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

My EVGA RTX 2080 Ti XC Ultra and 1600w T2 PSU we’re both delivered this morning. Luckily I have a camera at the front door and live in a safe neighborhood so no real worries but still. I was a bit nervous having $1800 worth of computer parts just chilling in boxes on the front porch until the wife got home to retrieve them.

Will be home tomorrow night to install both and test it out. Let’s hope this one lasts longer than a week. If this card somehow fails I honestly will be at a loss of what to do. AMD has nothing worth buying, and I don’t want to step back to a 10 series card nor do I want to downgrade or sidegrade. I’m hoping I’ll be good to go now but I’ve had 2 fail but both were from early batches.

 

I will say a prayer for you and your new GPU. I agree with everything you said. I cannot imagine you would have luck so bad that this one would also die. All of the alternatives suck. (Not necessarily so with 1080 Ti, but every other option you mention truly sucks.) If things do go south, you can buy both of the Kingpin 1080 Ti cards that @Rage Set has for sale in the NBR Marketplace and have money left over. And, 1080 Ti SLI would be more powerful for most things. Just saying, that's an option. No ray tracing, but who knows when that is going to actually matter at the rate things are going.

It is sad that there are so few things to be excited about for PC enthusiasts now. Any notebook you look at now is either poorly engineered in one or more ways, or it is a castrated turdbook, or both; Zen/TR CPUs suck at overclocking; Windoze 10 keeps getting worse and it's an absolute abortion... ugly, insecure, unstable and bloated beyond reason; AMD doesn't have a GPU that is worth a damn and probably never will; Intel's CPUs are too expensive for what you get in return for them; and, now everyone is getting brutally raped on pricing of GeForce GPUs that die suddenly and for no apparent reason. I cannot remember ever living in a more troubled time for PC tech.

 

Battlefield V Ray Tracing enabled! Finally get to put this RTX 2080 Ti to good use!

https://twitter.com/EA_DICE/status/1062635899786276864?s=20

They fixed DX12. No more stutters and better FPS according to posts.

 

Cool, let us know how it performs, on 1080p too as I'm interested to see how much they improved from the original 1080p 2080ti ray tracing enabled 60fps target.

 

Personally I wouldn't even game with Real time RT because when you play BF 5, looking at mouth watering VFX will get you killed! I feel RT will be fixed and not dynamic for wide range of objects and materials.
I hope you get the best experience and less downtimes due to hardware issues.

 

Right after I ordered an RTX 2080 then the news about RTX 2080 Ti catch fire and QA failure hit the ****...Dang it...

 

1.7K USD for 2080TI, and 1080ti is below 800 USD at this moment...

I'll be upgrading to 2080Ti when the 3080 or such comes out

 

I agree. The 9980XE is literally a relabel of my 7980XE with crap-implemented solder. Current GPU's are a non-starter (I ended up only getting the Titan V and canceled my 2080 TI purchase) and it isn't due to the lackluster QA either. I don't believe Nvidia is done with the higher end RTX cards. Why use 12nm tech when 7nm is available.

 

Yes, pretty disenchanting scenario right now. And, I know this is new stuff, but if ray tracing is going to be the performance killer it is shaping up to be with an almost laughable degree of graphic enhancements, I say who gives a flying rat's ass. I literally cannot see any difference in most of the examples provided, and where I can I have to pause and study the differences, paying really pay close attention and zooming in on the images. I would literally never notice anything in gameplay. Just a worthless gimmick if this is what we can actually "look forward" to. It might get better over time, but right now it looks like an absolute joke. Hopefully, it's not going to be the functional equivalent of NVIDIA GameWorks CrapWorks 2.0.

Battlefield V DXR Real-Time Ray Tracing Performance Tested | Finally a Look at the DXR / RTX Performance (Drop)

 

Sorry to be late to the party, but you under no circumstances try scrubbing the cpu pins. That is a big 10-4 negative! You should use something like Super wash or isopropyl high pressure sprays. Or my favorite, Birchwood Casey Gun Scrubber Cleaner. Man they use to get a crack out of what I was using it for.

 

Just in case anyone doesn't know...
10-4 is trucker lingo. Really made popular with the Smokey and the bandit movies.

 

I realy miss the old school clean humor without all the potty-mouth nonsense. They don't make them like they used to.

This clip is a classic. It's still funny all these years later. LOL

 

@Mr. Fox

This is true!

 

That <strike>Sum b***ch</strike>Scum Bum!

 

Pretty mild by today's potty mouth standards, but it was pretty edgy back in the day. The uncouth filth we hear in casual conversation today was reserved for things like accidentally smacking the daylights out of your thumb with a hammer, having your car vandalized, etc. But, now it's either viewed as normal or even humorous for some reason.

 

A little more RAM tweaking this evening. Have a nice weekend.

 

OOOPS. Forgot to include this one... slightly looser timings seems to get better read speeds, but a slight increase in latency.

 

have u done multiple runs and averaged values to counteract the usual fluctuations? i usually do 10 consecutive runs when checking ram speed and latency

 

Which DRAM IC does the Corsair vengeance use SK Hynix or Micron or Samsung?
Isn't tRFC quite high on AIDA64 memory benchmark? How did you manage to go around 350(tRFC) at those insane XMP profile of 4000MHz?

 

Here is my timing, very tight for this type of setup, an notebook of course, and at low memory voltage, ppl are holding world records with their overvolted memory, maybe I should start a new league for overclocking, best record at stock voltage

 

Yes. There is very little variance between runs. Enough to measure in a benchmark, but too small to influence real-world performance.

Samsung B-Die. tREFI should be set to the maximum capacity of your installed RAM for the best results. This allows the full capacity of the memory to be filled before a refresh. The tRFC can actually go lower, but it provides minimal benefit unless you decrease CL further. I can actually run 4000 CL14 with tRFC like around 250, but that takes 1.665V and I don't want to run it that way all of the time. Some of my recent benchmarks were run that way. The current settings require 1.500V to avoid memory errors, and I am comfortable running 1.500V. That used to be "normal" voltage for DDR3, so I'm pretty confident it's not going to hurt anything. (Some high end enthusiast DDR4 modules run 1.500V on their XMP profiles.)

Why not just bump up the voltage? Are you able to, or is there no setting available to change it?

---

Finally got sick of the dust build-up and cleaned my office again. There is definitely no shortage of dust in Phoenix. It will be dirty again by tomorrow, but seeing it clean again made me happy enough to snap a few photos to capture the rare moment of a dust-free man cave, LOL.

 

Here are some examples of CL14 benchmarks that I previously posted. If you viewed them without looking at the images full size it might have been missed. The tRFC was 276 (not 250).

Spoiler: Previous Benchmarks

 

Same settings seem to be working decent at 4200 as well.

http://hwbot.org/submission/3985792_


Edit: Windoze OS X doesn't like them at all. Windows 7 didn't miss a beat, but Windoze OS X is not stable using them. What a piece of crap OS, LOL.

 

You got your wires crossed here briefly Mr Fox, you've confused tREFI and tRFC, you yourself told me that tREFI should be set to max RAM capacity (that JayZtwocents video), I think the cleaning products got to your brain! ;-) tREFI performs better when set to a larger number, tRFC performs better at a smaller number.

 

Now you confused me.

 

Ha, don't know if you're joking or not, but tREFI is the refresh interval, so a larger number for tREFI means that the time between refreshes is longer, which means that the RAM can spend more time doing something useful. tRFC is the length of time the actual refresh process takes, so a smaller number here means the RAM can spend more time doing something useful. RAM has to be 'refreshed' because it's volatile storage and has to get this regular periodic 'refreshing' to remember whether it's a 1 or a 0 for each bit - ie it has to 'refresh' in order to not forget the data stored in RAM. So, tREFI and tRFC are just housekeeping tasks that the RAM has to do in order to not corrupt the data stored in it's RAM - by setting tREFI & tRFC to their optimum, then you're just minimising the amount of housekeeping that needs to be done, allowing more time for the RAM to be used for Read & Write = greater performance.

 

Yes, sorry for the typo. I fixed it.

 

http://hwbot.org/submission/3986952_

 

I switched my 1080Ti FTW3 for a K|NGP|N - now to retune the CPU and RAM and get the GPU dialed in (these results were a quick and dirty overclock to ensure the card was working, currently it's doing 2115 on the core and 12Ghz on the vRAM without touching the power limit).

 

https://valid.x86.fr/bench/yvc4hb/16

 

I want that chip back!

 

Yeah you'll get it once he gets new Intel CPUs and RTX cards.

 

Welp... Happy Thanksgiving, Fox... LOL.

Spoiler: Dia de los Muertos

I had been out most of the day running errands. Dropped off a box of computer parts at the post office for Brother @tanzmeister. Got home and decided to jump on the computer. Pressed the power button and instant flames. Immediately yanked the power cord. Not a great Thanksgiving present, but I am very thankful I did not leave it on. I might have come home to a burned up house instead of a burned up motherboard. So, definitely, thank you Lord for that.

Hopefully, nothing else was damaged. Probably not.









@Donald@HIDevolution

 

dafuq???

 

Parts can be replaced bro Fox. Not the life. It sucks, but happy this went well and not started a fire.
Seems you need a replacement. Maybe not as powerful, but finally you can find BGA laptop with the more correct prices label But will you support Dell... That's the big Q

Spoiler: $129 Dellbook

 

Yeah, very odd. Everything was working great. I was running round for about 4 hours, came back and pressed the power button and *boom* it fried. I wish I knew what circuit that is, but even the underside of the PCB is fried. It was actually shooting out about a 6-inch flame and sparks. Crazy.

Yes, stuff is just stuff. No loss of life or home is something to be very thankful for.

Literally the first thing out of the ASUS rep's mouth was "warranty does not cover physical damage" and I am like "Huh, you're kidding me, right???" After I explained that physical damage did not cause the problem and any physical damage it has now is a direct result of the failure, he repeated himself that they would not be responsible for any physical damage. So, I asked for a manager. Manager took my info and said someone from engineering would call me to ask more questions. He gave me a case number and said they would have to determine whether or not it is covered under warranty once they receive it. So, I'm not impressed even a little bit. They seem predisposed to washing their hands of any responsibility.

Yes, a replacement is inevitable. https://www.amazon.com/gp/product/B079WCBP18

 

That's mad, they've gotta cover that, I mean you can't just have products bursting into flames on their own, if they don't cover it they don't acknowledge the problem & worst case scenario they're putting customers lives in danger by ignoring such things - it's unthinkable behaviour for a company, negligent, and I suppose feels like a lawsuit!

 

Maybe the board was really bad and it was your sheer luck that nothing disastrous happened.

 

I guess we will see how it goes. I also don't know how they can not cover it under warranty. It is inconceivable to even think they might go down that road. Hopefully, none of the other components were harmed. I doubt it, but I guess I will find out when the EVGA X299 Dark gets here on Saturday. Based on my experience with them blowing me off when the Intel ME firmware update cancer causing a memory performance impairment and their refusal to downgrade the ME firmware to make me happy, I'm definitely expremely unimpressed with their service at this point. Offering no Advanced RMA/cross-ship options on their flagship motherboards like EVGA does also sucks. After seeing their reaction today (or lack thereof) that's why I ordered the EVGA board. If they send me a replacement, then I will sell it unless the X299 Dark doesn't overclock as well. (Hopefully, it won't turn out like the mediocre ASROCK Chai Tea.) If the X299 Dark is as good or better, then I'll probably be done with ASUS due to the lackluster service.

That's the part that has me scratching my head. It had worked flawlessly and wasn't having any issues that I am aware of. But, electrical parts usually die suddenly and unexpectedly. It reminded me of when my M18xR2 motherboard caught on fire when the Soundblaster Audio amp chip blew out. That was also sudden and unexpected. It was working great, no issues whatsoever, then I pressed the power button a *poof* it caught fire. That was a replacement motherboard that cost me nothing. It was only like around 90 days old and it caught fire about 30 days after my 5-year extended warranty expired.

 

Looks promising...

EVGA X299 DARK Motherboard Thermal Review: Someone Finally Gets It | GamersNexus
Awards: Best of Bench, Editor's Choice

 

yep i was thinking about those exact GN clips when u mentioned ordering the EVGA Dark mobo

I think "Something that isn't garbage" and "Ridiculous overkill" is a good match for your hardware tastes Fox

 

Yeah, that in-BIOS overclocking stress test utility also looks nice.



I also just noticed that one of the dual triple BIOS chips is socket-mounted for easy SPI programming.



27. Removable BIOS1 Chip

The ability to remove and replace the BIOS1 chip without requiring special tools is a key feature on many of our boards, by adding a level of protection against BIOS failures, bad BIOS flashes, BIOS corruption, etc. This feature also allows EVGA to work with end-users if something happens to render the BIOS chip unusable or a BIOS update is needed for CPU compatibility; rather than replacing the whole board, Customer Support can send a pre-flashed BIOS chip to get you up and running without requiring the end-user to send in the motherboard. Please see the walkthrough on Page 130 for further information.
​

Edit: I just learned it has TRIPLE BIOS (not dual)... https://xdevs.com/guide/e299ocg/


@Johnksss

 

They can because at 5GHz, power draw on a 7960X exceeds the ratings on the VRM. If you look, it started either around the Dimm.2 or with the traces going to the probe points (like a defect in PCB that shorted the trace and caused a heat buildup, which would be covered as a latent defect) (on cell, so haven't examined on larger screen, will tomorrow). The VRM, IIRC, for the R6 is an 8-phase 60A setup. At 5GHz, you are pulling a lot. But then comes the question of why it didn't pop a component of the power delivery system. That leads to questions of if the memory VRM popped due to voltages to stabilize the memory OC (didn't examine yet to see where those components are located yet), or something related to powering the DIMM.2. I'll look more in the morning, but if ran beyond spec where they can prove it was overclocked beyond spec, they wouldn't have to honor anything.

Meanwhile, the EVGA Dark and FTW boards both have better power delivery and VRM cooling. Not saying that was the issue, but is a nice to have. . . .

Edit: relooked at pics. Memory VRM looks fine and is a little ways down. I'll try to find clear photos telling what chips were in the specific area of the board later on one that didn't go up in flames. Trying to tell if the discoloration on the backside tracing back toward the CPU is due to the fire on the bottom side or if due to shorted traces for the probe points. Shorted traces is something they should have to replace.

Sent from my SM-G900P using Tapatalk

 

The EVGA X299 board is an excellent board and package all around, but then so is the APEX VI.

That removal bios is a nice feature. Only needs to be dumped then stored or technically not even stored in the machine, but then the whole purpose of having a dual or triple bios is to leave the original on one or the other.

Yes, Vince & Tin had quite a lot to do with how the dark came out, but it is missing some things that EVGA would not allow them to put in. But Still a great board.

 

I think you need to dial the OC settings way down.. Its depressing but it seems Windows asked too much voltage after BIOS allowed it w/o any fail safe mechanisms which is sort of a bug.

I wouldn't trust that stress test because 90% of the time those OC will be stable in minimal GUI mode and unstable on Windows 7/10. Then again, Triple BIOS chips sounds cool.

 

@Mr.Fox are you able to set BIOS defaults before sending the board back to them so they can't use any settings as ammunition? Maybe do a CMOS reset?

 

Speculations of course.

Well, after taking a closer look...The chip that seems to have started all the trouble is a Nuvoton 3949s 714GA. This chip seems to be in control of a fan header.

Website needs translation, but you will get where i'm coming from with this.
The margin of error is that theirs is a 645GA
http://www.itfiles.ro/2017/04/asus-...-pentru-pc-uri-de-gaming-in-format-compact/3/

And taking a closer look at the Apex VI
it has these same chips near every fan header on the board. Two where there are dual headers.

Now, this can pretty much be proven and a case made.....

 

Yep, EVGA are well known for their excellent customer service, I've never experienced/owned any of it, but definitely heard about it. Ha, the M18xR2 also 'blew up', you've not had the best luck! I was gonna say do you get lots of power surges where you are, but I think I remember you having a UPS which would smooth that out, and PC power supplies are supposed to smooth it out too.

 

I'm guessing deterioration of the IC caused it. Nothing to do with power surges. As to why it deteriorated is something else all together.

Edit: Because I do not fully disagree.

 

No way... No point in owning it if I have to "dial back" anything. This board is made with the knowledge and intent that people will overclock the snot out of it, and it is marketed in that manner.

It happened before POST, literally 1-2 seconds after pressing the power button, so we cannot blame Micro$lop.

I did just in case some nutjob cannot connect the dots with their marketing propaganda.

Thanks for checking that out. Should be pretty tough for them to wiggle an excuse out of that.