*Official* NBR Desktop Overclocker's Lounge [laptop owners welcome, too]

I set my memory, 3466 CL16,18,18 @1,350v to 3600 cl 16,18,18 1.375v

Edit; Now that I have worked on one, I think I want a custom water loop as well. I blame you @Trafficante, I was perfectly happy with what I had till we met.

 

Looking at my system to go with a custom loop I would have to go to a full size case. The mid tower as it is now is pretty well stuffed. When putting together the list of parts a custom loop was never a consideration and now I guess I get to pay for that decision.

 

@TANWare what case are you siding with? As you probably know, the new Lian Li PC-O11 Dynamic co-developed with der8auer offers plenty of mounting options for water cooling parts and comes in at a very reasonable $100 due to the ongoing $30 pre-order discount, and because it's a steel affair rather than aluminium. Of course, the square form factor is not for everyone, even if it's fairly compact for what it can accommodate.

 

Yeah, this one:
https://www.gamersnexus.net/hwreviews/3285-lian-li-o11-dynamic-review-benchmark-der8auer-case-design

 

I have to research it a bit. I went with the case I have as I need 2 external access 5.25 bays, one for a BR burner and the other for an internal USB 3.0 card reader. The card reader though could sit in a 3.5 bay though. There are also internal 2.5 1tb SSD, 2.5 120gb SSD, 3TB 3.5 HDD.

 

u have a 5.25" HDD?

Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk

 

ooops, overtired, 3.5 and others are 2.5. Doh........

https://www.newegg.com/Product/Product.aspx?Item=N82E16822149633

 

This is an old post of mine, but I have new news, it was my motherboard voltage regulators overheating - the combined load of 4.7Ghz CPU Prime95 & Firestrike GPU load proved too much. I found this out by running the latest version of Prime95 on it's own, as opposed to the older non AVX version of Prime - it eventually throttles the CPU with "IA VR Thermal Alert" and "Ring VR Thermal Alert" in HWinfo, and then eventually crashes (a sustained 138W CPU Package load). So, I've gone back up to the 4.7Ghz CPU overclock, because it's not CPU instability but motherboard instability, and I never push the system hard enough for "VR Thermal Alerts" during practical usage. I've noticed there's a coating of quite a bit of dust around the CPU socket area on the motherboard, I'll probably blow it out with a can of compressed air to lower the temperatures of the voltage regulators on the motherboard.

 

what kinda cooling do u have on the vrms?

Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk

 

That's a good question, I was about to post up here asking where the VRM's are located on my motherboard. I was gonna ask if the VRM's were located under the 2 heatsinks near the CPU socket, or if they were part of the uncovered components? Here's some pics of my motherboard, anybody know where the VRM's are?
https://www.msi.com/Motherboard/Z170A-KRAIT-GAMING-3X/Gallery

EDIT: I've got quite a lot of air flowing round that area I imagine. There's my Noctua NH-D14 CPU cooler with it's 2 fans, then there's the rear panel top case fan (the usual standard place for exhaust fan), and then I have a roof panel rear case fan too in that area above the CPU cooler (those are the two exhaust fans).

EDIT 2: After researching I've concluded that the two large black & white heatsinks on my motherboard located around the CPU socket are the VRM heatsinks. I'm concluding that one of my components under those heatsinks was getting too hot & causing the throttling. I'll blow the dust away from that area. I was surprised to see it getting too hot though, as there's a lot of airflow around that area, unless 138W sustained CPU Package Load is too much to ask for a Z170A Skylake board (i7-6700K, 91W TDP - so that 138W load is 50% more than the regular TDP, too much to ask of the board??).

 

At least here is the released overclock results from Amd's new High-end chips - 7 2700X. Awful reading!!
http://forum.notebookreview.com/thr...ga-polaris-gpus.799348/page-437#post-10714616

 

https://www.extremetech.com/computi...00x-reviewed-can-amd-cream-intels-coffee-lake

If you know you need single-threaded performance above everything else, the Core i7-8700K is still the top CPU on the market. But in well-threaded code, especially rendering applications, the Ryzen 7 2700X is superior. The fact that it’s $20 cheaper and ships with a decent stock cooler hurts nothing, either. Intel may already be planning its next counter-stroke, but until those rumors become reality, AMD has our nod for top overall CPU.


https://www.techradar.com/reviews/amd-ryzen-7-2700x

Thanks to the higher-frequencies Ryzen 2nd generation is able to achieve, the 2700X overtakes the Intel Core i7-8700K in all our single-core benchmarks. The processor also continues to hold onto AMD’s crown for the best multi-core performance. Both of these figures translate to better overall performance, as evidenced by the Ryzen 7 2700X being able to encode files with the highest frame rate out of all the processors we tested in this review.

As for gaming, the Ryzen 7 2700X doesn’t beat the Core i7-8700K, but it has eroded Intel’s lead by a mere one to two frames per second.
The AMD Ryzen 7 2700X is undoubtedly the best consumer processor on the market right now. It’s only fair competitor ,the Intel Core i7-8700K, is slower on the single- and multi- core front and doesn’t offer much better gaming performance to justify its slightly more expensive price tag.

I don't base my purchase solely on overclockability, but overall performance. Sure it would be nice to have some headroom for playing, but I love the competition AMD is putting up after so many years. If it wont clock to 4.5 or 6, I will be a bit disapointed, but the other enhancements to the silicon should make it worth the purchase. At only $329 USD, for an 8 core 4.3ghz multi threading beast, I think it is worth the cashola.

 

what you could do is take off the vrm heatsink and exchange the stock thermal pads with proper ones (arctic or fujipoly for example), that should take care of any thermal throttling. airflow is indeed important btw (as shown in the initial x299 dilemma of crappy vrm cooling, especially during overclocking), check and see if an additional fan directed towards the vrm heatsink would make any difference.

Sent from my Xiaomi Mi Max 2 (Oxygen) using Tapatalk

 

Amazing bable. It seems all depends on what reviewer is out to post it's bable... And this was compared with a wimpy overclock stock clocks on i7-8700K
AMD Ryzen 7 2700X 3.7 GHz Review-Techpowerup.com

 

New dedicated 240mm graphics card radiator quick run of Super Position. Never hit 40c during the run.

http://hwbot.org/submission/3837558_

 

It counts on the test etc. as to the winner. At stock the 2700x is overall a better CPU than the 8700K. I doubt you will find this contested. The problem then lies on if yo are an overclocker then the 8700K reigns supreme and this is not contested either. Fact being 95%, or way more, of users are not and this is where AMD is a superior choice.

Also true the fact that if you have a 1700 already on a x370 board the 2700x may not be worth the money to upgrade. This is a personal choice however. Me, I am just waiting to see what the near future holds.

 

Thanks for the ideas. I don't think it's worth taking apart the heatsinks & putting in new & better pads because the throttling only happens at 140W sustained loading (Prime95 latest version), and max CPU Package Wattage during normal usage (gaming, virus scanning) is only 85W (even wPrime1024 is only 90W, so the latest Prime95 is just an insane beast!). I was surprised to see that throttling though, and I'm fairly certain that was the cause of the instability I saw when running Prime95 (non AVX) when combined with Firestrike loop - kicked up the temperatures in the case & then overheated the VRM's, caused the instability, and led me to reduce overclock to 4.6Ghz because at the time I didn't realise the CPU was throttling & didn't realise the VRM's were getting hot (CPU temps were always low, below 65 degC). Given that it's an artificially high load, I'm not gonna base my overclock on such an extreme situation, so I'll keep it at 4.7Ghz (the highest I can get it). I was very surprised to see that VRM heat issue though during that extreme testing, perhaps with doing the things you suggested would prevent such throttling & instability in such extreme testing - but it's just not worth it for practical use, I will blow the dust out though from that area. (there's no temperature sensors readable in HWInfo that relate to the VRM temperatures, so no idea what temperature they're running during normal use, but should be a lot lower than the crazy 140W induced throttling I'm guessing!)

 

EEK'd a few more points put of my set up. I just cant get mine to hold 2100mhz anymore. Drops to 2088 and sometimes on down to 2076, of course it doesn't get over 46c though.
5.1ghz on the CPU

https://www.3dmark.com/fs/15369907

 

How come it won't hold 2100Mhz anymore?

 

Benching party today with Brother @Johnksss

Spoiler: Bench Station

https://www.3dmark.com/3dm/26230055 | http://hwbot.org/submission/3838771_

https://www.3dmark.com/3dm/26230239 | http://hwbot.org/submission/3838780_

https://www.3dmark.com/3dm/26230416 | http://hwbot.org/submission/3838788_

https://www.3dmark.com/3dm/26230661 | http://hwbot.org/submission/3838799_

 

I can't get these MSI 1080 Ti GPUs to run stable over 2038 on core (yet). They need to be modded like my SC2. I'm sure that would help yours as well.

 

Good question, I guess driver updates, or could be the ambient temps being about 10f higher than the last time I ran benchmarks causing my ram to get hotter?

 

I don't know if there are temperature sensors on your VRAM? We all know that as temperature of the GPU core increases then you get a step down in the clocks, but you're below even the first throttle point right, at your 45 degC and below? Like you said you used to be able to hold 2100Mhz, weird. Maybe if some other areas of the card are heating up then it's decreasing the power efficiency of the card and you're banging up against the power limit & causing your throttling (that would show in GPUz)?

 

It started out at 2100 and ran all the way to the 9th scene then dropped to 2088 then, It got to 46c and that's when it dropped to 2076 on superposition. It drops to 2088 immediately on firestrike within the first 3 seconds as it's still in the 30c range (drivers, benchmark itself?)

I didn't have GPUZ open but I will the next time. Only performance problem I ever see on there is voltage rel.

 

@KY_BULLET My 1080 ti drops to 2088 and dont hit 40c?

First Crappy overclock on the 2700X
http://hwbot.org/submission/3838932_raiderman_wprime___32m_ryzen_7_2700x_2sec_739ms?recalculate=true

 

Here is a GPUz screen shot during Super postion, Blue is Vrel.
It shows 2100mhz as max but was only there for the first couple seconds.

I probably need to work on a good voltage curve to keep it from dropping core. Last time I tried flattening it out at 1.062 (2100mhz), it didn't like that.

 

You can stretch the GPU graph window to make it bigger so you can see more history of results (if you didn't know that already), but you can only enlarge the window when it's on the Sensors Tab. Vrel is proving that you are indeed seeing clock drop due to GPU temperature - in as much as the driver & GPU combo has run out of allowed supplied voltage for any given frequency and temperature point (Pascal trys to add more voltage up to allowed amounts to enable boost to remain at the same level as temperature increases). If you didn't want to see any 'temperature throttling' then you'd see absolutely nothing listed in the PerfCap reason of GPUz - I just know this from testing my own card. Looks like you need to lower GPU temperatures if you don't want to see that 'temperature throttling', or perhaps if you can raise the allowed voltage level then it will have enough voltage to enable max boost clocks at the higher temperatures you're seeing. I don't know if anyone else knows other workarounds, Mr Fox mentioned something about a mod being needed, he did a power mod (not the problem you're seeing I don't think) and I think he allowed his GPUs to receive up to 1.2V (could help you I think, in which case you'd then probably have to do the power mod anyway due to increased power consumption with the increased voltage).

 

Well from the reviews I've seen you've done pretty well for an overclock on the 2700x, I wouldn't expect more than that really. From the reviews I've read I think it makes sense (and sometimes from a performance point of view too) to have maximum CPU cooling and then run the 2700x at stock clocks, and then let XFR2 kick in with the boost - often times the XFR2 boost is higher than the highest stable all core manual overclock, and efficiency goes out the window with manual overclocking. AMD look like they've really optimised & maxed out the 2700X already at stock.

 

The Wprime 32m score is really bad for 2700x running 4.350MHz. But Im sure it will be better

https://hwbot.org/submission/3699699_gulftown_wprime___32m_ryzen_7_1700x_2sec_705ms

 

I thought XFR2 is only on the x470 chipset, and I'm still on the x370 Yes it must be maxed out, unless you are using extreme cooling. Hwbot is full of LN2, and Dice runs with the 2700x. I will try some auto runs in the morning to see if there is much difference.

Sent from my SM-G960U using Tapatalk

 

Not much better than my 1700x, CB 11.5 is better than any run I've had though. Problem is getting it to finish a run.

Sent from my SM-G960U using Tapatalk

 

If you test Wprime 32m several rounds in a row, rather close Wprime and fire up again and test one or two times.

What with Cinebench R15 ?

 

Part 2: Benching party today with Brother @Johnksss

Spoiler: Bench Station

https://www.3dmark.com/3dm11/12751605

https://www.3dmark.com/3dm11/12751557

https://www.3dmark.com/3dm/26232109

https://www.3dmark.com/3dm/26232251

https://www.3dmark.com/fs/15371358

https://www.3dmark.com/spy/3653216

http://hwbot.org/submission/3838818_

http://hwbot.org/submission/3838819_

 

Part 3: Benching party today with Brother @Johnksss

Spoiler: Trómos on Chilled Water

https://valid.x86.fr/inp83u | http://hwbot.org/submission/3839184_

http://hwbot.org/submission/3839183_

http://hwbot.org/submission/3839182_

32M - http://hwbot.org/submission/3839175_ | 1024M - http://hwbot.org/submission/3839177_

https://www.3dmark.com/3dm/26239317 | http://hwbot.org/submission/3839199_

https://www.3dmark.com/3dm11/12751881 | http://hwbot.org/submission/3839171_

https://www.3dmark.com/3dm11/12751866 | http://hwbot.org/submission/3839167_

https://www.3dmark.com/3dmv/5688642 | http://hwbot.org/submission/3839166_

http://hwbot.org/submission/3839197_

 

7960x at 4.9 ghz with ddr4-4000 ram *LOL* i love it

 

Ah yes, you're right about X470!

 

I noticed that you overclock your GPU to different levels in some of the benchmarks, and in some benchmarks you barely overclock them at all?

Cool, chilled water for your CPU & 5.6Ghz on 8700K! What temps you seeing with that, I imagine low even at that 5.6Ghz?

 

The MSI 1080 Ti GPUs Brother @Johnksss has right now are not the greatest and they do not overclock that well. It wasn't easy to get them stable to the point seen in the screen shots. I suspect he will be getting rid of them soon, replacing them with EVGA or ASUS.

My modded EVGA 1080 Ti SC2 is the one you are seeing running stable at 2200MHz on core and 12,000MHz on memory with 1.200V.

With the 8700K on the chilled water I am seeing 18°C idle temps and 55-60°C max temps under extreme load, with most of the 3DMark benchmarks maxing out around 45-50°C.

 

Good stuff @Mr. Fox I may sell this nascar stuff and go Intel, so I can have some fun too. I dont like the everyone has the same results crap with AMD right now. It's just not fun. I spent all this money on a custom water loop, and someone with a stock air cooler can achieve the same results.

Sent from my SM-G960U using Tapatalk

 

The problem is the 2700x hits the wall way to early. this most likely is the process being so young, the TDP does not really get to be pushed before silicon failure. This is why I feel that 6 months down the line as chips are being made the process will mature and refine. This will start yielding tighter silicon that will allow further pushing of the thermal limits. 12 and even 18 months will probably be even better.

 

I dont want to go Intel really, as I support AMD, because they are still fully supporting Windows 7. I wont ever use Windows 10 as my main OS, so I will stay with who ever bucks the establishment.

Sent from my SM-G960U using Tapatalk

 

You know, AMD has really made amazing progress and for the first time in a very long time, Intel has had to up their game due to competition. That's good for everyone, but what you're feeling now about it is exactly why I have not adopted Ryzen. It's a powerful product, but when you really enjoy overclocking and doing something others do not with your hardware, there isn't anything to enjoy with it because of that. This has been an issue a long with their GPUs as well. GCN was great at first... really gave NVIDIA a run for the money at launch, but due to limited overclocking potential they quickly fell behind.

I hope so. I would love nothing more than to see a serious upset in the enthusiast ranks. I think they will need to make it a goal and a priority in order to do well with it. If they focus on doing a great job with stock performance and do not give it any attention it may not turn out as well. Maybe they need to have some die hard overclockers holding jobs in positions of authority that understand what it means, how it looks, how to do it, and why it matters a lot to some potential customers.

 

This is a stock run of CB 15, it beats my previous 1700X at 4.13ghz. Dont look at what CB says for my 1700X, it is at 4.13ghz
I wonder if having a board that ajusts the FSB could tweak a little more out of it, as messing with the multiplier only creates instability.

 

I do not see a serious threat just better results. If you enjoy overclocking AMD is not the chip for you. I would be happy with no overclock if they were at 4,5 for XFR and 4 4 for all core. I am more like the other 98% that does not care for extreme overclocking.

Even with 7nm there is a good chance AMD will use up all the overclocking headroom stock. This Is probably why you see Intel switching to solder for the IHS. They are getting ready for the onslaught. On this note with Intel also looking to use all of their headroom, extreme overclocking may be in 'Danger, Will Robinson'.

 

I was thinking the exact same thing! That we may be looking at the end of overclocking as we know it.

Sent from my SM-G960U using Tapatalk

 

Stock CPU run of fire strike beats the best I could do on my 1700X

 

...at least the mainstream overclocking on air and water, yes. extreme overclockers dont need to be worried though i think

 

I wonder what this is going to be?

 

hmmm....precision boost overdrive sounds like manual overclocking including clock boost. maybe its a way to keep XFR2 as well while doing manual overclocking

 

Also note the gold and silver stars. Do those denote the strongest cores?