2 X 7950 - Hashrate getting dropped in 4 to 5 hours after mining is started

[cellone]

Hi All,

I have 2 7950 (Sapphire 11196-09-40G HD7950 3GB Vapox-X Graphics Card) cards and my entire system configuration is given below.

The problem is initially while starting the mining task the hash rate is between 1060kh/s and 1080kh/s. But 4 to 5 hours later the hash rate is getting reduced and sometimes dropping to 880kh/s

Can anyone help me in solving this problem.

Mining Software :

• Reaper

System Configuration:

• Asrock Socket AM3+ 970 Extreme4 5200MT/S Motherboard

• Corsair AX Series AX860 860W Platinum Power Supply

• Sapphire 11196-09-40G HD7950 3GB Vapox-X Graphics Card (2)

• AMD SDX145HBGMBOX Sempron 145 Single-Core Energy Efficient Processor

• Generic Hard Disk Drive 320GB SATA 3.5"

• Corsair CMX8GX3M2A1600C9 XMS3 8GB (2x4GB) DDR3 1600 Mhz CL9 Performance Desktop Memory Kit

• Sharkoon T9 ATX Midi Tower Case - Red

GPU 1 and 2 TRIXX Information:



GPU 1 and 2 TRIXX Overclocking Settings :



GPU 1 and 2 TechPowerUP GPU-Z Graphics Card Details :



GPU 1 and 2 TechPowerUP GPU-Z Sensor Details :



GPU 1 and 2 Mining Details :



Coin conf Settings:

worksize 256 ; aggression 18 ; threads_per_gpu 1 ; sharethreads 18 ; lookup_gap 2 ; gpu_thread_concurrency 20992

Reaper Conf Settings :

kernel reaper.cl
save_binaries yes
enable_graceful_shutdown no
long_polling yes
device 0
device 1

[l3jmr]

Maybe the VRM is overheating i had the same problem with my card or you are getting HW errors due to higher intensity. I suggest you try cgminer instead of reaper and with settings -g2 -thread-concurreny 8192 -I 13

[cellone]

I tired CGMiner

When I start cgminer I am getting the error "cgminer.exe has stopped working" .
I am using Windows 8 Professional 64 bit Opperating System

[l3jmr]

That's probably because you used a wrong thread concurrency number.

Make a .bat file and enter the following text to try it out

Code:

setx GPU_MAX_ALLOC_PERCENT 100
setx GPU_USE_SYNC_OBJECTS 1

cgminer --scrypt --shaders 1792 --gpu-engine 1100 --gpu-memclock 1500 --temp-target 75 -I 13 -w 256 --lookup-gap 2  --thread-concurrency 8192 -g 2 --auto-fan -o stratum+tcp://stratum.wemineltc.com:3333 -u username -p password

For pool, use whatever pool you are mining at.

[cellone]

Whether also I need to have the conf file or its just enough to create a .bat file with these settings and use it.
 As I dont have that with me.

Also I am doing solo mining for paycoin now. So will it work for scrypt solo mining

[bitcoin-world.de]

your gpu temperature is too high, so your gpu down clock. Thats the problem.

[l3jmr]

Whether also I need to have the conf file or its just enough to create a .bat file with these settings and use it.
 As I dont have that with me.

Also I am doing solo mining for paycoin now. So will it work for scrypt solo mining

You dont need config file. Just create a .bat file and enter that. It works for scrypt mining since you have the --scrypt tag in the line i gave you.

[zackclark70]

the problem he is having is that the gpus don't run at 99% all the time they sit lower than that ( I used furmark to test and gpu usage still dosent hit 99%

[chondrite]

OP, try adding this line to your cgminer config:
--gpu-powertune 20


Looking at your VDDC values in GPU-Z before and after your miner started it seems like some of your card voltages are being down tuned.  This is a common feature on AMD cards to help prevent overheating. By pushing the powertune up to max it should help keep your hashes remain stable over time.

And I would suggest cgminer 3.3.1 as a good replacement miner for reaper. 

[l3jmr]

That is probably due to overheating since you have your cards in a case. Look at that VRM value after a couple of hours of running cgminer. Try with an open case for testing purposes or smth.

[zackclark70]

its not the temps as its still the same when fans are at 100% and the cards are a lot cooler

its not the voltage as its just voltage sag and its the same as all other rigs I have made

its not power tune as I have tested that

its not the overclock as I have tried it without

its not the psu ( ax860 )


( I built the rig for him and it got damaged in the post he had it repaired now it does this )

[hendo420]

One of my rigs is a bit unstable and if left to run 24/7 will eventually end up with a card that reads sick or dead.

I fixed this by scheduling a reboot every 3 hours and setting cgminer to auto start at boot. This has completely cured the problem. I suggest similar measures.

[zackclark70]

One of my rigs is a bit unstable and if left to run 24/7 will eventually end up with a card that reads sick or dead.

I fixed this by scheduling a reboot every 3 hours and setting cgminer to auto start at boot. This has completely cured the problem. I suggest similar measures.

thanks but the problem he is having is that the cards wont sit at 99% they sit at around 90% after only 1 or 2 mins

I have never had it happen before and I have made lots of rigs

[cellone]

That is probably due to overheating since you have your cards in a case. Look at that VRM value after a couple of hours of running cgminer. Try with an open case for testing purposes or smth.

I opened both the side doors in the case and tried it. But still the problem occurs