Forums first comp: CPUz Benchmark

[Mr. Fox]

10 hours ago, J7SC_Orion said:

 

How are you cooling the DDR5 modules ? I've been following several DDR5 threads and it seems that DDR5 is quite sensitive to heat, including the onboard PMIC.

I don't have a good way to cool them as I did the DDR4 modules. They do get hot under certain circumstances (I have seen a max of about 52°C, which is very warm for memory) and start throwing TM5 errors due to the thermals. The same was true of my DDR4 modules. If they got hotter than about 40-45°C it would produce errors in TM5.

 

Until heavy duty DDR5 heat sink jackets (like those used for water cooling) are available, I have a 120MM fan blowing on the modules to try to keep them cool.

 

I may try some of my DDR4 heat sinks on these modules, but the jury is still out on whether they actually fit correctly. I tend to trust Igor more than a reddit users.

 

DDR4 RAM cooler to DDR5? Not a good idea! The differences in detail including 3D scan and instructions

WWW.IGORSLAB.DE

---

Today it's about non-existing coolers for unavailable RAM, so once again pure unicorn cinema for all who are only missing the RAM for the new Intel system with

 

[EHW Ai]

1 hour ago, Mr. Fox said:

I don't have a good way to cool them as I did the DDR4 modules. They do get hot under certain circumstances (I have seen a max of about 52°C, which is very warm for memory) and start throwing TM5 errors due to the thermals. The same was true of my DDR4 modules. If they got hotter than about 40-45°C it would produce errors in TM5.

 

Until heavy duty DDR5 heat sink jackets (like those used for water cooling) are available, I have a 120MM fan blowing on the modules to try to keep them cool.

 

I may try some of my DDR4 heat sinks on these modules, but the jury is still out on whether they actually fit correctly. I tend to trust Igor more than a reddit users.

 

DDR4 RAM cooler to DDR5? Not a good idea! The differences in detail including 3D scan and instructions

WWW.IGORSLAB.DE

---

Today it's about non-existing coolers for unavailable RAM, so once again pure unicorn cinema for all who are only missing the RAM for the new Intel system with

 

 

Interesting, including Igor's input.

 

I happen to have an unused set of four EK-RAM Monarch Modules (enough for eight sticks) that I bought back in 2017, as well as the water blocks to go with them:

 

 

Once the DDR5 ram I ordered arrives and I've had a chance to make sure they work, I might see whether that EK Monarch setup will keep the ram cool enough. EK claims these parts are compatible with a large number of Z690 motherboards, including my EVGA Dark; and I'm hoping to either confirm or negate that.

 

If I'm not mistaken, you just bought a set of the same ram as me (F5-6400J3239G16GX2-TZ5RK), didn't you?

 

Added edit: oops, sorry about that; that was @geriatricpollywog over on OCN.

Edited March 6, 2022 by iamjanco

[Pawelr98]

Lowest tier Alder Lake, the allmighty G6900 at stock speed with a single stick of 4GB 2400MHz DDR4 ram.

 

And then my daily driver, Core i7 6950X clocked at 4.2GHz with 2x16GB 2666MHz DDR4 ram.

 

By single core score, you can tell how much has improved over the years IPC wise.

[ENTERPRISE]

1 hour ago, Pawelr98 said:

Lowest tier Alder Lake, the allmighty G6900 at stock speed with a single stick of 4GB 2400MHz DDR4 ram.

 

And then my daily driver, Core i7 6950X clocked at 4.2GHz with 2x16GB 2666MHz DDR4 ram.

 

By single core score, you can tell how much has improved over the years IPC wise.

 

Nice & welcome to EHW  

[bonami2]

20 hours ago, Mr. Fox said:

I don't have a good way to cool them as I did the DDR4 modules. They do get hot under certain circumstances (I have seen a max of about 52°C, which is very warm for memory) and start throwing TM5 errors due to the thermals. The same was true of my DDR4 modules. If they got hotter than about 40-45°C it would produce errors in TM5.

 

Until heavy duty DDR5 heat sink jackets (like those used for water cooling) are available, I have a 120MM fan blowing on the modules to try to keep them cool.

 

I may try some of my DDR4 heat sinks on these modules, but the jury is still out on whether they actually fit correctly. I tend to trust Igor more than a reddit users.

 

DDR4 RAM cooler to DDR5? Not a good idea! The differences in detail including 3D scan and instructions

WWW.IGORSLAB.DE

---

Today it's about non-existing coolers for unavailable RAM, so once again pure unicorn cinema for all who are only missing the RAM for the new Intel system with

 

 

 

Well just learned something!

 

Did not know ram was heat sensitive gonna check my ddr4 and add a fan

[Storm-Chaser]

Just applied the new liquid metal to my 9600KF rig. 

 

Insane temp improvements. I shaved  25*F off peak temps with the fans on idle. And 30*F with all the fans running full throttle. 

 

Can this really be true? Anyone else getting results like this? Just crazy I've never seen gains like this. 

[Mr. Fox]

4 hours ago, Storm-Chaser said:

Just applied the new liquid metal to my 9600KF rig. 

 

Insane temp improvements. I shaved  25*F off peak temps with the fans on idle. And 30*F with all the fans running full throttle. 

 

Can this really be true? Anyone else getting results like this? Just crazy I've never seen gains like this. 

100%. I have been a huge fan of liquid metal for roughly a decade and frequently see a ~20°C improvement in load temps. And, nothing beats bare die + liquid metal. On my 7980XE it meant passing Cinebench R20 at 5.3GHz all core with no AVX offset and no thermal throttling versus thermal shutdown.

Edited March 8, 2022 by Mr. Fox

[Storm-Chaser]

1 hour ago, Mr. Fox said:

100%. I have been a huge fan of liquid metal for roughly a decade and frequently see a ~20°C improvement in load temps. And, nothing beats bare die + liquid metal. On my 7980XE it meant passing Cinebench R20 at 5.3GHz all core with no AVX offset and no thermal throttling versus thermal shutdown.

Damn I wish I had found this stuff earlier in my overclocking career. 

 

Never going back to thermal paste, that's for sure.

 

I do have a copper water block but the research reveals the liquid metal will in fact do okay with a copper water block. While they do react, you will only get a discoloration on the copper base plate itself, and it does not compromise performance in any way. This is according to a study done by Gamers Nexus. Getting a new water block next month anyway.

[J7SC_Orion]

1 minute ago, Storm-Chaser said:

Damn I wish I had found this stuff earlier in my overclocking career. 

 

Never going back to thermal paste, that's for sure.

 

I do have a copper water block but the research reveals the liquid metal will in fact do okay with a copper water block. While they do react, you will only get a discoloration on the copper base plate itself, and it does not compromise performance in any way. This is according to a study done by Gamers Nexus. Getting a new water block next month anyway.

 

I've used LM with great success before on both delidded CPUs, and also GPUs. Apart from making sure to do the proper insulation around nearby components, the only other thing to keep in mind is that LM is not suitable for sub-zero. 

[Storm-Chaser]

2 hours ago, Mr. Fox said:

100%. I have been a huge fan of liquid metal for roughly a decade and frequently see a ~20°C improvement in load temps. And, nothing beats bare die + liquid metal. On my 7980XE it meant passing Cinebench R20 at 5.3GHz all core with no AVX offset and no thermal throttling versus thermal shutdown.

In that case, I'd like to have my 9600KF delidded for bare die cooling. 

My heatkiller is also ported. There is 90* turn, just inside the waterblock itself.

So I drilled and machined and removed that bottleneck. However, I would not do this unless you have more than one pump. Because of drilling that out, I get substantially better flow rates but I lose the 90* route that hits all the fins. So I just have to hammer it with good flow rate.

[Storm-Chaser]

Do we have any delidding pros here? I'd like someone to do this for me if possible. 

[Storm-Chaser]

7 minutes ago, J7SC_Orion said:

the proper insulation around nearby components

Warning: 20 questions below. lol

 

 I should probably get some fun tack or something similar. But in all seriousness, Im assuming you put a layer over the area around the CPU. 

Why? because it's unruly when you apply to the CPU? Will it short out the motherboard? And do you have to leave it that way? How much LM are you guys applying to the CPU?

[J7SC_Orion]

3 hours ago, Storm-Chaser said:

Warning: 20 questions below. lol

 

 I should probably get some fun tack or something similar. But in all seriousness, Im assuming you put a layer over the area around the CPU. 

Why? because it's unruly when you apply to the CPU? Will it short out the motherboard? And do you have to leave it that way? How much LM are you guys applying to the CPU?

 

...lots of threads on liquid metal (LM)...re. insulation, since LM is conductive, you want some sort of conformal coating around the die area / anything which might come into contact with the LM. El Cheapo modus operandi is to use nail polish (colour is up to you ). For GPUs (ie. vertical ones in a test-bench) I also added an additional outside ring w/ MX4 or other non-conductive pastes as extra insurance, over and above conformal coating.

 

Also, as mentioned above, LM and sub-zero is a big no-no

 

 

 

Edited March 8, 2022 by J7SC_Orion

[Storm-Chaser]

2 hours ago, J7SC_Orion said:

 

...lots of threads on liquid metal (LM)...re. insulation, since LM is conductive, you want some sort of conformal coating around the die area / anything which might come into contact with the LM. El Cheapo modus operandi is to use nail polish (colour is up to you ). For GPUs (ie. vertical ones in a test-bench) I also added an additional outside ring w/ MX4 or other non-conductive pastes as extra insurance, over and above conformal coating.

 

Also, as mentioned above, LM and sub-zero is a big no-no

 

 

 

I was curious about that. I would really love to bring the chiller back online if I could see gains like this. But u are saying this is a problem? About how far is it good down to?

 

[Storm-Chaser]

Well gentlemen it looks like I have some bad news. 

I bricked my MSI MEG z390 ACE while working on my cooling system. 

 

I had the second reservoir out for repair and I was putting it back into the loop today. Well, it looks like as I rotated the computer, one of the coolant lines started dripping onto the corner of the motherboard. I let it sit for a couple hours in hopes to would dry out, but thats a negative. I don't get any boot codes, it doesn't even turn on, even though the power button on the motherboard itself is illuminated. 

 

This is actually not really a bad thing, however, because this particular board has a bad primary PCIe slot, so I'm just going to find a used one in working condition and swap it out since this motherboard has served me very well and allowed me to reach 5.8GHz with this chip. 

[ArchStanton]

@Storm-ChaserThe motherboard was powered when the leakage occurred?  Either way, I am sorry to hear about your misfortune.

[J7SC_Orion]

19 hours ago, Storm-Chaser said:

I was curious about that. I would really love to bring the chiller back online if I could see gains like this. But u are saying this is a problem? About how far is it good down to?

 

 

...sorry to hear about the leak onto your mobo - I guess you're getting a new mobo ?! In any case, LM such as Thermal Grizzly Conductonaut recommend +8 C as the lower limit.

[Sir Beregond]

On 06/03/2022 at 09:59, Pawelr98 said:

Lowest tier Alder Lake, the allmighty G6900 at stock speed with a single stick of 4GB 2400MHz DDR4 ram.

 

And then my daily driver, Core i7 6950X clocked at 4.2GHz with 2x16GB 2666MHz DDR4 ram.

 

By single core score, you can tell how much has improved over the years IPC wise.

Think your Celeron single core beat out my 4.6GHz OC single core 4790k.

Edited March 9, 2022 by Sir Beregond

[Storm-Chaser]

6 hours ago, J7SC_Orion said:

 

...sorry to hear about the leak onto your mobo - I guess you're getting a new mobo ?! In any case, LM such as Thermal Grizzly Conductonaut recommend +8 C as the lower limit.

Its actually still under warranty so Im hoping MSI will replace it for me free of charge, we shall see. If they don't, I will just buy a used z390 ACE on eBay. Again I'm not overly concerned about it because if I do need a new motherboard I will be able to do some cable and loop management / wire routing to make it look better.

 

6 hours ago, ArchStanton said:

@Storm-Chaser

 

6 hours ago, ArchStanton said:

@Storm-ChaserThe motherboard was powered when the leakage occurred?  Either way, I am sorry to hear about your misfortune.

No it was powered off and the switch on the PSU was off. So I took a hair dryer to it as well but apparently that didnt help either. Going to do some more troubleshooting today but it doesn't look good for this board

 

But it's not that big of a problem. I have an HP z820 on the way for another overclocking project. I'm running the Xeon 1680 v2 CPU, which has an unlocked multi.

 

But yeah, going to get the same motherboard since it's worked very well for this project not to mention I know it inside and out 

Edited March 9, 2022 by Storm-Chaser

[Storm-Chaser]

8 hours ago, J7SC_Orion said:

 

...sorry to hear about the leak onto your mobo - I guess you're getting a new mobo ?! In any case, LM such as Thermal Grizzly Conductonaut recommend +8 C as the lower limit.

Good news!

 

Went down this morning to give it one more chance to come alive and it did! After trying and failing to get it to post for the first 20 minutes I had just one more card to play b4 I was going to pull the motherboard. The trick was I had to hold down the power button while flipping the power switch on the PSU to on. It immediately sprang back to life. I don't know why this worked, there are no recommendations to do this online, I just thought I would do it since every other combination attempting to get it running had failed up until that point.

 

Running stock settings for now to make sure everything is going to operate like it should. Then I will ramp it back up when the time comes for the next competition  

[Storm-Chaser]

So I am up and running! I didn't come through this unscathed, however. Looks like one of my memory modules, which was closest to the leak, has a problem. I discovered this because none of my OC profiles would load correctly (and they all have OCd memory clocks). They would always fail on the re-boot. So I set defaults and found that module wont go XMP anymore. It only works on stock non XMP settings.

 

Luckfully, I just bought an identical kit of Patriot Viper Steel 4133Mhz (8GB x 2) so I can swap that in place of the failing memory module. I don't think the slot is bad, I think it's the memory module itself.  Will report back with results.

 

 

 

 

[ArchStanton]

Glad to hear the patient has made it to the recovery room!

[Storm-Chaser]

Just now, ArchStanton said:

Glad to hear the patient has made it to the recovery room!

Yup that did the trick! Now I can run all my OC bios profiles no problem. Curious to see how this identical kit will perform ... (it's pretty decent b-die)... the water killed kit would top out right around 4733MHz C15, but that's just for cold weather benching. I run at 4266MHz CL15 24/7.

 

 

 

 

[Storm-Chaser]

The new Z820 arrived today! This is a long time coming.

 

I installed my new 1680 v2 overclockable xeon chip and did a fresh install of windows 10. Fast and dirty run up to 4.1GHz, still not entirely sure how to set a proper overclock on this hardware. For now I can only get to 41x. Which is weird because I have the multiplier limits  set to 43x. So clearly it's going past the stock 3.9 single core turbo now, but it's stuck on 41x, I can go down from there not up like I should be.

 

Edited March 10, 2022 by Storm-Chaser

[bonami2]

46 minutes ago, Storm-Chaser said:

The new Z820 arrived today! This is a long time coming.

 

I installed my new 1680 v2 overclockable xeon chip and did a fresh install of windows 10. Fast and dirty run up to 4.1GHz, still not entirely sure how to set a proper overclock on this hardware. For now I can only get to 41x. Which is weird because I have the multiplier limits  set to 43x. So clearly it's going past the stock 3.9 single core turbo now, but it's stuck on 41x, I can go down from there not up like I should be.

 

Maybe XTU would work

 

 

Just saw this if you go down the thread he talk about your cpu and it run at 4.2

 

Solved: Z620 XTU on Windows 20H2, overclocking E-5 1680V2 and Turbo ... - HP Support Community - 7956104

H30434.WWW3.HP.COM

---

Hi,   I just bought a Xeon E5-1680v2 and I realized that it is not possible to overclock the CPU under Windows 20H2 with XTU anymore. You can set the Ratio Multiplier and also the additional Turbo Voltage with no...

 

Edited March 10, 2022 by bonami2