ShadowChaser

The recent F@H sprints have allowed me to revive my PC of Theseus that has existed in some form since 2019. Currently the only things remaining from the original build are just the fans and case, but it's fun to see this thing still hanging around.
 
The latest iteration is just a mish-mash of spare parts I had laying around to get some extra points and warm up my apartment:
CPU: Ryzen 5 5600
Mobo: AsRock B550 PG4
RAM: Corsair Vengeance 2x16GB 2666
GPU0: EVGA RTX 3060Ti FTW3
GPU1: Nvidia Tesla P40
Storage: Samsung 970 Evo 500GB
PSU: Revolution D.F. 650W
All this built into a Raijintek Ponos and uses Noctua Redux 140 and 120s as case fans.
The Tesla P40 has a dual 40mm adapter I designed and is cooled by Arctic 40mm server fans.
 

Pretty pleased with the cable management in this iteration. For a budget case the Ponos is pretty good on cable routing options and that has always come in handy whenever I build in this chassis.
 
I can't quite get the P40 to dissipate more than 200-ish watts without the fans going screamo mode but I haven't repasted it yet. Hopefully that'll allow me to drop a few degrees and let me fun the 40mm fans at a more reasonable rpm. It is living in a hallway where the ambient temperature is about 15C so that does help matters a little 😛 

To be completely honest you should use one of these: https://www.newegg.com/silver-raijintek-0r100006/p/2T3-0007-00003
Plus some cheapo heatsinks to slap on the memory ICs if you wanted to cool this gpu the way you want to. But for that amount of money you probably could've gotten a better GPU to begin with... such is life.

Noctua 4020s are inaudible - Get about 150W of TGP at 80C
Arctic 4038s @5k rpm are drony - Get about 180W of TGP at 80C (This is what I daily drive as F@H rarely needs more power than this)
Arctic 4038s @15k rpm are jet engines - Get the full 250W TGP at ~65-70C
 
Originally I got this for some local LLM but 24GB of vram is getting borderline for some of the larger creative writing models and the token generation rate of Pascal is just not it chief. I think I'll be getting something Turing based for inference in the future. In the meantime, this is generating science and keeping my apartment warm 🙂 

do you have room to attach two 40mm fans on the end? Due to the heatsink design it isn't possible to flow air in the way you've described. The finstack is enclosed at the end that you want to attach fans to. It's harder to describe but if you look through the fins it'll be pretty obvious what I mean.

I recently designed this adapter to put dual 40mm fans on my P40. Regular 40mm fans can screw in and deeper fans clip on. When I tried with regular noctua 4020s it is somewhat borderline but with proper server fans it can be cooled but is loud.

The recent F@H sprints have allowed me to revive my PC of Theseus that has existed in some form since 2019. Currently the only things remaining from the original build are just the fans and case, but it's fun to see this thing still hanging around.
 
The latest iteration is just a mish-mash of spare parts I had laying around to get some extra points and warm up my apartment:
CPU: Ryzen 5 5600
Mobo: AsRock B550 PG4
RAM: Corsair Vengeance 2x16GB 2666
GPU0: EVGA RTX 3060Ti FTW3
GPU1: Nvidia Tesla P40
Storage: Samsung 970 Evo 500GB
PSU: Revolution D.F. 650W
All this built into a Raijintek Ponos and uses Noctua Redux 140 and 120s as case fans.
The Tesla P40 has a dual 40mm adapter I designed and is cooled by Arctic 40mm server fans.
 

Pretty pleased with the cable management in this iteration. For a budget case the Ponos is pretty good on cable routing options and that has always come in handy whenever I build in this chassis.
 
I can't quite get the P40 to dissipate more than 200-ish watts without the fans going screamo mode but I haven't repasted it yet. Hopefully that'll allow me to drop a few degrees and let me fun the 40mm fans at a more reasonable rpm. It is living in a hallway where the ambient temperature is about 15C so that does help matters a little 😛 

The recent F@H sprints have allowed me to revive my PC of Theseus that has existed in some form since 2019. Currently the only things remaining from the original build are just the fans and case, but it's fun to see this thing still hanging around.
 
The latest iteration is just a mish-mash of spare parts I had laying around to get some extra points and warm up my apartment:
CPU: Ryzen 5 5600
Mobo: AsRock B550 PG4
RAM: Corsair Vengeance 2x16GB 2666
GPU0: EVGA RTX 3060Ti FTW3
GPU1: Nvidia Tesla P40
Storage: Samsung 970 Evo 500GB
PSU: Revolution D.F. 650W
All this built into a Raijintek Ponos and uses Noctua Redux 140 and 120s as case fans.
The Tesla P40 has a dual 40mm adapter I designed and is cooled by Arctic 40mm server fans.
 

Pretty pleased with the cable management in this iteration. For a budget case the Ponos is pretty good on cable routing options and that has always come in handy whenever I build in this chassis.
 
I can't quite get the P40 to dissipate more than 200-ish watts without the fans going screamo mode but I haven't repasted it yet. Hopefully that'll allow me to drop a few degrees and let me fun the 40mm fans at a more reasonable rpm. It is living in a hallway where the ambient temperature is about 15C so that does help matters a little 😛 

The recent F@H sprints have allowed me to revive my PC of Theseus that has existed in some form since 2019. Currently the only things remaining from the original build are just the fans and case, but it's fun to see this thing still hanging around.
 
The latest iteration is just a mish-mash of spare parts I had laying around to get some extra points and warm up my apartment:
CPU: Ryzen 5 5600
Mobo: AsRock B550 PG4
RAM: Corsair Vengeance 2x16GB 2666
GPU0: EVGA RTX 3060Ti FTW3
GPU1: Nvidia Tesla P40
Storage: Samsung 970 Evo 500GB
PSU: Revolution D.F. 650W
All this built into a Raijintek Ponos and uses Noctua Redux 140 and 120s as case fans.
The Tesla P40 has a dual 40mm adapter I designed and is cooled by Arctic 40mm server fans.
 

Pretty pleased with the cable management in this iteration. For a budget case the Ponos is pretty good on cable routing options and that has always come in handy whenever I build in this chassis.
 
I can't quite get the P40 to dissipate more than 200-ish watts without the fans going screamo mode but I haven't repasted it yet. Hopefully that'll allow me to drop a few degrees and let me fun the 40mm fans at a more reasonable rpm. It is living in a hallway where the ambient temperature is about 15C so that does help matters a little 😛 

Sorry I'm late guys, but I finally built my new PC with my 408-
I missed it!?!?!?!?! FUUUUUUUUUCK
Well at least I can still comment about the eve-

I would too if most of my computers weren't complete hot boxes. I'd drop my power limit by 20% and call it a day. Miss the days when gpus topped out around 200W

Well i lost track of time We have 6-8" of snow here in Nashville. The State and City are basically shut down and im watching Daniel Craig kick some @$$ lol.  
 
Neeless to say stats are here Birthday Bash Daily Ranks - Google Sheets
 
One more day to go.  we have hit 4.3B We may get close to 5 Billion. 
 
We have had 61 people fold every day of this event out of 71 that have produced points and 85 that signed up.

The cold snap hitting us has me considering bringing the Tesla P40 I have back online. An extra 250W might help with the temperature at night as this apartment has terrible insulation 😛 only question is what computer to chuck it in and how to cool it since the server that used to house it no longer exists.

sometimes some f@h WUs are just really hard on hardware. My 3060Ti crashed out of nowhere on one specific WU despite that undervolt being rock solid for literal months playing games or rendering/simulating
 

Don't, it's not worth it, CPU folding is horribly bad. I'm just not paying for the power and they aren't in my home. I do it just to find out how much you need to be equivalent to a particular GPU. You end up using between 6kw and 9kw just to match a middle to high end GPU using only 150W-250W, as you can see super not good.

oh yes, I was just using that as an example
my nvidia gpus tend to run ~60% of their max power draw so there's not really any point tweaking the PL and stuff but the amd ones tend to run closer to their limits so I will drop it as much as the software will allow.

I would too if most of my computers weren't complete hot boxes. I'd drop my power limit by 20% and call it a day. Miss the days when gpus topped out around 200W

oh yes, I was just using that as an example
my nvidia gpus tend to run ~60% of their max power draw so there's not really any point tweaking the PL and stuff but the amd ones tend to run closer to their limits so I will drop it as much as the software will allow.

I would too if most of my computers weren't complete hot boxes. I'd drop my power limit by 20% and call it a day. Miss the days when gpus topped out around 200W

wow 30 whole watts huh?

I know the 6950XT is pushed a little far beyond the optimal spot on the VF curve so for my daily use I simply dropped the PL by 10% and let it chug away happily at the 255W TGP that results. I got interested in seeing how much further down the VF curve I could go considering that a 10% drop in power is the maximum that AMD supports on this guy compared to the 40%+ on other GPUs.
 
I would be interested in seeing how other 6950XTs compare but I doubt people bought this card just to kneecap it with ridiculous limitations 😛
 
I used Afterburner to make changes though Adrenaline would be fine as well, Time Spy to verify stability, and Furmark (1080p 8xMSAA) to get worst-case power draw. My testing was mostly limited by how much time the Time Spy stress test takes to run as well as how this card simply refuses to run all core loads at lower power (as is to be expected given the original PL limitation)

Undervolt Target Clock Power (W) Core Clock Core Voltage Time Spy FPS Avg (20 pass) %diff FPS %diff Power Furmark Power Furmark FPS %diff FPS %diff Power -10PL Stock N/A 255 2235 1006mv 131 --- --- 255 114 --- --- 1080mv 2400 245 2330 956mv 137 4.58 -3.92 272 121 6.14 6.67 1090mv 2300 220 2235 900mv 132 0.76 -13.73 245 117 2.63 -3.92  
The baseline is the -10% PL that I have been dailying and so the goal was to match performance while pulling less power. Should be easy, right?
My best result had the card running at 2235mhz @ 0.881v but it was borderline unstable so bumping it up to 0.9v cleared things right up. A power reduction of over 30W in a traditional rendering scenario is pretty solid IMHO and even under a torture test it still drew less power than the original power limit change.
 
This GPU really doesn't like to be worked hard under 0.9v I've found. Not sure why that is, perhaps it's so low that the vf curve just falls off. Maybe someone with more knowledge on how AMD's powerplay tables works could educate me on this as I've only really UC/UVed Nvidia GPUs in the past and they tend to be just as happy running at 0.85v as 1.05v.

sometimes some f@h WUs are just really hard on hardware. My 3060Ti crashed out of nowhere on one specific WU despite that undervolt being rock solid for literal months playing games or rendering/simulating
 

sometimes some f@h WUs are just really hard on hardware. My 3060Ti crashed out of nowhere on one specific WU despite that undervolt being rock solid for literal months playing games or rendering/simulating
 

I did manage to scrounge around and get enough parts to get my EVGA card up and running. Figured that this was an event that deserved having this gpu work a little instead of living in a display next to my desk 😛 
 
Too bad AMD cards aren't great for crunching numbers, but tagging out my 6950XT for another 3060Ti will be great for my power bill XD

I did manage to scrounge around and get enough parts to get my EVGA card up and running. Figured that this was an event that deserved having this gpu work a little instead of living in a display next to my desk 😛 
 
Too bad AMD cards aren't great for crunching numbers, but tagging out my 6950XT for another 3060Ti will be great for my power bill XD

I did manage to scrounge around and get enough parts to get my EVGA card up and running. Figured that this was an event that deserved having this gpu work a little instead of living in a display next to my desk 😛 
 
Too bad AMD cards aren't great for crunching numbers, but tagging out my 6950XT for another 3060Ti will be great for my power bill XD

I did manage to scrounge around and get enough parts to get my EVGA card up and running. Figured that this was an event that deserved having this gpu work a little instead of living in a display next to my desk 😛 
 
Too bad AMD cards aren't great for crunching numbers, but tagging out my 6950XT for another 3060Ti will be great for my power bill XD