Mr. Fox

Tony contacted me about the clear plastic on the thermal pad. He was sure that was why the memory chip had failed and I explained that the clear plastic was on it for storage (keep dust off the pads) and that almost the whole life of the GPU was run on a full block.

@tps3443 - so, I sanded down the underside of the IHS until the IHS was sitting sloppy loose on the die with an air gap so it would rock slightly due to the lip around the perimeter not making contact with the PCB and that helped. Temps improved ~5°C.

I also investigated the copper IHS more closely and I think I can fix it by taking a tiny bit of material off the length of the wing using a Dremmel tool on the side where the CPU frame holds it captive. It is just a tiny bit too long to fit into that notch. I can make it fit using force, but that won't be good for mounting. It should fit loosely, not tightly. Should work fine after that. I will attempt that project another day.

6 hours ago, Mr. Fox said:

Interesting rambling. Might be useful info.

This was useful. Adding 120mV (as he demonstrated) over the 0.090mV default CPU PLL, Ring PLL and SA PLL voltage (set each to 1.020V versus "Auto" default of 0.90V) and slightly increasing my VCCSA to 1.300V and bumping VCCIN from 1.900V to 2.000V added stabiliy to my memory overclock. I did this on both the Strix and Dark, and oddly enough while helping both it decreased my memory temperatures a couple of degrees somehow. Maybe "Auto" was causing something to adjust to compensation for those PLL voltages being too low at the default 0.090V.

The Blueanon platform losing their minds is a good thing as far as I am concerned. I can't really see a downside to it. Elon is a fixer. He takes over broken companies and makes them useful and relevant. The loss of control of the narrative for the left is what the world needed. A "free-speech platform" needs to be a free-speech platform. It wasn't.

LOL... hilarious meme

Interesting rambling. Might be useful info.

LoL. I agree with Chris. This needed to happen. It evolved into a scummy mess. Somebody needed to clean it up. It had become a political tool of misinformation for the extreme left. It's only a tragedy if you were part of the problem and now you don't have a platform.

15 minutes ago, Etern4l said:

 

If there is any justice in the world, they can also lose some business to AMD due to bad pricing and inventory shenanigans, and I hope they do.

I fear there is not any. We see lots of examples suggesting there is not. Being evil is rewarded and nothing good goes unpunished.

I think there are lots of people that are like me and not willing to forgive AMD for their past failures. I might, eventually. Maybe. I harbored hate for them almost a decade. Gave them a second chance with 5950X and they totally blew it. Silly me. I guess it serves me right for listening to fanboys and thinking we valued the same things. Might be another decade before I allow them a chance at redemption. Mind you, that doesn't mean that I love Intel or NVIDIA. No sir... only that they haven't failed me as hard or disappointed me so severely. The truth is, they suck as well. None of them deserve our money or forgiveness. 

2 hours ago, tps3443 said:

 

I use my computer for work everyday managing a team of agents, so I use that as an excuse lol. Nothing wrong with running the fastest cpu/ memory available.  It certainly helps. 

100%... Makes work a lot more fun. And, that's the story of Banshee's life. 40-60 hours a week is doing work and the fact that she is overclocked, overvolted, loving every minute of it, makes it more fun to work. I've actually gotten more than a few reactions during teleconferences with desktop sharing and people see the AIDA64 Sensor Panel and some are like, "OMG, you're nuts man." Especially so of those that know what clock speeds their own computers run.  Same with sharing my webcam and they see Wraith over my right shoulder and they're like, "What is that huge glass thing behind you with the clear water hoses?"

2 hours ago, Etern4l said:

 

TBH 4080 looks like a stawman that's been put out there to get people to spend even more money on a 4090... 

Maybe so, but I really do suspect the Green Goblin is just so self-absorbed and greedy that they were hoping more people would be stupid with their wallets. They can always lower the price. If people are as stupid as they had hoped, they make even more. It would be difficult to imagine these shysters allowing themselves to get caught in a scenario where they did not win. It's more like how big, or how small, is your win this time GG? If they loved customers as much as they love money iit would be a beautiful thing. It's unfortunate that they don't love customers even a little bit. They only care how much moolah they can extract from them before they're bled dry. I'm not sure this isn't equally true of Intel and AMD, but they do a better job of pretending to care about what customers want. NVIDIA is just too conceited and self-centered to pretend. The company is a reflection of Jensen in the same way the piece of garbage Windows has become is a reflection of Sadsack Nutella.

Anyone that calls the purchase of technology an "investment" is either a lunatic or a liar. Sometimes using technology is necessary, but it is worse than a car at not holding its value. And, most of us would absolutely not tolerate poor quality and incompetence on the same order of magnitude with an automobile that we typically do with buggy crap technology. Yet, poor quality and incompetence are status quo.

Twitter and Facebook are as much or more of a political tool of the radical extremist left than legitimate businesses. Being a major contributing tool to global misinformation and agenda-driven opinion manipulators, I am happy to see something like this come unraveled. I do feel bad for people that lost their means of supporting a family, but that could be lessened to the extent they were actually drinking the Kool-Aid rather than merely enjoying a paycheck doing what they were paid/told to do. To the extent any were thrilled with the opportunity to contribute to the cause, I honestly have no more sympathy than I would for a hit man that didn't get paid for his last assassination.

55 minutes ago, Reciever said:

This is framed as if there aren't better options? There is plenty left in the world outside of Musks domain. 

 

As long as both parties agree to terms (free from duress) and engage then I don't see the issue. I'm pretty sure I am not in the majority though. I have to cite state law and verbiage of the employee handbook to argue my way out of a lunch break, if I need a minute to collect my thoughts and scope the day that's a 15 minute legally required break. 

 

My work is repetitive but requires a clear head, eating food dulls my focus in the later half so I just have some caffeine and water, neither of which requires a 30 minute time out. 

 

If I had to put a finger on the mentality its likely come out of my time in contracting as well as the framing of, my time on this earth is finite. I do not care that you are compensating me for my time, you are still wasting it. 

I remember someone asking about the rep icons being updated. I always wish I had one of those red "100" tags. 😉

11 minutes ago, Etern4l said:

Looks like this is just for quick CPU tests/heatsink tests to avoid messing around with the paste.

Carbonaut won't even work for that. It is too fragile. You have to use a generic (Panasonic) sheet or the IC Diamond option if you want it to hold up to remounts.

I have not tried Carbonaut (nor do I plan to waste time doing so) but I tried the Panasonic sheets from Digi-Key and IC Diamond had me test the product they offer before it went to market and both were absolutely horrible. Worst thermals I had ever seen before... to the point of making my overclocked system unusable.

16 hours ago, Reciever said:

I dont mind working extremely long hours as long as I am compensated for the time, but I have essentially, 0 social life, and introverted enough to be able to not engage in society beyond required elements. Im also aware that my perspective is not in the majority. 

Same here, mostly. If you mean majority in the context of the Twitter zombie horde you would be correct. In the context of mainstream, decent citizenry that have not submitted themselves for reprogramming, you're squarely in the normal majority. They want you to think you're in a minority, because people who are concerned about whether or not they are popular will be influenced and subdued by the fear that they are not.

They create a perception of power and greatness by gathering together and excluding or silencing those who think for themselves and won't be intimidated and won't back down on expressing an opposing view.

That's why the liberal social media dictatorship banned so many. Having rational members that are willing to go against the flow is a serious threat to their agenda.

I cancelled my Twatter account after the Thought-Nazi censoring started with all of the irrational orange man bad emotard haters. I created a new account after Musk stepped into oust them. I don't use Twatter or Facepoot for the reason previously mentioned, but wanted to show my support for his hostile takeover of the woke libtard echo chamber. 

8 hours ago, saturnotaku said:

A lot of the people butthurt about Elon cleaning house at Twitter don't seem to have much to say about Amazon, Facebook, Lyft, Salesforce, Cisco, Stripe, Microsoft, and others making significant workforce cuts.

All of the retarded Tech Giants get a pass for their intolerant wokeness and lopsided political correctness.

I am glad that I have no interest in purchasing a GPU right now.

It feels nice that I can say, "Nah, I'm good... no thanks."

4 minutes ago, Etern4l said:

 

They knew the top-side location of the socket + crazy adapter was bad in the 3090 Ti already, hence the attempt to mitigate using the angled socket. Seems like with the Ada consumer cards, someone at Nvidia said "f.. it!" and just left it at the right angle. The workstation cards typically have the socket at the rear of the card. I guess more space that way in most cases, and no need for weird cable routing. Like you said though, the socket location options are limited by the shorter PCB relative to the cooler, to accommodate the rear fan. In theory board partners could have mitigated this (unless they are forced to use Nvidia board designs verbatim), but I guess none of them did.

Yeah, I think the "blow-through" fan thing is also a really stupid gimmick. I know it is subjective, but I think that also detracts from the aesthetic quality of the final product. I find the openings on the backplate to be very unattractive. I also do not like the opening for the four screws and X-bracket on the modern GPUs. I'd much prefer a clean backplate covering up that ugly mess. Overall, I think the back of the newer GPUs is very ugly-looking.

10 minutes ago, johnksss said:

I think it was a good idea, just badly executed. And should have made it more idiot proof. Plugging in a single cable is far better than plugging in 2 to 4 cables. 

Quick search reveals plenty of power cable issues before this came about.

https://www.google.com/search?q=burnt+up+pcie+cables&rlz=1C1ONGR_enUS990US991&sxsrf=ALiCzsbmB0_u8QBJRb7ooANHDLykHsoILQ:1668804057082&source=lnms&tbm=isch&sa=X&ved=2ahUKEwjWmpbIy7j7AhWgLUQIHenxAzMQ_AUoAXoECAEQAw&biw=2037&bih=850&dpr=1.88

Definitely could have been executed better. I think putting it in a really dumb location at the top center on the PCB and trying too hard to make it a little connector reflect poor planning. They could have done something beefier, like a shorter 12-pin version of the 24-pin motherboard cable. I think making the PCB 2/3 of the length of the GPU cooler created design issues for them that had a cascading effect that caused other issues. Having the cooler be 1/3rd longer than the actual GPU seems really limiting and  restrictive on engineering flexibility. Doing that accomplishes nothing useful for the end-user when the total package is massive. It would be different if the entire GPU were smaller. Then they could move the connector to the end of the PCB facing the opposite direction of the I/O panel. They could have also put it on the back of the PCB through the backplate with a 90° cable so everything would be neat and tidy.

I bet NVIDIA is wishing they would have left it alone. The push for the new 12VHPWR connection was another idiotic change for the sake of change, purely for the sake of padding someone's, or some group's, wallet. There was, and is, no need for this stupid revision. Another example of "if it ain't broke, don't fix it" being the best advice. In a bit of a sadistic way, I find it gratifying to see self-serving schemes like this turning to crap. Unfortunately, they won't learn anything from it.

Just released... minutes ago.

On 11/14/2022 at 8:57 AM, tps3443 said:

I’m using Alphacool 17Mkw thermalpaste on the IHS.

I got a tube of this with the Kryos waterblock and it did not work well at all for me. It seems way too thick. My load temps were about 10°C higher than they are with KPX or MX-4. With KPX and MX-4 I end up with an almost invisible layer of thermal paste on the IHS after mounting the block. After remounting twice with the Alphacool paste it remained very thick in comparison. I think it might work really well for a laptop and offer resistance to pump-out because of that.

On the first attempt, I tried using the mounting pressure to spread it from a line down the middle. Big mistake. Did not compress well enough. I spread it evenly across the IHS with a spatula after remounting and that helped, but the temps were still north of 102°C in Cinebench R23 (easily a 10°C temperature increase).

4 hours ago, Ashtrix said:

 

It is a new Windows 10 standard and it makes the drivers integrated further deep into OS also it makes the OEMs use Store based UWP apps as in it's an option to them (add the bonus telemetry that comes along with the UWP trash) which is why so much of the modern Software packages have a ton of bloat and are more utilizing UWP apps (MSI, ASUS, AW for eg). NVCP with the DCH drivers previously did not ship with NVCP we needed to add it through Store or NVCleanInstall injection, they provide the app now though, still NVCleanInstall is mandatory as the driver has a ton of stupid bloat.

 

Below is their own documentation mentioning that UWP part more.

 

• Declarative (D): Install the driver by using only declarative INF directives. Don't include co-installers or RegisterDll functions.

• Componentized (C): Edition-specific, OEM-specific, and optional customizations to the driver are separate from the base driver package. As a result, the base driver, which provides only core device functionality, can be targeted, flighted, and serviced independently from the customizations.

• Hardware Support App (H): Any user interface (UI) component associated with a Windows Driver must be packaged as a Hardware Support App (HSA) or preinstalled on the OEM device. An HSA is an optional device-specific app that's paired with a driver. The application can be a Universal Windows Platform (UWP) or Desktop Bridge app. You must distribute and update an HSA through the Microsoft Store.

And Intel Graphics app is a UWP one not Win32, so lack of proper filesystem for UWP applications  and not easy to hook and mod etc.

 

Excellent technical breakdown on this filth. Thank you.

For those with a short attention span, we can cut to the chase. DCH drivers can be casually described as  Deliberately Crapified Hype produced by degenerate imbeciles for their zombie sheeple herd... which is assumed to always be the case for the projectile-vomited malware  distributed by Micro$lop through the digital cess pool they have dubbed the "Microsoft Store." We are expected to believe they are better just because they say so.

Took about 2 weeks for this to get here. CPU load temps are the same as the OptimusPC Foundation, but it is far less restrictive. My loop flow rate increased by about 40 l/h.

3 hours ago, tps3443 said:

You are keeping the 3060Ti over that 2080Ti? I guess they do look to be comparable. This probably makes the most sense for your use on this rig anyways. I have had so much fun with 2080Ti’s especially once you flash KFA2 bios and water cool it.
 

I was managing around 7,800 with my 2080Ti FE in Time spy Extreme. I was blowing even a 3070Ti out of the water by about 20%. But this was before Ampere matured to what it is now. Turing was in its moment of peak prime. 
 

You’ve had that 2080Ti for a while too so that’s understandable. Also, I think Nvidia made some nice improvements with RTX3000 series driver advancements lately, like faster DLSS, and also faster ray tracing performance than RTX2000 series. So I imagine the 3060Ti has probably really come along ways since it’s launch. 
 

Hard to believe it has already been over (4) years since the first Ray traced Nvidia GPU’s have launched!

 

Yeah, it was a hard decision. The 2080 Ti FTW3 is absolutely a better GPU in every way. Had eBay not shafted me I would have probably just kept both of them. I decided based on economics and tried to stick to common sense and logic rather than getting hung up too much on sheer performance metrics..

The 3060 Ti FTW3 is brand new and has a new block. The 2080 Ti FTW3 is pretty old now and needs a new block. Spending another $150-$200 on a new block for the obsolete GPU make little sense. It is unlikely I will earn any points benching the superior 2080 Ti FTW3. I've milked it for all it's worth pretty much. I can shunt mod the 3060 Ti FTW3, set some 3060 Ti records and gain some additional hardware points if I decide to. 

If high-framerate gaming was super important to me I would keep the 2080 Ti FTW3 instead. 

5 minutes ago, Reciever said:

I was wondering why my wallet was on fire this morning.

Do you want me to put it on hold for you? Sorry you missed out on the last deal.

13 minutes ago, johnksss said:

True, but there may be some basis to the argument but going over board about it is what some do.😂

Yes, there is often a legitimate basis to it, but it's the reaction that can be comical and very telling about certain individuals. Sometimes I wonder how they have managed to survive so long without a heart attack. 

34 minutes ago, johnksss said:

Best watchout posting aida64 over on OCN. They have a few haters over there. LOL

Sometimes haters like that are an incentive to post. It can be amusing to poke the (gummy) bear just so see what kind of shrill crybaby noise it is going to make.

1 hour ago, johnksss said:

Looks like my Microcenter still has plenty of cards.

https://www.microcenter.com/search/search_results.aspx?N=4294966937&NTK=all&sortby=pricehigh

Looks like everything is still a poor value and priced very stupid until you get into RX 550 and GTX 1650  ~$200 price range. It is expected that something current is going to be overpriced, but my goodness... some of the old garbage that nobody wanted when it was new tech is still priced for idiots. The fact that it is still rotting on a store shelf because nobody wanted it during the pandemic when people were happy to have something better than integrated graphics say a lot.