Ryzen 3600 upgrade

[Waldorf]

@faye
but thats an issue with those chips, not so much caused by LC vs air cooling/dumping heat in the case..

 

[faye]

@faye
but thats an issue with those chips, not so much caused by LC vs air cooling/dumping heat in the case..

Not saying you're wrong, but given people are suggesting X3D to the OP (and most building or upgrading AMD right now) it has to be taken in to account.

 

[Apocalypsee]

and the psu would be another point i would swap (first).
i dont know specifically about this one, but the CX isnt something i would want running +1K of parts (cpu/board/gpu)..

That's what I'm worried about. Maybe its the PSU that causing my 3080 not working anymore, it's an old PSU dated maybe 5 years back? Previously it have the same issue because as I'm daisy-chain it, but it goes away when I'm plugging it individual 8-pin but now it just won't work anymore. You suspect the PSU is the culprit?

 

[RJARRRPCGP]

It struggles with the 5800X3D. I suspect the block needs to be offset to rest over the main heat load of the chip, but so far as I know there's no compatible kit available.

The non-3D-cache variant has the same type of thermal issue. I have an AIO and with defaults, it's hard to get a heavy all-core load at less than 83C!

 

[Waldorf]

@faye
i know at least the Arctic AIO have the offset option with different mounting patterns,
and one of the best (lower $) units.

@Apocalypsee
good chance.
i think the RTX 2000 and newer have all more or less "issues" producing transient spikes.
i had 3 units from seasonic/Tt/Gb ranging from 700-850, but only on the RMx did the gpu not whine about insufficient power
when booting/under max load (status led for each power port).

i recommend starting with the psu, maybe add the Arctic 280 or (better) Alphacool Eisbaer (not sealed/fixed hoses, uses G1/4).
at least it would give you a pretty decent "base" for future (socket) upgrade/swap.

@RJARRRPCGP
not that i see it (past the "normal" ryzen behavior of non-16C chips)
the chip is supposed to boost higher if its cold, so it will always be close to the max allowed temp for that chip.
my bad sample (140/95/130) with PBO enabled will do ~85C under full (Prime) load.
then again, i always try to lower Soc/VDDx voltages (1 step above the lowest thats stable),
and my board does have pretty tight range/regulation, helping a bit as well.

 

[AusWolf]

@AusWolf
in general, dumping the cpu heat inside the case, is always the worst scenario, no matter the type of cooling used.

e.g. most of my board temps (no block, just heatsinks) stay around 50-60C maybe 70ish, under full load (dont remember exact nbrs for diff parts),
definitely not happening with the cooler dumping the heat inside the case, no matter if air or water, so on avg having 20-30C less is a noticeable difference.

and in case of Nv (at least 2000 series and below), drooping clocks at 43C, means it doesnt matter much how "little" heat it is,
it will still be enough to warm up everything noticeably, and have the gpu drop clocks, reducing performance (compared what it could be).

for most, its usually "easier" to go LC, to dump the cpu heat outside the case,
instead of trying to completely isolate the cpu air in (common) pc cases.
and with chipmakers going the temp route (run as fast as possible, as long as its cool), means in the future temps will impact perf even more.

Temperature isn't the same as heat, not to mention heat transfer. A small Ryzen chiplet running at 80 °C while consuming 70-80 W is not the same as a large Intel die running at 80 °C while eating over 200 W. Air inside your case will be way hotter in the second case.

AIOs are generally bad with X3D chips because of the bad CPU-IHS-coldplate contract. Dumping heat in- or outside of the case has nothing to do with it.

Offset mounting didn't improve thermals with my 280 mm Silent Loop 2, either - the CPU throttled with it either way.

 

[freeagent]

I have a Thermalright AIO that I am going to install after dinner. I also have my X3D installed. With my setup, I can get away with running my X3D semi passively with just the two stock fractal fans that come with my case. I smack talk AIO's pretty much every chance that I get.. now I have one from a brand I trust.. lets see how it does

Heck, I can run my 5900X semi passively up to 180w PPT. But it gets a bit warm.. 160w is better. But no fun.

 

[RJARRRPCGP]

@RJARRRPCGP
not that i see it (past the "normal" ryzen behavior of non-16C chips)
the chip is supposed to boost higher if its cold, so it will always be close to the max allowed temp for that chip.
my bad sample (140/95/130) with PBO enabled will do ~85C under full (Prime) load.
then again, i always try to lower Soc/VDDx voltages (1 step above the lowest thats stable),
and my board does have pretty tight range/regulation, helping a bit as well.

Mine actually usually does not have temps like that in gaming loads, where you get the highest boost clocks. I wouldn't have known that before I ran Maxon Cinebench R23 multicore.

 

[Waldorf]

@AusWolf
never said anything to contradict cpu temp vs heat inside the case between different cpus, never even mentioned intel.

dumping ANY heat outside the case is better, no matter the amount, as other parts will have lower temps, especially the gpu.
and the fact is, air coolers cant do that (in +99% of pc cases for consumer).

and while not helping those that have them, the problem is with the chip, not the AIOs.
or the AIO would perform bad with ANY chip from any brand...

not sure how you can mount the Bequiet with an offset, as the mounting plate is fixed.
im talking about this (Arctic AIO):

 

[AusWolf]

@AusWolf
never said anything to contradict cpu temp vs heat inside the case between different cpus, never even mentioned intel.

dumping ANY heat outside the case is better, no matter the amount, as other parts will have lower temps, especially the gpu.
and the fact is, air coolers cant do that (in +99% of pc cases for consumer).

and while not helping those that have them, the problem is with the chip, not the AIOs.
or the AIO would perform bad with ANY chip from any brand...

not sure how you can mount the Bequiet with an offset, as the mounting plate is fixed.
im talking about this (Arctic AIO):

View attachment 318107

I mounted mine with a Thermal Grizzly offset mount bought separately.

I'm not saying it's the AIO's problem, but it is a problem nonetheless, and you should be careful with your choice of AIO (or better, not buy an AIO at all) for Ryzen, especially X3D chips.

The heat inside the case argument is pointless as long as a 7800X3D is limited to 50-ish Watts and 4.2-4.4 GHz boost before overheating under a be quiet! Silent Loop 2 280 mm (my experience).

My other point is that however hot your Ryzen CPU runs, the heat dumped inside your case will be negligible due to the low amount of heat produced by it. The exhaust air off a Ryzen CPU is much cooler than that from an Intel one (my experience as well), that's why you shouldn't worry about it. Hot CPU doesn't equal heat produced.

 

[freeagent]

or better, not buy an AIO at all) for Ryzen, especially X3D chips.

This Thermalright AIO is ok..

 

[oxrufiioxo]

This Thermalright AIO is ok..

View attachment 318108

All depends on the coldplate some are more geared towards intels concave ihs some are not.

personally I have noticed square coldplates work better with ryzen than circular but that could down to the brands I've used.

 

[AusWolf]

personally I have noticed square coldplates work better with ryzen than circular but that could down to the brands I've used.

Definitely. My Silent Loop 2 has a square coldplate, and works like pure garbage with the X3D (it's awesome on Intel, though).

 

[oxrufiioxo]

Definitely. My Silent Loop 2 has a square coldplate, and works like pure garbage with the X3D (it's awesome on Intel, though).

I stick with Corsair for AIO my old H115 from like 2014 worked much better on intel than Ryzen but both the h115i platinum and h150i elite work great on ryzen. The latest NZXT and the artic aio also work great on ryzen as well. The EK aio also do a pretty good job. Those are all the ones I've tested vs NHD15 which I also like for ryzen.

I did not see any difference with a 5800X vs a 5800X3D other than the X3D chip running cooler on an aio but I've only worked with 1 5800X3D and 4 5800Xs so very limited.

On ryzen 7000 I have only worked with 2 7700X and one 7600 and all three ran fine on both air and aio.

Might try a 7800X3D next though depending on how far out Zen5 is.

I'm still not a huge fan of recommending or not recommending a specific cooling or cooling type my results are my own I think everyone should try out multiple options and use what works best for them.

 

[Waldorf]

@AusWolf
it that was the case, my first (air cooled) 2080S wouldnt have been faster than my friends 2080ti, both running 5800X (his a golden sample, mine bronze),
even when both had used the same cooler (240 AIO), with his blowing heat inside (front mounted), and mine exhausting to the top.

as i said i havent messed with AM5/7800, but again, it would mean its an "issue" with that type of chip, and not related to air vs LC cooling.

as long as you havent used more than one AIO, and can actually confirm its related to all AIOs,
i still say its more likely a problem with the bequiet AIO, not all of em.
so far reviews like on guru3d state, even a simple LC can be used, so...

and even if there is an issue with X3D running hotter on AIOs, the same way we now see "offset" mounts for AM4/ryzen,
to account for the "hot spot" on non-5950x, i expect brands (that care), to adjust for that (design/new mount) on AM5.
(another reason i always wait for 2nd chip, by then companies fixed/updated their stuff).

i see it similar to the tire "problem", when the Veyron came out.
it wasnt so much a problem with tires (as a tech), but the fact that there was no road legal production car with +1000HP/+230mph before,
so no need for tire manufacturers to provide one.
and since then not a single hyper car review, has talked about that "there is no tire for this kind of car".



btt:
may recommendation for parts, incl the AIO, still stands, as it was about the existing 3600, and later 5800x/3D,
as a cooler running 3600, especially with rad set as exhaust thus lower intake air temp for the gpu as well,
would make both boost higher, and might help holding out longer, and save up for a full swap (cpu/board/ram).

 

[AusWolf]

I stick with Corsair for AIO my old H115 from like 2014 worked much better on intel than Ryzen but both the h115i platinum and h150i elite work great on ryzen. The latest NZXT and the artic aio also work great on ryzen as well. The EK aio also do a pretty good job. Those are all the ones I've tested vs NHD15 which I also like for ryzen.

I did not see any difference with a 5800X vs a 5800X3D other than the X3D chip running cooler on an aio but I've only worked with 1 5800X3D and 4 5800Xs so very limited.

On ryzen 7000 I have only worked with 2 7700X and one 7600 and all three ran fine on both air and aio.

Might try a 7800X3D next though depending on how far out Zen5 is.

I'm still not a huge fan of recommending or not recommending a specific cooling or cooling type my results are my own I think everyone should try out multiple options and use what works best for them.

My AIO worked well on the 7700X, just not the X3D.

@AusWolf
it that was the case, my first (air cooled) 2080S wouldnt have been faster than my friends 2080ti, both running 5800X (his a golden sample, mine bronze),
even when both had used the same cooler (240 AIO), with his blowing heat inside (front mounted), and mine exhausting to the top.

Why would a 2080 Super be faster than a Ti? Don't be daft.

With different AIO mounts, we're talking about a couple degrees warmer or cooler airflow at the GPU, which its fan speed mostly compensates for. You lose maybe a boost bin or two, a negligible difference.

as i said i havent messed with AM5/7800, but again, it would mean its an "issue" with that type of chip, and not related to air vs LC cooling.

as long as you havent used more than one AIO, and can actually confirm its related to all AIOs,
i still say its more likely a problem with the bequiet AIO, not all of em.
so far reviews like on guru3d state, even a simple LC can be used, so...

I never said it's related to all AIOs. I said it's related to some AIOs on X3D chips. But as long as you don't know which AIOs are good, and which are bad for X3D, I recommend avoiding them. Not to mention that my Silent Loop 2 cost me £120, while the Dark Rock 4 that runs miles around it (on the X3D at least) is about £50. Overspending on a big fat maybe isn't worth it when you have a much cheaper option that works with 100% certainty.

 

[oxrufiioxo]

I never said it's related to all AIOs. I said it's related to some AIOs on X3D chips. But as long as you don't know which AIOs are good, and which are bad for X3D, I recommend avoiding them. Not to mention that my Silent Loop 2 cost me £120, while the Dark Rock 4 that runs miles around it (on the X3D at least) is about £50. Overspending on a big fat maybe isn't worth it when you have a much cheaper option that works with 100% certainty.

I think that goes both ways someone on this forum had an AK620 Digital which I think performs similarly to your Dark rock pro 4 and said it worked fantastic on his chip tried it on a 7600 and results where pretty meh compared to a $120 Noctua u12a doesn't mean I am going to recommend everyone who buys a Ryzen cpu get a 120 usd air cooler. Maybe I got a dud I have seen it work well on other peoples 7000 setups but it just didn't work as well as expected for me.

 

[AusWolf]

I think that goes both ways someone on this forum had an AK620 Digital which I think performs similarly to your Dark rock pro 4 and said it worked fantastic on his chip tried it on a 7600 and results where pretty meh compared to a $120 Noctua u12a doesn't mean I am going to recommend everyone who buys a Ryzen cpu get a 120 usd air cooler. Maybe I got a dud I have seen it work well on other peoples 7000 setups but it just didn't work as well as expected for me.

That might be related to heatpipe orientation and contact, but fair point. I've read somewhere that coolers with direct contact heatpipes are generally bad for Ryzen (I'm not sure what the AK620 has).

All in all, I think it's fair to say that just because your AIO (or air cooler) works well on one CPU, you can't assume that it'll work well on another one as well.

 

[oxrufiioxo]

That might be related to heatpipe orientation and contact, but fair point. I've read somewhere that coolers with direct contact heatpipes are generally bad for Ryzen (I'm not sure what the AK620 has).

All in all, I think it's fair to say that just because your AIO (or air cooler) works well on one CPU, you can't assume that it'll work well on another one as well.

The coldplate


I am not even necessarily disagreeing with you and honestly for the money the AK620 Digital is a pretty nice looking cooler and maybe some of my disappointment comes from typically using much more expensive cooling my only point if you can even call it one is people should try everything and see what works best for them within their means anyway if a no name aio works awesome for them great if a 30 usd air cooler gets the job done awesome.

I would avoid the Be quiet aio due to your experiences but would maybe give a dark rock pro a chance in the future so all insight is valuable.

 

[AusWolf]

The coldplate
View attachment 318127

I am not even necessarily disagreeing with you and honestly for the money the AK620 Digital is a pretty nice looking cooler and maybe some of my disappointment comes from typically using much more expensive cooling my only point if you can even call it one is people should try everything and see what works best for them within their means anyway if a no name aio works awesome for them great if a 30 usd air cooler gets the job done awesome.

I would avoid the Be quiet aio due to your experiences but would maybe give a dark rock pro a chance in the future so all insight is valuable.

Don't get me wrong, the Silent Loop 2 is an awesome AIO. It just doesn't work with my 7800X3D, that's all. I'd happily recommend it for any non-X3D CPU, especially for Intel.

Edit: That coldplate looks a bit thin. Maybe that's the issue, as the area above the CCD is only in contact with a few of the heatpipes in certain orientations (just a guess).

 

[Apocalypsee]

I managed to get my 3080 back online, after some cleaning and repasting, and flashing latest VBIOS from Dell. Temperature drop is massive, if you see previous temperature is 70+ now its around 50+ idk what kind of paste the seller use previously, but I'm using old trusty Arctic Silver 5. I really don't know what's the problem with the card but I really hope it will last at least another half a year.

 

[oxrufiioxo]

I managed to get my 3080 back online, after some cleaning and repasting, and flashing latest VBIOS from Dell. Temperature drop is massive, if you see previous temperature is 70+ now its around 50+ idk what kind of paste the seller use previously, but I'm using old trusty Arctic Silver 5. I really don't know what's the problem with the card but I really hope it will last at least another half a year.

View attachment 319383

At least till the super cards come out hopefully

 

[freeagent]

but I'm using old trusty Arctic Silver 5.

People knock AS5 all the time.. I still think its good.

 

[Macro Device]

I managed to get my 3080 back online

Congratulations, dude!

 

[Hyderz]

I managed to get my 3080 back online

now you can focus on saving up some money and upgrade your cpu/ram/mobo happy gaming