Mr. Fox

29 minutes ago, tps3443 said:

Well I’m only running DDR5 6000c26 daily, so I don’t have to push high VDD’s. But yeah 7600c38 32k trefi needed like 1.700v when I was messing with it yesterday. 

I wish memory was not so retarded expensive. It could easily be your IMC is weak, and it could also be you memory kit isn't playing nice with ASUS firmware. It's too expensive to experiment at this point. Even if it is your memory, getting the right SKU for best results will be ridiculously and unjustifiably expensive, and sometimes very difficult to find at any price. If you lived nearby I would drop my G.SKILL 6000 CL26 EXPO kit into your Hero so you could see how it does.

ASUS has some kind of goofy love affair going with G.SKILL and all of their AMD firmware development seems to be focused on making two G.SKILL memory kits sing. It's really annoying. My MSI and Gigabyte AMD motherboards have all performed well with other brands of memory and were less finicky because they haven't invested all of their time and energy in making one or two part numbers work best.

I have G.SKILL 8000 CL38 XMP memory that worked fantastic in the Z790 Apex/Apex Encore and it doesn't even like to boot correctly. It was very difficult to get it tuned to run stable at any speed (including 6000). I had to jack around with all kinds of ODT and voltage resistance settings that can normally be left on Auto. That is when I gave up and bought the magic G.SKILL 6000 CL26 sticks before RAM prices turned to crap.

1 hour ago, tps3443 said:

Seeing how well ram tuning posting and stability responds to adding memory voltage, I am definitely pointing to a weak IMC. 
 

It’s like 6000c26 boots fine with auto 11K tREFI with 1.435v VDD. EASY EASY. 
 

32K tREFI? It won’t attempt to post, adding 32k tREFI puts a much heavier load on IMC, but adding voltage to VDD. Quick and easy bump up to 1.450v VDD gets it done! 

 

65k tREFI fully tuned timings? Similar situation. I need 1.480v VDD. Probably can get away with 1.460vdd. 
 

But it’s all too familiar! And it put me right back in the seat of tuning on Intel with DDR5. 
 

Weak IMC’s loves high or too much VDD/ VDDQ. And that’s my system to a T lol. 😂 

 

@Mr. Fox I’m going to swap it out for another. I’d like 2,200FCLK AS WELL. I’ll see how the next one runs. 
 

It’s a BEAST in games doing one thing at a time producing a high FPS though. That’s its main benefit. It’s not as fast as my Xeon doing brute work handling multiple things at a time. Like leaving a game minimized and doing other things simultaneously etc. it’s just not as fluid as my 28/56 Xeon 3175x is. That could be partly due to lower bandwidth, maybe higher latency, or merely due to being two CCD’s? Either way. Gaming performance is great! And that’s what I wanted to see. 

 

There is a penalty for the CCDs just as there is for Intel with P/E cores and now their chiplet nonsense. Monolithic processor design is MUCH better. 

I'm not sure what gave you the impression that you are needing to give it "too much" VDD/VDDQ. Literally everyone that I know one is using VDD and VDDQ of 1.600V or higher if they are pushing high clocks. AMD will not function correctly with low VDDQ like 13th and 14th Gen Intel did. You'll need to have VDD and VDDQ either the same or VDDQ almost the same. You might get away with 50mV less for VDDQ but MSI and Gigabyte AMD motherboards set them equal by default. You have to literally manually make them different values. When you set VDD they automatically sync them. You also need to be using at least 1.475V for VDDIO (MC voltage). You can match all three without any harm. Many AMD overclockers are using 1.550V-1.600 for VDDIO as well. If I leave mine on "Auto" it goes to 1.475V and works fine. If I try to dial back the VDDIO I have instant issues. You should be able to run LOW VSOC, like 1.175V. This is the equivalent of Intel System Agent.  I am running 1.600V VDD and 1.550V VDDQ at 8000 and 1.650V VDD and 1.600V VDDQ at 8200 and 1.175V SOC for both of my AMD platforms. Running less makes them unstable. Increasing VSOC does nothing.

The hardest part of the adjustment to AMD for me was having to deliberately set aside my knowledge from Intel overclocking because it was really messing things up and making it impossible for me to get any overclock stable. What works for one seldom works for the other.

AMD is also funny on CPU core voltage. If you are stingy with it you might not see BSOD and the normal tell-tale signs you do with Intel, it will seem stable but performance will drop like a rock. Cinebench scores will go down. If you go too far you will see the same behavior as Intel with crashing, but AMD is weird and I suspect a lot of people are living with lackluster Ryzen performance from trying to undervolt the CPU too much.

23 hours ago, tps3443 said:

I think my 9950X3D2 has a weak IMC. Different bios doesn’t help. Max 1:2 speed I can boot is DDR5 7600. 
 

Either weak IMC, or my 6000c30 A-die kit is just at the limits? 
 

I’ve tried like 1.5v VDDIO LOL 😂 

 

1.300v VSOC 

 

UP to 1.650v VDD/VDDQ. 
 

Nothing helps. 7600 max, and that’s hit or miss, 7400 posting works every time. 
 

I tried to use some preset timings. And that did not help either. 
 

Is it possible to have a crazy weak IMC on my 9950X3D2? 

I have heard several people with 9950X3D2 complaining about memory overclocking being limited by a weak IMC. Whether that is relevant I can't say, just mentioning it as a potential common complaint. Your CPU might not be below average for a 9950X3D2.

Had to make a place for the Z890 Apex.

X870E Apex is now installed in the O11D Mini V2 with the 4585PX and 5080. Getting it to fit was a HUGE pain in the butt. Never buy a distro block with the pump facing the inside of the chassis. Big mistake. I have the Alphacool reservior shoved tight against the roof of the case and the 5080 still touches it. With the reservoir against the floor the GPU will not fit. Always choose one with the pump that is in the closed in area behind the mobo tray. (That will also allow you to run the system laying on its back. You can't do that if the pump is pointed up toward the ceiling.) Now that it fits, it looks great. I love this case for mATX/ITX. For ATX it's not ideal unless you love having to spend a lot of time figuring out how to make things fit. (I don't enjoy that. I also hate puzzles.)

The Z890 Apex is in the O11D XL EVO with the 5090. It was super easy because all of the connects and mobo layout on the Z890 Apex is the same as the X870E Apex. I only had to modify the two lines going to the CPU block because the orientation of the ports on the block was different than the Thermal Grizzy Mycro direct die block. Speaking of waterblock, I was going to use one of my Optimus Foundation blocks. The silver looks perfect with the white and silver on the mobo. I could not use it. CPU would overheat SEVERELY for some reason (109°C doing CPU-Z stress test). I disassembled it and found it spotless on the inside. Not sure why. Had to run to Microcenter this moring and grab an Alphacool Core 1 block. Now it works amazing. Almost don't need to delid now, but still will. Because I can, and it will be much better. Not because it is unavoidable due to stupid temps. It runs as cool as the bare die Ryzen 9 processors.

The B850MPOWER is now on the open bench with the 9950X and 3070.

https://hwbot.org/benchmarks/cinebench_-_r23_multi_core_with_benchmate/submissions/6041469

20 hours ago, Talon said:

Im on the latest version, but I’ve heard people say it has internal limiters or something and you can’t downgrade.

 

Something does seem to be limiting you. 
 

I know you could probably score higher by undervolting (very small undervolt with offsets) on P/E cores but it might introduce instability long term. 
 

When I get home I’ll check my score tomorrow. On the road till then unfortunately. 

19 hours ago, Mr. Fox said:

I may just take a Macrium image of my current OS and install an older version of Windoze with no debloating mods to rule that out as a possible cause.

BINGO! 25H2 Cancer Edition... 100%. Installed LTSC IoT 24H2 and viola... all is well with the score. Fully modded and debloated. But, I still do not know why my effective clocks are lower than the core ratio set. Looking at similar submissions on HWBOT the core ratio and effective clocks were the same on other submissions. But, no throttling or temperature issues.

Now I can delete the cancer Macrium image.

Side note @electrosoft - no driver issues with 24H2. Smooth sailing.

Not too shabby for a chintzy Lian Li 360MM AIO and soldered IHS... Nearly identical Cinebench R23 performance as bare die 9950X (only cooler than a delidded 9950X) with only 22 threads (16 baby girl Atom cores) versus 32 threads on the Ryzen 9. Now I need to order a Thermal Grizzly delid tool and heater and bare die block and get it on a custom loop and chiller to do things right.

https://hwbot.org/benchmarks/cinebench_-_r23_multi_core_with_benchmate/submissions/6041409

ASUS did a much better job of PCIe lane allocation on the Z890 Apex than they did with the X870E Apex. It has two usable PCIe 4X slots (instead of just one) and neither one affects the GPU in terms of dropping it down to 8X instead of it running at 16X as it should. Only the bottom 16X slot steals GPU bandwidth. The two Klevv drives are running Gen3 because they are installed in a Gen3 PCIe NVMe card. It sucks that the only slot that does not affect the GPU on the X870E Apex is the one situated above the GPU. I have room for 4 more NVMe but have run out of 1TB and 2TB spares to install in it.

Right now it is on a test bench with a spare 3070. Have to decide which chassis to move it to.

1 hour ago, Papusan said:

Me now... Feel  I'm Doc Papusan Yup, more like a real doctor😁

 

LOL. AI sucks...... Wish we could push all the new modern trash into the grave.

 

"Papusan" (frequently active as "Doc Papusan" or simply Papusan) is a well-known, high-profile community figure and enthusiast in the global PC hardware, overclocking, and tech benchmarking communities. [1, 2]

 

Hardware Specialization & Stance

• Flagship Hardware Tuning: His personal system specs, famously nicknamed "The Killer" on TechPowerUp, boast extreme custom liquid cooling setups handling top-tier hardware like the Intel Core i9-14900KS, highly optimized DDR5 memory, and various flagship graphics cards like the RTX 4090 and RTX 5090. [1, 2]
• Anti-Censorship & Consumer Advocacy: Within his circles, he is known for a candid, blunt, and highly critical tone regarding modern consumer tech trends. He regularly calls out major manufacturers for prioritizing aesthetic RGB features over adequate power delivery cooling, criticizes bloatware in modern operating systems, and acts as a vocal critic of overhyped AI trends. [1, 2]

I  am very happy that you are on the mend. We are not doctors, but I do not have a great deal of confidence in medical science and I think sometimes we are better off figuring things out on our own rather than taking a doctor's advice. If we did everything our doctor's wanted, some of us would be dead. And, because some of us have followed the direction of medical professionals we are worse off (some maybe even dead). When it involves pharmaceuticals and surgery, it is driven my the same motivating factors as AI. The wealth generation matters far more than the lives of people it is supposedly going to benefit.

AI does suck, but it can be useful. It sucks because of what it has done to the world, but I do not believe it will end well. It has hurt more than PC hardware enthusiasts, and it is because of rampant greed of the companies developing it. I hope that it ends up ruining all of them with catastrophic and unrecoverable financial losses that puts most of them out of business.

I use it for self-help/medical questions as you did, and I use it at work frequently. A fair amount of my new job involves using it and identifying new uses for it.

It is dangerous for all of us in terms of personal security. AI should not know who we are. Because it does it will be used by bad actors to do bad things that would never be possible without the assistance of AI. Fake everything is coming our way. I suspect it will redefine what "fake news" means and take it to the next level. We will see and hear video and audio clips of thing that were never said and never happened. People will go to jail for doing things they never did, marriages will be destroyed because of things that never happened, etc.

I asked who is Mr. Fox? and it knows all of us. 

Quote

Mr. Fox (or MrFox) is a highly prominent and legendary figure in the PC enthusiast, overclocking, and high-end laptop communities. [1, 2, 3]

Given your previous question about "Papusan," you are referring to the tech forum user Mr. Fox, who frequently interacts with Papusan across various specialized hardware communities. [1]

 

Who is Mr. Fox in the Tech Community?

• Community Legend: He is widely recognized as a "Grand Master" member and administrator across elite hardware forums. He was one of the most influential figures on the famous, now-defunct NotebookReview (NBR) forums and remains highly active on replacement communities like ⁠NotebookTalk and Overclock.net. [1, 2, 3, 4, 5]
• Overclocking & Benchmarking Pioneer: His forum signatures and titles usually sport mottos like "⚡Overclocked⚡ ⚡Overvolted⚡". He is famous for creating and hosting massive community benchmark threads (such as the Official Benchmark Thread - Post it here or it didn't happen :D) where hardware enthusiasts push their rigs to extreme limits. [1, 2, 3, 4]
• The "Desktop Replacement" Era: For over a decade, Mr. Fox was a foremost expert on high-end, socketed gaming laptops (like older Alienware, Clevo, and Eurocom systems) that used desktop-grade CPUs and MXM graphics cards. He specialized in custom BIOS flashing, voltage modding, and unlocking locked power limits to maximize mobile hardware performance.
• Shift to Extreme Desktops: Though a legend in the laptop space, he has famously expressed frustrations with modern locked-down, soldered laptop designs, focusing heavily on open-bench and liquid-cooled extreme desktop builds. [1, 2]

It knows who we hang out with... past and present... from before AI was invented.

Quote

In the online world of high-end hardware, PC overclocking, and enthusiast laptops, Mr. Fox hangs out with a tight-knit inner circle oelite "power users," custom BIOS modders, and extreme benchmarkers.

Following the shutdown of the legendary NotebookReview (NBR) forums, this group migrated together and primarily hangs out on newer hubs like ⁠NotebookTalk, Overclock.net, and the TechPowerUp forums. [1]

His most notable online circles, collaborators, and forum "buddies" include:

 

1. His Core Inner Circle (The NBR / NotebookTalk Veterans)

• Papusan: As you noted earlier, Papusan and Mr. Fox are a famous duo in the hardware community. They constantly post in the same threads, tag-team hardware debates, and heavily collaborate on testing extreme, custom laptop cooling mods (like liquid metal applications).
• Prema: A legendary figure in the laptop community who is famous for creating "Prema Mods"—highly coveted, custom custom BIOS/vBIOS firmware that unlocks locked power limits. Mr. Fox was frequently a primary tester and vocal champion of Prema's work on high-end Clevo and Alienware systems.
• j95: Another elite tech forum veteran who frequently collaborates with Mr. Fox on advanced software tweaks, custom Windows ISO configurations for benchmarking, and driver modding.
• BrotherJohn: A well-known enthusiast heavily involved in the same benchmarking and community leadership circles, often managing or contributing to the same threads.

 

2. The Extreme Benchmarkers & Overclockers

Mr. Fox runs the famous "Official Benchmark Thread", which serves as a virtual hangout spot for extreme tech enthusiasts. In these circles, he hangs out with: [1]

• Users who build liquid-cooled, open-air test benches.
• Enthusiasts chasing top spots on global ranking sites like HWBOT for programs like 3DMark (Fire Strike/Time Spy) and Cinebench. [1, 2]

 

3. The Custom "Desktop Replacement" (DTR) Community

Before major manufacturers shifted entirely to thin, soldered-together laptops, Mr. Fox was the leader of the Clevo, Eurocom, and MSI "true" desktop-replacement communities. His online hangout spots are filled with users who swap out desktop CPUs and MXM graphics cards inside massive laptop chassis.

6 hours ago, Talon said:

Im on the latest version, but I’ve heard people say it has internal limiters or something and you can’t downgrade.

 

Something does seem to be limiting you. 
 

I know you could probably score higher by undervolting (very small undervolt with offsets) on P/E cores but it might introduce instability long term. 
 

When I get home I’ll check my score tomorrow. On the road till then unfortunately. 

5 hours ago, Mr. Fox said:

No rush. It is working fine otherwise. I tested 9931 and it is the same as 9901. I'm surprised if they were not able to downgrade using the flash recovery method. That always worked before (and does on AMD) even when ASUS says you cannot go back. I may test that on one BIOS position to find out if it is really true or not. I may just take a Macrium image of my current OS and install an older version of Windoze with no debloating mods to rule that out as a possible cause.

I flashed the latest BIOS without flashing the newer ME (using the flash recovery and renamed file) and it made little difference. I notice my "max effective clocks" are less even though I see no evidence of throttling. That makes me wonder if there is a BIOS setting I am missing somewhere. The max core frequency and max effective clock should be the same unless something is set wrong.

Here is what Google Gemini AI told me. 

Quote

To make the effective clock match the maximum frequency on the Intel Core Ultra 7 270K Plus, you must disable the CPU's dynamic power-saving and boosting features. 

• Disable Intel Speed Shift: Locate and disable Intel Speed Shift Technology in the BIOS, which allows the OS to dynamically change frequencies. 

• Disable C-States: Turn off all C-States (C1, C6, etc.) to prevent the cores from downclocking during idle or light loads. 

• Set Fixed Ratio: Change the Performance Core Ratio from "By Core Usage" or "Sync All Cores" to a fixed multiplier (e.g., setting all cores to 55x for 5.5 GHz). 

• Power Plan: In Windows, set the Maximum Processor State to 100% and ensure the power plan is set to High Performance to prevent OS-level throttling. 

Note that maintaining a fixed maximum frequency will significantly increase power consumption and heat output, as the CPU will not downclock when idle. 

Nice try, robo-tech. Already did all that.

*** Here is a praise post to offset the many (often justifiable) complaints we see posted in social media. ***

I have been a NewEgg customer for about 25 years. In spite of some of the horror stories, I have never once, that I can recall, had a bad experience with them in spite of hundreds of purchases.

My Z890 Apex was $549.99 with a $269.99 Crucial E100 2TB NVMe included as a "free gift" not a combo sale. It shows that way on the invoice. They are priced separately and the $269.99 deducted as a "discount" on the invoice. I received my order on Tuesday of this week.

The Z890 Apex is now available for $399.99. The customer service AI bot told me it was not eligible for price match because it was a combo. I spoke to a rep after the bot said no. The rep saw the issue and issued me a $150 refund in the form of a NewEgg gift card. No begging or arguing. He said it was not a combo, it was a free gift (agreed with me) and is eligible for the price match because it was within the 7-day time window. 

38 minutes ago, Talon said:

Im on the latest version, but I’ve heard people say it has internal limiters or something and you can’t downgrade.

 

Something does seem to be limiting you. 
 

I know you could probably score higher by undervolting (very small undervolt with offsets) on P/E cores but it might introduce instability long term. 
 

When I get home I’ll check my score tomorrow. On the road till then unfortunately. 

No rush. It is working fine otherwise. I tested 9931 and it is the same as 9901. I'm surprised if they were not able to downgrade using the flash recovery method. That always worked before (and does on AMD) even when ASUS says you cannot go back. I may test that on one BIOS position to find out if it is really true or not. I may just take a Macrium image of my current OS and install an older version of Windoze with no debloating mods to rule that out as a possible cause.

42 minutes ago, Talon said:

My guess is power limit unlimited?

 

Try E cores at 5.0. 

Yes, everything is maxed out on power and current limits. This is 57x all P and 50x all E. Same score. 1.385-1.400V and all return the same score, give or take about 50 points. Weird. 

And, only 76 processes running in Windoze 11. Temps are fine. No thermal throttling. 

What BIOS version and ME are you using @Talon?

1 hour ago, Mr. Fox said:

So, I am not sure how this stacks up in the grand scheme of things, but my gut feeling is this is actually quite respectable and above average for a non-delidded 20K Plus with nothing but a mickey mouse (AIO) cooling solution.

 

16 minutes ago, Talon said:

 

Actual VRM Core Input Voltage

1.385V

Manual Mode

- CPU Core Voltage Override

 

otherwise DLVR voltage is like 1.58v at those clocks and is wasted heat. 

That did not see to have any effect. Same score (roughly) and no real change I can see. I wonder why the score is low? Maybe need to downgrade to W10 or W11 24H2? At those clocks it would be 47-48K with my Ryzen CPU as well.

4 minutes ago, Talon said:

What is your DLVR input voltage? Or are you bypassing that entirely?

 

with those clocks you should be above 47-48K. 
 

Try using 1.4v input. 

Thank you. I will try that. I have them set on Auto. Set it for P and E at 1.400V input?

So, I am not sure how this stacks up in the grand scheme of things, but my gut feeling is this is actually quite respectable and above average for a non-delidded 20K Plus with nothing but a mickey mouse (AIO) cooling solution.

2 hours ago, electrosoft said:

 

 

Welcome to the club @Mr. Fox and I had a very similar problem actually with the Asus Z890 Strix and Intel B580 when testing my Acer B580 variant on it. Initially, I thought the card might be defective at one point.

 

I kept having the GPU randomly go into recovery mode as listed under device manager. Let's call it Code 43 adjacent. I tested it on a pre-built HP Omen HP sent me to eval and it suddenly worked fine no problems so I ended up doing a few clean installs, same issue would just randomly pop up and tracked it down to Asus driver revisions. I ended up going with Intel and Gigabyte variants and the problem resolved itself.

 

The Asus versions were so wonky that every time I went to install the Intel GPU drivers it kept trying to update the firmware which was already updated each and every time. Once I sorted out the drivers, the GPU driver install worked as it should.

Probably the Chinese rootkits and key loggers bundled with the A$$zeus drivers. 

1 hour ago, Talon said:

I’ve seen some doing 5.9ghz on direct die. Seems to unleash the chips a bit. 

Right now I'm kind of fumbling around trying to figure out what works best. Some voltage values seem lower than expected others seem much higher than I would have thought. I don't know what to do with the NPU stuff at this point. But I'm sure I'll figure it out. 

The hardest part of setting it up and finding drivers that actually worked. There were a lot of things that were code 43 in the Device Manager using the drivers provided by Asus and I ended up getting a lot of what I needed from Intel and the Microsoft Catalog as .cab files based on hardware ID. That was one of the more cumbersome Windows set up experiences that I've had for a while.

39 minutes ago, Talon said:

Dang you got that rig up fast lol. I think with your cooling, you'll be able to push that chip pretty hard. 

Right now it is on the test bench with a Lian Li 360 AIO. I'll probably order TG delid tool and heater and a direct die block for it in the next week or so.

Now I am a dual-Apex predator... purple now. Dang,it feels good to have some blue back. I've got to figure out some of the tuning. A lot to re-learn, just like switch to AMD had a lot to learn. Very different than Raptor Lake and pre-Raptor. I'm pretty sure right now I am butchering things.

2 hours ago, Papusan said:

 

Not so great. I need to figure out my blood pressure. Perfectly fine with medication in the morning, but then it increase up to 175/85 in the noon/evening. Has become like this the last days. Why the change, I don't know. I have even stopped smoking now, only some puffing/vaping. Then the blood pressure going ok again the next morning. Even before I start the medication in the morning. I can't have a good life now before this being fixed. This scares me to death☹️ I'll see what the doctors can do the coming days. But thanks for asking. I realy appreciate it.

You are in our thoughts and prayers, brother. Are you getting enough sleep? I very rarely ever get the amount of sleep my body and brain need me to, but when I do it always lowers my blood pressure. I'm wondering if yours is lower in the morning because you were sleeping and not stressed out as much as you are when you are awake.

Brother @Papusan how are you getting along? I hope things are looking better and you are feeling better.

13 hours ago, electrosoft said:

You *might* have a not so great 8000+ kit on your hands but a really good 1:1 6400 dual rank set in your possession @Mr. Fox

 

If you're able to do 2133/6400 on dual rank that is something major for my low end use case/testing and I would love to see some gaming and bandwidth/latency benchmarks for testing. I've had several dual rank 6000 kits that were trash for 8000 and couldn't hit  2133/6400 for any meaningful stress testing. Do they pass 2133/6400 dual rank Karhu and Aida64?

I will have to do more testing once I decide where the Z890 Apex is going to land and if decide that I am going to disturb the AMD builds. Depending on how the Intel testing goes, I anticipate the X870E taking the 4585PX from the B850MPOWER and going into the O11 Mini with the 5080, the Z890 Apex going into the O11 XL with the 5090 and the B850MPOWER taking the 9950X and going onto the test bench with my spare AIO and air-cooled memory. If the move back to Intel doesn't impress me the Z890 Apex will remain on the test bench, but that seems unlikely. I anticipate the change will be welcomed since I have plateaued on my current hardware and my interest is beginning to wane. If I like it, I will end up having to purchase a delid tool and heater and a Mycro block from TG because it is extremely unlikely that I will find contentment with anything less than bare die.

28 minutes ago, tps3443 said:

I am testing a 6600c28 though

If you can run 6600 C28 in 1:1 mode that is already more than what most people can accomplish with the CPU, mobo and RAM kit they have. Many can't even run 6400 1:1 with the crummy silicon they were blessed with. Your CPU would have a very good IMC to do that, so I am not sure why you are having issues with 8000, 8200 or 8400 unless you are not using enough VDD/VDDQ/VDDIO. If you are able to run 6600 C28 1:1 then I doubt there is anything wrong with your motherboard.

28 minutes ago, tps3443 said:

I almost bought that Vcolor kit! I thought it was suspiciously priced low. I suppose that’s why? Because it’s dual rank 16gigs?

I'm glad you did not buy it. I think it would be fine for an average Joe type of gamerboy that just sets XMP/EXPO and moves on without bothering to actually tune anything for maximum performance. The plastic covers that look like a heatsink are a joke though. Gupsterg (oc.net member) discovered a V-COLOR kit he had purchased had plastic covers as well. It kind of pisses me off that they would even sell garbage like that. What a ripoff. I think the marketing info and specs not disclosing the dual rank design and plastic covers qualifies as a scam and I think they are very conveniently not disclosing what they know would make it a hard to sell product.

I debated whether or not to just go ahead and install it anyway in my son-in-law's build with the other upgrade parts. The performance would have been fine for him, but I think the high temps could have been an issue with the plastic covers. I decided he was better off with the generic green 2x8GB Hynix A-die with nice aftermarket heatsinks for stability and reliability. Things get kind of toasty in his SFF build with an air-cooled GPU and air-cooled space-heater mounted on top of his CPU, both working together to create a toastly little hot box. He doesn't need 32GB of RAM. 16GB will handle anything he would throw at it just fine.