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

https://www.3dmark.com/3dm/24504728?

This thing is pretty damn good. Especially this is with stock core clock! Wait.... got change drivers.....


https://www.3dmark.com/spy/3102961

THERE WE GO! ZERO core OC and 17th on HOF for timespy extreme.

https://www.3dmark.com/hall-of-fame-2/timespy+graphics+score+extreme+preset/version+1.0/1+gpu

Timespy regular

https://www.3dmark.com/spy/3103005

@Mr. Fox
@Papusan

 

AMD All But Relinquishes High-End Desktop Graphics In 2018

A Gaping Hole In AMD’s 2018 Graphics Roadmap
It’s public knowledge at this point that NVIDIA plans to introduce a new lineup of gaming GeForce graphics cards based on its brand new Volta architecture this year. Unlike its competitor, NVIDIA has been following a strict yearly product roadmap since the beginning of the decade.

Like clockwork, every year since the introduction of Kepler in 2012, NVIDIA would deliver a high-end desktop graphics product that surpasses what precedes it.

On the other hand, AMD’s Radeon product cycle has seen far less consistency and as a result we’ve seen the company and its fans genuinely suffer at times. 2014 is one clear example when the company appeared to simply concede competition and resolved to a rebranding scheme that sent its market share tumbling to as low as 18% a year later, the lowest it has ever been since the acquisition of ATi a decade earlier.

AMD Radeon RX Vega 64 and Vega 56 Left To Fend For Themselves Against NVIDIA’s Volta In 2018
If we’ve learned one thing from AMD’s recently published 2018 graphics roadmap is that there’s going to be another product gap this year and we could perhaps be on track to see a repeat of 2014.


According to AMD’s most recently published graphics roadmap, 2018 is going to see a continuation of existing Vega 64 and Vega 56 graphics cards, with no new addition to the high-end desktop. Vega, which hasn’t been able to compete with NVIDIA’s GP102 based products like the GTX 1080 Ti, continues to be in short supply to date. What makes overall matters worse is that it will have to compete with a brand new roster of Volta products later this year.

 

I like the "average" column in HWInfo64 - if you reset the sensors just after starting a CPU stress test like Prime95 or GPU stress test like Firestrike loop for instance, then it will give you a very good representation of overall cooling capability over a long run as well as the average power consumption that the cooling system has had to deal with over that time frame, which is the factor that pushes a cooling system to it's steady state limits. I'm also a fan of that timer at the bottom of HWInfo sensors (if you reset it just before or just after starting a long stress test), because in post run screenshots it will help paint a more accurate representation of the length of testing done when you can see the length of testing time at the bottom HWInfo Sensors window, which is important in understanding if steady state has been reached - particularly in water cooled setups where it takes time to saturate the heat capacity of the water within the loop. Both the average column & the visible timer (if reset just before testing), really help to give context in screenshots on the temperatures & testing seen.

 

Money money money, must be funny...




Good scores!

 

3DMark 11 - https://www.3dmark.com/3dm11/12595237 | http://hwbot.org/submission/3757209_

 

OK, I will give you that. I can understand your point if you reset it right before a prolonged stress test or immediately prior to starting a prolonged gaming session. and I can totally understand how and why you would find some value in it. Fair enough.

But, my point still applies about the average column as well. Even if you do not know what the average temperature is, knowing what the maximum temperature is will always be critical because the max is what results in performance and/or stability problems and that is where the most effort deserves to be concentrated. If you're benching and pushing the system extra hard for short bursts to try and achieve higher scores, that matters most of all and the average is essentially irrelevant, which is what @Trafficante is shooting for.

The average is also dependent on being reset to be meaningful. If you leave your system on all day and it is idling and cold, that will also skew your average. It will be a mathematically accurate average based on the amount of data collected, but it will not be a relevant average that adds to your understanding of system behavior unless you reset it.

 

Yep, I agree, maximums are important too, and averages are meaningless on short runs, and you do indeed need to reset the timer just before a 'session' for it to hold any value.

 

3DMark 11 Physics Test Ranking - http://hwbot.org/submission/3757243_

 

Cinebench R11.5: http://hwbot.org/submission/3757262_

 

Please add all of your scores to HWBOT to help Team @Prema overtake the EVGA Enthusiast Team. We're less than 6 points away from pushing them down a notch on the leaderboard. Be sure to use the right links and correct content in the screen shots so they do not get taken down or banned. (Rules for each benchmark are linked at the bottom of each submission page.) Let's open up that can of i9 Extreme whoop-ass that you have hiding over there in Jersey and show the kiddos how it's done. Maybe Brother @Johnksss would have some time one of these days to remote connect to your monster with Team Viewer and show you some things you can do to clock the CPU a little higher that I don't know about and help us get in the top 75 ranking with your beast.

 

Is CB11 still valid for benching? I thought it was replaced by v15.

 

Which benchmark result are you submitting?

 

Do you have benchmark results from XTU? If so, use pre-populated fields from XTU.
Sometimes HWBOT doesn't have some CPU/SSD models so you must use some another name and post correct CPU from CPU-z.

 

It is not recognizing the CPU. Type 7940X and then select the option that appears in the drop-down menu that automatically appears. You cannot do free-form entry in the product or brand information fields.

Once you have already submitted benchmarks and a photo of your system, the easiest thing to do is select a previous submission and let it fill in all the blanks and add the photo automatically. Once you do that you can simply change anything that is different than the last benchmark, like CPU, GPU or RAM clock speeds if they are different than the last one.

They have his CPU as I submitted benchmarks already and the entries were accepted. Sometimes the XTU pre-populated values are rejected because they do not match how it is listed in the HWBOT database. It has to be character-perfect in every way so they do not end up with multiple categories for identical hardware.

 

You already joined back around Dec 26...

 

Mobius 1 has joined team prema.

 

Posted the new team banner here:

https://twitter.com/PremaMod/status/950323863409254401

 

Time Spy: https://www.3dmark.com/3dm/24520450# | http://hwbot.org/submission/3757694_

 

Updated the opening post...

Anyone that is serious about benching that would like to become a contributing member of the @Prema HWBOT benching team, let me know (or @Johnksss) and we will provide you with a special password to join Team PremaMod. I have PM turned off here, but you can message me or add a message to my profile wall at HWBOT or John at HWBOT and mention your forum name here.

http://hwbot.org/team/premamod/

 

@Mr. Fox @Johnksss @Mobius 1 We have overtaken EVGA enthusiasts. Congratulations.

 

Hallelujah!

Brother @Trafficante played a big part in that with his MegaBeast desktop. Will need a lot more with proper (qualifying) screen shots with that wicked machine to keep the @Prema Team advancing against our competition. Any submission that earns points matters.

 

As of right now, I am a member of xtreme systems, but don't know how to switch.

Sent from my SM-G935T using Tapatalk

 

Go into your HWBOT Profile and switch there. I will send you the password.

 

Password sent.

Click on My Account up at the top, then go down to Teams and select PremaMod from the drop-down.

Spoiler: Joining Team

 

Switched, although I have not done much as of late. Water cooler showed up today, and it is of great quality from what I am seeing so far. Water block alone must weigh a pound, as it is pure copper. I need to wait for my three RGB fans from newegg to arrive before I dive into installing it. Maybe I should have stuck with the stock 170mm reservoir, as opposed the 270mm I ordered. That thing is huge, but I will find a way to fit it in the case. Everything seems to be of good quality, so I hope it works as well as it looks.

 

Very nice. Congrats on the sweet parts, and thanks for joining the PremaMod family.

 

HWBot Aquamark3 @ 5.2GHz: http://hwbot.org/submission/3758857_

 

Do not have everything buttoned up yet as far as the water cooling, but ran a quick cinebench @ 4.1ghz.



Had the reservoir mounted in a different spot, and did not like it. I need to run another tube off the left side of the water block, as it is too short now.

 

Has it increased your GPU temperatures since having watercooled your CPU? Thinking reservoir is in path of intake fans, but also you're pushing all the CPU heat into the case in a dispersed fashion rather than using an air cooler that is in line with exhaust fan (by the looks of it?).

 

3DMark Fire Strike: https://www.3dmark.com/fs/14679857 | http://hwbot.org/submission/3760019_

3DMark Vantage: https://www.3dmark.com/3dmv/5668861 | http://hwbot.org/submission/3759964_

 

Gpu has cooled down since water cooling, at least at idle. You have to remember that I don not have the added heat of the heat pipe style air cooler air traveling around the case anymore also. I have not had a chance to test much yet. Case flows from front to back, and radiator from outside to inside. Everything exhausts to the back. Over clocked to 4.1ghz, running cinebench the temp topped out at 57C. I would say that it performs fairly well, since I was topping out on air around 71C.

 

Passmark cpu score @ 4.1

 

Cinebench [email protected] 60C

 

I suppose it's the load temperatures that matter, but idle temperatures should indicate load temperatures to some extent. I'm surprised that GPU temperatures have not increased with that layout due to the reservoir looking like it's blocking cold air flow to the GPU fans from the intakes at the front of the case, as well as all that hot air from the CPU just entering the whole case from the top panel radiator. With my setup the CPU load doesn't increase case temperatures at all because the CPU air cooler is in a direct line with the exhaust fan, and the exhaust fan is only an inch or so from the rear of the CPU cooler - so all the hot air just goes straight out; if you've got your CPU radiator on the top panel and it's spewing heat into the case in a diffuse manner there's no direct line to the exhaust from that heat, so the hot air is more likely to mix in the case (rather than being immediately exhausted) thereby raising case temperatures & consequently GPU temperatures. Maybe you'd have to do load testing with Prime & Firestrike before & after to reliably see the difference - it would have to be the same run & test methodology before & after CPU liquid cooling installation, so I'm not sure if you've got that baseline data from your previous air cooled setup.

(You say you "don't have added heat from the air cooler flowing around the case anymore", but your CPU radiator is acting as an intake right, if that's the situation then you're putting the same amount of heat into the case - so it just comes down to how efficiently that hot air can be exhausted from the case when comparing your air cooled to liquid cooled setup)

 

It might not make that much difference in the grand scheme of things, especially considering the type of heat sink on that GPU. The GPU is always going to suck in whatever air is present, blow it out the side and recirculate it. If the ambient temperature in the case is no warmer than before, which is most likely the situation, the air the GPU fans are circulating will be at or near the same temperature. If it were a blower style GPU that were pushing all of the hot air it generates out the back, then it might be a bit cooler since that type is not recirculating its own hot air. Ideally, when he buys his next GPU it will also be water cooled and both radiators will draw from cooler outside air to more effectively cool the liquid as opposed to blowing pre-heated air through the radiators. The ambient temperature inside of the case should not any be hotter, and perhaps a little cooler, than an air cooled solution because the CPU and GPU are producing the same number of watts worth of heat.

I have 11 fans in total, 10 of which pull in cool outside air and push the hot air out everywhere there is an opening. The way this case is made, (View 71 for those that are new to the thread,) there are openings literally everywhere. Nothing is sealed off anywhere on the front, sides, top, bottom or back. Every glass panel has between 0.25 inch and 0.5 inch air space space between it and the chassis to pull air in or push air out. The air inside of my chassis feels about the same temperature as the air in my room. I haven't compared the internal and external ambient temps with a thermometer, but it would be interesting to do that.

 

MAXXMEM: http://hwbot.org/submission/3760543_

AIDA64 Disk Read Test Suite

 

It goes to show how weighted PM is Compare mine and yours and see I double almost all your scores with the TR but because single thread, 1/9th of the scores is the same, I only total 26745 instead of like 34,000+.

Edit; if it were truly equal weighted.

 

The front case fans are two huge 200mm fans, so the reservoir doesn't really block air flow, it is simply redirected. The air coming in from the radiator is not hot by any means, and is also exhausted. Like Fox, my case is full of vents and holes, so everything is pushed out the back. I've run 3dmark 11, and not noticed any temp change. The gpu is a rog strix so it has a massive heatsink, and three fans. I wish I could find a 1080 worth a **** in stock that isn't $800!! Stupid crypto miners have bought everything up, and driven the prices to idiotic levels.

Sent from my SM-G935T using Tapatalk

 

Well it's good that your GPU heatsink can tolerate whatever the case temperatures are with ease, so that's good. You say the air coming from the CPU radiator is not hot - well it doesn't really matter what the temperature of the air feels like - your water cooled CPU is outputting the same heat into your case that the air cooled CPU was doing, although if your CPU is operating at significantly lower temperatures due to water cooling then this can reduce power consumption a bit, which is a win.

 

On the topic of mechanical keyboards & key chatter. I was getting some key chatter, which is when you press the key once, and then you get it registered as two key presses. I used some free software to diagnose which keys were being affected, called Switch Hitter ( https://elitekeyboards.com/switchhitter.php). I used a keycap puller to remove all the keys from my keyboard, then used can of compressed air to clean the keyboard generally, as well as blowing compressed air into the affected keys, I also put a couple of drops of isopropyl alchohol into each of the affected keys & worked them - after leaving to dry I've cured most of the problem keys. Tomorrow I'll have another go at the affected keys that have resisted my initial fixing!

I've got a "Thermaltake POSEIDON Z Brown Switches Mechanical Keyboard", but I heard that key chatter will affect all mechanical keyboards at some point - I've had mine for 1 yr 2 months, and it's just started to show those issues - anyone else had similar problems and their solutions to the problem/fixing? Who's had a rock solid mechanical keyboard with no issues for years of use, and what make/model is it? I may purchase a new keyboard if this one resists my attempts at fixing.

**EDIT: Because this is off-topic, apart from being desktop related, I've moved this to a new thread, so feel free to respond there if you like! http://forum.notebookreview.com/threads/mechancial-keyboards-key-chatter-malfunction.812824/

 

Just ran Firestrike to test Graphics temp, and the fans didnt even spin up to 20%, and top temp was 57c? Do I need to run firestrike extreme? I ran at stock cpu maxed at 36c.
I will run Oced, and use the extremes test, and see what temps are.
Physics score of 19200 Stock clock

 

I was interested in a before & after regarding GPU temperatures since you moved to liquid cooled CPU, but your temperatures are obviously pretty darn good anyway on your GPU. To test your GPU properly you need to run a loop of Graphics Test 1 of Firestrike Extreme for about 40mins - this will show you max stabalised temperature of the GPU. (If you buy Firestrike rather than use the trial version, then you'll be able to select an infinite loop of Graphics Test 1).

EDIT: @Raiderman , you could use a infinite loop of Unigine Valley at max settings & max AA instead (to aim to get frame rate of about 60fps) if you haven't bought Firestrike. Probably not as hot for your GPU, but will allow you to loop it for 40mins or however long you want).

 

Yes, I have it purchased. I will test in a minute. Thanks

Sent from my SM-G935T using Tapatalk

 

Well I am not going to run for 40 minutes...lol. Max temps overclocked are cpu 51c, gpu 69c fans never got above 40%. I also ran a physics test.
I hope Ryzen 2 will have more OC headroom, because temps are not a problem with this chip. Not quite sure how to get beyond 4.1 yet, as everything I have tried has failed above that. That is only 300 mhz above turbo. Hopefully as it matures it will have better overclocking, similar to Phenom, and Phenom 2.

 

Temperatures are ok, but a 15 minute run is not steady state; in my testing 40min is steady state for my GPU in my system.

 

Crap, I dont want to kill my gpu yet Have you seen prices for them right now?? I want a 1080, and should have pulled the trigger when they were $589 on Newegg, now you cant find one in stock below $700. I should have bought a 1080 when I built the system, but oh well.
Even My Rx580 is selling for $600...lol

 

Ha, you won't kill your GPU by running a Firestrike loop on it! ;-) Here's a proper stress test, I normally just run Firestrike loop, but I tried running 4 threads of Prime95 while running a Firestrike Graphics Test 1 loop for 40 mins, here are the results (2 screenshots to capture all of the HWInfo sensor data):




Both GPU and CPU at 71 degC maximum during that run, but in fairness that's a 360W system load as measured by a Killawatt wall meter, and during gaming max consumption is around 270W, so it's 90W above a normal gaming load. During that run, CPU at 96W average power consumption; GPU at 190W average power consumption.

In fact, previously I had said that CPU load doesn't increase GPU temperatures in my system, but this test proved that it does when compared to my historical data I have for GPU only testing using Firestrike loop - raises temperature of GPU by 2 degC when my CPU is also highly loaded. I've also seen here that CPU temperatures are also increased by the simultaneous GPU load, when compared to CPU only testing, again when compared to some historical data of mine. Temperatures when gaming are substantially less, GPU in the mid to high 60's, and CPU temperatures in the 40's & 50's.

EDIT: Ha, for some reason for my second pic in this post I ended up doing a screenshot of a screenshot in my haste to capture the information immediately after testing before the 'average' column started to become unrepresentative! The info is still there though.

 

I will try in a little while. Right now, I am creating a custom Windows 7 install disk with all the updates, and numerous new hardware drivers integrated. Going to be a large ISO, as I am integrating to all 4 indexes.

 

@Mr. Fox could you pm me, or tell me what I am doing wrong in some benchmarks. I submitted a wprime score @4.1 ghz, and I am last for Ryzen 1700x. Ive seen a sub 3 second score for the same speed with the same cpu, and slower ram??

 

PM sent.