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

Well call me a lie, because I just loaded up CP2077 and it's now using 14GB+.

@yrekabakery
Not sure I get what your saying or why the link to a 2080 Ti benchmark.
Just asked if it helped make the game run better was all.

Edit:
I also posted that I played BOCW on a 2 n 1. (HP Spectre x360 to be exact)

@Robbo99999
Not as an instead, but only for clarification.

 

The numbers they had were using the beta version of AB that shows per process VRAM allocation.

You need to scroll down to the bottom.

 

Got cha. I did the noob move and only looked at the top portion.

I also notice if I turn off DLSS I get this.

@tps3443
That's the gun you want my friend. Kills enemies through solid brick walls and pretty much one shot kills everyone if you add the mods to it.

 

https://community.hwbot.org/topic/2...-list-for-2021/page/4/#elControls_578041_menu

Edit.
And yeah, I couldn't grab the new golds if I used the latest and greatest OS versions from Microsoft

The newer Win 10 versions on the old with weak hardware is a disaster for benching.


The old with ended support from Microsoft.

https://hwbot.org/submission/4643252_papusan_superpi___32m_core_i5_5200u_11min_51sec_953ms/

And some still think their old machines from the beginning of this decade get an performance boost for each new screwed up Win 10 feature update, LOOL

Edit. Happy New Year for all



Some changes and more golds for @Prema Team



https://hwbot.org/hardware/processor/core_i5_5200u/

 

There, I fixed that typo for you. All versions of W10 are a screwed up POS, but the newer versions do suck a lot more. Nothing is good about W10... garbage. The only thing nice I can say about it is that it is better than Windows ME or Windows 8.X but that's not saying much. Any number multiplied by zero is zero.

 

So I ordered and installed the Cooler Master "Cyro" AIO cooler and so far I can't recommend it.

If you don't turn on the TEC/Cyro cooler then it will heat rapidly. It managed to crash my perfectly stable 5Ghz OC that I haven't had crash a single time since 10900K came out. The CPU almost immediately went red in CB23 followed by a crash to protect itself. Great, lets turn on the TEC cooler features and see what happens then shall we.

Turned on the TEC cooler and it immediately cooled down the CPU to around 10c idle temps in "Cryo" mode which is the recommended and daily safe operation mode. It heats my 10900K to about 65C at just 5Ghz. My h115i platinum 280mm AIO would run about 60C under same conditions so I am confused how this is supposed to be better considering the damn thing was draw well north of 100w just to run the cooler.

At low CPU loads like gaming it does run cooler around 35-45c at 5Ghz, but my h115i did around 45-55c so not really a huge improvement. When you consider it doesn't offer any higher headroom and the price, avoid. Unless I discover a glaring installation error this thing is headed back to Newegg.

Edit: I should add there is an "unregulated mode" that greats you with a warning that this mode can cause damage as it has the ability to cool below dewpoint and can cause condensation and we all know that isn't good. I did try the mode and it does cool the CPU to 0c (actually it's below 0c, but the temp can't be read accurately at 0 and below) but I don't think you can run this mode at all times.

 

Dang, that is worse than I expected. I'm glad I cancelled the EK Cryo waterblock thingamajig. I got to thinking about it and it didn't make good sense. It's not big enough. It would need to be at least 3 or 4 times larger to work as well as a water chiller because the part that does the cooling can only make a tiny amount of water cold at any given time. Then the radiator will just turn right around heat it back up to at least ambient temperatures almost immediately.

 

Run it in that mode and boot into windows. Lower your cpu clock speed to 5.0 ghz once in windows and leave everything else as is. Open hwinfo sensor page, click the gear, put a checkmark in "Accept negative temperatures" click ok. Your motherboard will also show negative temps.

That means -11. You can also set the debug to show this as well. Not the negative part but it goes back to zero and then starts counting up. You know this by watching the oled screen.

Wait a bit to see if it starts to go negative while at idle. Once done looking at that, open up cpuz and run it's stress benchmark. See how high the cores go. Open your over clocking software and go up 1 multiplier and watch and see if it keeps running. wait for a bit then do it again. The reason for this is to see if it will hold your cpu below 50C as you move up to your max over clock. if at anytime you start to over run the cooler, then stop test and let it cool back or go ahead and change the mode back to normal.

 

Sounds like it is great for benching with unregulated mode but for day to day clocks it is equal at best and subpar when pushed with much more fuss and muss.

Maybe try some big benches with it in unregulated mode to see how that works, but overall for the price erformance in day to day, it seems it may be a big, steaming mug of "meh."

 

Thanks will go give this a try.

Before I headed to bed last night I did some more testing and adjusted the hoses off the cooler a bit to see if anything was causing an issue. But I think I'm just learning that this cooler is great for short and light burst loads like gaming. What it does allow for is high clockspeed are pretty insanely low temps. I compared to one of my older Battlefield V videos and the temperature was actually about halved. I'm not sure how I misremembered that, but feeling better a bit better after discovering that. I will continue to test this cooler over the next week or so before making any decisions.

https://imgur.com/a/EutPA5C

Battlefield V Ultra Multiplayer Conquest @ 4K 160Hz. This is in the standard Cryo mode.

 

In case anyone that installs an Ampere GPU starts having an annoyingly loud and rapid 4-beep BIOS code when UEFI with CSM or Legacy BIOS is enabled, this is normal. It's not indicative of an issue with your mobo or GPU. It is an annoyance related to Ampere and not a mainboard or GPU defect. It does not affect functionality.

Beeps with 3090 FTW | EVGA Forums and TUF RTX 3090 4 RAPID BEEPS ON BIOS STARTUP | ROG Forums

Placing a piece of electrical tape over the hole in the motherboard's speaker dramatically reduced the magnitude of the beeping. It is still audible, but very faint and no longer annoying.

 

Win 8.1 is much better for benching on the old than the bloated Win 10. You'll never be able to remove everything from Win 10 to make it run better than it's predecessor. But Win 8.1 is only tasty for benching and nothhing more.

Win 7...


https://hwbot.org/submission/4643318_papusan_cinebench___2003_core_i5_5200u_1467_points/


https://hwbot.org/submission/2995499_pasatoiutd_wprime___32m_core_i5_5200u_16sec_552ms/


https://hwbot.org/submission/4644286_papusan_wprime___32m_core_i5_5200u_16sec_466ms/


I will probably take most of the golds for https://hwbot.org/hardware/processor/core_i5_5200u/ with Win 8.1

 

Escapee World Champ. LOL

 

Bah, I can top that.... (Loads up GTX 1060 6GB)....

Seriously though, 2700+ on core? Sheesh.

 

This dude just worked with the non 360 version and wasn't doing backflips over it and definitely doesn't recommend it I hate myself for watching his videos, but I can't help it.

 

That was really hard to watch.

 

Should add in this one as well He is number 1 right at this moment and broke port Royal WR on youtube since some of his scores were removed and one account ban at FM / UL.


"he is trolling himself with a youtube video of someone running Port Royal lol"

 

Still don't know if it's H20 vs LN2 aka Slinky aka Slinky-pc... As the rumor mill goes...

 

HaHa, We will see.

Win 8.1 vs. Win 10

https://hwbot.org/submission/4646165_papusan_pcmark_vantage_core_i5_5200u_14909_marks

 

Honestly all of his findings were exactly mine. I find it funny that my 360 AIO performs nearly identically to his standalone block and the cable management is actually good on the AIO setup. It also looks really badass IMO.

It's great for gaming and light loads, but terrible for anything heavy. It also is completely useless as a cooler in standby mode for really anything but the lightest of loads. The biggest issue of the TEC is that it doesn't run until you can quickly boot and select the TEC to Cryo mode, it gets stupid hot and then can lead to instability if heavily overclocked. I don't understand why they would design it this way or not give you the option to set the TEC to a marginally powered state upon boot. My guess is that the software is required to read the sensors to drive the algo for the TEC. Still I think they could have been safe with a lighter TEC power to allow for overclocked boots. All that said I have no issues getting into Windows with my 5Ghz all core and 1.22v with LLC5 which is my daily driver clock. This lead stupid low temps for Warzone and Battlefield V.

Unfortunately my pump has a really annoying and loud whine and for that reason alone it's going back to Newegg as soon as they approve my RMA.

It's a cool concept, and it really does cool off gaming loads, but it needs more time in the oven. I think by gen 2 or 3 this will be a great product and I hope they continue to work on it.

 

After taking a closer look at this...It's your memory. Somehow it's not running optimal so more than likely you may need to change something there. Once I fixed my memory, my scores went up.

 

Bingo. Not sure why the 3090 changed this, but it did. All of my same RAM and CPU settings were getting me into the 27K~ range on the physics tests with the 2080 Ti. Thanks for suggesting that. I'm not sure why the GPU changed this, but is obviously did. After slowing down the RAM clock and tightening the timings I am back to 27K~ at only 5.4GHz.

Are you having problems with the SystemInfo crashing when 3DMark and 3DMark 11 are launched? I'm having nothing but headaches with that. Every time I get a score that makes me happy, that piece of crap stops working and the benchmark is invalid with neither my CPU nor GPU identified. Both show as "unknown" every time.

 

So sick of buggy Futuremark sh*tware... https://www.3dmark.com/3dm11/14193762 < "unknown" CPU and GPU... loser bastards

 

I know right? He's like watching Morton Downey Jr. back in the day (for us older folk in here). I can't look away.

I'm glad you took a chance so we could see. Living vicariously through your purchases is great.

 

I had the opportunity to grab a RTX3080 for a really good deal. So I almost did, I spent hours comparing it to my current setup. And I came to the conclusion of it not being worth it lol.

So I ran this test to compare Ampere to Turing architecture. Control does extremely well on Ampere so this example would be the absolute best case scenario with a RTX3080 at 2560x1440P. In other words, Control usually shows the largest gap in performance between Turing and Ampere.

And! This is OVERCLOCKED VS. OVERCLOCKED!


I thought my “Control” performance was spectacular. So I started investigating it a little. So I found a video on YouTube of a overclocked RTX3080 running 2560x1440P resolution, I paused the video and did my best to replicate the scene to where I was at. And here are the results!


RTX3080 @2,085Mhz Core/ 20,204Mhz memory overclock. "CONTROL" 106FPS W/ 9900K @5.1Ghz


RTX2080Ti @ 2,115Mhz Core/ 16,222Mhz memory overclock "CONTROL" 91FPS W/ 7980XE @ 4.8Ghz


Both settings are mirrored identically, resolution, DLSS, scene, even GPU usage is solid on both, everything!! The overclocked RTX3080 at 2,085Mhz is 16.5% faster than my 2080Ti at 2,115mhz


Hope this helps!


Please reply away, I would love to hear everyone's thoughts on this.

 

Finally making some progress...

https://hwbot.org/submission/464676...geforce_rtx_3090_52045_marks?recalculate=true

https://www.3dmark.com/hall-of-fame...core+performance+preset/version+1.0.132/1+gpu
All 3 of the top 3 single GPU are on Team @Prema now, LOL...

 

Weird! I'm wondering & have wondered briefly in the past if these large Ampere GPUs are somehow more taxing on the CPU regardless of framerate in the game, so instead some additional overhead associated with the NVidia driver. I think I remember seeing some graphs where the big Ampere GPUs were underperforming some previous gen GPUs in CPU limited conditions or vs smaller Ampere GPUs. I'm very vague on this, but I do remember thinking this when I read & saw some performance results in some reviews. Since you have 3090 and if it is placing more strain on your CPU, then perhaps there's some RAM configurations that enhance or mitigate that "platform limitation". Vague I know, but a hunch.

 

Mr. Fox`s 3DMark - Time Spy score: 22352 marks with a GeForce RTX 3090 (hwbot.org)

 

Depends on what a good deal it was. You may have been able to sell your card and come out with a free upgrade.

 

For me it's still the memory that will flake out 3d11 at times....

 

I suspect Brother @temp00876 ( sweepersc) is getting such great scores because of that really high end memory he is running. It is still kind of puzzling that the 3090 changed my physics scores so dramatically and now it requires totally different memory settings to be almost as high as before. I guess that's just how the GPU interacts with the system. Weird. But, thanks for the tip. I never would have thought to mess with the memory because it was so dialed in (for a 2080 Ti, but not a 3090).

Speaking of 3090... I moved the fans to the opposite side of the radiator and added a short dual PWM extension and now the stupid fan cables are not nearly as wonky. I moved the fans to suck cold air in from the open window instead of blowing air toward the window. I also made a small base with rubber feet to keep the radiator from sliding around on my desk.

I sure wish we had a nice waterblock available now. I am still tempted to chop off the radiator and add QD3 fittings to integrate it into the loop. No telling how long we're going to be waiting for a waterblock. Probably at least another 60-90 days. I am hoping I can avoid having to settle for a Hydro Copper.

 

Indeed! He does have some very nice scores across all boards.

It took me a minute to understand it. I was like what is different, then I checked my memory. I could do it with 4 sticks, but not with 2. Then it hit me...The timings must be different and sure enough, that's what it was. And then I tried Timespy and sure enough it went up as well.

I did that about the 3rd day, as I don't have a case and that was very awkward turning the radiator around to face the window. Not to mention I caught my finger in one of them.

I'm gonna hold on the chopping of the line as it's not really that many people above us once you get down to it. And you still need a full water block so you can go higher on the memory. Not sure how that is going to fair with the memory on the back though, but with a chiller....It soaks through the board thus allowing it to run similar to the front. Speculations of course.

 

Not worth the hassle of removing my card from my custom loop, then water blocking the new card and putting that one in the loop.


The card was $800.00. So I would still come out of pocket for sure after selling my 2080Ti

I considered it though. I really did.


I think the 3090 is much better. I just wish pricing and inventory would normalize.

 

LOL, I did the same thing almost immediately. I stopped what I was doing and got three wire fan guards from my hoard of miscellaneous computer stuff that I can't bring myself to throw away and put those suckers on right then. (Glad I didn't toss 'em out already in this case.) Fortunately for my finger, the fans don't seem to turn very fast. It didn't feel great, but it didn't break the skin. I was actually more concerned about something else hitting them and breaking one of the blades.

On the subject of not spinning very fast, it doesn't seem there is a way to increase the fan speed on the radiator fans. The fan controls don't seem to affect them. Only the GPU power fan seems to be controllable. I am thinking about running a line to the fans from one of the empty channels on my Lamptron CF525 and see if I can make them run faster if I give them more juice. I don't know if they will push more air if I can make them run faster or not, but I am thinking it might be worth checking out.

It is pretty unfortunate that the AIO hybrid setup is not modular. For the price we paid, it should be made a whole lot better. Some QDCs or removable compression fittings (with G1/4 threads at the radiator end) and a fill port, at least, should have been standard on this setup. It's disappointing that it's made like a cheap consumer AIO. The way it is made, it could be awkward to manage even with a big tower case to mount it in. We have the same flat Wetbench and I don't think I will ever want to go back to a normal case. Even the inexpensive aluminum frame open bench that the CPU binning rig is mounted to is better than an enclosed case. I actually kind of like that simple design even better that the Wetbench, except for the fact that there is not much way to handle peripheral devices. It would be perfect for messing with LN2 or DICE.

 

@Mr. Fox

I don’t know why they don’t just sell a bare PCB Kingpin with a waterblock voucher, once available down the line.

But it looks like 75% of the Kingpin card sales cater to the “Average cool kid overclockers” and they probably think the card is extreme out of the box as it is lol. Technically it is I guess

But in all honestly, I think it’s a lot of added cost. And a (360) rad, fans, pump, tubing is probably more expensive that just a GPU block.

 

Or better yet, why doesn’t EVGA just sell the Kingpin as a Waterblocked card only??? I don’t get the whole AIO thing. This is a $2,300 GPU! Why are the pushing a AIO with this thing??! Haha.


Is this like, ohh fantastic! I can buy a king pin and run it with my stock Ryzen wraith cooler lol. No custom loop needed!!!

The Kingpin is a serious GPU. And that 2080Ti Kingpin waterblock fiasco what a mess! Those waterblocks were selling for $750 on eBay alone. Because it was the only block that actually fit the card. But still that’s crazy. I guess there will be a repeat of this.

 

Yeah, I don't know. They could have used an air cooler as well, but I think it runs cooler with the AIO than it ever could with air. I suspect the reason they didn't offer a Hydro Copper version is because the vendor that builds the Hydro Copper blocks probably didn't have them ready in time.

They usually do. They also sell the FTW3 and SC GPUs with a Hydro Copper option. I think the issue is what I mentioned... most likely the company that makes them probably doesn't have them ready yet.

I have my 2080 Ti FTW3 Hydro Copper block and the brand new EK Quantum Vector block and backplate for sale, hoping they sell before a block is available for the 3090 KPE. Maybe between the two of them I can offset the cost of that. They're getting a lot of views so far, but no offers.

 

@Mr. Fox how much for the 2080Ti’s? My older brother needs one. He just bought a 10900K, board, and memory. He can’t find a GPU at all.


Please PM me, and let me know. Only need one. Thank you!

 

For the blocks? I wasn't planning to sell the GPU. I am using it in the other desktop and put the hybrid cooling setup back on it. But, if he makes me a really sweet offer I can't refuse I would certainly consider it. Remember, it's shunt-modded and overclocks like a spotted-assed ape, so it's not a gamer-boy GPU. I'd kind of hate to part with it... I have a soft-spot in my heart for that bad boy, LOL. But, I can put the 780 Ti Classified back on that setup.

 

I didn't do anything with the card the first few days as I was waiting on my memory to show up. I was still using my 2080 TI playing CP2077.

I already tried this and it doesn't work as the max speed of the fans can't be increased beyond it's limits.

That would have been nice, but I think it may have fell under a liability issue. True, quite a few folks know what they are doing, but it comes down to the few that don't. Catch a leak burn some parts and it falls under faulty parts when it was 100 percent user error. they may do it in the future, but hard to say with all the complaints going on this very minute.

We are benchers so open benches are going to be the way to go, but most dont have the space for it and want their parts inclosed with all that is going on in their homes. Speculations of course.

I'm not sure I understand? Maybe I'm not seeing it, but who wants a card that isn't usable until a water block shows up? That may take 3 months to show up at that?

Most kingpin owners this round have to be elite members to get a card. Not talking about the elite members that sell to the cool kids at scalper prices.

And the card can be ran at stock up to LH, without doing 1 single hard mod to it. This is what makes the card appealing.

And if you are an EVGA fan, then 200 bucks really isn't that much considering the price of the Ultra, Hybrid and the Hydro Copper.
Of course I would rather pay 1K than 2K, that's for sure, but it is what it is. You either buy what you want or what you can afford. This has never changed....

 

It's a 1999.99 GPU my friend.

Edit: before taxes of course. Unless you are tax exempt.

 

Most of the Hydro Copper blocks and GPUs are selling on eBay for as much or more than they cost new. The Hydro Copper block for a Kingpin GPU wasn't out of scope with the price of the same block for other GPUs. It's made a little different, but not remarkably different. At the end of the day it is just a block of metal with plastic. The people on eBay are scalpers and they are selling an EOL product for more than top dollar because somebody is going to be dumb enough to buy it. Somebody that has more money than good sense and they want it regardless of the price because they can't get it new from EVGA anymore.

I believe there may have been another option for the 2080 Ti Kingpin block, but I am not 100% positive since I never owned one. If I have to settle for a Hydro Copper block for the 3090 Kingpin I will, but I definitely won't unless there is not another option available.

 

I don't know if I'd call $800 a deal. I guess with current market conditions sure, but overall....meh.

And yeah, the 3090 is a monster. I'm playing WoW @ 4k on a EVGA 1060 6GB and to say I'm missing 3090 4k @ Ultra 10 w/ RT on shadows max is an understatement. It was just financially well
out of my reach. As for the 1060....

Pros:

1060 6GB *IS* playable @ 4k settings 5 getting 60fps+ in many areas.

Cons:

Those dips into the mid 30's and sometimes high 20's
It's still Setting 5

I'm on the hunt for a 3060ti at this point. It will do WoW good enough and is faster than my old 5700xt and allow me to at least find a sweet spot with some RT enabled at an economical price.

 

Hi brother @Mr. Fox , @Johnksss , I was initially using the 16GB Patriot Viper 4400 single rank a couple of months ago until I ran into this video by Luumi


So I swapped them for a G.Skill Trident Z Royal 4000C16 (16GB x2) on my Apex XII. These perform really great but when switching to the 3090, running them at 4600 is unstable for me although it used to work for the 2080ti. The Win10 OS version used also contribute to the physics scores 162.99 (Win 1709) is also higher by 100-200pts at the same memory frequency but 162.99 no longer works when using the 3090.

 

Yes, 1709 doesn't work. You should go with Windows 10 LTSC 2019 for the best physics and combined performance from Windows 10. Isn't it really weird how changing the GPU changes the RAM behavior? I never would have thought so, but that's all three of us now.

 

LOL I switch to water chiller and lost 2k Points and the mem settings that worked for SS do not work for chiller. At the same exact speed.

 

Those are really fantastic modules and they are super-tempting, but (to me) they are so damned ugly-looking. I'd be temped to mask them off and rattle-can them with some flat-black Krylon or something. The silver and gold shiny heat sinks with the crystal plastic RGB diffuser is a major turn-off. I wonder if they make a 100% identical module with no RGB and just plain old black heat spreaders, or if they are cherry-picking the ICs for the Royal and you have to take what you get even if you hate how it looks? It looks like maybe the Trident Z F4-4266C17D-32GTZRB is the same module, but it still has RGB... so, not my favorite idea. I wonder how hard it would be to remove the heat spreaders and RGB and just run bare modules? I've never attempted to remove heat spreaders and I am not sure what all would be involved with removing the RGB parts.

 

I have the Royal z 4000 CL15 4x8 modules. And while I purchased this memory solely for pure performance. I thought they looked ok when I bought them. I bought them in September when they first came out, and I just don’t really like they way they look anymore. They literally clash with everything in my build. But I think you guys are talking about the dual rank versions.


Take them out of the system, sitting on the desk, and they look good. But once in the build it’s just kinda tacky. I’m not very good with building pretty systems either. I’m really bad at matching colors lol.


Either way, the memory is far better than my system will even let them be. Which is what everyone dreams of I suppose.

 

If it was only the gold or silver heat spreaders I would just paint them and be done with it, but the RGB thing really sucks. Especially with Windows 10 because it tries to install bloatware for the RGB whether you want it or not unless you are using a version like LTSC or a mod with Windows Store removed and Windows Update removed. I tried a couple of different RGB modules and RMA'd them for that very reason. Without bloatware installed you can't control the lighting and I am only interested in two options - white and none - but the kiddos have got to have their pretty Christmas tree lights.