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

I cleaned and reposted everything and no display. The power button just stayed white . I tried NVRAM reset and that didn’t work.

I took the CPU out of the socket and noticed a very very small chunk of hardened thermal paste under one of the pins. I cannot get a good picture because of how small it is. None of the pins look bent, but how the hell do I get this piece of hardened thermal paste out?

 

You can use a small paint brush dipped in alcohol. I’ve dropped thermal paste on cpu socket pins before a few times.

To remove the alcohol, just blow the cpu socket. I’ve done this on numerous boards. Never had an issue.

 

Alright, I got it out. I booted and everything was fine . I should have probably checked on that before putting the CPU back in whenever I took it out...

 

You should see the power cable on my 1600w T2. It's insane.

 

They are both C19’s. This was one of the main selling points for me on this unit. Other titanium/Platinum Evga power supplies that are 1200-1300 watts don’t have a C19 sized power cable. Evga decided to give the new 1300G+ the same sized power cable as the 1600 units.


For $210 it is an incredible value.


The PSU is made in Taiwan as well. Build quality is literally amazing! I like it so much better than my Seasonic Prime Platinum 1200.

I couldn’t afford the 1600 watt model. But, now I do have that full EVGA build I’ve always wanted.

 

12900KS exclusive to pre-builts only? Ugh, I hope not.

https://hothardware.com/news/intel-core-i9-12900ks-alder-lake-cpu-55ghz-stock

"Speaking of which, it sounds like the Core i9-12900KS will be exclusive to Intel's OEM partners. If so, the only way to get one would be to buy a pre-built system. We'll have to wait and see, but fingers crossed it will be offered as a standalone chip too."

 

Wow that would be complete BS. Another way to milk consumers I suppose. Just read that EVGA changed their step up rules and their extended warranty. While I am fine with a moderate increase in price, they went full 400% increase in price, and reduced maximum coverage from 10 years to 7. Yeesh, what's next?

 

They increased the extended warranty price too. Apparently GPU’s over $1,500 extended warranty is $240 and up now. Which, it was maybe like $50-60 bucks before?


I kinda wish my GPU had the extended warranty. Especially if I decide to just keep it for years and years as a back up card.

 

So, my benchmark numbers have gone up with my new power supply. I know my PC like the back of my hand.


I ran 3DMark Profile test with the same settings. And I beat out Sugi0lover’s 11900K, (Which was a GEM)

I jumped from position 98 to 87.

http://www.3dmark.com/cpu/348343

 

I'm sure many of the upper OC'ers have a lot of them and bin as Hell now. They did that for the 12900K as well. I wonder how many of the chips with lesser bins will reach out on the 2nd or 3rd market if most of the chips is for the OEMs.

Referesh page after your posts.
http://forum.notebookreview.com/thr...e-ltsc-2019-2021.833196/page-47#post-11136338

 

https://valid.x86.fr/t9sh5x

 

As of June 2021 they also eliminated their "advance RMA" feature. They still have the cross-ship RMA option, but they charge you full price for the replacement and refund after the part being replaced is returned. The "advance RMA" was an extra fee paid at the time of purchase to avoid being charged and waiting for a refund in the event of an RMA. That sucks, but it is still better than ASUS because ASUS doesn't cross-ship. BUT, I also found out that because I did not purchase my Z590 directly from EVGA that I am only entitled to "standard RMA" which means send it and wait for the replacement and be without a computer while that happens. Same as the deal with ASUS., so that sucks. I guess going forward I won't purchase anything in the EVGA brand unless it is directly from them. Otherwise, might as well save money buying something else with a crappy RMA process. So much for getting it swapped for the failed LED Q-code diode failure. I guess I won't bother with it.

 

https://hwbot.org/submission/4898936_mr._fox_wprime___32m_core_i9_10900k_1sec_800ms/
https://hwbot.org/submission/4898934_mr._fox_wprime___1024m_core_i9_10900k_50sec_995ms/

 

https://www.3dmark.com/3dm11/14782396


Second place is 5.7GHz, so I don't think I will be going that far. Brother @Johnksss was that on phase change or cold water? Awesome Physics and Combined scores.
https://www.3dmark.com/compare/3dm11/14190754/3dm11/14423748#

 

Thanks, i was looking at both types, but the water block would be a hassle.

Single stage.

 

Is that the same memory you had before or did you replace it with something better? Nice scores brother.

 

Intel 13th Gen Raptor Lake CPU Surfaces With 24 Cores, 32 Threads tomshardware.com

More E-cores en route. I wonder how this will work vs AMD's Zen 4 5nm Ryzen 7000 series that may launch at same time coming fall. It seems Intel put all the ammo on the E-cores and higher IPC. Maybe we will see an chips also with more P-cores?

@Johnksss nice see you back. What you think about 12900K vs Ryzen 5950X ?

 

Hey brother. It's the same memory from before, but finally figured out how to get it to act right.

Thanks!

Thanks!
Still trying to gage it, but does seem better at a few benchmarks over the AMD, (Which I still have. Just got rid of the board) the 5950X still has it's place in my ranks.

 

What did you have to change to get it to behave correctly?

 

I had not fully understood how the new memory worked so I was not using the right amounts of voltage. I was just assuming it would auto set it self.

 

Brother @Rage Set if you are still looking to buy a chiller, you can get one for a fairly decent price here. This is like mine and the one @tps3443 just bought. The face panel looks newer, but same basic model.

https://www.sea-of-green.com/active-aqua-chiller-1-2-hp-boost.html

 

I have the 1hp version of this. got it for 1100 new along time ago.

 

Nice scores, bro. Did you have any issues getting Windows 7 installed on Z690? The only way I could do it on the Ryzen X570 system was from DVD and using PS/2, then later installing USB drivers and what-not. I never could get it to work trying to install from USB. Slipstreaming the drivers to USB W7 installation media never worked for me. It was really easy on Z490.

Now that I have a PS/2 mouse and keyboard and SATA DVD (which I did not have before) installing it is easy on Intel and AMD, but it was a real battle trying to use USB. I did it that way on the Z590 as well just because it was less hassle.

My NVMe ports don't work with Windows 7 on the Z590 Dark. The PCIe ports for the NVMe slots have code 43, so I have to disable the M.2 ports in the BIOS running W7 or it has tons of BSOD problems. I am really temped to go back to Z490 because of it. I wish I had stayed on Z490 now. Life was better for me before X570 and Z590, LoL. Both brought nothing but regrets for me.

 

This was delivered today. It was $290.00 with shipping and taxes. It weights quite a bit. So, the seller couldn’t have made much more than $120-$150 bucks after shipping cost.


This is a 1/2HP chiller. Unfortunately it didn’t turn out to be the 1HP in the photo the seller published. But, I’m good with that.

It looks very clean inside. It looks like shipping cracked the front grill one little spot.

But, the unit is very clean inside! And, it also looks to be unused almost. It’s from late 2015 era.

This is incredible for $250! I have wanted one of these for years!!


the exterior comes clean with 100% alcohol. So, I can wipe it back to clean condition.

I am going let it sit for 24 hours, and try to hook it up to my system, after flushing the tank out.

 

Other than the adhesive labels/branding and white colored plastic, that is identical to my Hailea HC500A 1/2 HP chiller. They must be manufactured by the same company. You can probably fix that crack in the grille to make it nearly invisible. The outside panels are very easy to remove.

 

Thanks! still working out a few things with this...

Yes, mine works for installing windows 7 on my Z690 and it worked for my Crosshair VIII Extreme.

I also still have my PS2 PCIE adapter. Now I would like to have that with USB 3.0 ports.

Hummm, I have never use the NVME ports on board. Only add in cards or the Dimm.2 slot.

 

I think the issue with the M.2 slots not working under Windows 7 might be an EVGA engineering mistake. I don't have anything else to compare with, but I think it is their implementation (or lack of implementation) of ACPI. I may just buy another Z490 Dark, RMA this one and sell the replacement that EVGA returns. There are too many things I do not like about it. The Z490 Dark worked flawlessly for me on any OS.

 

So with an unaltered copy of Windows 7 you still get error code 43?

 

Yes, installing from the stock DVD with no modifications. In fact, it will throw tons of 0x05 BSOD while installing Windows 7 if I do not disable all of the M.2 ports in the BIOS. After installing Windows 7, if I re-enable them it will boot and show code 43 on PCI Express Root Port #9 and #21 (which are for two M.2 PCI-e 3.0 ports) and the NVMe SSDs are not visible. However, it still throws a BSOD during shutdown every time. If I disable the M.2 ports in the BIOS it works fine. And, if I use a PCI-e add-in card in one of the slots, NVMe works fine. It is only the onboard NVMe ports that are screwed up. And, they all work fine in Linux, Windows 10 and Windows 11.

I also checked with the M.2 slots vacant, thinking there was something with the NVMe drives and it does the same thing even if nothing is installed in the M.2 slots.

I messaged back and forth with canonkong (a firmware/driver modding expert) on win-raid forum and he said he thinks it is something EVGA did with ACPI that is not recognized or supported by Windows 7. Everything else except that seems OK. Here is a snippet from that PM.


If I install Windows 7 to NVMe in another system, if I attempt to boot from NVMe in the Z590 Dark it throws the same 0x05 BSOD attempting to boot. Everything is fine in the BIOS, it is only in Windows 7 that everything goes nuts.

All good when disabled in BIOS...

And, when enabled in BIOS...

I just bought a new factory sealed Z490 Dark on eBay and initiated the RMA, so I will sell the Z590 Dark that EVGA sends back as a replacement for this one with the Q-code diode failure.

 

Now that you mention it, I never upgraded to the Z590 board. I just used my Z490 board, but it was able to run PCIE 4.0 with 11900K CPU.

Wow, that sucks big time!

At least you were able to find a "new to you board" that you know works with Windows 7. My Z690 Extreme has lots of missing drivers at the moment. wish I could get those to work correctly.
.

 

Yes, I got lucky to find it, and the price was decent. If I could go back in time and do it over again, I would still be running that setup. Selling it and moving to 5950X was a terrible mistake and started an avalanche of bad luck for me. The Ryzen 9 experience was frustrating and disappointing. I should have trusted my gut and ignored it, but the CPU benchmark scores were tempting enough that bad judgment got the best of me.

I think I still have all the original files and stuff that I was using on the Z490 Dark that I sold to Brother @tps3443, like the XOC system BIOS saved on a HDD. I didn't really see any point in 11900K. It can't keep up with 10900K in an all-core scenario. Has a little better IPC on single core, but missing 2C/4T and memory overclocking is kind of goofy with the Gear 1/Gear 2 foolishness. The new 11900K that I sold to @tps3443 turned out to be a superior silicon sample, so I am glad he got it. It is serving him very well. Even had I known it was so great I still would have sold it because I still look at it as a minor performance downgrade. It's too bad it wasn't 10C/20T... then it would have been a minor upgrade.

 

I don’t think the 11900K is a down grade to the 10900K. I just don’t think it’s an upgrade at all. Which really made no sense on Intel’s behalf again lol. The IPC is definitely there. And you can certainly make the 11900K compete with a 10900K. But, doesn’t really make sense either.

I just don’t understand why Intel released something with 8 cores to just barely match something with 10 cores in multithreaded applications. Clock for Clock the 8 core 11900K can only amount to about about 96% of the 10900K’s 10 core performance.

So, I’m really not sure why Intel released a so called “flagship” processor capable of only 96% of the 10900K. That 4% is quite large in the end. Your gonna need a 5.5Ghz all core 11900K to match the whole 100% of a 5.3Ghz 10900K. And how many 11900K’s can run 5.5Ghz?

Now, if you got a 5.4Ghz or 5.5Ghz 10900K. Then, I certainly don’t think there is an 11900K out there that can even compete anymore.

While this 11900K is just astoundingly good. Not only is the silicon golden, but the IMC is golden too. However, even at 5.5Ghz it’s only going to just match an average 5.3Ghz 10900K in multithreaded loads. Which kind of brings us back to why did Intel even release it lol. And, I have no idea why.

It sure is a fun CPU. But I think Intel should have made them all as good as this one is. And just called it a 11900KS. Because, for gaming alone and IPC gains the chip certainly makes sense. But that’s about all it makes sense for. But a 10900K running 5.4-5.7Ghz certainly does just as well.

 

I Was not impressed with the 11900K at all. I should have just kept my 10900K, but I know where that chip is sitting so if I wanted to swap it out at some point I could. Right now it's just running web browsers.

The AMD over clocking experience was a mixed bag. Totally forgot a lot of tips and tricks when using it, but it's a great chip for multitasking and vmware sessions though. The chips seem to max at about 4.875 on anything hotter than say 60C under load I think it was. Something like that. So with single stage it would allow for the chip to run higher than 4.9 ghz stable.

The memory was a a whole number of issues re learning. smh. I may get a lessor board and re task that chip to a server or vm workstation machine....still up in the air.

 

Intel could implement 25% more cores but that would reduce the profits and the all core boost would drop to match specs (Equal for AMD). Remember 11th gen die is huge vs previous and 12th gen chips. And it would screw up their binning process.

 

I think that’s why the Delidding is so
tough. That’s a lot of surface area for solder to stick to

 

Personally, I'd happily take the increased single core performance at the cost of reduced cores. For the vast majority of use cases, single core performance matters more.

Intel's big-little approach on the desktop side makes sense for this reason. Applications that love single core performance do not scale beyond 8 cores, and you want each of those cores to execute tasks as fast as possible. Applications that benefit from lots of cores will be well served by lots of small cores since throwing more and more cores at them keeps increasing performance. Therefore, having 8 high performance cores and many more efficient cores makes sense since you have 2 sets of cores, each of which is good for the type of task it's engineered for. This makes the CPU good at everything.

Of course we'll have to see how this all turns out in practice. All big cores is definitely the way to go if there are no limitations restricting you from adding more.

 

I think for many of people that are keenly interested in the subject of this thread the most important thing is overclocked benching performance on more than a couple of cores. The majority of use cases don't even require a powerful CPU. That's why turdbooks can do an acceptable job for most use cases. As long as they have enough RAM, an ordinary quad core CPU and a snappy SSD, they can do an acceptable job at almost any common use scenario. Single core performance is certainly important, but it doesn't get you very far in the sport.

 

I’m flushing out the waterchiller. It looked to have some sort of brown slime inside that comes out looking like wet leaves almost. It was certainly used in some sort of hydroponics growing setup, maybe even a salt water fish tank. The brown stuff could be old algae. I flushed with cold water to see how it would go, and a good bit of stuff came out. The water was clean. Only it has little pieces in it. So, I went with hot 140F water. I filled the tank and put a hose in it, and started blowing to get the water sloshing and mixing around to work anything loose. I did this about ten times over and over again. I shined a light in, and it wasn’t quite perfect so, I decided to just go with a strong 70/30 vinegar mix. (With hot water) letting it sit for now. And, I will flush it like crazy after it sits for 30 minutes.

I don’t know what was circulating through this tank, but it’s gotta be debris free.. Or else the Optimus Sig V2 will collect every single piece. (One of the annoying parts about that cpu block) you’ve gotta have a water loop in top top shape for a Optimus Sig V2, or your cpu temps and water flow will both gradually go down hill overtime.


Should have the chiller up and running tomorrow though.

 

1/2HP Water chiller is 100% clean inside! I filled it up with distilled water. And, it’s ready to get hooked up. I used vinegar, bleach, hot water. And just flushed it to death. Then Flushed it some more, after flushed it so much until I was certain it was clean of detergents. I flushed it more with distilled water.

It’s glistening inside of the inlet and outlet within the tank.

 

Ghetto rigging at its finest.

 

Intel's Big little have internal middleware, which is a solution for the problem Intel created. Intel Thread Director. This is why Denuvo DRM for instance needed to have it updated. And same reason why E cores do not get used in gaming performance as well. Because it will slow down the processor, enabling more E cores = more hops to ring bus which reduces total P core performance.

I certainly do not believe that 8C is the peak mark for the ST performance back when CFL was out Steve from GN was like 9700K is best gaming processor, they recommended a CPU which did not have Hyperthreading at all look at that processor now, it gets hammered to ground. If 8C peak was the case 10900K would not have been chart topper on many benches and how 11900K loses out to it, it should have been opposite. Same for 5900X which AMD touts as fastest gaming processor for Zen 3 generation. Yeah its binned better than 5800X with higher 100MHz boost clock but base clock is 100MHz less than 5800X and has more heat density on top, but still it has the best gaming performance for the whole Zen 3 stack and has solid SMT performance too. Also most important factor is Cache. Zen has huge cache which offsets their poor IMC, MCM design and slower Uncore.

And the new Intel approach of biglittle adds more power consumption to the processor, this is why many disable E cores and rake up points at HWbot in the 8C/16T CPUs category, as the power plane is shared with P and E... once we disable E, P can shine brighter in clocks, cache, memory, temperatures. P cores are big power house of this design not E cores.




Also E cores are used in the SMT cases in a different way than a homogenous CPU does. Anandtech already explained clearly how the Work is allocated in the Processor.

They are splitting workloads, unnecessary step if you ask me. Also now because of this garbage, AVX512 is gone from the ADL processor officially. And RPL leaks show it won't have AVX512, with Igor's lab testing AVX512 enabled ADL shines more, no wonder they are disabling silicon in the damn CPU itself for this pathetic E Core drama as shared by @Papusan .

We do not need this ARM garbage solution in DIY market at all, why do we even need the E cores ? There's no need for efficiency or other crap, Intel Speedstep and Turboboost 2.0 and 3.0 are super advanced technologies, they ramp up and ramp down as we need on demand, plus more over Intel themselves admitted they cannot put more P cores because of the heat, ADL 12900K is already a very high heat dense processor.

If they put 2C P core more in that it will melt they have to reduce clocks if they do they will lose out to AMD.

Sadly they are going to double down on the E-Cores in Raptor Lake, and more in Meteor Lake just like pre Skylake era of doing same darn 4C design with mediocre improvements to adding more E cores... And they get away with gaming performance from the 8C design due to IPC boost in the new Golden Cove successor / P cores. Future is really made for BGA trash and mobile junk where they can advertise this more core formula in reality giving second class cores, pay for i9 12900K get more E cores than 12700K LOL...

I have one hope - Zen 4. Lisa Su already is challenging this Alder Lake 12900K in sheer gaming performance flat out as "World's fastest . A 5800X3D beats both 5900X and 12900K, that's an 8C16T CPU with cache boost vs biglittle hybrid processor with heterogeneous design and boosts beyond 5GHz with 16C24T on box sticker with DDR5 Quad Channel memory on top. I'm curious to see how the 3D-V Cache effects the SMT performance with the reduced clocks.

 

Putting hope in AMD is scary. It took me over a decade to forgive their incompetence. Imagine how disappointing it was to find them still selling buggy broken trash. I'm not sure that I'm actually even capable, much less willing, of giving them another chance at this point. The only real purpose I can identify that they serve is to help keep Intel and Nvidia kind of sort of maybe a little bit honest. Maybe they could start to earn back a little respect by getting rid of Lisa Su and replacing her with someone like Luumi, Kingpin or der8auer. And stop selling broken trash. Benchmark scores are important. Very important to me anyway. But, having a system that isn't marred by malfunction is slightly more important. I am not sure AMD is capable of delivering on that. Seems like they can't to me. I think they all suck at this point, but sucking less is preferable.

 

Now you need to add a QDC fitting on the line that is currently red (the "out" port on the chiller) and one on the "out" side of the radiator, with the male/female parts placed in opposite directions on each line. When you want the water extra cold, quickly disconnect the line going from the chiller to the radiator and the line from the radiator to the CPU and connect them together (which is why you put the male and female in opposing directions on those two lines). This will bypass the radiator so it doesn't "warm" the cold water.

 

Getting closer. I had to use the newer UI so it would allow me to get on the boards again..... Getting closer to that magical 16K

 

They disable the E-cores to be able to grab most possible Global points for GLOBAL 8X CPU RANK with an binned 16 cores Cpu: Gulp. Put all cores in work in 2D benchmarks has lost some of its meaning. But a nice way to grab most of the global points from the AMD Crowd running an real 8 cores Cpu. Or push all 8 cores Cpu's down to the utter bottom in the 8X Cpu leaderboard. Also the 12 cores 12700K will help on that.

Also MSRP has lost a lot of it's meaning/value.... Nvidia now washing their hands and don't want to be the dirty one. Smart move, let their partners determine MSRP/proper price point.

Don’t bother trying to hunt down a Founders Edition version of this graphics card. Nvidia told PCWorld to contact board vendors (like EVGA, Asus, MSI, et cetera) for more product information—including pricing. Yes, Nvidia didn’t even slap an official MSRP on this bad boy. Gulp. The original 10GB version ostensibly retailed for $700 but is currently impossible to find and going for $1,500 to $2000-plus on Ebay depending on the model thanks to the ongoing GPU shortage. Don’t expect it this upgrade to come cheap.
https://www.pcworld.com/article/564...-3080-packs-more-memory-and-a-bigger-gpu.html


MSI GeForce RTX 3080 12GB cards already on sale in Germany at 1699 EUR videocardz.com

The MSRP of the new 12GB version of the RTX 3080 SKU is not yet known. In fact, there is no confirmation that NVIDIA will even list the price on its website or communicate it to the media, as there won’t be a Founders Edition model. The upgraded RTX 3080 cards will only be available through board partners, who are reluctant to speak bout their inflated GPU pricing.

 

EVGA new Queuing system 3.0 is upon us (opens up tomorrow):

https://www.evga.com/support/queue3/