Rtx 5000 upgrade AW18 R1

[nicholas1020]

Hello everyone its been a little bit since my last update I have some benchmark results to post(: After extensive testing in pursuit of finding a problem and a workable solution I have followed ssj92 tutorial on driver installation, I can confirm my drivers do in fact work and are validated. I ran two games prior to running the time spy and fire strike benchmark. I ran MSI Afterburner to view gpu frames, power draw, vram usage and temps in re4 remake and the Dead Space remake. In each instance MSI Afterburner displays a prompt LIM Power, Voltage, and Temperature similar to what is display on a screen shot I took off of the firestrike benchmark. I though maybe the major downclocks I'm receiving could be attributed to a thermal throttle but I tried to eliminate this as possibility by ramping all my fans on max as well as placing my laptop onto of a fan blowing cold air on max, lol a lot I know but I had to know for sure. These benchmark results have been taking with these cooling conditions implemented to see what scores I may land in order to figure out if I have a bottleneck somewhere preventing me from hitting that 60fps in game or if its being caused by thermal throttling or even a lack of power somewhere. I am using a 330 watt power supply from dell, thank you everyone for taking the time and the effort out to help me in advance! If everything point to a lack of power I'm can certainly try what ssj92 and Falinov had suggested on implementing a resistor mod for an increase in power delivery if necessary and ill order the resistor and test it with the rtx 5000 if everything point to a power issue(: Additionally as its not listing in the benchmark I am running a standard hdd, no ssd at the moment and 16 gb ram sticks unmodified! Edit: I just ran re4 remake and lowered the resolution down to the lowest resolution and lowest graphical setting and I'm still not landing a solid 60 fps, solid 30 but not 60fps. Perhaps this could be point to a cpu bottleneck? I'm gonna try a cpu overclock and see if that makes any difference in fps!(:

 

 

 

 

 

[ssj92]

Do the resistor mod, it should fix the issue for you. 
 

here’s the one I bought :

 

https://www.mouser.com/ProductDetail/IRC-TT-Electronics/ULRG2.5-2512-R004-F-LF-SLT?qs=o5%2FBfEwRdcu3G0v%2B03nkPQ%3D%3D

[nforce4max]

On 2/12/2024 at 8:21 PM, Falinov said:

What??? I never said anything about chopping fin stacks that's a horrendous idea. No I desoldered the hot plate for the HP RTX cards so all the VRMs are cooled correctly as none of our heatsinks properly make contact with those 2 VRMs on the bottom. I don't have a proper photo of it but I didn't do a lot of modifications to make this fit. But it is 100% better to use the one HP made for the card. It creates less issues than trying to rig an alien heatsink lol. 

There is some real eye bleedingly bad heatsink mods in another thread to ruin your day lol.

[Falinov]

19 hours ago, nicholas1020 said:

Hello everyone its been a little bit since my last update I have some benchmark results to post(: After extensive testing in pursuit of finding a problem and a workable solution I have followed ssj92 tutorial on driver installation, I can confirm my drivers do in fact work and are validated. I ran two games prior to running the time spy and fire strike benchmark. I ran MSI Afterburner to view gpu frames, power draw, vram usage and temps in re4 remake and the Dead Space remake. In each instance MSI Afterburner displays a prompt LIM Power, Voltage, and Temperature similar to what is display on a screen shot I took off of the firestrike benchmark. I though maybe the major downclocks I'm receiving could be attributed to a thermal throttle but I tried to eliminate this as possibility by ramping all my fans on max as well as placing my laptop onto of a fan blowing cold air on max, lol a lot I know but I had to know for sure. These benchmark results have been taking with these cooling conditions implemented to see what scores I may land in order to figure out if I have a bottleneck somewhere preventing me from hitting that 60fps in game or if its being caused by thermal throttling or even a lack of power somewhere. I am using a 330 watt power supply from dell, thank you everyone for taking the time and the effort out to help me in advance! If everything point to a lack of power I'm can certainly try what ssj92 and Falinov had suggested on implementing a resistor mod for an increase in power delivery if necessary and ill order the resistor and test it with the rtx 5000 if everything point to a power issue(: Additionally as its not listing in the benchmark I am running a standard hdd, no ssd at the moment and 16 gb ram sticks unmodified! Edit: I just ran re4 remake and lowered the resolution down to the lowest resolution and lowest graphical setting and I'm still not landing a solid 60 fps, solid 30 but not 60fps. Perhaps this could be point to a cpu bottleneck? I'm gonna try a cpu overclock and see if that makes any difference in fps!(:

 

 

 

 

 

This is why I did the resistor mod and started looking how to fix that. Good to see my research and mod worked out. I'm interested to see how it'll pan out.

[Falinov]

19 hours ago, nforce4max said:

There is some real eye bleedingly bad heatsink mods in another thread to ruin your day lol.

This is why I am excessively picky about my heatsink quality. Lol

[TruenoG7]

On 2/20/2024 at 4:15 AM, ssj92 said:

Run 3dmark firestrike and show us the result. I’m assuming you also need to do the resistor mod. I just did it today on my 18 although it didn’t help against adlink RTX 3000 it should help with hp RTX 5000

What resistance mod are you referring to?

[Falinov]

46 minutes ago, TruenoG7 said:

What resistance mod are you referring to?

Check my single post please. It kinda shows you what it does. Allows these HP cards and many others thought to be limited use In these.

[nicholas1020]

Sorry fellas I know I haven't posted any updates in a hot minute. I did overcome my previous framerate issues by overclocking the cpu, my cpu was bottlenecking pretty hard. though it only solves part of the problem. Currently I am in the process of installing arctic thermal pads and paste to hopefully overcome the thermal throttle i'm getting in game. Also making minor adjustments to the mxm lifters and adjusting them to perfectly align with my gpu. Also creating more clearance between my heatsink and lcd because it does put pressure on the heatsink when I install it. The goal is to polish up my work in such a way you would have though the laptop was build that way out of the box. I still have plans on incorporating the resistor mod just haven't got around to it yet but I will keep everybody posted on any new development and hopefully everything works when I put it all back together(:

[nicholas1020]

On 2/22/2024 at 4:38 PM, Falinov said:

Check my single post please. It kinda shows you what it does. Allows these HP cards and many others thought to be limited use In these.

 

Hey Falinov I was wanting to ask a quick question out of curiosity in regards to the resistor mod. I understand it's supposed to allow hp cards and some other gpus to pull more power to the card. But what I don't yet understand is why some other gpus wouldn't receive the same added benefit from having done the modification? I appreciate ya I feel like I've learned a lot since I've carried out this project. It really has been a lot of fun and has certainly kept me busy(:

[Falinov]

4 hours ago, nicholas1020 said:

 

Hey Falinov I was wanting to ask a quick question out of curiosity in regards to the resistor mod. I understand it's supposed to allow hp cards and some other gpus to pull more power to the card. But what I don't yet understand is why some other gpus wouldn't receive the same added benefit from having done the modification? I appreciate ya I feel like I've learned a lot since I've carried out this project. It really has been a lot of fun and has certainly kept me busy(:

It can certainly be done. Yes you can theoretically do what us called a shunt mod. Me personally it's not beneficial. Defeats the whole purpose of a Quaddro card. If that's what you're asking me.

 

Also I'm not willing to track down which resistor I need to double over on a rtx 5000 that has a base clock of 600 iirc. That's why I went to the RTX 2080. That monster can and will draw the full 200W of power to the slot. 

 

Secondly. You should most definitely not be able to overclock the rtx 5000. Especially the HP cards they do not let you. If they do now then you have way more skills than I do lmao. But i digress. Overclock won't solve your issue. The card is a MaxQ architecture same as what we would define as a 2080 super "ish" thing is it is physically limited to the lower power band than its counterparts. Your scores weren't far from my own before the resistor I put in. 

 

This whole resistor mod is basically the inverse of a shunt mod. Instead of the gpu I'm doubling the capable wattage to the slot. We know the R2 and R3 can do 150w so in theory it should be around a 300w max by doubling this resistor. Keep in mind- that these laptops were built for cards that average 100W with a spike of 150 for a brief period that's why the 1080 has been a lucrative but previously unstable upgrade. The RTX 5000 is the same thing. Except its base wattage is 80W but it can spike to 130 or more. That's why when you're running a bench you can watch the wattage increase and suddenly it tanks and you're stuck at 930Mhz.

 

My only concern these days is the longevity of the mod. Will it put strain on the copper traces for the power rail over time? I have no idea. It's also why I took so long to post my findings because I didn't want to be branded "proof or it didn't happen" member and subject myself to unneeded scrutiny. 

 

@ssj92 im still not sure why your adlink card doesn't work... still convinced it's a miss communication between your Vbios and the laptop. I wonder if turning off the temperature monitoring on the gpu would help. 🤔 

[GuitarG]

On 2/21/2024 at 6:31 AM, nicholas1020 said:

Hello everyone its been a little bit since my last update I have some benchmark results to post(: After extensive testing in pursuit of finding a problem and a workable solution I have followed ssj92 tutorial on driver installation, I can confirm my drivers do in fact work and are validated. I ran two games prior to running the time spy and fire strike benchmark. I ran MSI Afterburner to view gpu frames, power draw, vram usage and temps in re4 remake and the Dead Space remake. In each instance MSI Afterburner displays a prompt LIM Power, Voltage, and Temperature similar to what is display on a screen shot I took off of the firestrike benchmark. I though maybe the major downclocks I'm receiving could be attributed to a thermal throttle but I tried to eliminate this as possibility by ramping all my fans on max as well as placing my laptop onto of a fan blowing cold air on max, lol a lot I know but I had to know for sure. These benchmark results have been taking with these cooling conditions implemented to see what scores I may land in order to figure out if I have a bottleneck somewhere preventing me from hitting that 60fps in game or if its being caused by thermal throttling or even a lack of power somewhere. I am using a 330 watt power supply from dell, thank you everyone for taking the time and the effort out to help me in advance! If everything point to a lack of power I'm can certainly try what ssj92 and Falinov had suggested on implementing a resistor mod for an increase in power delivery if necessary and ill order the resistor and test it with the rtx 5000 if everything point to a power issue(: Additionally as its not listing in the benchmark I am running a standard hdd, no ssd at the moment and 16 gb ram sticks unmodified! Edit: I just ran re4 remake and lowered the resolution down to the lowest resolution and lowest graphical setting and I'm still not landing a solid 60 fps, solid 30 but not 60fps. Perhaps this could be point to a cpu bottleneck? I'm gonna try a cpu overclock and see if that makes any difference in fps!(:

 

Hello ! I've been researching on subject and it seems my card works okay. Im using third gen i7 55W with rtx 5000 with 280W power supply, what displays in GPU-Z depends on nvidia control panel settings for power. When stressing on maximum performance or optimal/adaptive it clocks up to 1850mhz on core, 1500mhz memory. In PerfCap changes from Idle to PWR or Voltage or both  - I guess that is normal ? So it won't extend 80W supply limit by vbios ?
MXM 3.0 slot should be able to delive at least 100W of power, there are very wide copper traces on card - it's directly from power supply to GPU.

Maybe experimenting with vbios or nvclean settings should help ? Before I made mine work, I had code 43 error and constant high wattage draw on GPU - it overheated until I've correctly installed driver.
CPU is 100% not problem here, I would look into GPU directly and board BIOS/ KBC controller. Those are your culprits

EDIT: When using constant performance in nvidia control panel GPU stays on 600 mhz core, 700 mhz ram, PerfCap is "IDLE", core draws from 25W to -50W max

[nicholas1020]

Falinov thank you for taking the time out to explain how the mod works, I would really like to try out that modification it does sound pretty awesome. I didn't try to overclock the gpu but I though maybe overclocking cpu would help, my fault I know I didn't specify. I had 4710mq 2.5 ghz. I overclocked to 3.5 on all cores and almost gained an extra 10 fps in re4 remake. Though I am sure that my current performance is being hindered by the pwr threshold and I think maybe even thermals because my temps lock off at 84 degrees but never surpass that. It seems as though my wattage and base clock drop when I hit that 84. Granted it does say the operating Temps for the rtx 5000 is 93 degrees so it mainly just speculation at this point(:

[Falinov]

44 minutes ago, nicholas1020 said:

Falinov thank you for taking the time out to explain how the mod works, I would really like to try out that modification it does sound pretty awesome. I didn't try to overclock the gpu but I though maybe overclocking cpu would help, my fault I know I didn't specify. I had 4710mq 2.5 ghz. I overclocked to 3.5 on all cores and almost gained an extra 10 fps in re4 remake. Though I am sure that my current performance is being hindered by the pwr threshold and I think maybe even thermals because my temps lock off at 84 degrees but never surpass that. It seems as though my wattage and base clock drop when I hit that 84. Granted it does say the operating Temps for the rtx 5000 is 93 degrees so it mainly just speculation at this point(:

Looking at your previous post. You used a 2 pipe heatsink. This will hurt your temps. I usually max at 89 with a triple. There's no way for me to tell if yours will get hotter with the smaller 2 pipe design.

 

Second, if you can I would upgrade the CPU to a 4930 if you can this machine can handle the heat but the added performance will dampen the bottleneck. If you can get a 4940 even better. You can also get a kit of DDR3 1866 G.skill from Newegg for I think 72 bucks. I bought 3 kits all brand new. Your biggest upgrades gonna be the resistor, it's not a huge one but it's not a micro SMD. You should be able to solder a second resistor without *knock on wood* difficulty. It's a pleasure to offer my knowledge. 

[Falinov]

15 hours ago, Mashcar said:

 

Hello ! I've been researching on subject and it seems my card works okay. Im using third gen i7 55W with rtx 5000 with 280W power supply, what displays in GPU-Z depends on nvidia control panel settings for power. When stressing on maximum performance or optimal/adaptive it clocks up to 1850mhz on core, 1500mhz memory. In PerfCap changes from Idle to PWR or Voltage or both  - I guess that is normal ? So it won't extend 80W supply limit by vbios ?
MXM 3.0 slot should be able to delive at least 100W of power, there are very wide copper traces on card - it's directly from power supply to GPU.

Maybe experimenting with vbios or nvclean settings should help ? Before I made mine work, I had code 43 error and constant high wattage draw on GPU - it overheated until I've correctly installed driver.
CPU is 100% not problem here, I would look into GPU directly and board BIOS/ KBC controller. Those are your culprits

EDIT: When using constant performance in nvidia control panel GPU stays on 600 mhz core, 700 mhz ram, PerfCap is "IDLE", core draws from 25W to -50W max

If it's an HP RTX 5000 then I can tell you with absolute certainty, you cannot and will not have much success modifying the Vbios for that card. It will not work I've tried vbios from DELL, LENOVO, HP, ect. None worked and when the drivers installed it would hard crash the whole system. Stock Vbios is best- following on the power draw is the issue for this card. I assure you. I've already done a modification directly to the motherboard. As for the 280 PSU. I cannot speak for anyone else here but I've never heard of a 280W. I think you probably meant the 240. Either way the PSU you should be using at a minimum a 330 between the CPU and GPU draw I've seen pull in the 300w range before. I doubt your cards actually working to specification being limited by both those power issues. 

 

EDIT: you said 3rd gen i7.

 

You must be using an m18xr2. Yeah the 18xr2 doesn't necessarily require the modification but it does benefit from it. That is my mistake. We are talking about the m18xr3 the 4th gen.

[Naberius]

On 2/26/2024 at 10:46 AM, Falinov said:

If it's an HP RTX 5000 then I can tell you with absolute certainty, you cannot and will not have much success modifying the Vbios for that card. It will not work I've tried vbios from DELL, LENOVO, HP, ect. None worked and when the drivers installed it would hard crash the whole system. Stock Vbios is best- following on the power draw is the issue for this card. I assure you. I've already done a modification directly to the motherboard. As for the 280 PSU. I cannot speak for anyone else here but I've never heard of a 280W. I think you probably meant the 240. Either way the PSU you should be using at a minimum a 330 between the CPU and GPU draw I've seen pull in the 300w range before. I doubt your cards actually working to specification being limited by both those power issues. 

 

EDIT: you said 3rd gen i7.

 

You must be using an m18xr2. Yeah the 18xr2 doesn't necessarily require the modification but it does benefit from it. That is my mistake. We are talking about the m18xr3 the 4th gen.

 

 

(I seem to remember encountering a 280w PSU in the 330w form-factor but I suspect it was either an older spec or made by a third party)

[Falinov]

16 hours ago, Naberius said:

 

 

(I seem to remember encountering a 280w PSU in the 330w form-factor but I suspect it was either an older spec or made by a third party)

Huh. Didn't know that. 

[aldarxt]

11 hours ago, Falinov said:

Huh. Didn't know that. 

I did a google for 280w charger and 1 showed up but I say why when 330w is all you'll need

[Falinov]

5 hours ago, aldarxt said:

I did a google for 280w charger and 1 showed up but I say why when 330w is all you'll need

I agree. Why. Lol granted I use a 780W eurocom and a 330. But its to see if I need the extra horsepower. 

[aldarxt]

15 hours ago, Falinov said:

I agree. Why. Lol granted I use a 780W eurocom and a 330. But its to see if I need the extra horsepower. 

 

15 hours ago, Falinov said:

I agree. Why. Lol granted I use a 780W eurocom and a 330. But its to see if I need the extra horsepower. 

As PC enthusiasts Horsepower is the main ingredient! LOL

[DDAY3493]

I did the resistor mod today too but, I couldn't see any improvements in power or fps. I have the HP RTX Quadro 5000. I don't know if I messed up on the mod, I will take some pictures later on to see what you guys think?!

[Falinov]

1 hour ago, DDAY3493 said:

I did the resistor mod today too but, I couldn't see any improvements in power or fps. I have the HP RTX Quadro 5000. I don't know if I messed up on the mod, I will take some pictures later on to see what you guys think?!

M18x r2 or m18x R3? You won't see a massive leap like in the R3. But I know I saw some pretty significant increases in performance... I'm curious though, which resistor did you modify and how. 

[DDAY3493]

@FalinovI have the M18xR3 just like you! I have done the mod like @ssj92but I didn't see any improvements whatsoever. It's also worth noting I bought the same resistor from Mouser from the link provided by ssj92. The soldering iron gave me some trouble. I will take some pictures after I get home, I'm at work at the moment.

[Falinov]

1 hour ago, DDAY3493 said:

@FalinovI have the M18xR3 just like you! I have done the mod like @ssj92but I didn't see any improvements whatsoever. It's also worth noting I bought the same resistor from Mouser from the link provided by ssj92. The soldering iron gave me some trouble. I will take some pictures after I get home, I'm at work at the moment.

Strange. I immediately saw improvement after the modification. The 5000s performance was abysmal when I first got it. In the 10's to 20fps range an absolute stuttering mess on the most basic fire strike test (not even the ultimate) it would start at 60 and immediately crash and not maintain the core clocks due to power limits. Absolutely junk lmao. It was able to pull a rock solid 110 / 120 and above, after the modification. So it's strange to me you aren't seeing that uplift. I would show you a new set of tests, but the R3 is in pieces right now. 

[DDAY3493]

Yeah it's really strange, I have no idea what it could be but here I took a picture of how it looks with the resistor mod! 

[Falinov]

19 hours ago, DDAY3493 said:

Yeah it's really strange, I have no idea what it could be but here I took a picture of how it looks with the resistor mod! 

That's not correct. Thank God you didn't short it out. 😭