All Activity

Hello, That is an excellent result! I assume -3 is the offset for AVX. Z370 is the one which your clevo has or the result is from a PC? If the result is from a PC I think it is important how much was the power combsution and what temp?

So long time ago. Now I set it to fixed and full dGPU power off dynamic scheme. And it works. But not at full power. I have installed a 765m and in superposition it powered up only to 405Mhz. In the past I had also sometimes the same problems. Most of the time it was a driver issue. After installing all drivers it works. Now I have installed them all expect the Realtek Card driver. It gave me an error Any ideas why it powers only to 405Mhz?

I'm asking in case anyone knows the answer: should I replace VRAM C0 and D1 , or is there something else I should check?

The 1.3V VRAM supply rail had a problem because the PCB layers made contact with each other, which caused a short circuit. As a result, the two VRAM chips closest to the crack were damaged.

And now the short has been removed and the PCB crack has been repaired. Display data appeared on the screen, and I ran a MATS test to find out which VRAM chip is faulty and causing the artifacts on the screen.

No my friend, this is another graphics card with a different fault. It has damage to the PCB caused by a heatsink screw: a screw of the wrong length was installed on the arm of the modded adapter, which superficially penetrated the PCB and caused a short, permanently cutting off the display data.

The 5090 pricing is really crazy right now for sure. I have not been tempted to sale my 5090 Astral, but my brother sure has been pestering me about selling it lol. Amazing GPU! I like it so much better than my FE, and I really liked my FE. But I think most of the likeness for the FE came from the low price and ability to actually get one fairly quickly after launch. I have decided to keep the 9950X3D2 that runs the 2,200fclk. The IMC has no issues stabilizing tuned 6400c26 with a mere 1.500v VDD, and 1.400v vddq. Using the new to me G.Skill 6000c26 sticks, and reducing VDDP voltage to 1.050v, has made its IMC feel just as strong as the other one. And now I actually got its tPHYRDL’s matching on both ram channels with both at 35, which is great! I noticed on AMD that they are sensitive to waterblock mounting pressure. And maybe this last mount last night swapping back to the 2,200fclk chip, must have gotten something perfect. I swapped between these two chips and unknown amount of times. I think I ran each one at least (3) times. This one also boosts higher in games. No PBO+ or anything like that is enabled. It’s running higher clocks on average by about 70-100Mhz. The AMD CCD0 and CCD1 in Windows is also quite good! I see my games typically use CCD0, while windows services are targeting the slower CCD1. It’s pretty amazing to see this work so well. For now I am hanging on to that 5090 Astral.

Thanks for asking bro🙂 More like status quo. But something good has it been the last week... I got my driving license back after 4 months . But only for two years due the risk for new strokes. Then I need to pay yet another $150 USD for a new prolonged period of two years. This if I once again can manage all the tests to keep the driving license. Some part of the brain is gone, thats the reality of getting stoke. You can compensate with other parts of the brain when some parts fall out. But within limits. It's what it is. I try as best I can everyday. Not bad but it could be much better. But who know how it goes. The risk for new stroke is still high within the next 5 years😑 RTX 5080 with two old 8-pin connectors instead of 12V-2×6: Ronaldo's experiment or how much connector drama could NVIDIA have avoided? Former EVGA manager says NVIDIA pricing rules forced some GeForce cards to be sold at a loss

i would change probably only c0 and d1 as before u got error on a1 and b0 while not now. remember that samsung chip have like 13 omh reading... where do you have the short ? on the coil or on vram nearby caps ? shorted rail should give no display or also no power at all.. i see u are able to post: are u using a second vga ?

Thats good to hear your test proved a pass, so it seems to align the card is good. Definitely review your cooling solution as this may have been a bit tooo close a call to failure. Happy to measure my 8770W "half plate" GPU sink if you need heights verified. That will work for a 980M but not anything newer, if i recall @KbX has a 1070M in his (but suffers display output options) not sure if hes evolved his heatsink since. Can anyone advise if the 1080M has been tried, not sure why we stopped at 1070m as i saw one which didnt have the PSU plug so it must be downclocked to run via MXM slot.

I bought a 980m from Aliexpress about 2017ish...lasted until 2020 then suddenly died. Apparently the VRM's are underpowered on these models. I really should look at repairing it but I don't use my Ranger much at all Aliexpress listing will say if it has vbios for Alienware! I skipped the 1060 and tried an MSI 1070m...requires cutting your chassis a bit.

Yeah, it does. If a person had the soldering skills and DDR5 BGA template, that would be a cheap way of "making" RAM as long as they didn't need 8000+ MT/S speeds. Congratulations, brother. I am glad that your dedication to excellence, faithfulness and hard work is finally being recognized. I'm glad you are being facetious, otherwise I would be concerned.

thats the kind of stuff that reminds of the good ol' NBR days 🙂 very nice! no justifications necessary, JUST DO IT! 😄 at 6000+ bucks its a steal and totally worth it hahaha

Looks like my promotion is finally in the works. I am getting reports of multiple departments fighting to get me in their camp. Its a nice complement, but Im still treating it as non-existent until I get an offer letter. Due to my "detailed" approach, which is really just a really nice way of saying I demand details and ruffle quite a bit of feathers in the process. It has been deemed necessary for me to become an Engineer for Proof of Concept. Basically I work with onboarded clients and make the how-to for green IT'ish people to facilitate their orders. Still wont be able to justify a 5090 though lol

blackened death metal Haiduk - Archdevil

Hello thanks for answering. What settings should I use on the clevo? I only have the offset voltage, AVX, ratio and power limits. I have a 9900K that was tested at 5.2GHz all cores at 1.36V with AVX2 -3 with a Z370 motherboard.

@Maro97 Most likely all chips that report errors (6/8). Start with the 2 that appear to be physically damaged and report most errors then retest.

back from vacation gaiz! and damn... who wouldve thought that my hardware setup actually generates more returns than my actual stock portfolio?! LULZ but even at an average selling price of 6080€ (bought my Suprim at 3300€ incl. tax and shipping, so +84% markup now) theres literally zero incentive for me to sell. just goes to show that when it comes to my PC hardware hobby, the passion and fun i get out of it far outweighs any monetary benefit it might generate. since basically all our everyday expenses are pretty much on the frugal side (aside from travels), i guess u can call that my one splurgeful vice 😄 btw if anyone is looking for a game to max out the power envelope, im currently playing Resident Evil 2 Remake and it basically pegs the GPU at 700-800W, was surprised to see that! was also pleasantly surprised to see Control Resonant getting released, got it right away 🙂 games backlog is filling up very nicely with lots of awesome games im looking forward to! same way im looking forward to the AI bubble popping.... 😛

Thanks BIOS works fine, I tried BCLCK but my R3 hangs whenever I press ENTER key. There's no multiplier option in overclocking menu with i7 2760qm. Must be available only for XM cpu. Will try 970m & m3000m soon & let you know how it works (SG & PEG). It did not work with A12 stock bios btw, but IIRC that was a known caveat on the R3. I got 120GB SSD & WD black 750Go, both work fine in SATA3 mode.

My friend, the repair process has already been completed, and I have the MATS test report as the result. The card had a short circuit, and I already explained everything about it. My question is: based on the MATS report, how many VRAM chips need to be replaced?

I have news, very good ones, the P5000 arrived, i filed the radiator a bit more after i dremeled it yesterday, i put it in, and half-assembled the upper half of laptop. I got into Linux Mint 18.3 and it saw the gpu, with ./nvflash --list then i flashed the vbios just like DynamiteZerg did in 2018 and after the reboot it went straight into Windows10. After some fighting with NVCleaninstall i present you another example of P5000 working in Dell Precision M6800 I want to thank everyone for the provided help, will post some updates later after i reassemble the laptop fully and test it in some benchmarks.

Yes, you can use linux to flash the stock ES vbios. Overclocked ES vbios requires a hardware flasher. External keyboard is recommended yes.

from the second picture, looks like some trace is broken. always use ipa to make the pcb and trace more visible. if u can make a better picture with ipa on it...

The short circuit that happened on the MXM card has been fixed. But a short occurred on the 1.3-volt rail that powers the VRAM, which caused some of the VRAM modules to malfunction. I tested it using MATS. This is the result I got. Can someone tell me how many VRAM chips need to be replaced?