Topic: [4+ EH] Slush Pool (slushpool.com); Overt AsicBoost; World First Mining Pool - page 1053. (Read 4382653 times)

sorry it took me so long to respond to you guys I was away working on the new house we are moving to today.
Anyways

I have the corsair obsidian 800d case (5 fans)

ive included some pictures, ive got the 5870 on top and the 6870 on the bottom

(notice the lonely 6870 sitting on top of the case)

I was hoping to get all three cards in this case (there are 5 pci express 16 slots on the mb)
but with all three mining it gets way too hot

after much experimentation i found a solution to keep them running
the only way I was able to keep this stable is downclock the memory to 300 and only overclock the the cores slightly

im running the 5870 at core / 980 and mem / 300 - getting 410m/hash with it on guiminer poclbm
the 6870 is running core / 900 and mem / 300 - getting 265m/hash with it on guiminer poclbm
also both GPU fans are running manual at 75 % (noticed if I go much higher than that system becomes unstable)
anyways im getting 660+m/hashs out of this setup and the second 6870 stands alone (wish I could get that card in this rig!) any ideas? on how I can get this 3rd card in and keep system stable or how I can speed up m/hash and keep system stable/cool
also I think its important to note I am running Windows 7 Ultimate 64x and I changed the advanced video settings for 'best performance'
all comments are appreciated!
iMMUNE
18XiTbDoNuB51diAwCXajt9QDERTCRyqD4

I meant some disk controllers from higher class have an option to limit io.

I was going to go on about negative pressure vs positive pressure, but just 15 mins ago i broke my large 140mm exhaust (fixing a rattle) to much CFM blowing on the graphics cards upsetting the balance of the fan on the card.

Now i have positive pressure (more cfm intake vs exhaust) if by accident but surprisingly my temps are the same perhaps the rearrangement of the fans could be helping but now i have 152 CFM intake and poor 74cfm exhaust   

Anyhow @immune just try to create a wind tunnel effect prior to my accident it was 152 in 167 out, a  low CFM side and a powerfull front intake helps circulate the air between the cards as the bottom cards teds to heat the next card and so on.

PS if your in a tower i would lay it down as a desktop so the heat from each cards rises onto the roof of the case opposed to the card above it.

thank for answer my question man 

Slush - is the mining interface borked again this weekend?
3 hours of unreported mining..

Speaking of which - is there any way to change the update interval? (This is probably a poclbm question, but since you interface with it, maybe you can give me some info of how reporting works)
I have had 2 BSODs lately and both times wondered whether I lost mining progress and how much of it..

Also I really wish there were a stat that told me how many shares/BTC per hour I just mined(rather than using luck between my account updates, copypasta and a stopwatch). Or is there already such a function hidden somewhere in either client or pool?


Other than that, I am a fairly happy "customer"..

how many GPUs per motherbord?
what distance between GPUs?
how many fans around GPUs?
how many fans on GPUs?
did you downclock memory?

Immune what case you have currently? + what fans?

hey gents

since we are on this topic maybe you can help me out. I built a new rig and im trying to run more than one gpu's at once for mining, I have three a 5870 and two 6850's. I have the proper motherboard, case with plenty of cooling and a 1200 watt power supply, but I can not get even two of them to run well in the case together, they all start getting too hot
any ideas?

iMMUNE

if they are handles by separate instances of miners, then each must have different acocunt in the pool. If they are all handles by a single instance of your miner, then they don't.

for example phoenix must be run separately for each GPU.

I'm sure I read (and it makes sense) that it doesn't make any difference from a performance point of view. The only reason to use multiple workers is if you want different statistics for each one. It could be handy for example to be able to review the last time a share was submitted so you know if a particular machine has gone down.

I have 4 VGAs 5850 and 1 5970, I don't know which best when I combine them in 1 worker or let them run each other. Which best ?

I've come back in your pool today, i've noticed that the stale shares dropped, 15 days ago it was around 2% of stale shares, now it's roughly 1%, good job.

no it's just bitcoind badly written. I have raid10, 300MB/sec transfer, and when downloading blocks it's trashing the HDD like crazy.

Maybe the server has a controller with limit io features?

yes

Talking about IO you mean disk system IO ?

Instant payout means there is a button you can press and get your current balance sent to you immediately. You can limit it to being used only a few times per day. This is a useful feature I'd like to see implemented.

Definitely sounds like somebody is trying to exploit some sort of flaw in the bitcoin p2p connection as a method of attacking the pool.

Today I found similar abnormal activity on bitcoin p2p port like during weekend. I have no idea what was going on, because bitcoin logs say nothing. I decided to limit bitcoin p2p connections only to some range of IPs to eliminate chance that somebody is trying to do something nasty. Unfortunately it looks like those I/O problems caused two invalid blocks (#4371 and #4360). After limiting p2p only to several IPs, p2p and I/O activity stopped. I'll be watching this in following days even more.

Thanks Slush.  Keep up the good work.