Jump to content

FLDC


road-runner

Recommended Posts

I honestly think the "casual" miner crowed doing get rich quick will soon flood the market and you can grab a 1070 cheap(er) if you don't mind a little waiting.  Sadly I doubt Vega's going to be as good as we all hope comparing to the "workstation"(ish) card and with HBM2 costs I doubt the pricing will be as aggressive as AMD is usually known for.  With that said I'm just an idiot on the internet and unless I'm talking about what's out and known now I can end up looking just as stupid as anyone. ;)

 

As for the Folding OC thing,  my oh so scientific method of OCing my card mid WU netted me some clear gains in PPD/frame time going from my stock SC2 Black Edition speeds that I don't recall atm to the 2012/6003 it's at currently so it seems there's something to be had but I haven't tried adjusting just the core or just the memory (though I guess it may vary by WU anyway) For now I'm playing PUBG OCed and that's stable. (I spent the first 2 days with the game end of last week playing on my laptop because the main rig would crash to desktop after a couple rounds)

Share this post


Link to post
Share on other sites

+1,.. most projects use less than 864 MB memory,  with 3% up-to 45% memory controller load. Overclocking my cards memory has never shown any increase in PPD,.. but bumping up GPU core clocks, if you have the cooling can make a noticeable heat and PPD increase.   :yes:  

Edited by Braegnok

Share this post


Link to post
Share on other sites

Do you guys fold with your cpu?

 

I read that CPU folding can hurt the gpu scores, so I stopped cpu folding over the weekend. I haven't really noticed any change in ppd though. The CPU wasn't doing much points wise, and the GPU scores did't really improve.

 

I now have the cpu mining and the gpu ppd still seems to be essentially the same..., so just curios what you all do.

Share this post


Link to post
Share on other sites

Depends on your GPU, AMD doesn't really need much CPU power while folding, Nvidia cards on the other hand seem to need ~one core per GPU.  By default the folding client will adapt to this for the CPU slot and for example for an i7-4770K + GTX 1070 give the CPU slot access to 7 cores (HT counts here as a full core).  If the priority of other processes (such as mining) are too high and need the CPU core the folding relying on it can slow down.  With that said mining usually ignores HT cores so in the i7 example above there would be plenty resources for the CPU to mine and the GPU to fold. 

 

In other news, I decided to do some testing and typed up a big thing on it the other day about the GPU memory thing before OCC decided I needed a robot check and lost it all when posting so here's the short version of my testing (I can't be bothered to type it all out again) In some cases GPU memory makes a very small difference and GPU clock speeds should be prioritized, but OCing both GPU core and memory yielded a between .1% and 3% boost in performance and for a card currently making 1 million ppd that's a pretty big boost.  Testing was with my 1080 Ti and a mix of FAHBench and live WUs (tested by changing OC mid unit) with known stable OCs.

Share this post


Link to post
Share on other sites

Anyone else using the advance and 100% flags for folding? It seems I get about 800k PPD doing this and 1.2mill with the flags off. Anyone know if it matters since from what I could find about folding coins they are based on the amount of work your produce and not total PPD.

Share this post


Link to post
Share on other sites

Anyone else using the advance and 100% flags for folding? It seems I get about 800k PPD doing this and 1.2mill with the flags off. Anyone know if it matters since from what I could find about folding coins they are based on the amount of work your produce and not total PPD.

 

Advanced and 100% is best way to roll,..  :)

 

Under >Configure<, advanced tab, Checkpointing Frequency I set mine for Max 30-min. also. 

Edited by Braegnok

Share this post


Link to post
Share on other sites

So you all are using the 

 

Advanced

100%

Big 

 

flags under client type under :Configure--> Slots---Edit each slot? 

 

What kind of PPD increase does that net? 

 

No clue just read in several places for these new powerful nvidia cards you should use those three...  

 

client-type

advanced

 

next-unit-percentage

100

 

max-packet-size

big

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
×
×
  • Create New...