Overclock.net banner
2461 - 2480 of 2532 Posts

·
Registered
Joined
·
393 Posts
New here, finally upgraded to a 6900xt. First run and just playing with settings got me this score.
Card is a 6900xt toxic.
Max power curve, 1175 volts, 2800mhz max and 2700 min, memory at 2150 normal timings.

Nice one!

Is that the Extreme or the normal version?
 

·
Registered
Joined
·
64 Posts
A quick fyi: I noticed that MPT beta '4' was out. It still needs work, but I could oc my VRAM well past 2150. In fact, when I let the auto VRAM overclock in Wattman do it, it came up with 2260 MHz for the VRAM on its own :p ...I even 'seemed' to have 1.200v going on my non-H (Wattman accepted the 'lock') but could only run light rendering - crasheritis after...further, when you slide the VRAM above 2150, it will take out the 'max GPU MHz' setting, sadly. Still, some progress by the good folks behind MPT ! Clearly, they're on the right track.

View attachment 2518184


@ptt1982

While you had an 'unusual method' to flush (while already installed ?), I wouldn't panic about that if you subsequently flushed it thoroughly with distilled water and appropriate cooling liquid.

The Hotspot delta is very high, though it will increase significantly with the PL (ie 380 W) you mentioned anyhow. Re. flow, you might want to invest in a flow meter but even a visual inspection re. flow into the reservoir should tell you about blockage. I've only used D5 pumps for the last 10 years or so (usually two in series) rather than DDCs so I can't really comment directly on that setup as to what is a good flow for it. I have read though that that pump can be a bit restrictive.

Finally, is your GPU mounted vertically ? I ask because if you use thin (even highly-rated) thermal paste AND a block that has some stand-off tolerance issues, you might get an uneven mount and subsequently uneven die coverage / thermal paste displacement. I am not sure if that is the case with your setup, but the last year or so has seen a host of quality and tolerance problems with many GPU blocks (including my EKWB / 3090)...here is another OCN link which > alludes to that ...using washers (w/ plastic on the bottom re. avoidance of conductivity) can help, btw, but make sure not to over-tighten the block and also strive for even mounting pressure

If you still have a few weeks to re-do things, you might want to just give yourself some peace of mind by taking the loop apart, flushing each piece properly (as much as I don't think it is the problem) and also remount the block - perhaps after using some imprint paper and cheap thermal paste first for an imprint to see about die coverage.

I suggest you might also try some 'thicker' thermal paste such as Gelid GC Extreme (it is highly rated). Finally, if you are not sure about thermal pad size (or your block just has odd tolerances), you might want to try thermal putty (below, 10 W/mk) for the VRAM and power stages though obviously not the GPU die itself. Thermal putty is good stuff and it conforms to the available space w/o running out, even in a vertical GPU mount.

But it is worth noting again that a high PL such as 380 W will increase the delta hotspot temp no matter what...

View attachment 2518188
Thank you very much for the assistance and insights.

I've got the GPU in a horizontal position, using the original washers that came with the Alphacool block.

Something seems to be wrong. It did crash today once, first time ever, black screen, and had to manually reboot using power button. With further game testing, in Hitman 2 at 4K fps unlocked I saw a spike to 105C junction when the GPU Edge temp was at 60C, that's a 45C delta, and that was around 373W power consumption. This is late at night in a high rise apartment building air con targeting ambient temp of 22C. The delta is constantly getting higher whereas the Edge is as it was before at max 60C, so I need to do a remount. Maybe there's a spot the paste didn't reach, or the mount has loosened up. Ambient temp for GPU Edge is 30C and junction 35C, but the junction shoots up extremely aggressively under load.

Meanwhile, CPU temps haven't changed, but I think I will still reinstall the CPU waterblock as well, because I did change the motherboard without draining the loop and it was a bit chaotic and haphazard. I repasted the CPU then, but didn't have too much control over it due to loop being connected and full of water the whole time, and me and my wife in a strange position holding PC parts with both hands. I did a leak test afterwards for a couple of hours, and have used the computer for five weeks after that without problems, except for now. The problems are with the GPU though, and seem unrelated.

I have all that I need to reinstall everything: electric compressed air duster to drain the loop completely, paper towels, tapwater that I will turn into distiller water, 1g of Hydronaut, time, and a little bit of patience (the latter one I have the least of all.) I have to use the existing thermal pads that came with the waterblock because I'm afraid I will fall into a rabbit hole again if I'm trying to find the right aftermarket type (that is the reason I put this thing under water, the heatsink is incredibly hard to remount without the original pads, plus the temps were originally very high.)

I also might look into the thermal putty, it looks something that I would need. I want to try and fix this without spending a penny first, though!
 

·
Registered
Joined
·
92 Posts
Thank you very much for the assistance and insights.

I've got the GPU in a horizontal position, using the original washers that came with the Alphacool block.

Something seems to be wrong. It did crash today once, first time ever, black screen, and had to manually reboot using power button. With further game testing, in Hitman 2 at 4K fps unlocked I saw a spike to 105C junction when the GPU Edge temp was at 60C, that's a 45C delta, and that was around 373W power consumption. This is late at night in a high rise apartment building air con targeting ambient temp of 22C. The delta is constantly getting higher whereas the Edge is as it was before at max 60C, so I need to do a remount. Maybe there's a spot the paste didn't reach, or the mount has loosened up. Ambient temp for GPU Edge is 30C and junction 35C, but the junction shoots up extremely aggressively under load.

Meanwhile, CPU temps haven't changed, but I think I will still reinstall the CPU waterblock as well, because I did change the motherboard without draining the loop and it was a bit chaotic and haphazard. I repasted the CPU then, but didn't have too much control over it due to loop being connected and full of water the whole time, and me and my wife in a strange position holding PC parts with both hands. I did a leak test afterwards for a couple of hours, and have used the computer for five weeks after that without problems, except for now. The problems are with the GPU though, and seem unrelated.

I have all that I need to reinstall everything: electric compressed air duster to drain the loop completely, paper towels, tapwater that I will turn into distiller water, 1g of Hydronaut, time, and a little bit of patience (the latter one I have the least of all.) I have to use the existing thermal pads that came with the waterblock because I'm afraid I will fall into a rabbit hole again if I'm trying to find the right aftermarket type (that is the reason I put this thing under water, the heatsink is incredibly hard to remount without the original pads, plus the temps were originally very high.)

I also might look into the thermal putty, it looks something that I would need. I want to try and fix this without spending a penny first, though!
Yeah I'd say putty would be your best bet. Also, are you torquing your screws tight enough?
 

·
Registered
Joined
·
92 Posts
A quick fyi: I noticed that MPT beta '4' was out. It still needs work, but I could oc my VRAM well past 2150. In fact, when I let the auto VRAM overclock in Wattman do it, it came up with 2260 MHz for the VRAM on its own :p ...I even 'seemed' to have 1.200v going on my non-H (Wattman accepted the 'lock') but could only run light rendering - crasheritis after...further, when you slide the VRAM above 2150, it will take out the 'max GPU MHz' setting, sadly. Still, some progress by the good folks behind MPT ! Clearly, they're on the right track...

View attachment 2518188
Can you give a link to beta 4? I can't find it.
 

·
Registered
Joined
·
60 Posts
...here you > go (scroll down a bit and you see MPT 1.3.7. Beta 4 and it's download link)

What did you change to get the memory to 2260 ? If I change the overdrive limit of 1075 maximum memory frequency to 1200 , and set the memory to run at even 2160 it cripples my card :/
 

·
Invalid Media
Joined
·
4,400 Posts
With mem at 2160 I get a graphics score of 5000 :D After resetting sppt I get 22k...
...yeah, as mentioned in my earlier post, the moment you go over 2150, it will drop your max GPU down to safemode speed (this after just increasing the VRAM frequency in MPT beta 4)

...I also had 'some' luck with loading the Aorus Xtreme Waterforce bios, then dropping GPUv from 1.2 to 1.175 before activating...After reboot, Wattman opened up all kinds of new / different options. But still way too buggy and crash-prone.

Still, MPT is clearly drilling deeper and getting to hitherto unavailable parameters

EDIT> I haven't got enough time re. wok projects right now, but I wonder if uninstalling Wattman and just using MSI AB might work - weekend project perhaps...or just wait for MPT beta 5
 

·
Registered
Joined
·
60 Posts
...yeah, as mentioned in my earlier post, the moment you go over 2150, it will drop your max GPU down to safemode speed (this after just increasing the VRAM frequency in MPT beta 4)

...I also had 'some' luck with loading the Aorus Xtreme Waterforce bios, then dropping GPUv from 1.2 to 1.175 before activating...After reboot, Wattman opened up all kinds of new / different options. But still way too buggy and crash-prone.

Still, MPT is clearly drilling deeper and getting to hitherto unavailable parameters

EDIT> I haven't got enough time re. wok projects right now, but I wonder if uninstalling Wattman and just using MSI AB might work - weekend project perhaps...or just wait for MPT beta 5
Was thinking that. I actually have an MSI card as well. Downside to MSI AB is you can't enable the "fast timing" setting for the memory :(
 

·
Invalid Media
Joined
·
4,400 Posts
...yeah, as mentioned in my earlier post, the moment you go over 2150, it will drop your max GPU down to safemode speed (this after just increasing the VRAM frequency in MPT beta 4)

...I also had 'some' luck with loading the Aorus Xtreme Waterforce bios, then dropping GPUv from 1.2 to 1.175 before activating...After reboot, Wattman opened up all kinds of new / different options. But still way too buggy and crash-prone.

Still, MPT is clearly drilling deeper and getting to hitherto unavailable parameters
Was thinking that. I actually have an MSI card as well. Downside to MSI AB is you can't enable the "fast timing" setting for the memory :(
...given the latest entry fields in MPT B4 and the fact the Radeon bios are loaded into Win registry, there must be a switch somewhere for a registry hack...then again, if I can pick up 100 MHz on the VRAM w/o other efficiency issues, sacrificing fast timings 'may' be worth it, depending on the app...
 

·
Registered
Joined
·
60 Posts
...given the latest entry fields in MPT B4 and the fact the Radeon bios are loaded into Win registry, there must be a switch somewhere for a registry hack...then again, if I can pick up 100 MHz on the VRAM w/o other efficiency issues, sacrificing fast timings 'may' be worth it, depending on the app...
Agree. Did see someone stating that it's the driver that detects too high voltages/frequencies. If that is really the case then going with MSI AB will not fix the issue of it downclocking to 500 MHz :(
 

·
Invalid Media
Joined
·
4,400 Posts
Thank you very much for the assistance and insights.

I've got the GPU in a horizontal position, using the original washers that came with the Alphacool block.

Something seems to be wrong. It did crash today once, first time ever, black screen, and had to manually reboot using power button. With further game testing, in Hitman 2 at 4K fps unlocked I saw a spike to 105C junction when the GPU Edge temp was at 60C, that's a 45C delta, and that was around 373W power consumption. This is late at night in a high rise apartment building air con targeting ambient temp of 22C. The delta is constantly getting higher whereas the Edge is as it was before at max 60C, so I need to do a remount. Maybe there's a spot the paste didn't reach, or the mount has loosened up. Ambient temp for GPU Edge is 30C and junction 35C, but the junction shoots up extremely aggressively under load.
(...)
...doesn't surprise me that you got a shut-down...per below, a bios readout (H bios in this case) with temp limits. A remount might definitely be a good idea and I reiterate that you might want to do the paper-imprint test mount first to see about coverage (and if your block has stand-off issues).

..another quick 'free' test is to use a credit card and/or another perfectly level item and check for concavity or convexity of your GPU die and the corresponding area on the block.

2518233
 

·
Professional Lurker
Joined
·
207 Posts
Many informations about your GPU troubles
Howdy! Let's check a few things first, because SOMEthing is preventing your water block from contacting the GPU's die properly. It could be one of many things, or a combination of all of them

- First off, did you use the thermal pads that came with the water block? If not, why? If not, do you still have them? If you have them, dismount and use them.
Rationale: With the reference card and EK's Quantum Vector blocks, for example, one can NOT use aftermarket thermal pads like Thermal Grizzly Minus Pad 8, because the thermal pads that come with the EK block are VERY compressible, where TG's MP8 pads are not. Using TG's pads on a reference card with any cooler will prevent the die from contacting the block properly

- Thermal Paste Application. I personally use Gelid GC-Extreme for bare-die applications. Perhaps read up on how to apply it, then give it a try before going with thermal putty
Rationale: I learned this one the hard way with prior hardware, and thankfully haven't had any trouble with my RX 6900 XT reference w/ek block, nor has the b/f had trouble with his 3080FE on water. A thin paste and/or oil-based emulsion pastes like NT-H1, Kryonaut, etc.etc.etc., don't work well on bare-die applications that get really hot, as the heat cycles will "pump out" the grease over time. The thicker, non-oil-based emulsion of pastes like Hydronaut, GC-Extreme, Ectotherm, et. al, are much more resistant to pump-out than their oil based counterparts. However, you can't have clearance issues as described above when using thicker pastes, for you have to use a spreader tool and apply a layer of GC-Extreme over the die, with enough paste that you can't see the die anymore, but no more. An uneven spreading is unforgiving when using a water block, for the paste doesn't flow very much at all, so it won't fill in dips in the paste, especially if one is left in the center of the die.

But! If you get the application of GC-Extreme right, you will be blessed with 15C delta-T at 400W card power draw, and (my) max hotspot temp of 80C in TS #2, which it never gets close to in 4k120 gaming loads.

TL;DR: Take a step back. Take stock of how you've mounted your water block, and which thermal pads and thermal paste you've used in that process. Think logically about temperatures: There's either a flow problem, or a die-contact problem (more-likely given your delta-T). Figure out why you have that contact issue, take some of the advice above and in other posts to heart, and go from there.

I wish the best of luck to you. Let us know what you find 💗
 

·
Registered
Joined
·
393 Posts
Digging today for couple of hours, managed to hit 23800+ Graphic Score on a custom run with GT1+GT2 only, then tried to replicate on a "leaderboard" run and managed a 23739 Graphic Score.

Anyway been playing with the new beta and with the Strix LC XTXH bios.

Basically, in case you didn't know the XTXH bios increases VRAM Voltage to 1.406V from 1.356V of the XTX cards. This might help some of you gain a bit more performance if you're not completely stable at 2150 Mhz which I wasn't (ECC kicking in and performance drops a bit at least on 7.1 driver).

What I did was load up the XTXH bios, input the XTX values under voltage / memory DPM etc, basically match 1:1 all the XTXH specific settings to XTX.

I also disabled deep sleep, increased VSOC min a bit, increased fCLK and fCLK boost to 2000, increased vCLK and dCLK by 20 Mhz, enabled FT2 but left the driver at FT1 (rumour says it still tightens timmings not sure if true).

Also the card is very finicky with min/max clockspeed. It's not the same for everyone. Some cards like super high min clock speed, others super low, others inbetween. I did my benches at 2450-2735. Going higher or lower too much either dropped performance or made the card boost past 2730 Mhz during GT2 and crash the benchmark.

So for those of you looking to bench/squeeze more performance, test all these things.
 

·
Registered
Joined
·
132 Posts
The latest drivers look really good so far. Managed a new personal best of 20 308 points in Time Spy, compared to my last best of 19 243. GPU score is 23 662 vs 22 102. Not too bad considering I only have an i9-9900K, or?
 

·
Registered
Joined
·
608 Posts
I saw someone mention that Superposition 8K was the go to test for checking memory performance gains. Below are my results from Superposition.

Superposition:
2100Mhz=6774
2112Mhz=6764
2124Mhz=6797
2150Mhz=6847
2162Mhz=6859
2170Mhz=6832

2770/2162Mhz 21.7.1 (not my highest possible clocks on core)
2518343

As a comparison in my Timespy graphics score testing, I start losing performance at 2162Mhz, with 2150Mhz offering the best gains. So for me at least, Timespy seems a better gauge of memory performance gains.
 

·
Registered
Joined
·
7 Posts
I know not many people have them but has anyone changed the fans on the 6900xt toxic radiator. The plugs from the video card do not allow normal fans. Is there anyway to control the pump?
 
2461 - 2480 of 2532 Posts
Top