Mr. Fox

So, I pulled one of the two dual NVMe cards from the Encore for the X79 Gene. It already had eight, LOL. One less won't hurt anything. I sacrificed the Windows 11 drive since I almost never use that sucky POS cancer operating system anyway. Set that one on the desk for later, kept the one with Linux and put it on the single NVMe card in the Encore.

Hearing about you building a home network for your kids brings back a lot of fond memories. I built each of my five kids their own desktop. Correct on the Titan being too slow. It can still work on low to medium setting, but no ray tracing sucks. It's basically the same as 780 Ti. Nearly identical performance and in the same generation. 2080 Ti totally obliterates it. I had the 2080 Ti already. I bought it like about a year ago for Windows 7 benching. Just converted it to air cooling so I could use it in this system (no custom loop SFF). I actually really like this antique. It performs surprisingly great. After the BIOS mod for NVMe support (which did not exist when X79 was released) it because a pretty compelling machine. And, the best thing about it is... No driver mods required. That YouTube channel is pretty worthless. Besides just being dead wrong about a lot of stuff, the guy is about as entertaining as watching paint dry. Not really any better than Gamer Meld. Lots of nonsense from some of these bobble-heads. Ignorance isn't a sin, but it is contagious. Simple fix for it is knowledge, but know-it-all folk are hard to cure.

I decided to go ahead and install this in Methuselah and put the Titan on the shelf. I don't have the fan cable yet, so this is running without fans. Just doing mundane tasks like web surfing and email until I get the cable. But, doing that the GPU core has never gone above 35°C core and hotspot 48°C. The fans mounted in the chassis floor seem to be providing enough air flow for low intensity normal use. I think it looks great. I lost access to one PCIe slot due to the extra slot being taken by the girth of the air cooler, so I will have to install a dual NVMe card in the bottom slot rather than using two single NVMe cards. I bought some faster RAM for it as well. 2666 vs current 2400. Got it cheap on eBay.

All of the other stuff in your post is sad, but Buildzoid probably has no idea that 0x129 fixed the SA bug. So, for that reason only I would say it is not worthless, but extremely valuable.

I am still waiting for the adapter cable for the fans to complete the Xtreme Waterforce conversion to Xtreme Air. I went ahead and installed the assembly. To install the cable I only have to remove 4 screws from the shroud. This looks really nice. I am glad that I ordered the kit that includes the backplate because it is different than the Waterforce. It has additional screw holes. I am eager to see how hot/cold it runs with this thermal unit.

Yeah, those prices are ridiculous. Given that they seem to have no meaningful difference in performance, then I would definitely go with the more affordable option. It would be silly to spend more for one or the other. The Kryonaut Extreme obviously works great and for me to purchase it again would be fiscally foolish unless it becomes less expensive than KPx. Simple logic. Since I use liquid metal on my bare die setups I do not use very much thermal paste (or liquid metal since it does not wear out). I wish I had a good way to delid this Xeon. I believe it is soldered. I know it is not as popular with everyone as it used to be, but I really do not like anything that is not delidded and bare die. I will always do that when it is an option.

Don't believe it. People like to state preferences as though they are facts. Here is proof. Just did the comparison today. Kryonaut Extreme is the same as KPx in terms of how it works. It costs A WHOLE LOT more. Good stuff, but not worth it. I will use it since I bought it, but I won't ever buy it again. No benefit unless the goal is to spend $10.00 per gram more for the pink color instead of blue. Literally no difference. Newer is always better newer, right @Papusan?

I think one reason it works well for laptops is they don't really have high TDP on anything. There is probably nothing more durable. You can't re-use it like you can Kryonaut or KPx, but if you do not disturb it the stuff will last for a super long time. Also a reason why it works great for laptops. If you had a low TDP desktop CPU and ran it stock it would probably do fine. The thermal conductivity is pretty low and I think that is why it worked poorly for me with my overclocked CPU pulling over 300W. This is basically the same stuff that comes pre-applied on many heat sinks and AIO water blocks. If it has the clear film you peel off and bolt it down, it is this Honeywell pad.

Worse. I've already tested it. Three times, in fact. It will be worse than the Corsair paste you got from Walmart. It works good for turdbooks with crappy-fitting heat sinks because it doesn't pump out and it fills gaps, but it can say for certain it is the worst TIM that I have ever used on a desktop. Not anything crazy like 10-15°C hotter, but not nearly as effective as MX-4, MX-6 or KPx or TF7, etc. At least 5°C higher temps all three times. If you want to try something new. maybe test the TF9 paste. I have never tried it before. https://www.amazon.com/s?k=TF9+thermal+paste&crid=NA6AQT8BS3U9 The new BIOS works identical for me on the Apex and the Apex Encore. All BIOS settings set exactly the same on old and new firmware. Cinebench and AIDA64 scores exactly the same. No loss of performance or stability that I can identify.

Yes, I absolutely would have preferred that it be the motherboard over the CPU. All three of my desktops have extra-good CPUs that would be terribly difficult to replace, and cost as much as the motherboard. I still have two spare CPUs but their bin quality is inferior to the three I am using.

OK, I flashed 2503 to the Apex (white) and had a bit of a scare. After flashing I got Q-code "00" (almost always dead CPU). I cut power and waited, then turned it back on. No lights turned on. I pressed the power button and it came back with "00" immediately displayed again (first and only code, no other codes). I thought, crap... ASUS killed another one. So, I held the clear CMOS button down on the rear I/O, said a quick prayer "Please Lord Jesus, don't let it be dead" and then it booted normally. *WHEW* Running like a top... My IVR TX is set on Auto and that is usually around 1.330V. I will mess with setting it a little lower manually to see what happens. I use similar settings as far as the other ones go. Voltage Monitor [Die Sense] VRM Initialization Check [Disabled] CPU Input Voltage Load-line Calibration [Level 3] CPU Load-line Calibration [Level 7] Synch ACDC Loadline with VRM Loadline [Disabled] CPU Current Capability [Auto] CPU Current Reporting [Auto] Core Voltage Suspension [Auto] CPU VRM Switching Frequency [Auto] VRM Spread Spectrum [Auto] CPU Power Duty Control [Extreme] CPU Power Phase Control [Extreme] CPU Power Thermal Control [125] CPU Core/Cache Boot Voltage [1.05000] CPU Input Boot Voltage [Auto] PLL Termination Boot Voltage [1.15000] CPU Standby Boot Voltage [1.15000] IA VR Voltage Limit [1500] BCLK Amplitude [900mV] BCLK Slew Rate [Fast] BCLK Spread Spectrum [Disabled] PCIE/DMI Amplitude [900mV] PCIE/DMI Slew Rate [Fast] PCIE/DMI Spread Spectrum [Disabled] SA PLL Frequency Override [3200 MHz] FLL OC mode [Ratio OC] UnderVolt Protection [Disabled] Core PLL Voltage [1.05000] PLL Termination Voltage [1.15000] CPU Standby Voltage [1.15000]

Well, I forced myself into moving the excellent bin 14900KF with the SA bug into the Encore for testing. I am happy to report it is fixed for me as well. I set 1.300V on VCCSA in the BIOS and it runs perfectly fine now. I am going to leave it in the Apex Encore because it is just a little bit better than the 13900KS SP117 monster. That is running beautifully in the Z790i Edge now. Leaving it alone. Here are the SP ratings and VF curve on the CPU that used to have the SA bug. And, the same for the 13900KS SP117 monster now in the SFF system (for comparison). Very similar VF curve and MC SP. The 13900KS runs as flawlessly in the Z790i Edge as it did in the Encore. Yeah, I also find him extremely annoying. His arrogance and know-it-all potty mouth is over the top. Truly a legend in his own mind. It also annoys me that about 95% of the time I agree with him. He is right more often than not, but that gets lost in his caustic personality.

Funny you should ask. I wondered the same. I have been wanting to test it for a long time, but haven't really made it a priority because I haven't identified a legitimate need to. Coincidentally, I have some arriving tomorrow to find out. I will share before/after temps on the Xeon in the Rampage IV Gene. It already runs cool with KPx. I have a huge tube of it. It will take forever to use it all. It had better be as good or better because only 4 grams of the Kryonaut Extreme is priced about the same as my 30 gram tube of KPx. Even if it is better, I think it is grossly overpriced. Maybe because it has Extreme in the name that makes it OK to carry that over to the price. The main reason I ordered it was to try it out on the 2080 Ti Xtreme when I install the air cooler.

Thanks. Now I know not to test Corsair TM30. Did you remove the old Kryonaut Extreme and save it "just in case" the Corsair crap was no good? Still using that waterblock I gave you? It's a great block, isn't it? It worked extremely well for me.

Looks like the Orwellian nighmare our brothers are dealing with in UK's government is angling toward overstepping jurisdictional authority and looking to work its wicked ways into American life. These creepy clowns are insane. I hope the people of UK can remove their facist government before it is too late.

So, now the DNC is using AI (among other equally deceptive tactics) to make things appear differently than they actually are. The Deep State is frightening enough without the AI-driven fabrication. This is something every American voter needs to be aware of regardless of what side they are on. If you don't like this guy, set aside the bias for a bit and pay attention to the information. AI definitely has the potential to destroy us all. Newer is always better, right @Papusan and @Raiderman? They're even paying thousands of dollars to young people to post propaganda on YouTube, TikTok and Twitch. Listen to some of these young influencers talking about being bought by DNC PACs. Most of them are very upset by the effort to "buy them" (who wouldn't be).

LOL... totally agree. One of the most worthless PC tech YouTube channels. Not only because of his irrational fanboyish tendencies, but seems to promote some of the more outlandish rumors (like that one about the APU) and also leans toward exaggeration and pole vaulting over mouse turds. Other than what seems like an anti-NVIDIA and anti-Intel agenda, he seems like he might be a nice person otherwise. For the record, I tend to not value any of the YouTube channels that promote speculation and rumors, even when there is no brand bias or agenda. Graphically Challenged is another one that falls into that bucket, although not nearly as jaded as Gamer Meld. Seems like a nice guy, but spare me all of the what-if, maybe and some people are saying garbage. Don't care. Show me the real hardware and the real numbers, or hush up and go do something else.

That's great news.

That seems a bit far-fetched since it uses shared system memory and that would be much slower than the 4070 VRAM. I guess we shall see if/when it happens.

Just a suggestion, but I would recommend moving away from the opportunistic boosting of the preferred cores and lock them all down to what is stable and adequate voltage for the "not-preferred" cores and then the better cores will be cruising on easy street. I think the preferred core boosting is part of what created the mess they are sorting out.

Sometimes being future proof has the undesirable implication of holding onto status quo and a normalization of "meh" products. That is looking like what AMD is serving for dinner now. Socket longevity can be good or bad, depending on what happens (or doesn't happen). Sometimes moving onward and upward is better, but we see plenty of examples that newer is frequently not better. This just means that taking a "wait-and-see" approach is always the smartest way. Much smarter than listening to marketing propaganda (lies).

Yes, I agree. It should address it fine. If you go back to the early overclocking threads at oc.net and ASUS forums, Falkentyne and Robert Sampiro threads, they were saying setting IA VR Max to 1700mV to limit spikes, but that is too high even if it did come from Intel. You might even be able to set lower, like 1450mV or even 1400mV, depending on what core clocks you are setting. It will be interesting. Now I just have to muster the desire to mess with it. I always love it after diving in, but as I get older I find I dread the idea of diving in before I do. Maybe because I know the rewards are smaller than they used to be. I'm defintely a reward ≥ effort and cost for something to be worth doing type person. That is part of the reason I no longer have any interest whatsoever in making turdbooks less turdy. Huge effort with minimal reward. That sucks. In the end it is still a compromised pile of crap. I believe you are totally right. There are way too many 14th Gen silicon lottery losers to set the voltage limit lower. There would be too many that need more than 1.500V or need to be clocked under spec to use less than 1.500V. Update: Gigabyte RTX 2080 Ti Xtreme Waterforce conversion to Xtreme (air) Received the air cooler and backplate from AliExpress today. I do not have the 12-pin to 4-pin adapter cable they sent in a separate package yet. I will need to wait for that to swap out the fan wiring harness before installing it. These parts are brand new OEM, not used pulls. This will be a massive upgrade over the GTX Titan Black (Kepler) GPU in the legacy SFF build. They even included a package with all of the OEM screws and stock thermal paste.

Exactly. And, safety aside, using 1.500V to 1.550V on a water cooled CPU just isn't feasible. It will get too hot and thermal throttle with voltage levels that high, even with a delid and bare die. That just creates too much heat to remove through such a small die. Contrary to the modern myth, shrinking die size has some negative strings attached. Even using chilled water and bare die cooling a CPU running that much voltage in a workload like Cinebench is no small undertaking. This should explain why performance is not being impacted. If someone owns a CPU that needs 1.500V to 1.550V to run stock or eTVB boost clocks it is a silicon lottery loser than needs to be RMA'd. It will never be a good CPU. I am happy that, at least for now, Intel is going to leave the possibility of overrides in place for those playing with sub-zero extreme cooling. Otherwise, a 1.550V hard cap would ruin everything for them.

That makes good sense because they were also originally saying that you should cap your VR max to 1.700V, which used to be Intel's safe maximum. That's just way too high.

@electrosoft I will have to wait for a BIOS update for the Z790i Edge, but using the Windows MC that Brother @Prema provided I can increase the VCCSA to 1.225V now without a hard freeze, so it does look like it may at least improve the SA bug. Will need it in the BIOS probably to know for sure if the SA bug gets eliminated. This is up from 1.190V, so not a lot but still more. I did test 1.250V and it will pass Cinebench without a lockup, but not AIDA64 memory benchmark without locking up. Maybe if I feel motivated enough tomorrow I will yank that CPU from the Edge and see if the SA bug is gone when installed in the Encore. Hopefully the effort to find out will not prove to be a waste of time.