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

I think I got it from Tech Power up. There is a stand alone version floating around. If you need it, let me know, I will post it on Google drive or something.

Edit: I got it from Guru3D, and I think it was taken down. I will upload it for you, and give you a link.

 

The standalone version is from saving the BIN folder after installing it in Windows 10. If you have it in Windows 10 you can drag and drop a copy of that folder elsewhere to make your own "portable" version. @Ultra Male uploaded his in the Cinebench 20 thread. It's pretty deplorable that the dullards at Maxon agreed to only have it hosted on the Micro$lop store, especially since their own web page says it is for Windows 7 forward. I've lost a lot of respect for them because they made such an absolutely retarded business decision.

 

4.3Ghz on Ryzen, that's quite impressive I think? What's the max that people have been able to get on water, and what is the most common max on water?

 

I sent you a link in your inbox, if you need it.
@Robbo99999 I think its fairly common to reach that on water. With AMD cool and quiet disabled it runs at 4.25. 4.4 and above are kind of rare, Not alot of headroom which isnt as much fun, but hoping for better things in the future.

 

Well, that's good still, I mean you say above 4.4Ghz is rare and you're at 4.33Ghz - in reality there's not any significant performance difference there.

 

Got it. Thank you!

 

I have emailed Maxon support regarding their so called Supported Systems. I will post the response back if I'll get it



Meet the new, tougher Cinebench R20 benchmark: We test it on Xeon and Threadripper Pcworld.com

From the link...

Windows Store only
One thing that’s certain to raise some blood pressure is how you can—and can’t—get Cinebench. Previously, anyone could download Cinebench R15 and install it as a standalone Win32 app. With Cinebench R20, the free benchmark can only be downloaded directly through the Windows Store.

You don’t have to log in or create a Windows Store account to install it. But you do need Internet access, and you don’t have any control over versions. If Cinebench R20.1 came out in six months, it could invalidate your previous tests, because you can’t run the older version.​

We asked Maxon officials about a standalone installer, and the answer wasn’t encouraging. “It’s doubtful Maxon will be releasing a standalone Win32 installer version at this time. Cineware is a derivative of Maxon’s flagship product Cinema 4D, used to create ‘high-end’ 3D animation,” Maxon officials told PCWorld. “Cineware is meant to test the hardware and OS artists will actually be using in production. Creatives are constantly pushing the boundaries of hardware and OS performance. Therefore, Cinebench too must push those boundaries in order to provide accurate comparisons.”

They gave the middle finger to the benchers!!!

 

This is why I'm glad I got the stand alone before taken down and stashed it multiple places. But this mentality is horrible. No one wants anything from the Microsoft store EVER!

Sent from my SM-G900P using Tapatalk

 

I emailed them asking the same question about where I could download it for Windows 7 and never got a response. Great way to handle product support. Maybe they only respond after they take your money for a product. Or, maybe not even then.

 

How funny I had emailed them the same thing a few days ago... no response.

 

Very first Cinebench R20 run with Linux on the Desktop...



@Mr. Fox this should be interesting with some tweaks..

 

How did you get it to run? I haven't figured out how to make it work on Linux so far.

 

I did a clean install of linux onto its own dedicated drive, installed wine via command line and not manually with the deb file ... It's still a little wonky, but once I get it dialed in I'll let you know... Once it's stable this could be really exciting in terms of a big leap to total migration. ie... ditching M$. I'll dig up some goodies related to the advancement of gaming support that's very exciting as well. Just got home from a trip and pretty beat .. I'll dive into this a lot more tomorrow for sure.

 

New GPU-accelerated, but hardware and API agnostic, ray tracing demo by Crytek called NEON NOIR. Would be sweet to have it as a benchmark! We'll hear more at GDC. Hope to see this tech trickle down to Amazon's Lumberyard and Star Citizen.

https://www.tomshardware.com/news/crytek-real-time-ray-tracing-amd,38837.html

 

That demo was run on a Vega 56. Between that and DX12 being ported to Win 7 for some titles (it is a title by title implementation is my understanding), and with DX machine learning and DXR coming with upcoming Windows releases, we should see more and more like this (although mileage will vary depending on hardware, obviously).

In three days (March 20th), we will hear not just about Zen 2 in more depth (noon event or around there, I believe pacific time), AMD will be doing a similar optimization event for their GPUs that evening at either 5 or 6PM. We should see Nvidia having some good stuff as well, and potentially watching out for some of the Intel presentations (hoping they give more about their dGPU around then, even though a year out).

I'm loving the competition coming back slowly on the GPU side (even though next year is when it really hits) and CPU side, along with RAM prices dropping and NVRAM dropping (SSDs). Either later this year or next year may be a darn good time for upgrading or building a system!

 

So, new news. Zen 2 will officially allow the Infinity Fabric ratio to be clocked 1:1 with the memory clock, instead of 1:2 like Zen and Zen+. That explains double the bandwidth, but if anyone wants to do the calculus to find the limit on reduction of latency, let me know so I can give data points for them to calculate the latency curve as speed of the fabric is increased.

 

Cinebench R11.5 run on the 9900K @ 5.3GHz (non delidded one)...

25.48 points!



Hope everyone is having a great Sunday! Cheers....

 

delid that thing already!

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

 

I have another 9900K Delidded on the tarmac waiting orders to take off as soon as this one is exhausted from its dog fight.

 

Thats next on the shopping list! BTW, I suck at math.

 

I think I'm going to sell my current GPU's for the KP cards.

 

I know right? ...I saw that earlier and drools... First World Dilemma.

Here's a new Cinebench R20 PR -- 5572cb @ 5.3GHz



Cheers

 

9900K / 2080Ti

Fire Strike: 31461


Highest Graphics Score so far has been 40300+ on other runs.. this just happened to be the highest overall PR.

 

Okay. That's what I'm going to aim for!

 

Yeah, I saw this earlier today and immediately thought, "Yeah, I might have to return to the Nvidia fold."

I've been slowly building out a custom water loop because I want to position the reservoir to directly benefit from my central AC and switch to a remotely placed open bench config (if it was next to me I would go bonkers). Nothing would top it off better than this beauty.

 

Sounds like it could be a somewhat unique and interesting setup. Feel free to show it off here once/if you get it sorted.

 

got my new RAM kit in yesterday, woooooo see updated signature

did some quick and dirty playing around, managed to bump VDIMM to 1.40V without issues (previously on the 7700K and older PremaMOD even 1.35V was semi-stable), VCCSA and VCCIO to 1.30V also without issues. didnt go higher since i was just fooling around a bit. also managed to up CL to 22+, previously maxed out at 18. so nice to know that ive got more wiggle room (i know i know! who would WANT such abysmal timings, anyways, right?! just testing the waters here, people!)

so before my limited tweaking time was over, I managed to do some quick boot tests and for the first time ever i saw DDR4-3200 on my machine, yehaw!

It'll sure take me a while to tune in the higher RAM clocks, some of you might remember how long it took me to get 3000 Mhz stable on my Corsair kit

Btw, I'll likely offer the Corsair kit on the marketplace soon, so watch out for that

cheers!

 

Cool! What are you at now for timings on your 3200Mhz? What kind of Mhz do you reckon you can aim for, and how much timing loosenage can you live with?

 

LOL youre asking all the right questions, but unfortunately i didnt have enough time for tinkering yet! i barely got in two hours before going to bed last night, and wont have any time today, all that annoying real life in the way of my virtual pleasures!

those first few tests at 3200 Mhz was at 18-18-18-38, i basically just matched the XMP timings for 3800 and did a quick test what clocks would boot right up at the first try. unfortunately, there are a few secondary and tertiary timings that ill need to read out of the XMP profile first before i can input them manually.

naturally, just slapping in the sticks and loading XMP didnt amount to anything, wouldve been WAY too easy right? haha

what could I live with...... hmmmmmmm....

well my corsair sticks did 3000 Mhz at 16-18-18-39, so at 3800-18-18-18-38 we're talking about a 25% bump in raw bandwidth and latency. would be nice if i could get at least HALF of that stable, say 3400-3600 at acceptable timings

in the end, i fear that my motherboard traces will be the bottleneck in this! but lets see what i can squeeze out of them... it will sure take some time!

 

Sounds interesting, keep us posted! It's strange that laptop RAM seems to lag so far behind desktop RAM, even when we're talking about the same voltages - maybe it's those motherboard traces you talk about, or increased heat, but don't know how much increased heat affects RAM stability specifically. (I think my RAM in my desktop is in the 40 degC's when gaming). Maybe power supply cleanliness is the difference between laptop & desktop, is that what you mean about your motherboard traces? I know that on my MSI motherboard I remember the marketing blurb harping on about isolated memory traces (I think).

 

yeah, lots of factors putting laptop RAM at a disadvantage compared to desktop:

- small form factor = more heat (although this should be the smallest negative factor)
- small form factor = less voltage headroom (just compare SODIMM with DIMM voltages, ull see what i mean. this greatly puts a lag on RAM speeds, since we need to wait longer for more mature RAM chips to be capable of reaching higher speeds at low voltages)
- laptop motherboard = generally lower quality PCB when it comes to memory traces compared with desktop boards. the small form factor means you cannot put the RAM sockets as close to the CPU socket as on a desktop board (= longer traces negatively impact signal)
- laptop form factor = all components are way closer together, means more potential interference from other components on the signal going through the memory traces

taking these factors and looking at that AsRock MiniITX X299 mobo, that only leaves one single factor: motherboard quality namely in terms of trace lengths and potential interference.

the AsRock board is easily capable of running both the G.Skill 3800 and Corsair 4000 32 GB kits at DDR4-4000 at 19-19-19-38 or even at DDR4-4160 (G.Skill) and DDR4-4120 (Corsair) at 19-24-24-48 timings.

Of course, there are always differences in overclocking capability between different kits of the same brand and SKU, or even between the sticks inside a single kit. but were talking about 8 different sticks in two different kits from two different vendors. and they both reach very similar clocks and timings, only common factor being....wait for it...the motherboard and its traces!

 

Cinebench R11.5 run on the 9900K @ 5.3GHz (non delidded, yet)...

New PR... 25.53 points!

 

Can anyone tell how to save result for HWBOT submission. I saved this screenshot but when I go to submit it it's telling me invalid file

 

Hmm I see that you're version is 1.0.3, was that a beta version? Try using v1.0.2?

 

Thats odd. Its grayed out/

 

Doing some RAM tuning today.. hoping to post 4533MHz with decent timings... Fingers crossed.

In the meantime...

Fire Strike Extreme run: 18732

 

LOL sweet! what kinda sticks do you sport? somewhere in the range of at least 4200+ I hope

Oh wow, I just saw that DDR4-4800 sticks became available, dayum!

Link: https://geizhals.de/corsair-dominat....html?hloc=at&hloc=de&hloc=eu&hloc=pl&hloc=uk

900 EUR for 2x8GB at 1.50V VDIMM

 

I have the G.Skill Tridentz 4266MHz sticks ... 8GBx4 kit

I was tempted to get the 4600 set, but I had faith in these sticks to do ~4533

Wow those 4800 sticks..monster!

 

Wow it's faster than my car LOL, too bad it's not a 6-speed

Although I liked installing things on my car myself

 

New Fire Strike PR...Finally broke the 31(500's)!

9900K @ 5.4GHz / 2080Ti
Fire Strike: 31737

Max Graphics Score: 40466 on my other runs.

 

First test results on my G.Skill RAM kit are in I'm playing it safe with the timings thus far, wanna go up fast with the frequency before diving deeper into tuning of the timings. A RAM Timings calculator and SPD/XMP timings read with HWInfo are a great help to quickly dial in stable timings

TM5 stable thus far:
3000-17-17-17-39
3067-18-18-18-40
3100-18-18-18-40
3200-19-19-19-42

Next up: 3300-19-19-19-43

I think now is the time to check if upping VDIMM/VCCIO/VCCSA would cause any hiccups

wish me luck, biatches!

 

For VCCIO and SA, of possible, try keeping to 1.2 or less. 1.3 is toward the high end for air.

As for Vdimm, 1.5 is about as far as I'd push DDR4 for a daily driver, with 1.4-1.45V more preferable, so long as DIMM temp allows. For many DIMMs, errors will crop up more around 50C, mainly due to thermal radiation causing decay which can lead to bit flips and data loss.

But, some so-dimm MFRS use ram chips made for higher temps. If you have those, in theory you can go to the 70s or 80s.

Ultimately, it comes down to errors popping up on your set as to what to use and check.

Happy trails!

Sent from my SM-G900P using Tapatalk

 

New Time Spy PR...Finally broke the 16(100's)!

9900K / 2080Ti
Time Spy: 16146


-

Night Raid: 72841

 

So, posting this article here and the AMD thread. It is a Ryzen mem overclock guide.
https://www.techpowerup.com/reviews/AMD/Ryzen_Memory_Tweaking_Overclocking_Guide/

 

Good good, I'll stay tuned! Once you've worked out how high you can go with the frequency you might want to test different frequencies for real world performance to see if compromising on looser timings is causing performance problems - to work out the optimum config.

 

thx for your input guys and no worries, i got it all covered so far

SA/IO voltage at 1.25 for testing purposes, will dial it back down once i got my clocks n timings all set
as for VDIMM, ill start with 1.35V for the time being, since that is also the setting in the XMP profile.

good info on temp ranges @ajc9988 in any case, TM5 stock is only for quick stability testing, ill move to TM5 1usmus 10 cycles (single rank, thx for the tip on that btw!) and memtest86 overnight later on.

@Robbo99999 thats exactly the plan, you read my mind

ive found that getting a frequency to even boot with semistable timings is the most frustrating part of RAM OC. once u got it booting, dialing in the timings is much more time consuming, but at least u can go about it systematically. thats why im starting with clocks first at loose timings, determine what i can reach max on my mobo. then i tighten up the timings on the uppermost 2-3 or 4 frequency straps and check which one would give me the best bandwidth/latency package

btw, 3300 mhz booting woooo

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

 

Got a portable AC unit coming in so I can OC deep into the summer. While it is not a water chiller, the advantage is I can use it with my various rigs and their radiators. My basement is becoming crowded, from my new FT2 Functional Trainer and various weights/benches on one side of it, to my wall of rigs on the other. I think it is time to let some of the rigs go.