Raiderman

3 hours ago, Mr. Fox said:

X570 and 5950X had poor core overclocking, absolutely repulsive memory overclocking, ludicrous latency issues, WHEA errors up the kazoo and extremely frustrating issues with USB dropouts even when using the Agesa cancer that they said "fixed it" made zero improvement for me. Ran MUCH hotter than my 12900K/KS or 13900K at lower clocks and lower voltage. The pairing of one good CCX with a second horrible CCX was also more idiotic than Intel putting Atom cores on 12th/13th Gen.

 

The USB drop-out problems made it unreliable even running BIOS defaults. I literally had to disconnect and reconnect my mouse and keyboard at every cold boot or they were not detected, and they would randomly stop working during normal use.

 

The only thing I liked about it was fluffy Cinebench scores. Otherwise, I would not have wished it on an enemy. It was the most regrettable "upgrade" I can remember in my lifetime. Very expensive garbage.

  

 

Sweet! What did you have to do to make W7 work right? Or, is it only Sky Diver and 3DMark 11 still not cooperating with you? 

 

If you haven't already done so, try setting the minimum core clock 50-100 MHz less than max. And, also try diabling "Advianced" core and memory overclocking. Set core minimum at 100% and core maximum at 110% and memory at 110%. Doing either of those things dramatically improved my 3DMark 11 scores and destroyed W10/11. 3DMark 11 is extremely demands and Windows 7 makes it more so, and left to its own devices I think the GPU downclocks to try to avoid using as much power as Windows 7 wants it to use. I found that setting the max core clock only allows the GPU fall to stock clocks under severe load.

 

EVGA will probably sell you some of those Velcro straps if you email them. I love them... best ones I have found anywhere.

  

No, but it hasn't seen enough use yet for me to comment. It has been in use only a couple of weeks. It was brand new before I installed it. Seems like a very high quality GPU block to me. I really like it.

I agree with that. Probably is a one-off. I have found everything stays looks extra nice using automotive coolant in the loop. I never plan to use anything else. Learned that from Brother @johnksssand when you stop and think about it, kind of like common sense. We don't have issues with our automotive cooling systems when we use it. 

I hadnt actually tried sky diver on Windows 7 until today...haha. No, it is 3Dmark 11 giving me troubles by not clocking as high as Windows 10. Wattman kind of sucks, but I will try your settings. I see those settings get your vram up to 2334mhz @ 112%. What is the stock speed on that 6900XT? It is 2000mhz on my 6800, with the slider only allowing 2150mhz. This is what I get in Wattman, which gives me the results displayed in gpuz.

55 minutes ago, Rage Set said:

 

I have a Bykski block on my 6800 XT in my all AMD build, with zero issues. I have performed a full cleaning/flush of my rig after six months. From my years of custom watercooling, if a block (CPU or GPU) are discoloring, it is usually the result of mixing metals or the coolant mix. 

 

What I find suspicious is the nickel plating is failing but his temps remain good. Those nickel particles have to go somewhere and usually it is clogging the coldplate fins...worsening temps. 

 

That being said, his problem could be a one-off. All brands have them.

I had a Bykski water block on my Radeon VII, and it worked flawlessly. I was a little worried about buying it, but at the time they were the only outfit making one.

If anyone knows where I can get some of those EVGA velcro straps (again) let me know please.

Windows 7 on the left, Windows 10 on the right.

4 hours ago, tps3443 said:

6.1P/4.8E/5.1R@1.315V
 

 

Cinebench is a weak test of stability. I found this out the other day when running Indigo Benchmark. Cinebench used to tax the cpu to the max, but it no longer does, at least not on my AMD chip. It doesnt even make it sweat compared to Indigo. So either Cinebench is not optimized to run correctly on AMD, or its just a BS benchmark now.

I have to have a little color, so I swapped the plain black Lianli case fans for some 140mm rgb Phanteks. They move a good amount of air and look nice as well. Not a big fan of the "breathing" unicorn vomit, but just a steady single color.

22 minutes ago, Etern4l said:

 

Very sorry to hear man, well benchmarks can make things come out of the woodwork. FWIW, it looks to me like we cannot exclude a hardware issue, possibly RAM (remember that out of bounds error?). In your shoes, I would ensure cooling is 100% in order and run at least one full cycle of memtest86 to begin with. I'd also try running Linux from a live USB. If issues continue there, it's likely hardware (would be an unwelcome conclusion, but at least you'd know where you stand)... Good luck bro, this puts my 13900K issues in perspective.

 

 

In this case we can assume the IHSes we are comparing are made of the same material. Are you saying that an IHS 10% larger will provide zero additional benefit, or that the benefit will be small? The latter I would agree with. Surely cooling efficiency is not linear in the area, but still in principle more IHS area is better, with diminishing returns. 

 

By the second law of thermodynamics, heat flows from hotter to colder objects (whoa!), but what that means is that initially heat will prefer to flow to that excess copper, and from there - through the extra contact area - transfer to that coolant at a higher rate than it would have otherwise.  This is clearly preferable to a situation where this extra route of escape for the heat does not exist. We haven't been bashing BGA manufacturers on account of their small heatsinks for no reason. That's the exact same idea.

 

Now, clearly there are diminishing marginal returns, but that depends on the initial size of the IHS. Imagine the IHS is infinitely large, you are not going to gain anything by adding an extra 10% lip around it, as heat will never get there in the first place. In contrast, take a look at that pitifully small LGA1700 IHS being asked to pump 300W+ lol

 

Going back to the original question: what improvement would an additional 10% IHS size on LGA1700 yield. I am afraid, we are not going to be able to derive the exact formulae here, nor will we be able to test this ourselves unless we find an aftermarket original 12th gen Rockit IHS, but yeah - if I were to guess probably somewhere around 5% (but not zero!).

 

 

I mean apparently even Linus managed to use it correctly (then proceeded to do his usual thing: dropping things etc. lol)

 

But then of course that's an ad.

 

 

What does that even mean? Do you mean a mod expressed a concern? It's not like has been publicly tainted somehow, there is nothing on his profile. On the other hand, a VAC ban is for life 🙂

 

Edit: nm, Reciever clarified.

 

 

For not observing a prior warning about fairly strong 1930s style content he posted (from what I gather).

I used Windows 10 built in memory tester, and zero errors were found. It runs two passes, and takes about 30 minutes to complete. TDP on the Ryzen 7950 is 95c, and while running Indigo I did not see it above 92c. Im hoping its more of bug in a first gen platform.

Sounds to me like someone was pushing the envelope, and received their due consequences.

Anyway, @Mr. FoxI would like to see windows 7 running on that card!

5 minutes ago, ryan said:

his name was hertzian. hes done alot for us. we owe him a thanks. not a warning. he will most likely be banned for that name

What was the warning for?

7 minutes ago, Mr. Fox said:

Check the W7 physics score, with NO performance tweaks, at a lower clock speed compared to Winduhz 10/11 filth.

Hmm... that forum name is... unacceptable. I am surprised it was allowed to exist.

Wish you could get windows 7 going on that 6900. I get 3300 more physics points in Win 7, but can not come close in the graphics.

29 minutes ago, Mr. Fox said:

No, I cannot flash it. Still has the stock vBIOS. It runs out of core voltage to go higher that 2800 (1.200V is not enough). It is not a 100% stable product when overclocked beyond a little bit past stock. Sometimes it runs smoothly and other times using the same MPT profile it is totally unstable. Very odd and frustrating, but this has been what I have always experienced with AMD products, so I am not complaining or surprised by it.

 

What does make me surprised and want to complain is lack of Windows 7 functionality. If I attempt to boot Windows 7 with the 6900 XT it not only fails miserably, it wreaks all sorts of havoc with my Windows 10/11 installations. After Windows 7 fails to boot, I have a somewhat difficult time getting the other two OSes to boot again. Something with the ASROCK vBIOS is really messed up in this respect.  It takes multiple chkdsk operations on each OS (individually when they are attempting to load) to get them to boot again.

 

Out of curiosity, I pulled out the 6900 XT and dropped in the 3060 Ti and *BOOM* Windows 7 runs like a top again. No issues whatsoever. This really rubs me the wrong way. ASROCK did something wrong with the firmware, but I don't know what or why.

If you have a dremmel tool grinding off a small portion of the "wing" on the left side of the IHS allows it to fit perfectly with the Thermalright CPU frame. The new copper IHS design is *almost* identical to stock. Had they made the left side wing exactly the same dimensions as stock it would not have any issues serving as a direct replacement with no modding needed to use the CPU frame.

 

Holy crap batman....You just described what I go through on a daily basis with this build. Constantly having to do chkdsk on the Windows 10 install. I was going to post what happened to me yesterday, and decided not to. Now that you described what you are going through, I think I will.

@Etern4lhad me try out the Indigo benchmark, and I was impressed with how it stressed the cpu, however, shortly after running it, I experienced BSOD's in Windows 10. I couldnt stay in the OS for more than a few seconds, so I reduced everything to stock in the bios, and still had the same issues. So I thought I would try the Windows 7 disk, and it was BSODing before entering the OS, so I was left with nothing to boot to. I thought for sure I cooked the cpu or the ram, so I did a memtest, and it came back normal, I was like wtf! So, finally I went into the bios and disabled XMP, and bam, it booted to both OSes??? After booting back into both OSes, I have since enabled XMP, and my undervolt curve, and have not had issues yet. Another thing is, this build does not like to restart, or shutdown from either OS. The OS itself shuts down/restarts normally, but I think its hanging via the bios or something. Windows will shutdown, screen is blank, but all the power is still on, and the EZ debug LED is stuck on yellow, meaning its hanging. Usually at start up, I get two LED's, red, and Yellow. Yellow off first, then red off, then post... To say the least, I am frustrated. In order to get it to power down completely I must hold the power button, and push it again. Not sure if its the MB, or a flaky bios.

But I feel your pain when it comes to chkdsk...I have to run it from within windows 7 to fix my windows 10 disk.

Uploading video

11 hours ago, Mr. Fox said:

https://hwbot.org/submission/5230241_

Looks like you flashed the vbios? I have not done anything yet, been too busy. How'd it go for you? No issues? 2800 boost clocks are pretty good arent they? Not sure what the defaults are on the 6900xt, but in wattman the sliders go to 2150 mem, and 2800mhz clock. I have never been able to bench anything above 2670mhz. Anything above that will artifact heavily and crash.

The Sapphire Pulse version,  I think,  is a slightly lower binned chip. For most runs, it has to be around 2600mhz.

Indigo is being a bit flaky for me. It will render the bedroom scene, then just quit. I must say, it is a very cpu intensive benchmark. Cinebench cpu usage is pathetic compared to indigo. I disabled my undervolt curve to see if that would help the bench render the super car, and it would not. It just quits. I was getting 5.3 on the bedroom scene before lowering it to the screenshot below. As you can see, it tasks the cpu much harder than CB.

3 minutes ago, Etern4l said:

 

Sorry this was on Linux, so there might have been a penalty due to the emulation layer (I'm not sure, my Windoze 11 died trying to upgrade itself against my will). 

Also bear in mind max power draw was around 200W, not 236W.

 

Do you want to have a go at Indigo Benchmark? Would be interesting to see how well 7950X does there.

Oh I see, you were running in a VM. I am trying the Indigo bench now. Ive been getting an "out of bounds" error. I will keep trying and post back.

6 minutes ago, Etern4l said:

I wrestled with my 13900K for a while today, and managed to get it to a usable state by taking a few hints from derBauer's video on tuning this CPU for efficiency, as well as your advice guys, @Mr. Foxin particular. Fixed ratios(51/42), fixed voltage ((ring 0.975V, Vcore: 1.125V). On top of that a new AIO with 25% larger coldplate, 50% faster pump, and T30 fans, but 11mm thinner radiator (cold during tests anyway). Used some throwaway paste there, don't remember MX-5 or TF7.

 

The result: 38K in CB23 at around 200W... Very usable. Temps still reach 90C so something is wrong and deliding will be necessary. 

 

 

Indigo is also not far off 300W results:

 

 

GPU is doing OK vs @Mr. Fox's OCed 6900XT on Linux, roughly in line with the spec advantage. conservative OC +70/+2000 

 

How are you running Cinebench 23 on Windows 7? With 8 more cores than me, you should be above the 40k mark.

3 hours ago, Mr. Fox said:

. My cynical self believes that this is by design to prevent Windows 7 from exposing what botched up messes Windows 10/11 are.

 

 

 

 

I almost guarantee that it is by design. Just as the last AMD driver for windows 7 cripples my overclock. 2560mhz in windows 7, and well over 2675 in Windows 10??? Hmmm

1 hour ago, Mr. Fox said:

Brother @RaidermanI really dig the DG-86. I like it more than I expected to, and I expected to like it. Thank you. I did some polishing on the acrylic and it's not quite as scruffy. I think if I spend more time with a power buffer or flame it that will be resolved.

 

I have it on a wheeled cart, so I can just disconnect audio, video cables, etc. and unplug the water lines with the QD fittings and roll it out into the middle of the room to work on it... spin it 360° to tinker. Super convenient being able to do that. 

  

 

 

I am really glad you are happy with it! I wish shipping was a bit cheaper than it was, but I suppose it is because it is so huge. I wonder if you could replace the plastic window with a piece of tempered glass? I never went as far as you have with that case. It was just too big for my taste. I think it is kind of designed to sit with the wiring side to a wall, so it sits like a TV almost. Again, very glad you like it. 🤙

2 hours ago, Mr. Fox said:

Not an option. They are built the way they are built. Both are made the same. All of my pumps are external. I get the air out by connecting the components, removing them and tilting them around to work the air out, while the pumps are running, then putting them back in. Been doing this for years without issues. The problem is the EKWB Velocity2 direct die block is not made correctly. 

 

But, that's all about to change.

Exactly. Not an issue. People that can't get the air out of the loop are just not knowing what to do.

Spoiler alert. Still lots of tedious work to do before I can use it. But, it's getting closer.

After I am done playing around with the 6900XT the 3090 KPE will probably go in. Then it will be almost entirely EVGA. Case, PSU, mobo and GPU.

 

Although it is not necessarily larger than my Level 20 XT, this sucker is gigantic. Much larger than it seems in videos and pictures. Don't know what I am going to do with the Level 20 XT now, LOL. Might make a good miniature greenhouse, LOL.

 

That looks sick bud! The 6900 XT looks like a gtx 1050 in there!

@Mr. Foxis busy swapping his build into the gigantic EVGA DG case he recieved Friday 🤙

I am so glad someone wanted that thing! One less thing I have to pack into a Uhaul. Thanks bro @Mr. Fox🏚️

12 minutes ago, Etern4l said:

 

BTW noob Q: this is just needed post-setup, and then presumably once in several months, right?

No, bleeding is only required if you expose the pump to air. That would require running out of fluid. I think its more of a maintenance thing for people like us, because we are always changing crap. New water block, new gpu cooler, swapping cases, etc. Once bled, it will only require bleeding if you open the system to change something , or running out of fluid (unlikely)

2 hours ago, MyPC8MyBrain said:

 

this and few other pics you posted of your system are making my OCD acting up.
you should raise your external res at least above the highest line point inside the case.
you have huge air pockets you will not be able to get rid of, which interns reducing your cooling efficiency significantly.

your external rad from the limited views i was able to catch is also setup incorrectly, your inlet outlets should be at the bottom not the top (it would be wise to flip it upside down).

There is not much difference between an external, and internal reservoir, and an internal reservoir is never above the highest line point inside the case. Air bubbles aren't  a thing when the pump is part of the reservoir. All you have to do is open the fill port, and tilt your case one way or another. Once bled, your done. Bleeding water cooling systems is part of the maintenance with an open loop system.

So I double checked all the settings in 3Dmark 11, and they are the same with win 7 and win 10. I cannot achieve the clock speeds in Win 7 that I can in Win 10. Screenshot does not show, but I am set at 2577mhz in Win 7, and Win 10 is set at 2624. It crashes in Windows 7 at anything close to 2600. It has to be the driver causing this.

https://www.3dmark.com/compare/3dm11/15610712/3dm11/15649756#

3 minutes ago, electrosoft said:

 

 

 

The actual PCB design and power delivery of the Asrock trumps all other cards including XFX. Factory issued boosts never tell the actual story of the silicon or board design.

The asrock card has a P bios, and a Q bios listed in the specs, with the P bios having a boost clock of 2495mhz

8 hours ago, electrosoft said:

 

Plus that is the best of the best class 6900xt card. The Asrock Formula OC 6900xt/6950xt have the best PCB design of them all (nevermind the beautiful aesthetics).

 

 

Haven't checked the 6900xt, but the 6950xt XFX card has ~50mhz higher boost clock 2324 vs 2368 compared to the Asrock card.

1 hour ago, Mr. Fox said:

I have never seen an example of that with 3DMark 11 before. Something must be wrong with your Windows 7 installation.  That, or since my last benching spree UL have released a software/system info malware update to deliberately mess with Windows 7 so that Windows 10/11 don't get molested severely. Also, check your Windows Update vomit trail. If KB2670838 is installed, remove it. It will ruin 3DMark 11 performance and cause problems with some games and benchmarks. You have to have that update for IE11, but nobody cares about that anymore. If you have the GUI mods, getting the GUI back to stock will also all Windows 7 to shine bright.

I'll have to check that, plus I did notice that it runs full screen 720p in win 7 as opposed to a cropped/centered 720p in win 10. I wonder if I accidentally ticked a box in the settings of 3dmark.? Yes I run the windows classic gui, and set it to performance in advanced system properties. Looks like windows 2000 when it's all configured.