AMD Ryzen /Threadripper OverClocking V1.1, big or small, share your results here...
|
|
 Aug 20 2019, 12:12 PM
Return to original view | Post
#1
|
 
Junior Member
398 posts Joined: Jul 2014  
|
pbo is pretty much broken across platforms, that's not limited to b450
if you're fancy to OC at this moment, do auto oc only. actually if I were you, at this moment I would simply adjust PPT, TDC's, and EDC's. This post has been edited by nrw: Aug 20 2019, 12:14 PM |
|
|
|
|
|
Aug 22 2019, 12:14 AM
Return to original view | IPv6 | Post
#2
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(ah_khoo @ Aug 21 2019, 04:50 PM) pretty much max freq can be squeezed out from d x470 + CJR... Even though I hate mem (benchs) on settings that are far from being stable for long term that's a pretty good 2666 rated kit you got there despite your timings are suffering big time to achieve the frequency.   That cpu however o.O 4.125GHz with 1.456V … damn. Mines doing 4.225 more with 1.415 (AVX ST) (which translates to 1674 mc cb in r15) |
|
|
Sep 1 2019, 02:44 AM
Return to original view | IPv6 | Post
#3
|
|
Junior Member
398 posts Joined: Jul 2014
|
one stable light overclock on my small ryzen system then
 3600 @4.125Ghz, 1.344V (no lcc so its @1.31 after vdroop) B450 Tomahawk 2x16GB E-Die 1:1 FCLK, 3800, 16-19-16-38 solid cpu but far from outstanding. maxes out at 4.225 @1.39 with llc. I like my system silent and more efficient however. E-dies as expected, tRCDRD and tRFC (a bit) higher than what one would see using B-dies. other than that it's on par with 'regular' binned B-dies. This post has been edited by nrw: Sep 1 2019, 02:46 AM |
|
|
Sep 1 2019, 08:25 PM
Return to original view | IPv6 | Post
#4
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 1 2019, 06:38 PM) Karhu 10000% passed at 3800 16-19-16-38-56? didn't ss that, I'll run over night again n share, npMind sharing your ss? can just give you the log for now QUOTE [2019-08-27T00:54:55.7314524+08:00] Started testing 27181 MB with 12 thread(s). [2019-08-27T08:02:08.8908140+08:00] Stopped testing after 0:07:07:13 with 11247 % coverage and 0 error(s). This post has been edited by nrw: Sep 1 2019, 08:31 PM |
|
|
Sep 2 2019, 05:18 AM
Return to original view | IPv6 | Post
#5
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 1 2019, 06:38 PM) Karhu 10000% passed at 3800 16-19-16-38-56? morning, there you goMind sharing your ss? This post has been edited by nrw: Sep 2 2019, 05:25 AM |
|
|
Sep 3 2019, 11:22 AM
Return to original view | Post
#6
|
|
Junior Member
398 posts Joined: Jul 2014
|
really calm around here
 -> question: anyone running his b-die kit on 3733/3800 c14 stable? Gonna try coming weekend. Unfortunately my 3700x (unlike my 3600) doesn't do 3800 on 1:1 so I am eying 3733 14-15-14-35-42 for now, maybe can drop ras a bit. should drop latency to or slightly below 65. |
|
|
|
|
|
Sep 3 2019, 02:39 PM
Return to original view | Post
#7
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(ah_khoo @ Sep 3 2019, 01:59 PM) very2 nice timing & frequency considerin they are more of budget sticks (correct me if i'm wrong) bro... yup, it's def. a better than average set I got there.for just slightly over RM800 for 32GB with these capabilities is a steal. if ur referrin to 32GB kit, i believe it's close to 0 here (from observation). as for 16GB, pls wait for bro owi... results comin pretty soon i guess...  if I didn't have some very nice b-die kits I would keep them for sure. let's see what your sr b-dies can do  I have a sr hyperx 4000c19 rated kits here which are OK. soon will get back my 3200c14 g.skill ones, despite having only used those on intel some time back they should be better... lets see |
|
|
Sep 9 2019, 07:34 PM
Return to original view | IPv6 | Post
#8
|
|
Junior Member
398 posts Joined: Jul 2014
|
Samsung started selling a-dies ddr4 32gb today. 10nm class3 (1z 12nm), 16GBit/chip, rated 2933 cl21. Doubt they will be any good for OC but we shall see.
This post has been edited by nrw: Sep 10 2019, 12:03 AM |
|
|
Sep 10 2019, 12:02 AM
Return to original view | IPv6 | Post
#9
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 9 2019, 10:03 PM) B-die is 2666 CL19, pretty much the same true latency but less bandwidth. can't judge from absolute latency to oc capabilities 2666 CL19: 2000/2666*19 = 14.25ns 2933 CL21: 2000/2933*21 = 14.32ns Ya, curious about A-die's OCbility vs B-die. We gonna wait and see later...  the m-dies for that series was kind of underwhelming (however that one was 1Y(15nm) and not 1Z(12nm) ~RM650-700 fo 32GB stick btw. This post has been edited by nrw: Sep 10 2019, 12:04 AM |
|
|
Sep 10 2019, 10:44 AM
Return to original view | IPv6 | Post
#10
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 10 2019, 12:43 AM) Continuing with RAM OC, so far hasn't been so impressive mainly due to high latency issue. Gj, looking at your latency I am fairly sure your power down mode is enabled.It could be that the latest DRAM Calculator v1.6.2 by 1usmus has decided to set the subtimings too loose that eventually hitting the RAM performance. Maybe need further tweaking to optimize the bandwidths and latency. RAM OC with SK-Hynix AFR: 3466 16-19-19-40-1T @ 1.41v:  3533 16-19-19-40-1T @ 1.45v (not stable):  RAM OC with Samsung B-die:  3600 14-15-15-30-1T @ 1.45v:  3800 16-17-16-32-1T @ 1.41v:  4000 16-17-16-32-1T @ 1.45v:  4133 16-17-16-32-1T @ 1.45v:  Again, still need to figure out how to reduce improve the efficiency. Aim for 3800 CL14 stable but I don't think possible with the current ADATA D80 sticks. Later I swap to my TridentZ RGB 4266 kit and see how much I can push. |
|
|
Sep 10 2019, 12:49 PM
Return to original view | IPv6 | Post
#11
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 10 2019, 11:48 AM) I need to recheck this later as the GDM and PDM are mixes of enabled and disabled as per recommendation by DRAM Calculator. Since few versions those recommendations are quite off, power down disabled is hardly a source for instabiliy, whereas the disabled gear down mode from the last version is not really advisable in my opinion for long term stability.I made the observation with my c7h that despite disabling power down mode it remains active with certain configurations. |
|
|
Sep 14 2019, 08:09 PM
Return to original view | IPv6 | Post
#12
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 14 2019, 12:57 AM) Latest AGESA 1.0.0.3 ABBA seems to have fixed my 3600X's boost bug that I had earlier, now I can see one core is boosting up to 4.4GHz as per AMD's original specs. CB R20 runs with everything default: good work! BIOS 0803:  BIOS 1001:  My 3700x made similar improvements and is within specification now. The 3600 I got still doesn't want to reach 4.2 single core. It comfortably does 4.25 all core with 1.3675V though  Guess there will be more improvements to come as 3900x' didn't seem as positively effected. The core rating in Ryzen master however is still off. What you haven't pointed out from your results is that (unfortunately in hwinfo64 the polling period is not visible) voltages under load and hence temperatures dropped for stock settings. |
|
|
Sep 14 2019, 10:58 PM
Return to original view | IPv6 | Post
#13
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 14 2019, 10:19 PM) Voltage is set by ASUS BIOS so nothing to do with AGESA, temperature is stock cooler vs custom loop LOL. then the testing method is quite off  as thermal throttling starts in low to mid 70s already.and ofc the smu in your new agesa controls voltage why you think its called system management unit? with its sensory sampling its controlling circuits on the chip. that includes voltage level, voltage supply to psms, thermal management, c-s boost. |
|
|
|
|
|
Sep 23 2019, 01:45 PM
Return to original view | Post
#14
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 22 2019, 11:43 AM) 3800 14-16-14-28-1T @ 1.480v nice primary timings but something is still way wrong there. both bandwidth and especially timings are way sub par. Your results should be more like high 56k / 30k / 56k and your latency should be >10ms lower than what you have right there. wanna take a screenshot of ryzen master listed timings & config? This post has been edited by nrw: Sep 23 2019, 01:45 PM |
|
|
Sep 23 2019, 07:57 PM
Return to original view | IPv6 | Post
#15
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(raydenex @ Sep 23 2019, 01:49 PM) Actually true also, 3800c14 should be getting sub-70ns latency.. sub 65 edit, for reference 3600 instead of 3600x same board 2x 8gb sr b-dies, likewise same primary timings this is what you should be looking at, ideally This post has been edited by nrw: Sep 23 2019, 08:20 PM |
|
|
Sep 24 2019, 11:12 AM
Return to original view | Post
#16
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(ah_khoo @ Sep 24 2019, 09:40 AM) very nice bw, latency & clock there bro... how u guys can have such high clock is beyond me... jelly... tbh, this is not my setup, I just used it as example for @owikh84 as primaries, board and basically cpu were same.  so looking forward for him to share his settings should he wish to figure what the limiting factors both in bandwidth and especially latency are. |
|
|
Sep 24 2019, 10:42 PM
Return to original view | IPv6 | Post
#17
|
|
Junior Member
398 posts Joined: Jul 2014
|
'quick' set up for friend, 4x8gb sr on a b450 aorus m paired with a 3700x.
have to admit gigabyte did a good job with stability and functionality, they're making progress. |
|
|
Sep 28 2019, 11:04 AM
Return to original view | Post
#18
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 28 2019, 10:06 AM) Bingo, not being able to do FCLK 1900 at DDR4-3800 with my 3600X chip has seriously killed my latency and bandwidths. There's a 10ns penalty for 3800/FCLK 1866. At 3733/1866 (1:1) I could get sub 70ns and better bw. you can't boot or can't get stable with 1900 if? tried adjusting vddg/soc? I also came across an article from an ex-ROG staff doing comparisons of memory performance on 3600X/1xCCD vs 3900X/2xCCD: https://elmorlabs.com/index.php/2019-07-10/...0-investigated/ We can see significant differences in terms of mem bandwidths with 3900X over 3600X. It is also suspected that dual-rank and 4x8GB DIMMs are capable of delivering higher bw. u'd kind of expect a better i/o die in an 3600x |
|
|
Sep 28 2019, 11:52 AM
Return to original view | IPv6 | Post
#19
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 28 2019, 11:31 AM) Cant post at all, stuck at debug code 07 displayed on mobo. Ok 1.1 vddg pretty high already, no luck with your IF then.I've tried playing around with VDDG up to 1.1v, SOC up to 1.15v but none of them seems to be helpful. I also tried BCLK 102 with FCLK 1866, getting the same error code. My BIOS settings as follow:  Is your bw really going up when you run 3733 synced? I used to think theres just a latency penalty. Btw, you might be able to drop vsoc. That one can be tested with stress cache only im aida64 quite easily/quickly. This post has been edited by nrw: Sep 28 2019, 12:21 PM |
|
|
Sep 28 2019, 12:35 PM
Return to original view | Post
#20
|
|
Junior Member
398 posts Joined: Jul 2014
|
QUOTE(owikh84 @ Sep 28 2019, 12:23 PM) ya I think 1.1v is quite high already, I even tried 1.2v SOC just now. thanks! quite a surprise.Not really want to go higher otherwise I will kill the IMC, I think I'm just one of those unlucky ones with FCLK capped at 1866. 3733 14-16-14-28, FCLK 1866:  3800 14-16-14-28, FCLK 1866:  yes, 1.15 with soc is my personal limit already, but so far all my settings run stable below 1.1. |
| Change to: |  0.0174sec
 0.84
 7 queries
 GZIP Disabled
Time is now: 27th November 2025 - 10:31 AM |
Quote