|
Post by scruffy1 on May 15, 2021 6:24:22 GMT -5
so i had a bit of a play with ctr2.0 and had no success at all
had another look at it yesterday and noted there's a newer version
anyhows, maybe i'm less ignorant than last time, or maybe upgrading to the b.die and finally tweaking the timings at 16-16-16-32 undervolted but stable was the magic
or even trading the hyper212 up to the noctua monster ?
the current rig is : ryzen 5 3600 /noctua nh-d15 / msi b450m mortar max / samsung 970 evo plus 500gb nvme/ wd 2tb / 2x8gb G.Skill ripjawsV @ 3600 c16 (b-die) / mx master3 / khadas tone1 dac + klipsch promedia 2.1 + alo pan am + focal elear / Lian Li PC-A17B & corsair hx750i & noctua fans / gigabyte gtx1060 oc 6g / dell u2515h
the weather here is cold for here - around 14c tonight, but "feels like 6c" per the weather site, and although you people may laugh at temps over zero, it is brass monkey levels here for the locals
anyway, with some instruction from a fairly useful video ( ), i threw the dice and came up win !
diagnostics revealed :
and so i couldn't resist applying a tune and saving, which resulted in :
oooeeee !
and for running on limited cores i hit 4.625 peak with all core watts peaking at 64.77w
idles at 3.8 on all 6 cores pulling 1.096v
amd is back in the catbird seat
|
|
|
Post by ShrimpBrime on May 15, 2021 9:25:59 GMT -5
The Bdie and cooler upgrade was totally worth it!
|
|
|
Post by scruffy1 on May 15, 2021 9:40:03 GMT -5
i still don't quite grok how it all works did a second tune and...
which is a major jump
not that i'm complaining
i hope when the freeze here subsides that the silicon is stable at more "normal" ambient temps
but hey! i'm overclocking like it's 2003
edit : only complaint is that hwmonitor suggests vid#5 hits 1.537.... but it must be a momentary thing if it's real, as the rest of the time it usually sits at 1.1v or less, with an initially recorded max of 1.375v is that a monitor error ? rather not fry the silicon, but ot makes no great sense that it's real
all that stuff has ceased to be - never seen again after a reboot, so i'm ignoring it
|
|
|
Post by ShrimpBrime on May 15, 2021 10:01:55 GMT -5
i still don't quite grok how it all works did a second tune and...
which is a major jump
not that i'm complaining
i hope when the freeze here subsides that the silicon is stable at more "normal" ambient temps
but hey! i'm overclocking like it's 2003
edit : only complaint is that hwmonitor suggests vid#5 hits 1.537.... but it must be a momentary thing if it's real, as the rest of the time it usually sits at 1.1v or less, with an initially recorded max of 1.375v is that a monitor error ? rather not fry the silicon, but ot makes no great sense that it's real
It may be better to use HWInfo64 for accurate readings. www.hwinfo.com/download/Generally if you are overclocking manually beyond defaults, you can ignore the VID.
|
|
|
Post by scruffy1 on May 15, 2021 19:03:29 GMT -5
thanks shrimp, i already also use that program hwinfo doesn't seem to show individual cores, which makes me think the reading on hwmonitor is either a sensor error or just a program glitch you know it's cold here (or the tuning is great) when i am idling at 3.8ghz and the cpu is at 27c - never usually under 40c at 3.5 - but i only just turned it on, so in an hour or so...?
i stared at hwmonitor to see if i could catch how long it hits that big voltage, and of course it's never when you are watching
ima live dangerously
|
|
|
Post by ShrimpBrime on May 15, 2021 19:53:20 GMT -5
thanks shrimp, i already also use that program hwinfo doesn't seem to show individual cores, which makes me think the reading on hwmonitor is either a sensor error or just a program glitch you know it's cold here (or the tuning is great) when i am idling at 3.8ghz and the cpu is at 27c - never usually under 40c at 3.5 - but i only just turned it on, so in an hour or so...?
i stared at hwmonitor to see if i could catch how long it hits that big voltage, and of course it's never when you are watching
ima live dangerously Looks like good temps to me. I found my 2700x to be pretty darn efficient up to about 3.9ghz. Past that, the voltage and heat start climbing. I try to keep it daily under 1.3v and it runs just great in silent mode. But you have an advantage on temps over me I think with mine in a cabinet so it's hidden and stuffy in there. I like HWInfo because it shows the memory temp and drive temps along with a whole lot of other information, some of which is useless IMO, but seems most accurate on AM4 I think.
|
|
|
Post by scruffy1 on May 15, 2021 22:48:13 GMT -5
and of course, there's no further sign of voltage aberration since i posted about it just some happy numbers
and
|
|
|
Post by freeagent on May 15, 2021 22:56:00 GMT -5
Nice man, she's a ripper for sure
|
|
|
Post by scruffy1 on May 16, 2021 8:22:47 GMT -5
dialled in, and the voltage weirdness is nowhere to be seen
boosts to >4.6 on 4 out of 6 cores, idles at 3800 on 1.096v while running significantly cooler, simultaneously running all cores at 4.375 on 1.219v stress stable (if the half an hour using the ctr stability test is accurate)
good timing to revisit this software for me given the new release was only 3 days ago - i'm very satisfied with the effort
|
|
|
Post by Vinster on May 16, 2021 22:42:37 GMT -5
for me all the attachments show a stop icon. I don't see them. happy you got them dialed in.
Vin
|
|
|
Post by scruffy1 on May 17, 2021 2:53:29 GMT -5
sorry vin, dunno why that's the case
they are google pics
are they invisible to everyone else ? i'll need to find somewhere to host the photos if i am so keen to show off
|
|
|
Post by cbjaust on May 17, 2021 8:52:44 GMT -5
Yeah, I cannot see your pictures either. Great to hear you're having fun with new hardware!
|
|
|
Post by zila1 on May 17, 2021 17:08:50 GMT -5
Same here. No pictures.
|
|
|
Post by Vinster on May 17, 2021 20:26:28 GMT -5
sorry vin, dunno why that's the case they are google pics are they invisible to everyone else ? i'll need to find somewhere to host the photos if i am so keen to show off I had a look at the links and they all lead to a stop/fault icon. that's weird.
are the photos in your google drive? if so, you need to make the folder open/shared for anyone to see the content in the folder.
Vin
|
|
|
Post by scruffy1 on May 18, 2021 7:01:21 GMT -5
reapplied the links so hopefully you can all get impressed with the outcome i certainly am
let me know if i need to try again to get them visible
|
|
|
Post by Vinster on May 18, 2021 13:52:55 GMT -5
I can see Vin
|
|
|
Post by scruffy1 on May 18, 2021 17:25:25 GMT -5
|
|
|
Post by zila1 on May 18, 2021 18:23:43 GMT -5
Yup, I can see them just fine now.
|
|
|
Post by scruffy1 on May 19, 2021 0:56:04 GMT -5
it suggests there's a little bit more space to push given vcore maxxed out at 100% stress all cores is 1.256v
what's realistic upper limit ? the dialled in max is 1.375v, and it hits it obviously at some points momentarily on hwmonitor on every core, but i haven't seen it sit there ever... i'd monitor with zen master but ctr hijacks that to do its magic, and zen master won't allow two instances to run together (surprise!)
i'm still amazed that this thing hits 4.6 simultaneously on 2 cores on the best ccx, and the temperature cracks 46c briefly, but is generally at low to mid 30's... although it's still quite cold here presently compared to "normal"
so, any of you people using ctr, or is that cheating ? because it's working for me, though ymmv
|
|
|
Post by Mr.Scott on May 19, 2021 17:31:07 GMT -5
Max for 24/7 is 1.375v all cores all the time. It will most definitely increase your temp.
|
|
|
Post by ShrimpBrime on May 19, 2021 18:21:45 GMT -5
Max for 24/7 is 1.375v all cores all the time. It will most definitely increase your temp. That's the negative effect of these chips. Best for daily, BPO enabled and increased PPT and EDC limits. Let the Cpu do it's thing and monitor it's boosting based of self monitored temps. If running an Asus ROG board with "Performance Enhancer" and fairly decent cooling, I recommend lvl 3. Lvl 4 may require your chiller.
|
|
|
Post by scruffy1 on May 19, 2021 19:39:13 GMT -5
thanks, o learned colleagues!
i shall leave it unmolested further
if hwmonitor is to be trusted enough, the cores all momentarily hit 1.375v from time to time, but desktop use is mostly well under that, and temps are low under the noctua nh-d15
ctr is using zen master to drive the ship, and per instructions (for my msi mortar max), load line calibration is medium (3), pbo is disabled, global c-control is auto
sadly for me nearly every nerd video seems to favour asus rog as the board, and the msi bios is another animal altogether, although i am pretty good at finding the settings that equate (when they exist)
still, for the price and reviews of the power arrangements, it is a great (and so far for a year) reliable board
thanks again for your input; i can recommend having a play with ctr - i never would have got to my current tune manually, i am far too lazy and vague these days
|
|
|
Post by scruffy1 on May 20, 2021 9:27:48 GMT -5
well, goodness !
edit : but never seen again, so i have my suspicions it's another brainfart from hwmonitor
ah well... 90mhz in the scheme of things is not a disastrous deficiency
edit 2 : well dang ! i just had a peek and core #2 hit 4722 it says - maybe that's a feature of ctr to occasionally overshoot the max clock ? the volts never peak above 1.375, so i won't bitch about it
|
|