Asus Z790 14900ks throttling question.

[lexluthermiester]

would you mind telling me how to use Ycruncher on 64 GB? Lol

I've never used it.

Well it passed an hour of AIDA64 lol

Then you are well & truly golden. Consider your config tested and stable.

 

[D007]

I've never used it.


Then you are well & truly golden. Consider your config tested and stable.

Sweet! Thank you all very kindly.
Not bad for 64gb huh?

 

[ShrimpBrime]

Well it passed an hour of AIDA64 lol

would you mind telling me how to use Ycruncher on 64 GB? Lol

Passed the 10b thing anyway if it matters lol.

OCCT Linpack. You can manually dial in how much memory usage you want. Free version test is 1 hour. usually puts coolers to the test.

 

[lexluthermiester]

OCCT Linpack. You can manually dial in how much memory usage you want. Free version test is 1 hour. usually puts coolers to the test.

Nah, he's good. 60minutes of AIDA64 is enough to prove stability.

 

[ShrimpBrime]

Nah, he's good. 60minutes of AIDA64 is enough to prove stability.

And this point I ask my own self, should I beg to differ or leave that lie in the mud.

No, I do believe OCCT Linpack is a bit more intensive. It's been a long time since I've used AIDA64 stability test. I don't remember it being too intensive.

But if he's stable with AIDA for you, then I'm good with that too. Just wanted to give the man some options to work with.

 

[lexluthermiester]

And this point I ask my own self, should I beg to differ or leave that lie in the mud.

No, I do believe OCCT Linpack is a bit more intensive. It's been a long time since I've used AIDA64 stability test. I don't remember it being too intensive.

But if he's stable with AIDA for you, then I'm good with that too. Just wanted to give the man some options to work with.

Well, you do you. However, the OP wouldn't gain anything from an OCCT run.

 

[D007]

You guys have all been a great help. I like checking out all this different software so I'll look into OCCT as well.
Thank you.
I'll probably make a small OC video for this as I don't see a lot of them for 2x32 (64gb) kits.

 

[ShrimpBrime]

You guys have all been a great help. I like checking out all this different software so I'll look into OCCT as well.
Thank you.
I'll probably make a small OC video for this as I don't see a lot of them for 2x32 (64gb) kits.

OCCT will run AVX/2 instruction sets. Get ready for some actual heat. Use linpack 2021.

Well, you do you. However, the OP wouldn't gain anything from an OCCT run.

Sure he'll gain something. Have you ever used it?? If not I highly recommend playing around with the free version. Just that is worth the DL actually.

OCCT lets you fine tune your testing. You can litterally type in the exact amount of memory you want to use. Set which instruction sets to test with. Specific testing for AMD even. Monitoring temps, core usage, set which cores you want to test single core boosts with, logging. Pretty much everything you'd expect from a high end stress test. Not everyone's favorite, no. I get that. Sometimes just click a button and go is easy. Doesn't mean it's THE stability test to go to. I'd also recommend Prime 95 on top of this.

Also want to mention that OP should do a cool down period with the cpu, light loads and give a restart after each stress test. It's just some stupid practice I have, seems the system always likes it.

 

[D007]

OCCT will run AVX/2 instruction sets. Get ready for some actual heat. Use linpack 2021.


Sure he'll gain something. Have you ever used it?? If not I highly recommend playing around with the free version. Just that is worth the DL actually.

OCCT lets you fine tune your testing. You can litterally type in the exact amount of memory you want to use. Set which instruction sets to test with. Specific testing for AMD even. Monitoring temps, core usage, set which cores you want to test single core boosts with, logging. Pretty much everything you'd expect from a high end stress test. Not everyone's favorite, no. I get that. Sometimes just click a button and go is easy. Doesn't mean it's THE stability test to go to. I'd also recommend Prime 95 on top of this.

Also want to mention that OP should do a cool down period with the cpu, light loads and give a restart after each stress test. It's just some stupid practice I have, seems the system always likes it.

Fo sho

 

[Dr. Dro]

My opinion is that each and every person has their own concept of "stability" - speaking for myself, when I build or do major changes to my system configuration I run a full battery that many if not most will consider overkill:

- Overnight (~12h) run of bootable Memtest86

MemTest86 - Download now!

Download your free copy of MemTest86 here for USB.

www.memtest86.com

- 4 1h runs of OCCT's CPU test, free version works

OCBASE/OCCT : Free, all-in-one stability, stress test, benchmark and monitoring tool for your PC

Ocbase is the home of OCCT, the most popular all-in-one stability / stress testing / benchmarking / monitoring tool available for PC

www.ocbase.com

-4h of y-cruncher VST, download BenchMate for easier setup - Benchmate can also be used to post validated benchmarks (an anti-cheat of sorts)

BenchMate

benchmate.org

-4h of Prime95 blend mode (mixes small FFT which are super intensive on the CPU's integer and floating point units and large FFTs which overstress cache and memory)

GIMPS - Free Prime95 software downloads - PrimeNet

GIMPS has free software available for Windows, Linux, FreeBSD, Mac OSX. Contribute to the effort by using your computer's spare processing power.

www.mersenne.org

Finally, I run RealBench to catch any runaway transient-related issues that may arise from loadline or idle voltages, this is the closest thing to a "real world use" that you can automate:

Download ASUS RealBench - MajorGeeks

ASUS RealBench benchmark test can judge the real performance of your system allowing you to compare a new build to an old one or the stock speed an overclocked one.

www.majorgeeks.com

RealBench is old and is an automated script that runs a few things like Handbrake, Luxmark and GIMP, it is not suitable to test long-term stability but it catches the few "ghosts" that stick by under low and moderate system load, which is why I kept it around.

In general, I run tests for ~2 days on my new machines. Once that is passed, I consider it production-ready and fully stable for the long run.

 

[D007]

My opinion is that each and every person has their own concept of "stability" - speaking for myself, when I build or do major changes to my system configuration I run a full battery that many if not most will consider overkill:

- Overnight (~12h) run of bootable Memtest86

MemTest86 - Download now!

Download your free copy of MemTest86 here for USB.

www.memtest86.com

- 4 1h runs of OCCT's CPU test, free version works

OCBASE/OCCT : Free, all-in-one stability, stress test, benchmark and monitoring tool for your PC

Ocbase is the home of OCCT, the most popular all-in-one stability / stress testing / benchmarking / monitoring tool available for PC

www.ocbase.com

-4h of y-cruncher VST, download BenchMate for easier setup - Benchmate can also be used to post validated benchmarks (an anti-cheat of sorts)

BenchMate

benchmate.org

-4h of Prime95 blend mode (mixes small FFT which are super intensive on the CPU's integer and floating point units and large FFTs which overstress cache and memory)

GIMPS - Free Prime95 software downloads - PrimeNet

GIMPS has free software available for Windows, Linux, FreeBSD, Mac OSX. Contribute to the effort by using your computer's spare processing power.

www.mersenne.org

Finally, I run RealBench to catch any runaway transient-related issues that may arise from loadline or idle voltages, this is the closest thing to a "real world use" that you can automate:

Download ASUS RealBench - MajorGeeks

ASUS RealBench benchmark test can judge the real performance of your system allowing you to compare a new build to an old one or the stock speed an overclocked one.

www.majorgeeks.com

RealBench is old and is an automated script that runs a few things like Handbrake, Luxmark and GIMP, it is not suitable to test long-term stability but it catches the few "ghosts" that stick by under low and moderate system load, which is why I kept it around.

In general, I run tests for ~2 days on my new machines. Once that is passed, I consider it production-ready and fully stable for the long run.

Dam, now we’re cooking with fire lol.
I’ll check em all out.

 

[RootinTootinPootin]

Ahhh IVR transmitter.
What do you think I should set that to?
I’m thinking 1.4.

the TX is a direct voltage, you shouldn't be pushing that hard unless you're on LN2 or Liquid Helium, for 8600MT's I am just running that at 1.32v with the Apex Encore (should be lower in your case)

 

[phanbuey]

My opinion is that each and every person has their own concept of "stability" - speaking for myself, when I build or do major changes to my system configuration I run a full battery that many if not most will consider overkill:

- Overnight (~12h) run of bootable Memtest86

MemTest86 - Download now!

Download your free copy of MemTest86 here for USB.

www.memtest86.com

- 4 1h runs of OCCT's CPU test, free version works

OCBASE/OCCT : Free, all-in-one stability, stress test, benchmark and monitoring tool for your PC

Ocbase is the home of OCCT, the most popular all-in-one stability / stress testing / benchmarking / monitoring tool available for PC

www.ocbase.com

-4h of y-cruncher VST, download BenchMate for easier setup - Benchmate can also be used to post validated benchmarks (an anti-cheat of sorts)

BenchMate

benchmate.org

-4h of Prime95 blend mode (mixes small FFT which are super intensive on the CPU's integer and floating point units and large FFTs which overstress cache and memory)

GIMPS - Free Prime95 software downloads - PrimeNet

GIMPS has free software available for Windows, Linux, FreeBSD, Mac OSX. Contribute to the effort by using your computer's spare processing power.

www.mersenne.org

Finally, I run RealBench to catch any runaway transient-related issues that may arise from loadline or idle voltages, this is the closest thing to a "real world use" that you can automate:

Download ASUS RealBench - MajorGeeks

ASUS RealBench benchmark test can judge the real performance of your system allowing you to compare a new build to an old one or the stock speed an overclocked one.

www.majorgeeks.com

RealBench is old and is an automated script that runs a few things like Handbrake, Luxmark and GIMP, it is not suitable to test long-term stability but it catches the few "ghosts" that stick by under low and moderate system load, which is why I kept it around.

In general, I run tests for ~2 days on my new machines. Once that is passed, I consider it production-ready and fully stable for the long run.

I find that the most effective test of stability is game night with the boys, you're hosting, and you're at that clutch moment when it's a big fight.

100% finds instability every time. It helps if you haven't saved in a bit.

 

[RootinTootinPootin]

I find that the most effective test of stability is game night with the boys, you're hosting, and you're at that clutch moment when it's a big fight.

100% finds instability every time. It helps if you haven't saved in a bit.

true, aside from the stability from those apps, I also play games, and I find it at most times, you are still required to fine tune due to games acting/performing up not the way you wanted or it used to. in my case, since I frequently use USB C to transfer my stuff to my external storage, I have to increase the SA voltage a tad bit more as if I don't the external storage keeps disconnecting on me, I am normally stable at 1.23v for my settings, but due to that, I have to increase it to 1.26v for everything not to crap up.

 

[rethcirE]

A known stable game can be a great stability test if you play long enough and truly trust it. Benchmarks seems more purpose built and therefore output simply more accurate data, I think. All that said; I've found unlocking the power limits is a definite two edged sword and if you're not careful with these programs like OCCT you can push 300-400W thru the CPU and introduce instability. At least I found on certain instruction sets with fully unlocked limits it asks for scary power levels lol. A game wouldn't hit to the same degree I believe, but benchmarks hit the nerve so to speak.

 

[Dr. Dro]

I find that the most effective test of stability is game night with the boys, you're hosting, and you're at that clutch moment when it's a big fight.

100% finds instability every time. It helps if you haven't saved in a bit.

Very fun when your PC crashes during a game night

 

[lexluthermiester]

My opinion is that each and every person has their own concept of "stability" - speaking for myself, when I build or do major changes to my system configuration I run a full battery that many if not most will consider overkill:

- Overnight (~12h) run of bootable Memtest86

MemTest86 - Download now!

Download your free copy of MemTest86 here for USB.

www.memtest86.com

- 4 1h runs of OCCT's CPU test, free version works

OCBASE/OCCT : Free, all-in-one stability, stress test, benchmark and monitoring tool for your PC

Ocbase is the home of OCCT, the most popular all-in-one stability / stress testing / benchmarking / monitoring tool available for PC

www.ocbase.com

-4h of y-cruncher VST, download BenchMate for easier setup - Benchmate can also be used to post validated benchmarks (an anti-cheat of sorts)

BenchMate

benchmate.org

-4h of Prime95 blend mode (mixes small FFT which are super intensive on the CPU's integer and floating point units and large FFTs which overstress cache and memory)

GIMPS - Free Prime95 software downloads - PrimeNet

GIMPS has free software available for Windows, Linux, FreeBSD, Mac OSX. Contribute to the effort by using your computer's spare processing power.

www.mersenne.org

Finally, I run RealBench to catch any runaway transient-related issues that may arise from loadline or idle voltages, this is the closest thing to a "real world use" that you can automate:

Download ASUS RealBench - MajorGeeks

ASUS RealBench benchmark test can judge the real performance of your system allowing you to compare a new build to an old one or the stock speed an overclocked one.

www.majorgeeks.com

RealBench is old and is an automated script that runs a few things like Handbrake, Luxmark and GIMP, it is not suitable to test long-term stability but it catches the few "ghosts" that stick by under low and moderate system load, which is why I kept it around.

In general, I run tests for ~2 days on my new machines. Once that is passed, I consider it production-ready and fully stable for the long run.

This is is the very definition of "over-kill". No one and no machine needs this much testing, nor anything near it, to establish stability. If a user is having problems, then yes some of this would be in order after a repair is complete but not nearly this much.

Over-night runs of Memtest? Are you kidding? That's just silly. No. Two full passes and you're golden.

4hrs of OCCT(or AIDA64)? Seriously? Again, no. 20minutes would more than enough.

4hrs of Prime95? Again, no. 20minutes or 10 successful passes is enough.

Some of you seem to be wasting a metric ton of time doing things that are unneeded. You want to waste time and power, go right ahead, it's your time and life to waste.

@D007
If it's running well at this point, there are no reasons to be wasting your time doing the amount of stress testing suggested above in posts from other users.

Sure he'll gain something. Have you ever used it??

Yes, I have. They would gain nothing from wasting that much time.

 

[ShrimpBrime]

Yes, I have. They would gain nothing from wasting that much time.

I suppose so. usually instability shows pretty much right away (most of the time)

 

[lexluthermiester]

usually instability shows pretty much right away (most of the time)

Exactly. The vast majority of the time. And when a problem does show itself through normal usage, then more extensive testing is employed to troubleshoot.

 

[D007]

the TX is a direct voltage, you shouldn't be pushing that hard unless you're on LN2 or Liquid Helium, for 8600MT's I am just running that at 1.32v with the Apex Encore (should be lower in your case)

I’m not sure how much lower or really. There’s no way I can hit 8000 on this motherboard. I’ll be lucky to get one stick of ram running at 8000. Lol
You have two dims, I have four. And I have much denser memory. 32 GB sticks x2 64gb.
I’ll try to lower it though for sure.
Thank you for the suggestion.

 

[ShrimpBrime]

I’m not sure how much lower or really. There’s no way I can hit 8000 on this motherboard. I’ll be lucky to get one stick of ram running at 8000. Lol
You have two dims, I have four. And I have much denser memory. 32 GB sticks x2 64gb.
I’ll try to lower it though for sure.
Thank you for the suggestion.

Your board supports 8000mt/s+ overclocking, you just don't have the correct memory set to do so.

 

[D007]

Your board supports 8000mt/s+ overclocking, you just don't have the correct memory set to do so.

I tried a set. Nope. Maybe one stick or 2x16 but not 2x24. I tried the Gskill that is supposed to work and it didn’t.

 

[lexluthermiester]

I tried a set. Nope. Maybe one stick or 2x16 but not 2x24. I tried the Gskill that is supposed to work and it didn’t.

That sounds more like a motherboard firmware issues to me. Have you checked for BIOS updates?

 

[D007]

That sound more like a motherboard firmware issues to me. Have you checked for BIOS updates?

Yep, on the most recent. Did it last week when I got the mobo.
It won’t post with 2 sticks at 8000 and the 14900ks

 

[ShrimpBrime]

Yep, on the most recent. Did it last week when I got the mobo.
It won’t post with 2 sticks at 8000 and the 14900ks

Might have to work your way up with some training. Run Gear 4 so IMC runs 2000mhz. If they would post 7200Mhz Gear 2, then you should be able to get 8000Mhz Gear 4.

Still have the ram? Or did you send it back in? What was it? answers, we need answers!! haha kidding.