setting do not include Cores 0,1,2,3,4..etc just the CPU temp. I don't know where you got the idea that I was making voltage changes in windows. Wasn't me!  I make all my changes in the bios.
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 1:03am |
Ok I see what you are referring to now. I was never made aware of that so I will check into it but from what I see the single temp reference is fine as I have had confirmations from others with that board they were able to locate the tower watt load.
I'll inquire directly about it now.
I was making a general comment about the use of Windows software for voltage changes.
You mentioned at some point that were were installing Asus Intelligent Processor software in AI suite and I am just making sure you and no one one else is using that for any voltage changes.
|
 |
Avidean
Senior Member
Joined: February-04-2012
Points: 1349
|
Posted: September-08-2020 at 1:06am |
Fly happy wrote:
Interesting fan solution. I see you got your RAM sticks sorted too. So you got all three M.2 slots occupied?And are those drive temperatures from when running benchmarks in the Magician software using the magician temperature? As you may have seen, I was taking the reading out of HWInfo and there the max temp was 63C and in Magician 48C.
Let's what Nick has to say about OCCT, but maybe we can still use it but monitor the core temps manually in HWINFO?
It's totally fine and appreciated if you share your findings and knowledge here.
|
I had that G.Skill DRAM 2 fan bracket from my old Haswell build.
The 970 EVO+ temps are at Ideal for those temps.
I just checked it. Got up to 40 when benchmark was running
|
 |
Avidean
Senior Member
Joined: February-04-2012
Points: 1349
|
Posted: September-08-2020 at 9:55am |
Fly happy wrote:
|
I like this idea of removing the plastic ROG bling from here. I'll be doing the same. That piece of plastic is definitely shielding the chipset heatsink from the air flow from the fan I have directed at it.
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-08-2020 at 11:46am |
Agreed. It's in the motherboard box, and stays there  This little fella lowered the temperature of the drive from 48C to 36C after three benchmark runs. It runs at 60% speed, which is the speed where I can't really hear it.
It's in chassifan 3 now. I moved all three factory case fan connectors to the fan hub on the back of the case. There are four fans connected to the hub now which leaves me with one free chassi fan header.
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 2:26pm |
As long as you are sure those covers serve no other purpose than bling, that's fine. It looks to me from the picture if you find you need to use the lower slot for any card with a bit of size the fan may be in the way.
48c drive temp with 3 top benchmark load runs is well within limits. The system will never run the drives that hard.
The caps they placed in the center may be there to prevent some type of interaction between the video card and the board components beneath them so be aware of things like that and take a good look with that in mind if you intend to remove them.
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-08-2020 at 2:42pm |
Yeah I'll check for clearance once I have the video card. At this time I don't think I'll use any cards in the other PCIe slots. The thickness of that cover where it partly covers the top M.2 heatsinks and fully covers the chipset heatsink is 2.5 mm. And there is a "window" in it for beauty lights which I have turned off in BIOS.
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 2:57pm |
If there isn't going to be any cards down there and there is a side door fan it may be best to wait till the tower has all the final components in it and is closed allowing the natural airflow to be analyzed to see where any such modifications may or may not be needed.
The video card is going to create an a air draft too which is another reason the shotgun design, although harder to clean out over time, is a better solution as it draws air in and expels it out the back instead of blowing it everywhere in the tower.
|
 |
Avidean
Senior Member
Joined: February-04-2012
Points: 1349
|
Posted: September-08-2020 at 3:06pm |
Speaking of fans. I put this 200mm sucker in the top of the case which I
scavenged from a used HAF-X that picked up on Buy & Smell for $50
instead of paying $30 for a 200mm fan on scamizon! 
I'm just about ready to start watt load testing. I gave it a dry run yesterday at optimized defaults and XPM -400Mhz +0.02 v. System when through 1 hour of OCCT:Linpack without braking a sweat. I think I saw a max temp of 69c and the watts maxed out at 160 or 170 or something in that range. I was going to take the grill off the top of the case and replace it with a piece of Plexiglass with a 200mm hole for the fan but I think I'll skip that because I'd want to do it right and it would take me a day that I don't have. But I don't doubt that the grill is impeding the air flow from that fan a lot. I'd guess its generation a back pressure that would reduce the CFPM of that fan by at least 1/2. I have to go shopping with the wife so it will have to wait for now!  
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-08-2020 at 3:46pm |
Seems like Nvidia uses a push-pull fan system for the RTX 3080 and RTX 3090 coolers. The left fan pushes air onto the baseplate and exhausts out the PCI-E slot. The right fan pulls air through the finstack and heatpipes, which then gets exhausted right out the other side of the card and makes its way directly to your CPU cooler and RAM 
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 4:06pm |
That's a very weird design... I would have to get one and review it in use.
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 4:15pm |
If you are using a liquid head, I don't see a problem there but with a HSF? Then there is the memory itself. DDR4 runs naturally warmer.
Its going to be a apply and see what's up situation with that one.
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 4:20pm |
I can see something with that image.. the primary heat point hot spot which is the GPU is cooled with the rear fan. The other fan appears to assist but is targeted at a different set of components. That might not be as bad as it appears. The massive heat source is being ported out the back quickly and instead of the air traveling across a long deck of fins that tend to get plugged up with dust over time the areas to clean are far less likely to trap dust.
So it may actually not be such a bad design. Like I said it will take getting one and observe the design in action.
|
 |
Fly happy
Senior Member
Joined: October-10-2012
Location: Sweden
Points: 1053
|
Posted: September-08-2020 at 4:32pm |
It looks like they got half of it right 
My brain starts working on some kind of solution 
Mounting it vertically with the fan normally blowing air over the CPU and RAM facing towards the side panel, and a pull fan mounted in the side panel to extract the hot air out of the case.
Edit just saw your reply. We'll see how the card works in practice first.
|
Hans
W10 Pro, 10900K, ROG Maximus XII Hero, ASUS TUF GeForce RTX 3090 24GB Gaming OC, GSKILL F4-4000C15Q-32GTZR, Corsair H150i RGB PRO XT, Fractal Design 7XL, Corsair HX850
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 6:04pm |
Or if you are really cheap and lazy  you could buy something like this from home depot, cut and paint it, and use washers
Depends on how ridged it is I guess. If that's too flexible then something like this:
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 6:36pm |
Fly happy wrote:
It looks like they got half of it right 
My brain starts working on some kind of solution 
Mounting it vertically with the fan normally blowing air over the CPU and RAM facing towards the side panel, and a pull fan mounted in the side panel to extract the hot air out of the case.
Edit just saw your reply. We'll see how the card works in practice first.
|
Yea I think it might be time to back off on the mods till the box is built and all the components are singing together, then figure out any mods you may deem needed.
If you are sure there is no blinky light software installed into Windows and booting you can now go back into the BIOS and change the following:
ONBOARD DEVICES LEDS- When system is in working state [ALL ON]
But then this may appear right below it.. also make sure this is set to:
When system is in sleep, standbuy.. bla, bla, bla: [AURA OFF]
and leave the QLED on post code only
If the CPU at idle remains in a low state, then that gives you the pretty motherboard lights back without the hit. If the CPU is still being polled all the time then that wont work on your board... just shut it back off.
On my board I can set like that and there is no CPU hit. With my board it comes down to making sure there is no Windows blinky light software installed or running anywhere.
|
 |
Avidean
Senior Member
Joined: February-04-2012
Points: 1349
|
Posted: September-08-2020 at 9:00pm |
I was just about to start watt load testing and started HWmonitor just to check Vcore and the exe informed that a new veriosn was available which includes comet lake support!
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 9:18pm |
Well there ya go.. I would rather you use that anyway if it works because it is far more simplistic and easy to deal with for the clocking exercise.
I have confirmed it worked fine for CPU and other related voltages and temps in the last version as testing comparing HW Monitor and HWInfo matched for z4x/10 and z3x/9, just no watt load. I also confirmed the drive temperatures matched the SSD software, at least on my end.
HWinfo is nice but I think it's a bit overblown for our needs in a clock setup.
And just so everyone knows I gave Avidean a special CPU voltage/offset setup to start with that is not like the 9700K clocking guide.. everything else in the guide is the same with the exception of where some of the BIOS settings are located and I gave him that list too, but the voltage scheme had to change so don't follow that guide for additional CPU turbo voltage and + offset as I am still working that end of it out for z4x and 10 series.
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 9:33pm |
Avidean wrote:
As can be seen from the screen shot, the setting do not include Cores 0,1,2,3,4..etc just the CPU temp.
|
Did you try changing the dropdown at the top of the settings in OCCT to see if there was a difference in the list? Probably wont be but just confirm the cutoff for the CPU is set to 93c
|
 |
Avidean
Senior Member
Joined: February-04-2012
Points: 1349
|
Posted: September-08-2020 at 11:12pm |
Yes I check that but OCCT is the same. Anyway I am glad to be able to use HWMonitor for the sake of familiarity. Did my first run of OCCT at 5ghz and completed the one hour. Max temp was 90 and the power was 245w.
I suspect thats not so good since you ran the exact same setting and go 90c 203w.
Anyway on I go with an Additional Turbo Mode CPU Core Voltage bump of 0.01v 
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-08-2020 at 11:22pm |
No.. the higher the watt load the better the final clock. At least for the 9700K z3x board!!
All that means is your cooling system is able to dissipate more watts than mine and that's good.
MAYBE it depends on what Intel screwed with here in the end.
You are not looking for clock here!!! You are looking for the efficiency of the tower!!!
HOWEVER I am still working how Intel really messed with this chipset to make things look better than they really are so there is a lot more testing to do since I hacked a few BIOSs and Intel's z4x chipset register.
EDIT: From where I started you from, you might be able to speed things up moving up 0.02v instead of 0.01. Just watch that readout..
You are still running high support voltages and those drive watts/heat too but they also get you to where you need to be FASTER and then those might come down.. so get this stuff out of your head about the efficiency of your system.
You are not judging a clock or what the tower will do with it.. you are finding out with a stable setup the MAX WATT LOAD the tower can handle.
Clocking is a whole different ballgame after that
|
 |
Avidean
Senior Member
Joined: February-04-2012
Points: 1349
|
Posted: September-09-2020 at 12:06am |
Got it Boss!  I'm really enjoying this!  This is next level tinkering for me! 
|
 |
NickN
Certified Professional
Joined: November-21-2007
Points: 21104
|
Posted: September-09-2020 at 12:12am |
 This is a next level nightmare for me and be sure of reentry, deacceleration with course and pitch changes based on a set of changing environmental variables with a nose down perfect final drop to the runway right on... C/L.. without the damn hardware in front of me!!
I stepped into this cow pile and I will wash my shoes and finish it but after this, seriously I'm done.
There is something else you guys and everyone reading this thread need to seriously understand and this goes back years when it comes to clocking a system for games and typical use.
We are not and never were clocking computers for scientific research which round massive numbers in massive quantity for engineering be it mechanical, biological, Astrophysics (my 2nd skin), etc..
If that is the kind of stability you think are getting, that is false and let me be very clear here.. anyone who says they are running Prime95 small FFT's @ AVX2 and 5.1-5.3GHz on SAFE 24/7 voltages is a lying sack of sh*t even if they produce a (photoshop enhanced) image of the test results for bragging rights.
Now, it might be possible to stabilize 5.0GHz with these new procs and HT enable with that kind of stability but it will take voltages we are not working with for ALL CORE support and 24/7 use. That is why we don't clock a processor AT ALL for that kind of mission critical productivity.
We are not decompiling the human genome here or working out the size of Betelgeuse compared to Soul, we are clocking a system for gaming and typical use where errors in rounding to ............. have ZERO affect on the usage outcome, cause no CTD's or BSOD's.
If you think your system is stable for scientific research using these methods, you need your heads examined  
As long we all understand this... carry on. 
|
 |