Overclock.net banner

7261 - 7280 of 12850 Posts

·
Registered
Joined
·
3 Posts
My c7h is bugging with the 2406 bios a lot. Half the time it resets my xmp ram from 3333 to 2133 and 8/10 when I shutdown windows it does not shut completely leaving all fans working. Back to previous bios for me
 

·
DRAM Obsessor
Joined
·
1,034 Posts
3733 stable before tightening timings...
 

Attachments

·
Registered
Joined
·
1,752 Posts
3733 stable before tightening timings...
BTW Have You tested actual performance on 3600 vs 3733 vs 3800 ?? Cause 3600 turns out better than 3800 in my case with same timings. The error correction drawback is real :/ thats why im messing around with 3633 .

AAAA and do You still use Haxed cooling solution aka Air con :D
 

·
Registered
Joined
·
1,752 Posts
3733 stable before tightening timings...
BTW Have You tested actual performance on 3600 vs 3733 vs 3800 ?? Cause 3600 turns out better than 3800 in my case with same timings. The error correction drawback is real :/ thats why im messing around with 3633 . Memory errors and stability is miles ahead of Zen1 but im looking for actuall cb point from tighting timings. And seen diminishing returns on going TOO TIGHT !!! lost performance not gained !!!

AAAA and do You still use Haxed cooling solution aka Air con :D
 

·
Registered
Joined
·
199 Posts

·
Registered
Joined
·
199 Posts
BTW Have You tested actual performance on 3600 vs 3733 vs 3800 ?? Cause 3600 turns out better than 3800 in my case with same timings. The error correction drawback is real :/ thats why im messing around with 3633 . Memory errors and stability is miles ahead of Zen1 but im looking for actuall cb point from tighting timings. And seen diminishing returns on going TOO TIGHT !!! lost performance not gained !!!

AAAA and do You still use Haxed cooling solution aka Air con :D
I was getting better performance at 3800 at least shown in AIDA64 any other way to check that out?
 

·
DRAM Obsessor
Joined
·
1,034 Posts
3733 stable before tightening timings...
BTW Have You tested actual performance on 3600 vs 3733 vs 3800 ?? Cause 3600 turns out better than 3800 in my case with same timings. The error correction drawback is real 😕 thats why im messing around with 3633 . Memory errors and stability is miles ahead of Zen1 but im looking for actuall cb point from tighting timings. And seen diminishing returns on going TOO TIGHT !!! lost performance not gained !!!

AAAA and do You still use Haxed cooling solution aka Air con /forum/images/smilies/biggrin.gif
I havent tested different setups yet but ive noticed a few things. I think its going to come down to if you can get low timings. Single core dropped a bit on 3800 until i tightened the last bit up then it edged ahead of what I seen on 3733 but im about to fibish tightening 3733. I doubt 3600 will be better bc at 3733 im able to pretty much tighten to the maximum. Maybe 3666 at 14-13-13-22-36 or something could edge out the others. I could push to 3866 or 3933 but it seems the bios wont allow it at 1:1:1. It gives an error “C5” and i have to clear CMOS or it wont boot at all. Same thing if I try to adjust tRDWR at all, I get an automatic C5 error and have to clear cmos. Im hoping a bios update fixes it. I know i can get 3866 easy and more than likely 3933 at cl14, 4000 would be doubtful. I only use the AC when I bench on hwbot.org if I need it. Im undervolting this processor quite a bit so It barely hits 60c under load.
 

·
DRAM Obsessor
Joined
·
1,034 Posts
3733 stable before tightening timings...
could you share timings and voltages please? also what latency you get?
BTW Have You tested actual performance on 3600 vs 3733 vs 3800 ?? Cause 3600 turns out better than 3800 in my case with same timings. The error correction drawback is real 😕 thats why im messing around with 3633 . Memory errors and stability is miles ahead of Zen1 but im looking for actuall cb point from tighting timings. And seen diminishing returns on going TOO TIGHT !!! lost performance not gained !!!

AAAA and do You still use Haxed cooling solution aka Air con /forum/images/smilies/biggrin.gif
I was getting better performance at 3800 at least shown in AIDA64 any other way to check that out?
I shared timings in an above post from Ryzen Master. As far as testing performance of memory you dont ever go by Aida64 to judge that. Its more of a reference tool to kinda give an idea of what gains a change has made. You can guve a set up wildly unstable timings and it will show gains in Aida64 so it is misleading. Also, the latency numbers you see is Aida is not the effective latency. SiSoft Sandra or Passmark Performance Test can show effective latencies. In Aida Ive gotten it down to 63ns but the effective latency was 28ns. In comparison the 2700x best effective latency I saw was around 45-46 so even though this architecture may show similar latencies to Zen+ in aida64, in actuality Zen2 is a 20ns gain in latency performance over Zen+. Best Intel chips have effective latencies of 13ns for another comparison.
 

·
Registered
Joined
·
1,752 Posts
I shared timings in an above post from Ryzen Master. As far as testing performance of memory you dont ever go by Aida64 to judge that. Its more of a reference tool to kinda give an idea of what gains a change has made. You can guve a set up wildly unstable timings and it will show gains in Aida64 so it is misleading. Also, the latency numbers you see is Aida is not the effective latency. SiSoft Sandra or Passmark Performance Test can show effective latencies. In Aida Ive gotten it down to 63ns but the effective latency was 28ns. In comparison the 2700x best effective latency I saw was around 45-46 so even though this architecture may show similar latencies to Zen+ in aida64, in actuality Zen2 is a 20ns gain in latency performance over Zen+. Best Intel chips have effective latencies of 13ns for another comparison.
Iw settled for 4350 all core 1.5 hour aida and realbench stable. Maxes out around 80c after that.

Ibt sadly got some bug cant runnit on zen2 and it was my nr1 performance test :/ now im looking at 3 rins of cb20. Maybe could try cpuid bench cause gives quite allot of points to see difference just need 2 xheck gow repeteable it is.


I know i cant boot at 3733 with ky 14 15 14 28 settings no matter what i tried to change. Would need to go and teat out performance on the 14 15 15 32nypu are using.
 

·
Computer Enthusiast
Joined
·
2,904 Posts
I havent tested different setups yet but ive noticed a few things. I think its going to come down to if you can get low timings. Single core dropped a bit on 3800 until i tightened the last bit up then it edged ahead of what I seen on 3733 but im about to fibish tightening 3733. I doubt 3600 will be better bc at 3733 im able to pretty much tighten to the maximum. Maybe 3666 at 14-13-13-22-36 or something could edge out the others. I could push to 3866 or 3933 but it seems the bios wont allow it at 1:1:1. It gives an error “C5” and i have to clear CMOS or it wont boot at all. Same thing if I try to adjust tRDWR at all, I get an automatic C5 error and have to clear cmos. Im hoping a bios update fixes it. I know i can get 3866 easy and more than likely 3933 at cl14, 4000 would be doubtful. I only use the AC when I bench on hwbot.org if I need it. Im undervolting this processor quite a bit so It barely hits 60c under load.
What happens with tRDWR on Auto?

I'm still waiting for my CPU! - should be here tomorrow..
 

·
►¿¿¿
Joined
·
2,269 Posts
Can anyone, or everyone with this board please do a DPC Latency test with LatencyMon?

Considering this board, or the x470 Taichi/ultimate.. But as far as i know..

The asus has worse DPC Latency and Daisy-Chain RAM
But the ASRock is T-topology RAM

Seems really hard to find any dpc test anywhere now-adays, wheres r0ach when you need him?

Also.. can you disable HPET in the bios?
 

·
Registered
Joined
·
2,308 Posts
Can anyone, or everyone with this board please do a DPC Latency test with LatencyMon?

Considering this board, or the x470 Taichi/ultimate.. But as far as i know..

The asus has worse DPC Latency and Daisy-Chain RAM
But the ASRock is T-topology RAM

Seems really hard to find any dpc test anywhere now-adays, wheres r0ach when you need him?

Also.. can you disable HPET in the bios?
I believe this board is t-topology.

I haven't run LatencyMon in a while. I think since before Windows 1903 and definitely before the latest BIOS updates. I also never had problems with my mic before Windows 1903 so I'm thinking the higher results I'm seeing now are from Windows 1903. It's pretty bad.
 

·
Registered
Joined
·
2,308 Posts
I havent tested different setups yet but ive noticed a few things. I think its going to come down to if you can get low timings. Single core dropped a bit on 3800 until i tightened the last bit up then it edged ahead of what I seen on 3733 but im about to fibish tightening 3733. I doubt 3600 will be better bc at 3733 im able to pretty much tighten to the maximum. Maybe 3666 at 14-13-13-22-36 or something could edge out the others. I could push to 3866 or 3933 but it seems the bios wont allow it at 1:1:1. It gives an error “C5” and i have to clear CMOS or it wont boot at all. Same thing if I try to adjust tRDWR at all, I get an automatic C5 error and have to clear cmos. Im hoping a bios update fixes it. I know i can get 3866 easy and more than likely 3933 at cl14, 4000 would be doubtful. I only use the AC when I bench on hwbot.org if I need it. Im undervolting this processor quite a bit so It barely hits 60c under load.
Is this because of the halved 16-bit write lane?
 

·
►¿¿¿
Joined
·
2,269 Posts
The DPC latency is terrible right now, like 2~6k ntoskrnl.exe, but it seems is because of windows 1903

https://forums.geforce.com/default/...e-large-latency-spikes-/post/6114668/#6114668

And you can only disable HPET with modded bios sadly
Well damn, never had any good experiences with modded BIOS'es, but at least id be willing to use it, if theyre available.

Guessing no one is running 1809 to test, heh
I believe this board is t-topology.

I haven't run LatencyMon in a while. I think since before Windows 1903 and definitely before the latest BIOS updates. I also never had problems with my mic before Windows 1903 so I'm thinking the higher results I'm seeing now are from Windows 1903. It's pretty bad.
the spreadsheet and any info i could find says the Asus is Daisy-chain
 

·
Registered
Joined
·
1,170 Posts
I need to understand something, I'm putting 1.4125vcore through my Ryzen 5 3600 and the highest bench temps I've seen was 63c. Running Aida64 stress test it will climb up to 78c. My problem is that vcore reported by CPUz, Coretemp and Ryzen Master is 1.1vcore. I cannot run Hwinfo64 it won't open keeps hanging while starting up.
 

·
Registered
Joined
·
199 Posts
I shared timings in an above post from Ryzen Master. As far as testing performance of memory you dont ever go by Aida64 to judge that. Its more of a reference tool to kinda give an idea of what gains a change has made. You can guve a set up wildly unstable timings and it will show gains in Aida64 so it is misleading. Also, the latency numbers you see is Aida is not the effective latency. SiSoft Sandra or Passmark Performance Test can show effective latencies. In Aida Ive gotten it down to 63ns but the effective latency was 28ns. In comparison the 2700x best effective latency I saw was around 45-46 so even though this architecture may show similar latencies to Zen+ in aida64, in actuality Zen2 is a 20ns gain in latency performance over Zen+. Best Intel chips have effective latencies of 13ns for another comparison.
that some nice info thanks. Today i might try to test every with Sisoft and report back, will try 3600cl14 a 3733cl14 if i can or cl16 like i'm running it atm
 

·
Registered
Joined
·
316 Posts

·
Registered
Joined
·
2,308 Posts
I need to understand something, I'm putting 1.4125vcore through my Ryzen 5 3600 and the highest bench temps I've seen was 63c. Running Aida64 stress test it will climb up to 78c. My problem is that vcore reported by CPUz, Coretemp and Ryzen Master is 1.1vcore. I cannot run Hwinfo64 it won't open keeps hanging while starting up.

Have you tried going into HWiNFO Settings > Safety > Disable Drive Scan?

Or S.M.A.R.T. Support?
 

·
Registered
Joined
·
1,170 Posts
Have you tried going into HWiNFO Settings > Safety > Disable Drive Scan?



Or S.M.A.R.T. Support?
Hi nick, sorry my complete illiterate question, how do I do that.

Sent from my SM-G960F using Tapatalk
 
7261 - 7280 of 12850 Posts
Top