Mr. Fox

32 minutes ago, Talon said:

@Mr. FoxIt seems USPS came through a day early, shes out for delivery. 😀

My eyes are peeled and I am waiting with baited breath. 😄

1 hour ago, Rage Set said:

 

I have never registered that card and if I did, I've deleted it from My Products page on evga.com. The card was purchased on 2/4/19. I can send you the PDF of the invoice via e-mail. You should be able to register it.

Thank you. Yes, please email that to me. It may be out of warranty since 3 years have passed. If memory serves me correctly the standard warranty is 3 years, but I would need to check. I could still register it and find out that way, too.

Block installed. Man, what a huge improvement.

Here is some interesting information about A-die versus M-die timings and the impact of performance. A-die is only better when it can handle good timings, but apparently sometimes it does not.

10 hours ago, electrosoft said:

 

True, we need to figure out the parameters used to ascertain the rating vs Asus because if an SP101 is coming back 149 and Sugi's SP114 is a 108 something needs to be established that is consistent considering @johnksss SP99 came back at 136....their testing criteria either needs work or is much different than SP.

I suspect it is a new work in process with a lot of learning and data gathering that needs to happen before it can be viewed as accurate or reliable. It needs to be accurate or reliable before it can be related to ASUS SP ratings. 

Another example of where an industry standard would benefit both the seller and the buyer and improve the outcome for all involved. When we see things like this remain proprietary it really diminishes the value it can add. Building a reliance on ASUS SP ratings is also problematic.

It can also be misleading because it is sometimes not meaningful compared to the user experience. It better to be disappointed finding out later that an SP rating is low on a CPU that runs like a top than it is to have a CPU with a good SP rating that is mediocre, and we have seen both of those things.

20 hours ago, Talon said:

Stock 13900K undervolted with 420mm AIO. Quiet fan curve. I wonder how low I can get my vmin, this is my SP106 chip. Finally messing around with it a bit more now that the BIOS is fixed. 

I flashed v2.03 on my Dark KPE last night and it is working fine with the 12900K. Something about it did not play nice with my Winduhz 11 OS. All USB function was lost. Windows 10 LTSC 2019 and 2021, and Windows 7 were fine, but I had to clean install Winduhz 11 for some reason. I spent several hours trying to fix it with no success. I finally gave up.

CPU and memory seem to function flawlessly. Not sure what the heck was wrong with the cancer OS. That was super weird, but it is weird and crummy by any other measurement anyhow.

2 hours ago, tps3443 said:

I’m really happy with my 13900KF. I think

its a keeper for sure. I’m really glad I upgraded. Massive improvement over the 11900K. 
 

Looking at how these compare to 12th Gen it is just amazing! 
 

PS: DDR5 7200 will be here tomorrow Hynix A die.
 

@Mr. Fox

 

Delidding your 12900KS only builds that confidence for the 13900K delid. I think 6Ghz all core is very doable once delidded. Anyways, no matter how many times we’ve delidded, after that 11900K incident yeah I’d be nervous about Delidding too, I tore my 1st 11900K up pretty bad ripped of (16) small SMD’s. Confidence is put in check real quick after something like that with these small fragile components. But, I’m very glad it was successful on your 12900KS. Did you heat it up first? Or just pop the IHS cold?
 

Maybe you did too many good delids, and with the 11900K you’re like pshhhh walk in the park, I got this, delid # 398 coming right up, easy peasy lemon squeezy, I can do this with eyes closed, one hand tied behind back, “Crunch” oops. 😄

I delidded the first 12900K, the second 12900K and now the 12900KS. I heated the first one but not the next two. It did not seem to help anything other than making it more difficult to handle. Just a heads up, in all three cases the delid causes minor damage to the 12th gen IHS and will probably do the same to the 13th gen. The first one (that I heated) got the top surface slightly deformed the top contact surface and I had to lap it to make it flat. It works flawlessly. On the second 12900K and the 12900KS the little "wing" that the ILM presses on received minor scarring where the delid tool presses against it. In both cases I easily fixed it by sanding the bottom surface where the rubber seal makes contact.  This also accomplished cleaning off all of the old rubber gasket at the same time as fixing the burr on the wing.

I mention this so it does not come as a surprise. In all cases the delid tool screws were very tight, but it seems like the manufacturing is off a little bit on the tolerances. All previous gen Rockit delid tools popped the IHS off without a blemish. 

I had definitely grown complacent after doing about 20~ CPU delids and finding the process to always be quick, easy and painless. The 11900K disaster made me approach the process with much greater caution and care.  The delay in delidding the 12900KS was more being pure lazy than worry about how it would turn out. Of course, having a chip with a high  silicon quality is a good reason to be extraordinarily cautious. I used to delid all CPUs before installing them for the first time because I viewed it as a waste of time and thermal paste to install it before delidding. I have never seen a CPU that doesn't benefit from it, but I am more inclined to see what the temps are with the solder first and evaluate the potential benefit of a delid  against the risk.

2 hours ago, Rage Set said:

@Mr. Fox

 

I'm sorry I've been missing in action. Did you get the warranty straight with the 2080 TI? Lots to read through. (If you ever need me, you have my number, just call or text me bro).

Hi bro. We have missed you and I was planning to call you to make sure you were still happy and healthy. I suspect you have been busy with the new job, etc. I have not done anything yet on the 2080 Ti. It is not registered in my name, but I am assuming it might have a little bit of time left on the warranty. I was going to try to fix it, but it might make more sense to put the hybrid thing back together and RMA. Can you check the status for me and let me know how much (if any) time is remaining on the warranty and if we can arrange a cross-ship at my expense? I did not want to bother you with this since I suspect you have been very busy. Plus, I didn't want the warranty question to appear to be the only reason I would call to check on you.

1 hour ago, Talon said:

 

Works great. My flash went off without a hitch. I flashed from version 2.01 to 2.02 inside Windows with batch file as Admin. Rebooted and began tweaking. 

I updated to 2.03 and it's working fine, ready for 13900K. 

Weird thing though, now Windows 11 has non-functional USB ports. Older versions of Windows work fine. I tried flashing in Windows 11 on older BIOS and it gives an error about "Intel ME Device not found" and reboots. Lovely POS from the Micro$lop Mafia.

I figured I should stop being lazy and delid the 12900KS before it goes in Banshee. As expected, a nice 10-12°C drop in load temps and allowed me to lower vcore from 1.360V @ 54x all P cores to 1.340V. As many times as I have done this, it still makes me nervous since the first 11900K that got de-die'd rather than delidded. I never batted an eye before and viewed it as non-chalant as a repaste. 86°C core max on ambient water... not too shabby. Just under 400W. No special tweaks or tuning here... ran it on my crash test dummy cancer OS.

6 hours ago, Talon said:

Dark BIOS 2.02 is out! 

How is it? I heard there were reports of issues in EVGA forum, but haven't gone to investigate yet.

Just checked. There is a 2.03 beta out now to fix issues with 2.02.

BETA BIOS Updates for Z690 DARK K|NGP|N (2.03) / CLASSIFIED (2.03) [Fixes 53,54,7F loop]

Newer is better, right? https://imgur.com/gfRxA0E

1 hour ago, Talon said:

I'm convinced that guy bent that power cable trying to fit it into his case. It's extremely easy to mess that cable up, which is definitely on Nvidia and this crazy new cable, but if you follow the instructions included in the box or some common sense this should not happen. 

 

When putting the FE 4090 into my Lian Li O11 Dynamic (non-XL) the cable would not fit. I attempted a small angle to make it fit but I quickly discovered the pressure it was putting on those cables and the tip. Plenty of idiots are just going to smash it in there and then cry foul when their cables start on fire. 

 

Right now I think Nvidia should be thinking seriously about a cable adapter that offers an easier 90 degree angle for attachment. This would be a relatively cheap and simple fix for the end user. But I guess a few lawsuits might be even cheaper in the end. 

56 minutes ago, Papusan said:

Yep. And you don't need to be an idiot. There is many people out there that isn't so clever in tech. And many don't follow the manual properly. Their main goal is to cram it in to start use it. 

This means bad publicity. But they have to do something with this. There have been a lot talk about this the last months. And we can now see the outcome. Was doomed to happen. Even an bigger warning sticker in the Gpu package won't help. Because people are so different. 

Another wonderful example of change for the sake of change. It's new, so it must be better, right?

I wonder if NVIDIA gets a "license fee" on new PSUs sold with their crappy new proprietary connector?

4 hours ago, Ashtrix said:

 

Luumi's OC run is taking more than 300W, on Stock 5.5GHz and 4.3GHz, its 320W and with OC it goes to insane 390W with P cores at 5.7GHz and Add-on cores at 4.5GHz, Cache Ring at 5GHz with 1.35v. Then he increases the Add on cores to 4.6GHz and P cores to 5.8GHz same Vcore it goes to blistering 415W. P95 goes to 468W with AVX offset.

 

 

 

 

 

That's awesome. The thrust of focus should always be in the results. Being obsessed with power draw is silly and a distraction. The way I look at it is, "it is what it is" and what I should be paying more attention to is everything else. The right questions to ask are things like "how did I do?" and "what is my rank?" and "can I beat the person in front of me?" and "what do I need to do in order to do that?" and power draw is merely a natural byproduct of that pursuit. If I let it become a focal point, success will evade me.

Once you have a baseline on what is normal, increasing power draw while simultaneously increasing performance confirms you are on the right track with tuning. The only point at which it matters is whether or not you have the necessary hardware to tame it thermally. Determining success based on a preconceived bias about what that number is, including a conclusion that it is "too high" is arbitrary. It tells me that person is potentially misguided and lacks understanding. Yes, that includes lots of famous people on YouTube, many of whom are legends in their own minds. They may be worth putting up with if they are a source of useful information that I can leverage. A massive subscriber base is a measurement of popularity, not authority. 

The abortion we see with modern turdbooks is a great example of what the end result looks like when the wrong thing becomes most important. That is a mess, not a success. No thank you... It draws less power, but it still overheats. Somebody with an engineering degree sucks at their job.

29 minutes ago, tps3443 said:

Ohhhhhh! A cool fancy feature. 
Let’s see.

Drum roll please.

My Ho-Hum= 134

 

@johnksss

 

What is your Ho-Hum rating? 

It would be interesting to know how to translate that against an ASUS SP rating, or if doing so is even feasible due to lack of data to validate the conversion. 

@johnkssswhat is your MSI "CPU Force 2" rating?

On 10/23/2022 at 11:44 AM, Mr. Fox said:

Spoiler Alert:

I decided to do something constructive. When the 3060 Ti block and other parts arrive and I have time to tinker, I am going to be moving Banshee to an external cooling solution like Wraith. Both of the high quality internal radiators will be offered for sale in the forum marketplace.

Hasn't shipped yet...

42 minutes ago, electrosoft said:

 

Hopefully good stuff on the 13th gen....

 

And just a dud (ish) MB for memory then. Harkens back to the early days of Z690 fishing through motherboards and lamenting the lack of a 2 slot DDR4 motherboard.

 

Post by Sugi regarding SP bug reports:

 

 

I'm still on 1601 as I haven't found a reason to update but if/when a 13th gen chip miraculously wills itself into existence for my D4, I'll make sure to upgrade. 😉

Everything remained the same with my SP rating after the firmware update. I was even able to import my saved BIOS profile with all of the old settings without error, so that was a real time-saver.

I flashed from 1504 to 2103. I totally skipped 2004 and had no issues.

10 minutes ago, electrosoft said:

lol, it might end up like @Mr. Fox when he shipped that brand new 11900k to @tps3443 and he ended up with an absolute god tier 11900k....

That is what I am praying happens. Seems like the right thing to happen.

10 minutes ago, electrosoft said:

That's the luck of the motherboard with Z690 DDR4. As a test, have you tried going into G2 to eliminate the CPU as the potential bottleneck and see how far you can push your DDR4? Does it still cap at 4000? I know my board can push 4200+ and I run it rock solid at 4133 daily but I still didn't go in there yet and test it in G2 just to see where the memory/motherboard gives up the ghost.

Yes, I have. I have spent hours trying different things, including that. It is definitely something with the mobo. Even my custom 4500 XMP profile that ran as smooth as silk on the Z490 Prime, Z490 Strix, Z590 Dark, Z490 Dark is unbootable. It never trains and dumps me back on the boot screen with the F1 option to try something else.

5 hours ago, jaybee83 said:

make sure that brother @Talondoesnt just give u the worst binned of his lot 😄 

He would not do that to me, or anyone else, without full disclosure and letting them decided if that is OK.

1 hour ago, Talon said:

 

😄 He's getting a factory sealed 13900K haha. It was staring at me on the desk last night, mocking me saying how it's an SP120. 🤪 Shipped it off to him this morning, he should have it Thursday according to the nice lady at the front desk. 

Looking forward to getting it. Crossing my fingers. I already flashed the 13th Gen update to the Strix D4 to check the SP rating before moving it to the Z690 Dark.

Now I am wishing I had not purchased a DDR4 mobo, but can't cry over spilled milk. It is what it is. It's actually a good mobo, just limited on RAM clocking. It can't overclock DDR4 as well as the Z490 mobos I used these same modules in. They were stable at 4500 on Z490 Strix and Dark. Now they are only bootable at 4200, stable at 4000 on the D4 for some weird reason.

48 minutes ago, Etern4l said:

How did you catch wind of this? Sounds like an offer you couldn't refuse... 😉

Same place that @johnksssand @Talon purchased theirs with the same 2-year Micro Center warranty. 

@tps3443 - this is what I was referring to before about correcting the high PCH temperatures. I showed this to you once a long time ago, but in case you did not remember. See this post and change from "disabled" to "auto" and they should dramatically drop. If MSI changed the defaults to "auto" in the new BIOS you may not need to do that.

https://forum-en.msi.com/index.php?threads/crazy-pch-temperature-on-z690-carbon.370039/post-2096947

27 minutes ago, Papusan said:

I have also fan hubs. Also the command center from Corsair. But I really love have a few of these lying in the drawer. For powering 4 fans and then connect one of the connector to the MB. And I use only one fan for each MB fan header. Even for the high amp fan header (3 Amp). No need to max them out.  

Yes, I use those as well. They are very handy, and I always keep spares in case I ever need them.

I also use these. https://smile.amazon.com/gp/product/B0975XH3JS (new and improved design)

53 minutes ago, tps3443 said:

Are you getting a random brand new untested 13900K? Or is it pre-binned?

It is brand new from Chicago Micro Center. If the binning is no good it will get exchanged for a better one or I will get my money back.  Once that is settled, I will move the 12900KS to my work computer and sell the golden 12900K in the work computer to offset the cost of the 13900K.

53 minutes ago, tps3443 said:

Usually when I plug in (10) fans in to one single pwm plug the RPM’s were never the highest on EVGA motherboards.

46 minutes ago, johnksss said:

I would watch that trying to run a gazillon fans on 1 header.... You might be sorry for it later.....

33 minutes ago, Reciever said:

Even I was nervous for attaching 3 to one header on my x570 itx board 

That is very much a no-no. Good way to kill a mobo, or burn up a fan controller. A single Delta fan is enough to kill it.

@tps3443don't risk it. It's not worth it when $10 is enough insurance. It will get power from SATA and only use the speed control from the mobo. https://www.amazon.com/ARCTIC-Case-Fan-Hub-Distributor/dp/B0887VG14J/ref=sr_1_3?crid=IBF6J65W1NOV (This can handle a maximum of 1.0A per fan port. Most fans are less than 0.5A each.)

36 minutes ago, tps3443 said:

Do we max out tREFI like with DDR4? I’m guessing not all DDR5 can do this. I see it goes up to like 200,000 thousands lol.

 

Testing 90,000 tREFI. Huge BOOST! 😃

It's going to vary some based on a variety of conditions. The higher you clock the memory and the tighter you make the primary timings the more tREFI has potential to cause instability. Higher tREFI also takes more voltage  Luumi has a couple of good DDR5 videos that touch on this. Even if you can't get it to pass memory tests without errors, I have generally found it stable enough for benching when it is maxed out. I've seen some difference from one memory set to the next.

27 minutes ago, Etern4l said:

 

I guess you guys WC the RAM too? 50C+ is what I get with 1.1V under load lol

I do, but naked DDR5 sticks run cool enough with a fan blowing on them. I think the errors occur at lower temperatures as clock speed increases. It seems that way based on my own experience.

10 minutes ago, tps3443 said:

Some DDR5 modules have a voltage limit of 1.435V. It’s just a bios settings that unlocks the DDR5 PMIC so we can max out voltage for higher memory overclocking. Not all DDR5 sticks are like this though. Some do not need to be unlocked apparently. 
 

I’m using Corsair 6000 C36 currently, and it physically won’t let you send more than 1.435V. This is where this unlock comes in handy. Or not. I really won’t know until I test it a little further if there is a benefit with more voltage. But it does let me send whatever voltage I want now. 

If it does truly lock it then it absolutely will help. Samsung B-die, SK Hynix and Micron all overclock higher with added voltage. You know this already because you're used to spending lots of time overclocking Samsung B-die on DDR4 sticks and you have seen how things don't work if you don't give the RAM enough voltage.

Be sure to monitor the temperatures because those high clock speeds will actually fry your memory. DDR5 gets much hotter than DDR4 did. You'll also start to see errors running memory tests before the RAM gets hot enough to damage it. Errors begin to surface around 45-50°C and an unstable overclock can often become stable below 40°C. A fan blowing on them helps tremendously.

8 minutes ago, Etern4l said:

 

Thanks, gl&hf.  BTW I wonder if Raptor Lake also improved 4 DIMM support. Any word?

Good question. If I had to guess I would say probably so, but that's just speculation.

3 minutes ago, Etern4l said:

 

Could you at briefly explain how you are going to use this? A quick google didn't help me 😕

It's actually very simple. It's only a matter of turning it on. When you enable it, that unlocks the industry standard 1.435V cap on DDR5 voltage. For some idiotic reason MSI removed that menu option and it meant that certain brands of PMIC chips could not exceed 1.435V.

5 minutes ago, tps3443 said:

I found the PMIC unlock in the Unify-X bios!

 

I can really juice these modules now, it works.

 

 

 

 

That was missing in the previous BIOS. It was there in an earlier one and MSI removed it. Awesome to see they brought it back and now you can really get everything that memory is good for. You might even be able to return the Hynix RAM if this performs as well now that it is unlocked.