the 6870 pulls about the same hash rate as the 5830 - I have both. and even that CPU is way more than enough to drive your card and it should barely be ticking over while doing it.
do this: change your SDK to 2.4 and the driver to 11.6. see if that drops your CPU usage.
|
|
|
Don't confuse short-term and long-term. Current usability is mediocre, agreed. Make an attempt to see the future, Shinobi. The first automobile was considered a silly, expensive toy for the rich.
also the 1st land line phones were considers a fad for the rich that would go nowhere. everyone was like, "just who you gonna call, anyway??? hardly anyone else has em." course bitcoins may not ever get that popular (or even a small percent of that, or just fail anyway), but without someone giving it a shot and hanging in there through the tough times it has no chance at all.
|
|
|
RE: possible incorrect mem clock reading on ASUS directCU6870
Thanks for the feedback. I'd be very cautious of believing any other number than that one reported back by cgminer. There are different ways of querying the card that report back different values and cgminer is simply reporting back what the card says is its current running speed. On my 6970s it says the speed was set fine when I set it low, but then reports back at full memory speed. Looking at the temperature+fanspeeds confirms it is definitely running at the hotter 1375 (default).
interesting. GPU-Z reports mem temps (well I think its mem temp) so Ill play with the clocks and see if the temps move around enough to see what the clocks actually are. Ill let ya know what I find - if I can figure it out that is, heh EDIT hmm I doubt its mem temps its monitoring. crud.
|
|
|
for windows I use GPU-Z http://www.techpowerup.com/gpuz/kinda have to guess which sensor is the VRMs, though. play with the clocks and you can usually figure it out.
|
|
|
possible memory clock display bug? running CGMiner 2.0.2 on win7 64 bit. sdk 2.4, 11.6 driver. ASUS directCU 6870 non ref card. mem clock is displayed wrong in win7 build on the ASUS. "--gpu-memclock 300" is in the command line and GPU-Z and CCC both correctly show the 300 MHz clock speed, so it does get set. below is the "G" output. -------------------------------------------------------------------------------- [P]ool management [G]PU management [S]ettings [D]isplay options [Q]uit GPU 0: 74.0C 2807RPM | 311.3/314.8Mh/s | A:59 R:0 HW:0 U:4.70/m --------------------------------------------------------------------------------
GPU 0: 319.8 / 316.6 Mh/s | A:29 R:0 HW:0 U:4.30/m 74.0 C F: 71% (2840 RPM) E: 1010 MHz M: 1050 Mhz V: 1.200V A: 98% P: 0% this is a cosmetic bug only as the clock is correctly set. and ONLY the asus 6870. on my 6770 (2.0.2, vista 32 bit, SDK 2.4, 11.6 driver) the mem clock displays correctly. other than that, which may be specific to that brand/model card (its very non reference, 6 phase power etc.. and I actually dont care if it gets fixed, just figured youd want to know) FANTASTIC.
|
|
|
I can design this with 3 fans and have cooler temp... guarantee. Wanna bet?
Hint: check out server chassis designs.
then do it. but make your own thread please, and reply to this message in that. server chassis usually have straight as an arrow (much as possible anyway) enforced air flow.. how you gonna do that with side sucking and all around exhausting video cards (I assume by force feeding air to only ref design cards and using their fans to augment the other fans, but thats cheating)? with a multi level (cards on extenders) design? and what 3 fans? how quiet will it be? how expensive?
|
|
|
new version is sweet!
I like the cleaner output. the gpu and mem clock ranges rock. and the emergency temp cutout. and performance is better than any other miner Ive run before by an easy 6+ percent. and I could go on..
I see Im gonna have to redo my sig with updated figures heh
|
|
|
I have a newly installed 850w Kingwin P/S that replaced a 500W and the 350W that I was using to power the 5830. System seemed to be stable, going 7days + without a hiccup, but now seems to lock up (screen frozen, or blacked out when I turn the monitor on) at least once a day (of course, usually just after going to bed, or leaving for work).
Looking for some ideas on what might be wrong. Dirty power? [...] Just frustrating to have a system that ran so well becoming such a PITA.
I would seriously consider dumping that kingwin psu. they are junk. the fact that its worse with this psu (as opposed to your 2 psu setup) should tell you something. you are getting pure lockups right? nothing in the event logs? thats another indication its bad hardware, not a driver issue. lower clocks = less power draw from the psu of course, and less voltage swing on card loading/unloading (on a junk PSU that is, real ones have no problem). could be another reason milder clocks give less chance of lockups with your current setup. for a serious psu (one that delivers its rated power, cleanly, and at real world temps), get an antec, pc power and cooling, corsair, seasonic, or silverstone.
|
|
|
just gave this a whirl on my 6770.. NICE! I love the customization options. just running with autofan and temp set to 65 right now but looking forward to using the gpu engine clocking aspect which is my main reason for trying this. getting tired of clocking my 6870 up at night when its cool and down a bit during the day if I think it will be too warm. of course the extra few MH/s doesnt hurt either heh. gonna play with it and run it a few days on the 6770 and then once I get the hang of it stick it on the other 2 rigs. BTW running it with vista ultimate 32 bit, sdk 2.4, 11.6 driver on my HTPC. thanks for letting us winblows folk play with the cool stuff too ![Smiley](https://bitcointalk.org/Smileys/default/smiley.gif)
|
|
|
So to answer the questions. I am using MSI Afterburner to do the overclocking and I do have the unofficial OC enabled.
The Brand is Sapphire and the model I believe is 100297-2L Let me know if that doesn't make sense. (this one does vent into the case)
VGA SAPPHIRE|100297-2L HD5830 1G same as I have.. above is the part number from my newegg invoice. mine is in a dedicated mining rig with XP Pro so cant comment on aero with low mem clocks. I use trixx for overclocking. PSU is an Antec Basiq 500. not a high end PSU by any means, just a decent no nonsense one. using GUIMiner with "-v -w 256 -f 5" flags
|
|
|
huh. guess I got lucky and scored a good one. I thought all 5830s were good for an easy 1000+ on the core..
|
|
|
at stock volts I run 1040 core 354 mem for weeks on end and its never crashed at that speed. even at 1050 it runs for a few days before crashing. I dont like having to check it so I leave it 1040.
what are your temps? under 70? mine runs in the mid 60s. open case, with a 120v muffin fan aimed at it.
what PSU? card is on PEGs with no adapters (although adapters shouldnt make a difference)? wondering if your 12 is sagging.
mine is the sapphire 5830 xtreme, it has a non reference cooler. but from what Ive seen and read theres nothing special or "extreme" about it.
|
|
|
OK I must be missing something.
OP has more than one pcie card and slots, correct? if so why cant you swap cards around, boot normally with a good card as display, and flash the dead one?
|
|
|
not exactly the same card, mine is an asus directCU 6870. stock volts, cant help there.
sdk 2.4, 11.6 drivers, guiminer with "-v - w 128 -f 20"
@ 1010 core it pulls 300 MHs, temps are 70s with fan @ 75%, closed case. but I had to replace the stock thermal junk with artic cooling TX2. memory is about stock as this is the main "goto" computer in the house.
|
|
|
I just consider this an interesting experiment. a bit raw and loose, might make a few bucks, might loose em, but entertaining all the way.
I should add this: I also run folding@home, and that rig draws 650 watts and has no tangible (to me at this moment) reward. just the warm fuzzies that Im helping advance medical research. might help me personally in the future, maybe not.
now, bitcoins have made me a couple hundred bucks (after power costs) but has not quite paid for what Ive put into it. but I also have to consider Im contributing to an economics (probably a better term can be used here) and, to a point, social experiment. that in itself is worth taking part of.
so, making money or not Ill be part of this for a while. I try to stash 1/3 of my coins in an offline wallet (just in case it takes off big time), and spread the rest around.
|
|
|
I like this too. simple, clean layout, reasonable protection for the hardware.
perhaps a bit more closing off in front and (if it were mine) maybe some mesh filters for the fans. got cats here, hair can get anywhere it seems.
|
|
|
thanks!
Ive run the online ones on occasion but I too like printed tables. also makes it easier to explain things to the wife.
|
|
|
my dedicated rig is a hp/compaq dc710.. P4 prescott 2.8. 40 gig hdd, 2x 1 gig ram, no opticals.
with an antec basiq 500 psu powering everything, killowatt says 321 watts. thats 20% cpu load, 1050/354 clocks on the 5830 and it mining .
before the basiq psu I ran it on the stock hp/compaq 350 psu with the card @ 850/354, killowatt said 340 watts mining. cant remember idle draw. only did that for a few weeks though.
with the basiq, idle the rig draws 120 watts or so.
just tossing the numbers out as I have a similar setup as the OP. btw the basiq is NOT an 80+ psu.
|
|
|
dump that junk PSU. as you clock up the 5830 its power draw goes up, and the PSU cant meet it 12v draw. 12v drops. hence, crash. and that psu is probably over rated by 20-30% anyway, most all generics are.
silverstone, antec, corsair, seasonic, pc power and cooling. any 450+ watt psu by those will be fine, and will actually deliver whats it rated for at real world temps. but get a reputable brand like one on that list. most rosewill and cooler master are not (although a couple got good ratings at johnny guru I think, unsure on that). some use em and it works, for a while at least, but personally I dont trust em.
grab a 650+ if you plan a 2nd card at some point, otherwise 450 or so is good.
once you have a decent PSU you should easily hit 1000+ Mhz on the core.
|
|
|
for the redneck look, and a very low restriction, coarse filter, go the the hardware store to the sheetrock section. look for the self adhesive mesh joint tape. its an open mesh, sticky, cuts easy and can be layered to get smaller mesh openings.
works surprisingly well. not fine dust of course, this is more along the lines of SMOKEUs filter but even less restrictive. anything that gets through doesnt seem to bother or build up my gpu/cpu hsf assemblies anyway, and Ive been using it for months.
|
|
|
|