|
I'll AIO my CPU next build for sure with the big 280 double rads, why not. Also as someone with an AIO GPU solution on a 980ti hybrid I have to say that works excellently as well.
|
#
?
Mar 28, 2018 04:27
|
|
|
|
| # ? May 15, 2024 16:00 |
|
|
craig588 posted:Air if you have a big case that can fit it, AIO if you have anything small or awkward. Custom loops can perform better but still need maintenance, I haven't done a 560MM rad because of that, wouldn't want to trade out my D15 after trying a Corsair 280mm AIO though, similar performance but more noise. I'm simultaneously super impressed by the performance of the Scythe Fuma, and uncertain if I want one because I'd be missing out on AIO RGB bling. I am an idiot. Goddamn it. I may buy an AIO anyway because I'm probably going to get one of these poor airflow/sounds quiet cases. My current build is a high airflow case wherein I did everything I could think of (bigger 200mm fans running slower, for example) to reduce noise. EDIT: On the other hand, if I got the Fuma I could put it on my current CPU almost immediately and probably have better cooling than this H80. 
Craptacular! fucked around with this message at 09:20 on Mar 28, 2018 |
|
#
?
Mar 28, 2018 09:10
|
|
|
Alpha Mayo posted:Would liquid metal benefit me much with a 2500K? I know my chip can do 4.8ghz but the temps climb into the upper 80s at that point, so I leave it at 4.5. Using a 212 cooler with AS5. Some games are a bit CPU bound with my 1070Ti. The heatspreader itself is soldered. You'll need a better cooler than a 212, and also some better thermal paste.. the best non-electrically-conductive is Thermal Grizzly's Kryonaut.
|
|
#
?
Mar 28, 2018 09:47
|
|
|
HalloKitty posted:... the best non-electrically-conductive is Thermal Grizzly's Kryonaut. Yes by 1-2°C and only if you get a good batch. Going by recent (german) forum posts and amazon reviews they have QC issues that throw off temperatures because some batches have bubblegum-like consistency that just won't spread. It seems like I had a "good" and a "bad" 1g batch because the difference was very significant. I ended up using a different product. craig588 posted:Air if you have a big case that can fit it, AIO if you have anything small or awkward. Custom loops can perform better but still need maintenance, I haven't done a 560MM rad because of that, wouldn't want to trade out my D15 after trying a Corsair 280mm AIO though, similar performance but more noise. A main reason why AIOs are prevalent now is because overclocked prebuilt or X299 systems require good cooling and it is much harder to ship a system with a heavy Dark Rock Pro 3/Fuma/Cryorig R1/NH-D15 mounted to the motherboard than with a radiator bolted to the case. That and "watercooled" just sounds good. With acceptable case ventilation dual tower heatsinks cool as good or better than 240mm AIOs and you don't have to worry about catastrophic unattended pump failures and resulting leaks, pump noise/vibration, silent degradation due to liquid permeation, corrosion clogging up microfins or questionable windows-only software keeping your CPU from clocking down just to manage pump/fan RPM. 280mm AIOs may outperform good aircoolers by ~5°C, perhaps more with bad case ventilation, but for me they're not worth the downsides because I have an ATX case and don't move/ship my PC around. Custom watercooling is the best option if value for money is no concern and you thoroughly enjoy tinkering with your PC as a weekend project. btw same goes for GPUs: https://www.youtube.com/watch?v=K-9k5cRbrPU
|
|
#
?
Mar 28, 2018 11:28
|
|
|
I'd be surprised if the difference between Kryonaut and AS5 is actually meaningful, and the need to use AS5 instead of whatever paste shipped with your purchase evaporates a little more each year.
|
|
#
?
Mar 28, 2018 12:02
|
|
|
eames posted:Yes by 1-2°C and only if you get a good batch. Going by recent (german) forum posts and amazon reviews they have QC issues that throw off temperatures because some batches have bubblegum-like consistency that just won't spread. It seems like I had a "good" and a "bad" 1g batch because the difference was very significant. I ended up using a different product. OK, I didn't know about that. Fair enough. In that case, I'd recommend Noctua NT-H1; it's cheap, comes in giant syringes, and always performs really well.
|
|
#
?
Mar 28, 2018 14:46
|
|
|
Craptacular! posted:I'd be surprised if the difference between Kryonaut and AS5 is actually meaningful, and the need to use AS5 instead of whatever paste shipped with your purchase evaporates a little more each year. It was several years ago but HardOCP did a test and determined that American cheese was only about 6C worse under load than their best thermal grease, so I'd be willing to wager that the vast majority of thermal greases are going to perform similarly.
|
|
#
?
Mar 28, 2018 14:47
|
|
|
American cheese comes in a default state better than, from what I've seen, most people's attempts at spreading thermal paste. A lot of people would probably benefit from using sliced cheese.
|
|
#
?
Mar 28, 2018 15:38
|
|
|
Eletriarnation posted:It was several years ago but HardOCP did a test and determined that American cheese was only about 6C worse under load than their best thermal grease, so I'd be willing to wager that the vast majority of thermal greases are going to perform similarly. There is definitely such a thing as bad batches though. So while the brand doesn't matter, if you see poor results from one tube it might be improved by getting another one.
|
|
#
?
Mar 28, 2018 15:51
|
|
|
Alpha Mayo posted:I could definitely fit it, though I've read things that make me worry about AIO durability. Never owned one but can they last for years? If I bought one I'd want to use it with my next CPU when I eventually build a new machine (probably in 2019) I don't have data on failure rates but I've built a couple dozen systems with them and never had a leak or have a pump break. Sometimes if you get one that's been sitting on a shelf for years, it'll make noise from a small amount of air bubbles in the system, but the hydro refurbs I linked I've found to be topped off better than the "new" ones. I have some in my office here that have been running 24/7 for over 3 years now.
|
|
#
?
Mar 29, 2018 01:32
|
|
|
I wouldn't worry about the longevity of AIOs, I have one still going after a decade, originally bought to cool a QX6800 in 2007. It's mostly that they're louder for the same cooling performance, if you don't care about noise and have an awkward case get one by all means. It's not like going from 40 to 60dB, more like 40 to 45, it's a relatively small increase in noise for the convenience.
|
|
#
?
Mar 29, 2018 01:44
|
|
|
My 7600k @ 5ghz is score a lot higher than an 8700k on cpu-z benchmark for single core. Does this make sense ? It scored 608 while the 8700k is high 400's.
|
|
#
?
Mar 29, 2018 06:03
|
|
|
jonathan posted:My 7600k @ 5ghz is score a lot higher than an 8700k on cpu-z benchmark for single core. Does this make sense ? It scored 608 while the 8700k is high 400's. Was the 8700K also at 5ghz?
|
|
#
?
Mar 29, 2018 06:32
|
|
|
jonathan posted:My 7600k @ 5ghz is score a lot higher than an 8700k on cpu-z benchmark for single core. Does this make sense ? It scored 608 while the 8700k is high 400's. If it's a stock 8700k this makes sense. If they are both at 5GHZ, I am guessing the 8700k is getting throttled somewhere. The best test would be to just set 1 core to 5GHZ to give an apples to apples. Also, even if everything is fine I'd expect the two probably within margin of error of each other in single core. If the 7600k is just barely edging it out I wouldn't be shocked. Edit: High 400s is stock 8700k, I think it throttled to stock. 8700k @ 5GHZ will be closer to 600 http://valid.x86.fr/bench/a5wyah https://valid.x86.fr/lmdr5k Lockback fucked around with this message at 16:12 on Mar 29, 2018 |
|
#
?
Mar 29, 2018 16:10
|
|
|
My 8700K at 4.9 GHz but with an AVX of 4.7 and dipping into that often enough that I'm considering just handing it more volts and an AVX offset of 0 does 570 single-core. I would not be surprised if it at a solid 5.0 GHz hit 600+ no problem.
|
|
#
?
Mar 29, 2018 20:41
|
|
|
Try using OCCT to see the actual clock during the usage. The clock might run wildly especially if its getting too hot and didnt delid.
|
|
#
?
Mar 29, 2018 21:25
|
|
|
Yeah sorry I meant my 7600k @ 5ghz is benchmarking quicker than the reported stock 8700k. I thought the single core performance was a big bump from the 7600k but I guess not.
|
|
#
?
Mar 29, 2018 23:38
|
|
|
jonathan posted:Yeah sorry I meant my 7600k @ 5ghz is benchmarking quicker than the reported stock 8700k. I thought the single core performance was a big bump from the 7600k but I guess not. I don't think they added much in single core performance. The main upgrade is just the additional two cores.
|
|
#
?
Mar 29, 2018 23:44
|
|
|
jonathan posted:Yeah sorry I meant my 7600k @ 5ghz is benchmarking quicker than the reported stock 8700k. I thought the single core performance was a big bump from the 7600k but I guess not. They increased the single core stock speed, but not dramatically (about 100MHZ across similar levels). It is a big jump in stock speeds from the 7600k to the 8700k, but OCing covers that ground.
|
|
#
?
Mar 30, 2018 02:08
|
|
|
So in the few recent generations, the single core gains have been in clock speed stability rather than doing more at a given frequency. Gotcha.
|
|
#
?
Mar 30, 2018 03:36
|
|
|
jonathan posted:So in the few recent generations, the single core gains have been in clock speed stability rather than doing more at a given frequency. Gotcha. Skylake/Kaby Lake/Coffee Lake have the same architecture and IPC, iirc
|
|
#
?
Mar 30, 2018 10:07
|
|
|
jonathan posted:So in the few recent generations, the single core gains have been in clock speed stability rather than doing more at a given frequency. Gotcha. Some other features too that won't show up in all benchmarks, but yeah. The big performance jump is in number of cores.
|
|
#
?
Mar 30, 2018 13:29
|
|
|
It sucks that more cores doesn't really mean anything for average desktop user or games. I remember back in like 1993 you could get dual 486 motherboards, but it wouldn't do anything for anyone except maybe 3d studio for DOS with a special $500 plugin. And it's no different today. A $150 SSD shouldn't be bottlenecked by a 32 thread threadripper when running Windows updates or whatever. I'm talking out my rear end but it's annoying that there isn't good scaling in day to day tasks with high end CPU's and multi GPU.
|
|
#
?
Mar 30, 2018 20:47
|
|
|
jonathan posted:It sucks that more cores doesn't really mean anything for average desktop user or games. I remember back in like 1993 you could get dual 486 motherboards, but it wouldn't do anything for anyone except maybe 3d studio for DOS with a special $500 plugin. My one Pc regret (which is weird AF as it was years ago) is that I didn't buy a dual P3 board back in the day. Because that would have owned so hard at university for years.
|
|
#
?
Mar 30, 2018 21:41
|
|
|
Is buying a used AIO off Craigslist a bad idea for any reason? There is an H110i for $40 and some others in the 30-50 range. That's more in the budget of what I'd be willing to spend on my 2500K
|
|
#
?
Mar 31, 2018 01:23
|
|
|
I'd be a bit wary of a used AIO. They're probably the most likely thing in a PC to break and failure could be catastrophic. If it's from someone you trust that's a different story, but some random guy? ehhhh Edit: I'm so paranoid about being a guy known for passing off bad old hardware I give it all away. Not my fault if it breaks in 2 hours, worked fine here and it was free. I haven't had anything crazy break, but I have that paranoia. craig588 fucked around with this message at 01:35 on Mar 31, 2018 |
|
#
?
Mar 31, 2018 01:31
|
|
|
craig588 posted:I'd be a bit wary of a used AIO. They're probably the most likely thing in a PC to break and failure could be catastrophic. If it's from someone you trust that's a different story, but some random guy? ehhhh Agree with this because AiOs can age really badly with the moving parts and the fluid. The pump might be hosed, the coolant might be sludge, it could have 6 months of useful life left
|
|
#
?
Mar 31, 2018 01:37
|
|
|
Also if your AIO that you bought new fails you can have the company pay for the CPU replacement.
|
|
#
?
Mar 31, 2018 01:44
|
|
|
It really depends on the brand and model of AIO. EVGA guarantees their AIOs for five years and I called a tech on the phone to confirm that, like their video cards, the warranty moves with the product. My current H80 for five years is a Fry's open-box return. I've never purchased a cooler new.
|
|
#
?
Mar 31, 2018 01:52
|
|
|
 Good? Got my new rig, 4.9(5.0+ wasn't stable even at 1.4) i7-8700k and OC'd the 1080 Ti a bit, seems stable for now.
|
|
#
?
Apr 1, 2018 02:39
|
|
|
Akumos posted:
4.9 GHz all-core on the CPU is just fine yeah. GPU overclocking these days, especially with Pascal, isn't nearly as much of a big deal as it was five years ago. You should be able to get 5.0 GHz all-core stable if you put in an AVX offset of -2. That'll drop cores whose pipelines have AVX instructions in them to 4.8 GHz until the AVX instructions are out of the pipeline.
|
|
#
?
Apr 1, 2018 04:09
|
|
|
What's the deal with AF140 (Quiet Edition)? Some seem to swear by it while others report a really weird noise profile at various RPMs. Do Corsair have a really weird QA issues, or what's going on? I thought going by recommendations was hard, but even more so with fans.
|
|
#
?
Apr 2, 2018 22:06
|
|
|
ufarn posted:What's the deal with AF140 (Quiet Edition)? Some seem to swear by it while others report a really weird noise profile at various RPMs. Do Corsair have a really weird QA issues, or what's going on? The new hotness is ML140 and they're desperately trying to pretend they didn't push the airflow/static pressure dichotomy for case fans
|
|
#
?
Apr 3, 2018 04:08
|
|
|
mewse posted:The new hotness is ML140 and they're desperately trying to pretend they didn't push the airflow/static pressure dichotomy for case fans
|
|
#
?
Apr 3, 2018 13:08
|
|
|
Is there any decent monitoring and logging solutions that run as a service? I basically want to get a CSV of temps, fan speeds and so on but my computer is multi-user and I'd rather not leave an account always logged in (plus automated reboots, etc).
|
|
#
?
Apr 9, 2018 19:35
|
|
|
So up until recently I was running an i5 3570k that I'd overclocked to 4.7ghz (odd for ivy bridge I know) and I felt it was getting pretty long in the tooth, there were more and more games coming out where I was CPU limited even with that overclock. But I couldn't just move to a whole new platform since it's way too expensive at the moment (thanks DRAM prices you fuckers) so I felt kinda stuck. Coincidentally a friend of mine had an i7 3770k that he wasn't using that he just gave me for free because he's a bloody legend, and I set about overclocking that once I got it. Now I'm getting the true ivy bridge experience. I've got it stable at 4.5ghz but temps run at about 80c under load, which is high but acceptable imo. My current cooler is a hyper 212 evo, and the fan that came with it seemed to have worn out bearings and was making a bunch of noise, so I replaced it with 2 cryorig fans and mounted them both on the heatsink with a new application of thermal grizzly cryonaut which is apparently the best typical thermal paste but I have no loving idea. Anyway the idle temps improved a whole lot, like 20c in some cases, which is cool and all, but the load temps are identical? I gather this is the limitation of the crappy TIM they started using on these intel CPUs? I'm just confused since I got a pretty remarkable idle improvement but I can't get these load temps down at all. I was hoping I could get the temps down since it got to 4.5ghz stable without increasing the voltage at all.
|
|
#
?
Apr 10, 2018 07:01
|
|
|
cat doter posted:Anyway the idle temps improved a whole lot, like 20c in some cases, which is cool and all, but the load temps are identical? I gather this is the limitation of the crappy TIM they started using on these intel CPUs? I'm just confused since I got a pretty remarkable idle improvement but I can't get these load temps down at all. I was hoping I could get the temps down since it got to 4.5ghz stable without increasing the voltage at all. Delid?
|
|
#
?
Apr 10, 2018 07:19
|
|
|
Haha, if only I had a printer on hand. If I can't afford a new CPU/RAM/motherboard I certainly can't afford one of those. I figured delidding was the only solution, which is a bummer. I've seen one of those derbauer delidding tools but here in australia they're like $50. 4.5ghz will have to do I think, it's not like that's a bad overclock.
|
|
#
?
Apr 10, 2018 07:23
|
|
|
cat doter posted:Haha, if only I had a printer on hand. If I can't afford a new CPU/RAM/motherboard I certainly can't afford one of those. There's a site called 3dhubs where you can find people doing 3d printing services locally. The tool can be printed out of PLA plastic which is the easiest for 3d printing, recommended 40% infill. Should be significantly cheaper than $50.
|
|
#
?
Apr 10, 2018 07:35
|
|
|
|
| # ? May 15, 2024 16:00 |
|
|
Could use a razor blade.
|
|
#
?
Apr 10, 2018 08:46
|
|