Author

Topic: Stability with four Asus R9 280X (Read 1203 times)

newbie
Activity: 21
Merit: 0
March 13, 2014, 09:12:49 PM
#12
I don't know if anybody is interested in this topic anymore but I felt obligated to inform it in here about what happened in my situation;


As suspected, apperantly my problem was with the psu I was using(Xilence Redwing 800W)

I replaced my psu with Cooler Master Silent Pro Hybrid 1050W.

2 Asus R9 280X Tops are doing great with it.

Thanks for the help.
newbie
Activity: 21
Merit: 0
February 26, 2014, 10:22:16 PM
#11
I believe your problem is power, R9-280 at full tilt will take 330w, along with your core i7 and system overhead thats taking you way too close to your 800w top end of PSU. There are a lot of minor PSU brands out there that do not really deliver the wattage that they claim, use a mainstream brand (Corsair, Seasonix)

Yes, this is what I thought too but I'm not very eager about buying a psu before trying if the problem really is about psu. At the weekend I'll be able to try the whole system with a different psu. So until weekend I'm bound to look for other options.

Hi,

Don't be afraid of BAMT, I was a noob at Linux, actually BAMT was the first time I ever touched Linux. And I'm still a noob, but this tutorial is very detailed.
Also bear in mind that a USB stick consumes (and costs) less than a HDD...

An did I mention that it takes only 15 minutes to get a new rig up and running with this? (time to assemble the rig and copy the image on the flashdrive)

Oh, ok I'll look into that tutorial tomorrow. But really for some reason, I'd rather just buy a new psu and get it over with Smiley Simply because I'm very lazy with learning new stuff heh. Nevertheless, I'll definitely check out the tutorial and see if it fits my needs.

Thanks for the help guys.
hero member
Activity: 672
Merit: 500
February 26, 2014, 07:35:55 PM
#10
I have the same issues trying to run 4 280x's under clpts. 3 run fine but the fourth one causes either the bat file to crash or driver crash. All are on powered risers and using a Coolermaster V1000 psu. Moved around cards to different slots, reinstalled drivers, etc. So far am at a loss.
Again, power is your problem, 4 x 330W cards > 1000W PSU! You need at least a 1500w or 2 x 750w PSU

Son-of-a-beech...you are right. Some how i kept thinking that 1k psu was ok. I just bought a 750w for another purpose, i will grab another and give it a go.

Thanks!
newbie
Activity: 36
Merit: 0
February 26, 2014, 04:53:20 PM
#9
I'm having the same problem with 2 Asus R9 280X DCU2 Tops. With 1 card I have been able to mine for 3 weeks without a problem now. So with the money I made I went for the second one and now I'm having random reboots and everything.

The thing is, when I set the Intensity to a lower amount like 8-10 it goes on about 15-20 mins with low khashes, if I set it to 12, system reboots itself after 10-15mins and with intensity 13 it reboots almost instantly without showing me hash rates..

Both the new card and the old one runs without a problem when I use them solo.

My rig,

MSI P67A-GD55 (B3)
i7 2600k
8 gig ram
Xilence 800W psu
2 x Asus R9 280X DCU2 TOP
Bios is the latest version: 04.06.05

The reboot process is kind of interesting though, looks like a bios reboot in which all system looks like it had been shut down(all fans and everything stops for a second then restarts on its own)

I have been fighting with this for a day now, can you share what else could work besides going for BAMT, I am really a linux noob don't think I have what it takes to deal with the linux just yet.

Appreciate any help guys, cheers.

Hi,

Don't be afraid of BAMT, I was a noob at Linux, actually BAMT was the first time I ever touched Linux. And I'm still a noob, but this tutorial is very detailed.
Also bear in mind that a USB stick consumes (and costs) less than a HDD...

An did I mention that it takes only 15 minutes to get a new rig up and running with this? (time to assemble the rig and copy the image on the flashdrive)
hero member
Activity: 574
Merit: 500
February 26, 2014, 04:30:44 PM
#8
I have the same issues trying to run 4 280x's under clpts. 3 run fine but the fourth one causes either the bat file to crash or driver crash. All are on powered risers and using a Coolermaster V1000 psu. Moved around cards to different slots, reinstalled drivers, etc. So far am at a loss.
Again, power is your problem, 4 x 330W cards > 1000W PSU! You need at least a 1500w or 2 x 750w PSU
hero member
Activity: 672
Merit: 500
February 26, 2014, 04:23:15 PM
#7
I have the same issues trying to run 4 280x's under clpts. 3 run fine but the fourth one causes either the bat file to crash or driver crash. All are on powered risers and using a Coolermaster V1000 psu. Moved around cards to different slots, reinstalled drivers, etc. So far am at a loss.
hero member
Activity: 574
Merit: 500
February 26, 2014, 04:06:47 PM
#6
I'm having the same problem with 2 Asus R9 280X DCU2 Tops. With 1 card I have been able to mine for 3 weeks without a problem now. So with the money I made I went for the second one and now I'm having random reboots and everything.

The thing is, when I set the Intensity to a lower amount like 8-10 it goes on about 15-20 mins with low khashes, if I set it to 12, system reboots itself after 10-15mins and with intensity 13 it reboots almost instantly without showing me hash rates..

Both the new card and the old one runs without a problem when I use them solo.

My rig,

MSI P67A-GD55 (B3)
i7 2600k
8 gig ram
Xilence 800W psu
2 x Asus R9 280X DCU2 TOP
Bios is the latest version: 04.06.05

The reboot process is kind of interesting though, looks like a bios reboot in which all system looks like it had been shut down(all fans and everything stops for a second then restarts on its own)

I have been fighting with this for a day now, can you share what else could work besides going for BAMT, I am really a linux noob don't think I have what it takes to deal with the linux just yet.

Appreciate any help guys, cheers.
I believe your problem is power, R9-280 at full tilt will take 330w, along with your core i7 and system overhead thats taking you way too close to your 800w top end of PSU. There are a lot of minor PSU brands out there that do not really deliver the wattage that they claim, use a mainstream brand (Corsair, Seasonix)
hero member
Activity: 574
Merit: 500
February 26, 2014, 04:00:23 PM
#5
Yeah i've suffered identical problem with my five R9-280x (four Asus Direct CU and 1 Matrix edition). I'm finding stability a real pain in the a**.
So far the best i've managed is a max of 3 cards on an Asus p8z77-lx2 motherboard, with 2 Seasonix 850w PSU's and cgminer 3.7.2. This will happily run 24x7 for weeks. Adding one more GPU causes a full system freeze after only a few seconds.

Similarly on Asus p8p67 motherboard I can run four cards, at stock clock/memory but stability is terrible, can be 10 mins or 10 hours before system freezes, totally unpredictable. And it's always a power cycle to get it started, no reboot available as system freezes completely!

I've literally just upgraded to latest AMD beta driver tonight, but spent months messing around with latest bioses, versions of cgminer etc to no avail.

I should add that was BAMT 1.2 with different variations of cgminer and updated R9 drivers.

This thread shows same problem: https://bitcointalksearch.org/topic/4th-gpu-crashing-cg-miner-4x-gigabyte-r9-280x-oc-edition-448756
newbie
Activity: 21
Merit: 0
February 26, 2014, 07:28:27 AM
#4
I'm having the same problem with 2 Asus R9 280X DCU2 Tops. With 1 card I have been able to mine for 3 weeks without a problem now. So with the money I made I went for the second one and now I'm having random reboots and everything.

The thing is, when I set the Intensity to a lower amount like 8-10 it goes on about 15-20 mins with low khashes, if I set it to 12, system reboots itself after 10-15mins and with intensity 13 it reboots almost instantly without showing me hash rates..

Both the new card and the old one runs without a problem when I use them solo.

My rig,

MSI P67A-GD55 (B3)
i7 2600k
8 gig ram
Xilence 800W psu
2 x Asus R9 280X DCU2 TOP
Bios is the latest version: 04.06.05

The reboot process is kind of interesting though, looks like a bios reboot in which all system looks like it had been shut down(all fans and everything stops for a second then restarts on its own)

I have been fighting with this for a day now, can you share what else could work besides going for BAMT, I am really a linux noob don't think I have what it takes to deal with the linux just yet.

Appreciate any help guys, cheers.
newbie
Activity: 36
Merit: 0
January 16, 2014, 02:02:51 PM
#3
Might still be the risers, can you try 4 powered risers? Otherwise try to measure the power used by the PSU's

Hi gbrs,

Thank you for your message. I ordered a kill a watt, but it did not arrive yet. So what I did is that I installed BAMT on a flashdrive, well... my minion has been hashing solid for the past 19 hours!!!

So it was either a Windows problem or HDD related...

Anyway I find BAMT extremely efficient and cost effective, so I'll stick to that  Grin
newbie
Activity: 43
Merit: 0
January 13, 2014, 03:00:06 PM
#2
Might still be the risers, can you try 4 powered risers? Otherwise try to measure the power used by the PSU's
newbie
Activity: 36
Merit: 0
January 13, 2014, 01:52:16 PM
#1
Hi guys,

I am desperately looking for help. I've been looking for a fix for days on this forum and google without success yet.

The problem is that my rig cannot mine on four GPUs without crashing and rebooting randomly. The longest I was able to mine on four GPUs is 7 hours, but on average my rig crashes within an hour of mining.

Here is my component list:

Asrock 970 Extreme4
AMD Sempron 135
4GB of RAM (Hynix)
4x Asus R9 280X DCU2 Top
2x 850W Akasa Cobra 82% efficiency (two GPUs on each PSU)
3x powered riser 1x to 16x
1x unpowered riser 16x to 16x

Windows 7 x64 Ultimate
CGMiner 3.7.2
Drivers AMD 13.12

Here is my .bat file:

Quote
setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1
C:\cgminer-3.7.2-windows\cgminer.exe --scrypt -I 13 -w 256 --thread-concurrency 8192 -g 2 --auto-fan --temp-target 75 --temp-overheat 85 --gpu-engine 1080 --lookup-gap 2 --gpu-memclock 1500 --scan-time 1 --queue 0 --no-submit-stale -o stratum+tcp://(pool) -u (user) -p (pass)

Total hashrate is about 2.96 Mh/s

The funny thing is that it runs normally without crash with 3 GPUs.

At first, I thought it was a riser issue, as I had lost hair trying to make my rig detect my GPUs, just to figure out that it was due to faulty risers. I swapped them, but still crashing.

Then I reinstalled my drivers, no change.

Then I thought it was due to a temperature problem, although my cards run at 82° for the hottest. Still, I put a temp target to 75°, and it still crashes randomly.

I think that 850w x2 is enough for four of those cards, am I right?

Can anybody help me please? It is so frustrating...  Huh

(PS: I run the same .bat file on my main PC with one 280x, and it just rocks!)  Grin


Jump to: