Folding Community Board

[Jake_Fisher]

1 hour ago, TVwazhere said:

Unfortunately, yes. Even though it's for a very good cause  

Ok, now I know. At least you left my links for the F@H page about it and YouTube video explaining it, which have links to the donation site. Thanks for leaving those 

[Critaek]

Also smy gpu got stuck on one WU, everytime  she's got to 99.99% windows says fahcore22 stopped working and it restart from 0%, it really sucks.

[rodarkone]

yep happened to me 2 days ago .. one of my 980M GPU's reached 99% and the FAH was dead, on restart .. lost all progress  - in theory it should save the progress every XX minutes but ... in my case no save

 

[MISTERDIEABETIC]

1 hour ago, rodarkone said:

yep happened to me 2 days ago .. one of my 980M GPU's reached 99% and the FAH was dead, on restart .. lost all progress  - in theory it should save the progress every XX minutes but ... in my case no save

 

I had the same thing happen to me a few times when running it on my  craptastic laptop. No biggie though, would take 9 days running 24/7 at full power just for a 60,000 point work unit lol

[cbigfoot]

call the presses: my cpu runs circles around the Vega 7 lol

[ssmmdd]

9 hours ago, Kilrah said:

I had one on 13408 that did half the usual but I dumped it, and a few in the range that failed

 

Apologies for the n00bness, but how do you get those stats?

[rodarkone]

 

I'm contributing to Project 13409

Testing new core22 release v0.0.5 for our ability to roll out extremely large scale relative binding free energy calculations in support of the COVID Moonshot.

To hear more about the COVID Moonshot and how Folding@home is collaborating with them, check out this recording of a talk we recently gave, hosted by BioSolveIT.

 

 

13409 is the same low points / high utilization low PPD as 13408  not sure who setup those units so badly if an average GPU (Vega / RX 5700xt) needs to run it for  about 1 hour for 13K points when in reality should produce at least 5 times as much. 

 

[Critaek]

9 hours ago, rodarkone said:

 

I'm contributing to Project 13409

Testing new core22 release v0.0.5 for our ability to roll out extremely large scale relative binding free energy calculations in support of the COVID Moonshot.

To hear more about the COVID Moonshot and how Folding@home is collaborating with them, check out this recording of a talk we recently gave, hosted by BioSolveIT.

 

 

13409 is the same low points / high utilization low PPD as 13408  not sure who setup those units so badly if an average GPU (Vega / RX 5700xt) needs to run it for  about 1 hour for 13K points when in reality should produce at least 5 times as much. 

 

Exactly the same WU as mine, giving 4 times lower points than normal. Also really low gpu usage, mine never went higher than 45°C.

[TVwazhere]

13 hours ago, ssmmdd said:

Apologies for the n00bness, but how do you get those stats?

HFM.net I believe

[Jake_Fisher]

So, here I am folding along and I've noticed this trend a lot on my gpu. For the most part, every minute or so my usage drops to 0% (with core & memory clock falling to minimum most of the time) along with the obvious recurrent dips of temp and fan speed. It's an ASUS R9 290 DirectCU II (don't be jealous) so it's not exactly the top of the line as far as R9 290s go, but it's not a slouch either. It's not OC'd, in fact the memory clock is underclocked by 60mhz (I had weird monitor blackouts (with a hard freeze) and once an OC failed at boot a couple weeks ago so I took off all of my computer's OCing...everything stable since then).

 

Could this just be an odd quirk? Some throttling? Anything to be concerned about? I've had this setup for 5-6 years (I, being poor, have only my dreams). This is the most work my gpu has been doing in probably half a year. 

[Favebook]

4 hours ago, Jake_Fisher said:

So, here I am folding along and I've noticed this trend a lot on my gpu. For the most part, every minute or so my usage drops to 0% (with core & memory clock falling to minimum most of the time) along with the obvious recurrent dips of temp and fan speed. It's an ASUS R9 290 DirectCU II (don't be jealous) so it's not exactly the top of the line as far as R9 290s go, but it's not a slouch either. It's not OC'd, in fact the memory clock is underclocked by 60mhz (I had weird monitor blackouts (with a hard freeze) and once an OC failed at boot a couple weeks ago so I took off all of my computer's OCing...everything stable since then).

 

Could this just be an odd quirk? Some throttling? Anything to be concerned about? I've had this setup for 5-6 years (I, being poor, have only my dreams). This is the most work my gpu has been doing in probably half a year. 

That is totally normal. Those dips to 0% usage are either 3min gaps for manual checkpoints (this one you can change in settings) or they are automatic checkpoints (not actual checkpoints but rather one part of WU was done and it moved on to next). I've noticed a lot of this on my older GPUs, especially AMD. Nvidia seems to not have that 'problem'.

[Jake_Fisher]

34 minutes ago, Favebook said:

That is totally normal. Those dips to 0% usage are either 3min gaps for manual checkpoints (this one you can change in settings) or they are automatic checkpoints (not actual checkpoints but rather one part of WU was done and it moved on to next). I've noticed a lot of this on my older GPUs, especially AMD. Nvidia seems to not have that 'problem'.

Ah, got it. Thanks for the info! 

[ssmmdd]

8 hours ago, TVwazhere said:

HFM.net I believe

<snip>

Thank you!

[RollinLower]

1 hour ago, yaboistar said:

sorry it took so long folks

 

295x2 is online. 1.4mppd per core reported

those numbers are insane considering it's basically 2 regular 290(X?)'s! damn impressive!

[Bitter]

2 hours ago, yaboistar said:

sorry it took so long folks

295x2 is online. 1.4mppd per core reported

I have doubts, it doesn't look like it's been doing work very long yet. Let's see what the average is after 100WU's then see.

[Kilrah]

But the report was invalid anyway since it had only just started...

[GrockleTD]

New folding rig... @Gorgon I've definitely asked before, but is going with something like Ubuntu worth it in terms of PPD?

[Bitter]

Turned the Vega56 back on for a night to leapfrog me back up a bit. Need to fix the positioning of some fan wires on the P106 tomorrow.

[justpoet]

On 6/3/2020 at 8:52 PM, leadeater said:

Solar does a decent job but doesn't do jack in winter so looking to blend so I get more than nothing in winter 

Spoiler

 

 

 

.

Wind is really good here so a 600W-2000W turbine should do nicely, peak output of my solar is only 2kw and wind blows all day here so who know, might work out really well or be terrible lol.

What are you thinking of for a wind solution?  I've thought of looking into some of the smaller scale vertical wind turbines to put on some rooftops, but never went further down that path.

 

FWIW, here's my 3.1kw west facing array output this year, which I installed in December.

Spoiler

 

[leadeater]

5 hours ago, justpoet said:

FWIW, here's my 3.1kw west facing array output this year, which I installed in December.

Hmm my peak output is 2kw and the best month so far I've had is 445KkWh, most I'll produce in a single day is about 18kWh, so my best case is mid 500kWh. Looks like ours is very similar. Mine is actually 9x 325W panels but the micro inverters are only Enphase IQ6 so cap the panels at ~220W maximum continuous, wasn't worth the extra cost for IQ7 at the time.

 

5 hours ago, justpoet said:

What are you thinking of for a wind solution?  I've thought of looking into some of the smaller scale vertical wind turbines to put on some rooftops, but never went further down that path.

Not sure on how I'm going to link it in to my system due to my battery configuration and running voltage since it's powering a UPS, DIY solution with free batteries and UPS. I'm either just going to directly grid connect the turbine which will power the existing battery changers, but this has large efficiency loss, or I'm going to have to move the batteries to a 48V configuration and have the turbine charge those but I'll likely need boost battery chargers as well.

 

I'm looking at this since the price is fair, certified for sale and use in my country and is trustworthy in it's claims to be quiet when operating, unlike the AliExpress ones.

https://www.tesup.nz/product-page/atlas2-0-wind-turbine-made-in-europe-new-zealand

 

Generally speaking roof tops don't work too well as the air is rather turbulent which isn't good, there are ones designed for building mounting that handle that better, not sure if this is one of those or not. The other thing to factor in is that you get between 10%-40% Wh production for the rated W of the turbine, so if it's a 2kW turbine you'll get between 200W to 800W over an hour. These are average figures over a day so you'd be looking at 4.8kWh to 19.2kWh per day production on days with decent wind for the day, so expectation safety half that and assume many days of low wind so even less when averaged over a month in actual use.

 

Use Google and find weather stations near you and look at average wind speed per day in a given month, you want greater than 8 m/s (17.9 mph) 15 minute average a fair decent amount over a day and those conditions over a month otherwise payback is really long and usefulness becomes questionable. Below 5 m/s (11 mph) is where I'd say it's useless.

 

Higher you can mount the turbine the higher the average wind speed is, 10m pole seems to be the recommended best but that's pretty damn high (two story house roof height).

 

Residential solar will out produce residential wind rather easily but you need a lot of roof area and it's rather or very seasonal, main benefit of wind is when solar is bad generally wind is good, rarely both at the same time.

[chaozbandit]

6 hours ago, GrockleTD said:

New folding rig... @Gorgon I've definitely asked before, but is going with something like Ubuntu worth it in terms of PPD?

 

 

In my testing (on Linux), there's a noticeable improvement on nvidia GPUs (38% improvement on a 1080 via FAHBanch) and a slight bump overall in CPU. The only downside is there's no easy gui for things like controlling AIOs but you can still set fans and power limits via nvidia settings.

[Windows7ge]

This seems as good a place as any to ask but can I run BOINC and F@H at the same time? When I free up my GPU's in the coming months I want to install them in a BOINC rig. The project I'm running doesn't have GPU tasks so I figure let's give F@H a shot if I can run the two at once.

[justpoet]

1 hour ago, Windows7ge said:

This seems as good a place as any to ask but can I run BOINC and F@H at the same time? When I free up my GPU's in the coming months I want to install them in a BOINC rig. The project I'm running doesn't have GPU tasks so I figure let's give F@H a shot if I can run the two at once.

Yep, both can run together.  Just be careful not to over-do things like CPU and end up choking out the GPU (they still need a CPU core or so to run right).

[Windows7ge]

1 minute ago, justpoet said:

Yep, both can run together.  Just be careful not to over-do things like CPU and end up choking out the GPU (they still need a CPU core or so to run right).

I have it run 14/16 threads. I'd reduce BOINC down to 12/16 as I have 2 GPU's. I have relative confidence that that should be sufficient.

[marknd59]

36 minutes ago, Windows7ge said:

I have it run 14/16 threads. I'd reduce BOINC down to 12/16 as I have 2 GPU's. I have relative confidence that that should be sufficient.

I was running both BOINC and F@H during the Pentathlon and had no problems. If they are NVIDIA cards each one will keep 1 thread at almost 100% useage most of the time. If they are AMD the thread useage is no where near as high.