this card looks like it's in between having high density setups vs more mh/s per dollar. you can get around the same hashing speed with 2x 5830 for only around $250, although it would require more hardware space. this is not nearly as bad as the 6990 though.
Not quite as easy to find 5830s anymore unfortunately, although the Mh/$ ratio for the ones that recently sold brand new for $130 easily passed 2 Mh/$ making them very desirable cards. I bought two when Newegg had them for $130 and I can get each one to run ~330 Mh/s completely stable with added voltage, ~324 at stock. To get the same hash rate to price ratio out of the AX6870X2, you'd have to buy one for under $300! While the 5830s are about the best in terms of Mh/$, there is also something to be said for the AX6870X2's higher hash rate per card. At some point, performance per slot becomes more and more important. amzon has used 5830 at 139 and there's always ebay, seems there's always a good deal out there if you watch it. |
|
|
|
|
They have 61 cards left in stock.
|
|
|
|
|
I assume under linux, each instance of your miner pointed to the correct GPU number?
if you are getting 1600 from 3 6990s, it's a little low there at 266 Mhashes per GPU... you should be getting 400 Mhashes per GPU with very modest overclocking.
|
|
|
|
|
Re-apply thermal grease. People have used arctic silver and have reported better temps after applying them.
there's even a post about this subject here on the forums with pics and instructions. you'll have to search for it.
|
|
|
|
i had issues like this as well and ended up going to a 900w psu. those 5830s are power hungry especially when over clocked. i know lots og people say 600 or 700 is enough for two or the rule of thumb for btc mining is 200w per card plus 150 for mobo, memory... but for my 6 systems i needed about 250 or more per 5830. just my 2 cents
200+200+150=550 /750 =73.33 % load on a 750W PSU. if you have an 80% efficiency 750W PSU, it should be more that plenty. HOWEVER!!! if you have a 750W PSU and the 12V rail is not delivering enought amps for both cards, then it would be a problem. I'm not sure how you were drawing 250 /card, sounds like you were overvolting and then, yes you need more juice: 250+250+150=650/750= 86.66% here you'd need a Gold cert. 80% efficiency 750W PSU at least... |
|
|
|
I have been noticing something weird over the last few days. I usually keep the statistics page open to see progress and check on luck every once in awhile. Now, I've been having a few moments where I could have sworn my BTC balance at the top right said one thing but infact said another. Well, just a moment ago after about an hour I refresh and noticed my BTC balance dropped by .01. And if this is infact what was happening before, there were other instances where I say my balance drop by .01 or .02.
So it leads me to ask what is going on? Is the 3% fee taken out an hour or more later for a previous batch of rewards or something?
I saw exact same thing happening to my balance in the last 2 days... I thought I was just imagining. Glad to see that I was not. What does it mean? |
|
|
|
What everyone here forget, hashrate doesnt equal share submits.
If you process/submit shares slower than your hashrate, you can get reports of even half your actual hashrate on the poolsite. Pools only report hashrate based on submitted shares.
Clipse, We're not forgetting anything here. What you are saying is trivial and obvious. What we are pointing out is the timing of this observation. What I'm saying is that recently I started noticing wild swings. I fully understand how the reported rate on the website is arrived at, but I have not seen such large swings before. I wonder if this is tied to the recent release of 2.1 phatk |
|
|
|
|
BTW, I called XFX just now, and they said don't expect any to ship till the end of August.
|
|
|
|
I spoke with TigerDirect customer service earlier this morning to find out if these cards were actually in stock or not and was told the date they were given by the manufacturer for receiving them was next week and that they were expecting 30 in stock. I called again here just a little while ago to confirm that response and was simply told their system showed that if you ordered now it would be on backorder. So keep in mind that what you guys are buying is a pre-order for cards they may or may not get and that these are not actually in stock. TigerDirect is notorious for taking a bunch of orders for items they can't fulfill.
They were supposed to get 100 July 29th from XFX (according to TD, just got off the phone, checking the status of my order). That shipment has not come in yet... TD has about 30 orders for these cards right now. They offered to cancel the order or change to a "comparable" device.... well, there was no comparable device (i did ask if by some chance they had a 5970 collecting dust over there... LOL ), so I asked them to just let the order stand as is... will check with them on Monday. On another note, here's a link that shows more info on inventory: NOTE! this is not Tiger Direct's inventory AFAIK, don't get confused. http://www.shopblt.com/cgi-bin/shop/shop.cgi?action=enter&thispage=0110040015014_BGB5888P.shtml&order_id=!ORDERID! Give the page a efw seconds to load the inventory info in the middle of the page... 158 on order, 99 expected by 8/15 |
|
|
|
I am getting rejects and hardware problem errors since the modification of __init__.py. Diapolo's 7-11 version is the last stable version for me on my 2 5830's @ 1000/350 stock voltage. I switched back to 7-11 version last night and today on BTCGuild I am back to showing 4500 (31, 0.68%) on one card and 4324 (27, 0.62%). I am getting 320mh/s with the 7-11 version and was getting 324mh/s with your phatk 2.1. The number of stales/rejects showing on BTCGuild and in phoenix log after a short period of mining was up to and over 3% on both cards yesterday. If you would like I can let one card run each version to compare the difference. I can also provide a log of phoenix if needed. Not for certain about this, but I believe I was only getting the "hardware problem?" error from diapolo's versions after 7-11 and your phatk2.0. With phatk2.1 I saw the introduction of the rejected shares. I can provide any info if needed. I assume the kernel is pushing the card too hard as Diapolo mentioned earlier in his thread. I have no idea why it is affecting some and not others.
I had to reduce the OC by 10, kept same memory settings with my 5830s using 2.1 |
|
|
|
Limit 1.  |
|
|
|
|
Did you name the frog Mr. "Rig Marshall"?
|
|
|
|
? where's the nerd fight
Stop it, I'm tryin to get my troll on! Bad potato!!! LOL |
|
|
|
NEEERRRRRRRRD FIIIIIIIIIIGHT!!!
? where's the nerd fight |
|
|
|
My always sweet spot for 5870 is memory clock is equal to core clock divided by three.
mem= core/3 = 975/3=325
Would this hold true for 5850 and/or 5830s? I'd say it hold fairly true for my 5830's I get my best megahash as 1030/350. I've only done minor tests with my 5850's but I have a really hard time getting the mem clock down because it starts to crash my system when I do that. Mike, did you ever figure out that problem you had with MSI Afterburner? I think you posted on my thread as well about same issue I had... |
|
|
|
My always sweet spot for 5870 is memory clock is equal to core clock divided by three.
mem= core/3 = 975/3=325
Would this hold true for 5850 and/or 5830s? |
|
|
|
|
Mobo's definitly contribute to this issue.
I wonder if it's the mobo's ability to draw power to that slot or if it's bus timing.
to test if it's power draw related, the second card, the one that isn't able to reach higher speeds, a cable extender with molex (power) should be used. If it reaches expected speeds, then it's power issue through mobo to that specific pcie slot.
otherwise, it's something to do with bus timing. In this case, maybe under-clocking CPU and memory (if possible) can be explored to see if anything can be improved...
I'm not sure what else to consider here...
|
|
|
|
|
Do us all a favor and throw that card out and stop asking.
there's a page with hardware comparison: google "bitcoin hardware"
|
|
|
|
That is overkill of a power supply
I would also guess that you are looking at a system drawing somewhere around 900-1000w before overclock. Overclocking will jump you to about 1200w if you push the cards hard.
really? do explain how, if you were to draw 1200, a 1500 is an overkill??? |
|
|
|
|
like the man said: for future. the 8 cards in the pic are to show how it would look like with 8 cards. that's all.
|
|
|
|
|
Show Posts
