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

Well, that's good. I did not remember it being that high. I thought it was like 1.5W/mK. That being the case, there is no reason to worry about finding that Ziitek compound. K5 Pro will be just as good or maybe even better. I like using it, and applying it with a syringe is neat/tidy (no mess) and super easy.

 

Ah, okay. That's what I was talking about. The stuff Luumi is using. Although, quite a few of these products look like they are on many of the factory laptops I have seen. And some of it even looks like Fujipoly...

Never remember EVGA actually taking an iron to a pcb for repair and then send it back. They always replace it. at least that was always my take on it. I remember asking for repair on a 480 and they couldn't do it. I got a refurbished one instead, but that was a long time ago.

 

Looks like a whole relearn to get this stuff to act right....

 

Does it feel strange to run CPU and RAM at such slow clock speeds?

 

Not sure about the cpu yet, but the ram is killing me. Best time is like 56.2 in aida64 while on intel it's 36.7. When I first booted and it was 127 WTH!
So far, I have only used auto everything and no tuning.

 

56.2 is an improvement over my 3960X and other last gen Ryzen procs. The lowest I've gotten was in the low 70's. I'm curious to see how far you can get on the clock speeds.

 

Which AIDA64 benchmark? Are we talking about memory latency?

Yes, that should be interesting. It seems like most Ryzen overclockers can only achieve average Intel-like clock speeds using sub-zero cooling. I am curious how far it will go with phase change cooling.

 

On stock auto it wants to act like it can hit 5.0 ghz. With an average of 4.6 ghz. Looks like story telling to me.

Aida64 V6.25.5400
Yes, memory latency.

Won't be doing that for a while. Not till I have an understanding of how AMD works. They pretty much have different names for everything.

And you are right, I haven't seen very high over clocks of AMD for quite some time.

 

Okay, doing a little better. Those whea errors were killing me.

 

The longer I own one, the more meh GeFarts 30-series GPUs seem to be. This temperature-driven core clock castration is absolutely bull feces. Whoever designed that dumbass abortion needs to have their face(s) rearranged with a ball peen hammer. It doesn't make a damn bit of difference where you set the multiplier, how high you go on the voltage or what the power limit is. The core clock has its nuts cut off and is capped based on temperature. I've never been more convinced that NVIDIA engineers are stupid bastards. With 3DMark 11 I can run the benchmark over and over and it caps out at 52654 because the core clocks drop to 2190 based on NOTHING other than they've set that clock to coincide with the temperature... total nonsense. That sucks six ways to Sunday. I think this is the last GPU I am ever going to buy. I am so tired of wasting money on broken trash tech. This bullcrap really pisses me off. If set the core for 2250 it needs to run 2250 until it overheats and shuts down, or crashes if I don't give it enough voltage. I wish I hadn't wasted my money on this piece of trash... hard to believe it's the best 3090 money can buy. I can't imagine what kind of a piece of crap a normal one would be. Maybe we'll get lucky and NVIDIA's Falcon will get loose from its "security" cage and peck all of their eyes out in their sleep. The dumb-dumbs that drink the 3080 notebook Kool-Aid are going to get screwed so bad.

 

I'm pretty sure it's not the gpu that is killing you in 3d11, it's your combined score.... That's just my opinion.

 

No, I am talking about a firmware castration problem that is all benchmarks, all OSes. 2235 needs to be 2235, not 2190 or 2175. I can set it for 2250 and it runs 2190, or 2175, depending on whether it is above or below 25°C.

My combined score is good. Everything looks good except for the idiotic GPU core downclocking bullcrap.

 

hummm, ok.
https://www.3dmark.com/3dm11/14190754

But I do understand what you mean by the temp/clock situation which isn't going away any time soon.

 

Don't you think that your combined score is a little higher because it is running 5.7GHz instead of 5.5GHz?

I suspect you are right that downclocking crap is here to stay. If I could recover 100% of the total purchase price I'd sell this GPU right now, shunt mod another 2080Ti FTW3 and give up benching. Even HWBOT sucks now. Way too much stupidity to put up with at this point in the PC tech space, at every turn.

I apologize for acting like an ass. I am just really disappointed with all of the nonsense. It is frustrating to spend so much money on garbage that is broken by design (intentionally).

 

So, was updating the windows image I build for common usage before doing memory or getting too far into overclocking, so most of my day was spent on that. Then a work project came in (there goes my next week). But, I did more testing to try to make sure fully stable.



This is after 30 minutes of CB R23, after 10 rounds of CB15 Extreme to test the non-AVX workload temps and stability.
The files showing CB15 Extreme is too large due to one screen being at 4K, but here is the HWInfo



It took 1.36V. 1.34V crashed during the 5th run of CB15 Extreme. I only have the Ring at 4.5 currently and haven't even started on ram (although it did boot 4133 on 4x8GB G.Skill using the XMP profile without needing extra voltage, I just am not using the XMP here until I test for full stability with no errors).

So, I know I can suicide run some benches at 5.3, but without better cooling, I likely won't be going beyond that (which 5.2 on water cooling without a chiller with a single core hitting 77C is doing just fine using 1.36V, not the best chip in the world, but not a bad one at all). Plus, with this cooling, the score doesn't dive at all while running it 20 or 30 minutes. You can see that, in part, with the average temps.

Once I get fully stable figured out (and after work is finished this week), then I'll get to some fun benching and I'll be posting scores.

 

At that point you could very well be correct. And in going along those lines that is why the score is higher. Not because of the gpu score which is lower.

Heck, running Firestrike now and the best gpu score i have is 50.3k, but an overall score of 40.3k, with a combined score of 16k. The problem here is combined is suppose to be 20k + which uses that same lowly gpu and kicks the overall score in the 42k range. Now on the 10900K setup that gpu score is 54K+ and 13.9k combined and it loses.

No worries.
I totally get it, been there and done that quite a few times over the years. Now for the most part I just deal with it and adapt.

 

You do have some control over that variable. It will just require that next level of cooling to combat the green goblin castrations. Including potentially LN2. I’m still wondering why you haven’t dabbled at least with LN2. It would open up the next level for you both GPU and CPU. Now that dynamic boosting on CPUs and GPUs is getting more ingrained each generation, extreme measures are going to be the only way to realize meaningful gains in the ballpark of what you seek.

 

Nice visiting with you Brother @Johnksss and thanks for talking me out of my tree.

https://hwbot.org/submission/467201...geforce_rtx_3090_39438_marks?recalculate=true

 

https://hwbot.org/submission/4672077_mr._fox_3dmark___port_royal_geforce_rtx_3090_15757_marks

 

That's what friends are for!

And a great time was had.

 

https://hwbot.org/submission/4672071_mr._fox_memory_frequency_ddr4_sdram_2354.6_mhz?recalculate=true | https://valid.x86.fr/lbgm95
4700 CL18

 

Taken from reddit, but these prices are getting out of hand. Newegg or MSI has once again jacked up the prices of 30 series cards. A 3070 is now more expensive than what my FTW3 3080 was.

https://www.newegg.com/p/pl?d=msi+r...sFDgeimpKdrvvVJHvILc4&N=100007709&isdeptsrh=1

$2109.99 for a crappy Ventus OC 3090... Yeesh. Makes my $1999 KPE look like an insane value in comparison.

Maybe their plan to "fix" the stock situation.

 

Isnt KPE huge and have some not practical TDP?

 

Hey friend! Nice to see you. Long time no see. I hope life is treating you well.

I think the right answer is no to both questions. It's not really remarkably larger than my 2080 Ti FTW as far as I have noticed and doesn't seem abnormally large to me. Maybe a little bit, but not enough that I would notice and not remarkably different. I haven't measured it though and it doesn't really look huge to me. At the same time, I don't have any case space constraints to deal with and my answer might be different if that were true. If you're talking about the PCB itself rather than the complete package then yes, the PCB is huge compared to the goofball new design of an FE or reference 3090. Seems pretty normal otherwise in terms of size compared to what I'm used to.

The benchmarks I ran last night pulled in excess of 750 watts from the GPU. I think that much is great. It could be over a 1000W and that's still fine. I think the Kingpin GPU merely does what all 3090 cards would do if they were allowed to run wild and free and were built in a way that it doesn't blow up or malfunction as much. It's not crippled to the disappointing degree that an ordinary 3090 or other gamer-boy GPU has been.

In other words, I don't think there's anything special about the Kingpin GPU, it's merely the opposite of normal because normal means crippled and malfunctioning to a great degree. It sucks to have to pay extra to get something that malfunctions less. It also sucks to save money with the unfortunate trade-off that what you get is something that malfunctions more.

 

Just imagining you in a tree yelling out and bashing Nvidia, AMD, Intel, Dell....etc...

I know it gets frustrating being purposely limited for no discernible reason except to profit somewhere, somehow. I'm finding it irritating watching the clocks drop for no good reason even putting the radiator between my storm panels at 18 outside and letting it cool down and idle for 20-30 min and ending up with sub 40 ambient temps on runs and watching the clocks drop. When I switch back to a CPU AIO, I can see placing a small fan on the back to cool the memory, but this card is clearly asking to be better blocked or LN2 especially with those temp caps.

If I'm feeling this way and a casual's casual, I know you must be ready to pull your hair out.

 

It's getting insane and yeah, picking up two $1999 KPEs is looking pretty decent value wise relatively speaking.

I sold my extra 3GB 1060 for $150 shipped and it was sold within the hour of listing it.

6GB 1060's go for 230+ now.

I managed to score a sanely priced 1650 super SFF for my brother and it is now going for over $300.

I am using a KPE atm with the 3070 FE sitting on the shelf. Wife is using a 1060 6GB I actually bought B-stock from EVGA back in October with full warranty just to use up my EVGA bucks I had. I'm seriously thinking about keeping the KPE and either selling the 3070 FE for ~$650 shipped (which will basically sell close to almost immediately) or giving her the 3070 and selling the 1060 6GB for $230+.

I knew when I saw PNY 3090's selling out at $2200 on CDW a few weeks ago things were heading in a bad direction pricing wise.

Another friend who kept holding out to replace his 1080 thinking supply would catch up and pricing would normalize finally bit the bullet, waved the white flag and ordered an MSI 3070 from CDW a few weeks ago for $767 to just be done with it. He wants to play CP2077 with RT Ultra and games @ 1080p. Another friend in december was on an old AW laptop that was dying and while trying to build a system couldn't source a GPU and finally just bought a prebuilt 10900kf system from Amazon with a 3070 in it and called it a wrap.

Memory prices are creeping up too.

The $60 RGB 2x8GB crucial ballistix kit I picked up in early December is now almost $90
The G.Skill B-Die 4000 kit I picked up for $330 in October is now $450 (yikes)

 

My 5700XT Red Devil card seems to be running in the 750-1000 range and I bought it new for 400. Reminds me of the mining craze

 

EVGA keeping the price at 2000 is one of the reasons I picked the card up. I know that is a crazy amount but like you stated, it is of the best 3090's on the market selling less than the "stock" AIB cards.

Now I just need to get either a 6800 XT or 6900 XT for MSRP. Not likely.

 

Back in September/November, I don't think anyone could see how bad the perfect storm would manifest.

I still haven't even seen 6800XT or 6900XT come into stock even in passing since December. Newegg is using that new lottery system and they haven't had a tweet in awhile where everything is going up for sale (or a large swath of CPUs and GPUs) in awhile.

 

My Microcenter has 1 left and it's open box at 508. New it 564.99. And they have been having them cards.

Had a chance to get a 6900XT last week, but passed on it as I was too lazy to get down to Micro and wait in the rain. They had about 50 to 60 cards total in stock 6900XT's and 6800 XT's.

 

BenchMate 0.10.7.1 was released a few days before. The improvement & Fixes in the Changelog is on Microsoft's level Hope this won't be the normal.

Integrates PYPrime and HWiNFO, improves support for AMD Ryzen CPUs, adds new sensor values to a trimmed down result dialog and fixes several bugs.

• Added PYPrime, a single-threaded prime number benchmark written in python by Monabuntur
• Added HWiNFO 6.42 as additional monitoring tool. This version matches the integrated HWiNFO version for sensor measurement and can be run next to BenchMate.
• Improved HWiNFO sensor measurement performance by avoiding redundant CPU scans
• Effective Clock will no longer be shown if the CPU's TSC timer is skewed. It's unreliable due to the skew and will show wrong values.
• AMD Zen support: CPU temperature, wattage support, FCLK and UCLK
• Improved readability of raw CPUID brand string in result dialog by replacing (r), (tm) and (c) with the appropriate unicode character
• Added measurement of average CPU load during the run. It's a running average of all collected sensor measurements, so don't expect a precise value here. But it can help especially in continous multi-threaded workloads to see if the work was done.
• Renamed Mainboard to the correct term: Motherboard
• Merged CPU, memory and motherboard information in the result dialog to reduce its size
• Updated CPU-Z to version 1.95
• Versioning of benchmarks now relies completely on the internal version detection inside the benchmark process. If the version can't be detected, the result will be invalid.
• Improved workflow for CLI wrapper's "RUN" button by avoiding that you need to press it twice due to the application not having the current focus
• Launcher: Added a maximum operation timeout to the update check (5 seconds)
• Improved security for guarded benchmarks
• Improved security and performance of Super PI integration. For those that had a problem with capturing Super PI results, this is now fixed as well.
• Allowing the presence of an enabled Hyper-V hypervisor, but will show invalid results when used inside a virtual machine.
• Fixed a bug with wrong detection of Engineering Sample flag on AMD Ryzen CPUs
• Fixed DIMM slot count for an upcoming GIGABYTE motherboard
• Fixed a bug on Ryzen where the HPET timer was disabled after returning from standby
• Fixed a rare bug where previous run errors were overwritten
• Fixed a bug where the launchpad is not fully visible/correctly updated when the client is initializing itself
• Fixed a bug with GPUPI 3.3 results not being visible in the result list when BenchMate is restarted
• Fixed a bug for y-cruncher on Skylake-X not showing any output
• Fixed a rare bug in kernel driver where a specific error message could result in a system hang
• Fixed a rare bug in kernel driver where a process error of a child workload was not correctly inherited by the parent
• Fixed a bug with CLI wrapper showing "Client was not ready" although no run was started yet
• Fixed a bug with skewed running load average if multiple GPUs are in the system

 

I'd gladly swap your Microcenter for ours.....wow.

 

6900XT 999.99 I almost went for it just to test then sale, but was like.....i have other stuff to buy at the moment.

 

Nice, Galax HOF Edition 3090:

https://videocardz.com/newz/galax-g...clocked-to-3ghz-breaks-multiple-world-records

 

LOOOL Only two days after the release

 

Is it even worth bothering to download it? It sucks that you have to install it now rather than it being a portable app. I will try installing then copying the installation folder elsewhere and then uninstall it. Maybe it will run as a portable app in spite of what they have done to it.

 

Yes, it runs fine without being "installed" in the OS. I installed it on Windoze OS X Cancer Edition and ran it from LTSC.

Isn't that "special" now? 2nd place among 10-core CPUs and you get 2.5 points hardware points. OMG, that is so generous, how can I ever repay them?

https://hwbot.org/submission/467279...ore_with_benchmate_core_i9_10900kf_18474_pts/

 

Welcome to the world where almost none post benchmarks. 1th, 2nd or 3d will give you a huge amount almost Zero points Even if 10 post results, you won't be a rich man on points. Gold, silver or bronze doesn't matter. Yeah, so much is Gold or Silver worth on the bot if you are one of few.

They want Rockies, but if the new-one bench with rare hardware and get wimpy 1-3 points vs quadruple x3 or more for the elite with the correct hardware, they will give up before it begins.

Be nr 15-17 of a few more subs and you'll get the points.

Yeah, isn't it nice. If 8 more post in your leagues R23 bench (with KF) you won't get so much more than for 10900K


+ their cut off points system + calculation works in a weird way. Buggy + not fair. Damn nice being the 11th



Yeah, kind of stupid have to use a second machine as guinea pig to make it as portable

 

Based on what I seen, all 3090 PCBs are over engineered. Even my XC3 PCB should be able to pull 500w+ without too much issue beyond 2 x 8pin. I am like 95% EVGA just nerf the vbios for product segmentation. TBH, anyone can unlock the vbios via a CH341a flasher if one desire to. I already got my hand on an Ampere vbios editor and basically you just need a CH341a with SOP8 and 1.8v adapter to flash it based on what I heard.

I think KINGPIN dies are binned for high leakage though. So if you have the space for 360mm rad and want to go XOC, its a good idea. I have case constraints, I need a relatively smaller midtower I can travel with.

In todays market? A good 3090 is the one thats in stock!

 

Noticed they also deprecated CB15 points for CB20 for global points, along with a couple other things, like PCMark 10. I mean, seriously? PCMark 10? For global points?

ANYWAYS, even though I was going to wait, I couldn't. NOT happy with the results here and really have to figure out if it is Win 10 2009 cancer or platform or what.



Granted, not a benching optimized OS, but that score seems WAY too low. Even though the ring is at 4.5 and the Skylake ring was at 4.8, this score doesn't make sense compared to my 6700K from years ago. These are the 2016 G.Skill 4x8GB 4133 19-21-21 dimms, so the B-die isn't as good as some of the more recent ones able to hit tighter timings (and those timings are at 1.45V).



These scores are just so far away even with the same exact sticks that I am saying there has to be something more than 300MHz on ring. Granted, I still have more work to do (I was going to wait but couldn't help myself from starting).

Edit:
That is odd. I run it as admin and get the slower score shown above. I run AIDA without it being elevated, I get this:

 

Yeah, it's really sad how things are happening at HWBOT. It seems like the decision-makers have totally lost their minds. They excel at demonstrating their abject stupidity. They should all become politicians, because nothing they do makes a lick of sense.

Windows 10 performance is extremely erratic and anything after 1809 is going to have impaired CPU and memory performance to an exaggerated degree. You will notice it in memory tests and any benchmark that includes a combined test (3DMark 11, Fire Strike, etc.). To be blunt, it's a piece of trash. Perhaps there is some comfort in knowing that the downgrade in memory performance still looks amazing compared to anything powered by an AMD processor. For some reason, that's OK now and it needs to be forgiven for no reason other than it's newer and it's not Intel. And, we must not dare to question anything that is happening with Windows 10. Just accept it all because they know what is best for all of us. Okie-dokie.

Hardly anything makes sense anymore. We've never had so many glaring examples or irrational, evil, corrupt human folly, and the technology space is no exception. In fact, the technology space is possibly one of the driving forces behind it all. I've never been conceited enough to view myself as a genius, but it seems like common sense and decency are rapidly becoming such a novelty that rising to genius status requires nothing more than rejecting everything that aspires to be regarded as the new normal. I can't turn on the TV, radio or go on the internet without being bombarded from every side with nonstop examples of unprecedently human idiocy and evidence of a growing inability to think being demonstrated by self-anointed imbeciles that view themselves as the leaders and shapers of the rest of our futures. If I could run a secure erase on the heads of all media, technology, "educational" and governmental institutions currently in control of the world around us I'd do it without batting an eye.

Here is a nice example of corrupt idiots in control. They have an army of muppets drinking their Kool-Aid. https://arstechnica.com/gadgets/202...racking-to-ios-14-users-in-new-pop-ups/?amp=1

Now for something completely different...

 

Small groups of people tend to act in an intelligent manner. Unfortunately increasing the size of the group for some reason magnifies the degree of stupidity in the group, which makes no sense, but that's what ends up happening.

I don't plan on budging from LTSC 1809 unless a future version of Windows 10 is just as good (maybe this will happen) or better (unlikely).

 

I use LTSC 1809 on my desktop and so far I am fine but lately I have been building out a Precision 7810 tower to become my ESXi server, just grabbed around 60GB of DDR3 to distribute amongst the desktops I have, grabbed another 7TB of storage and still got like 5-10 128GB SSD's. All decom stuff from my client of course.

Though would be interesting having a 2x CPU platform, never used one before...

 

And it's here...



I'll try and get to it soon. Just debating on whether I want to turn the 7980XE into the main gaming rig. Or - to install it first on the 10900K rig that's suppose to be the new gaming rig replacing the 9900KS rig...

I hope that this card is a nice bin.

 

I think we should chip in and get @Mr. Fox a Galax 3090 HOF edition so he can dominate the leaderboards.

 

I'm not sure money could buy such a thing if I, or anyone else we know, are not on their list of predestined recipients for the good stuff. (They don't really sell the truly good stuff to peons like us as far as I know.) But, it's the thought that counts and that is very kind of you, bro. Thank you.

Overclocking is slowly transitioning from being less of a personal achievement to more of having the right connections, being selected to participate in the sport using custom hardware that others have no access to, and lots of free stuff that someone else is paying for. Today it is more of a staged spectator stunt than a hands-on event involving many participants on a somewhat level playing field. There will always be those that have more discretionary money to blow on things that will get them closer to the top, but even money doesn't offset secret weapons made purely for marketing efforts that leverage name recognition among an elite group of chosen individuals. There are still some guys like us out there that only achieve through personal cost and effort, but we certainly are a dying breed.

 

hwbot, 3DMark etc... needs to create a new category to isolate the privileged (those who receive the hardware that no one else can get) from the public to keep it fair.

 

Wow, this memory is no where near as good as it is on Intel platform when it comes to latency....

 

Sadly, I think what you are seeing is inherent to all Ryzen processors rather than the memory itself. The mere fact that you have been able to get it to run 3800 CL18 is quite a feat in itself, which speaks to the quality of the memory, the CPU sample bin quality (better than many) and your tuning skills.

 

So it appears my 3960X died or my Asus Zenith Extreme II Alpha took a leap. Getting Memory Code 00 on boot up. On Asus boards, this normally means it can't recognize the CPU. I don't have much time today but tomorrow, I'm going to completely break down the build and reseat the CPU. The crazy part is that I wasn't overclocking. I only do video transcoding with this rig now. Doesn't make any sense.